PDA

View Full Version : wireless range vs. power



Adam Chapman
04-02-08, 10:00 AM
Hi,

I've ordered a wireless webcam that will be out onboard an autonomous
aircraft for a university project.

The camera has a "2dBi Diversity Antenna" and transmits video over an
IP protocol in the 2.4-2.4835 GHz range.

Now the aircraft needs to operate in a 500x500m area, although I
expect that my ground station computer will be placed some distance
away from the operating area.

My problem is that I don't know how to calculate the achievable
transmission range from power and frequency. Is there a rule of thumb
calculation for finding the transmission range?

UK regulations say that I cant have a power above 100mW- does anybody
know what range is achievable with this power in the 2.4 GHz band?

Any help is heartfuly appreciated.

Adam

DTC
04-02-08, 10:37 AM
Adam Chapman wrote:
> My problem is that I don't know how to calculate the achievable
> transmission range from power and frequency. Is there a rule of thumb
> calculation for finding the transmission range?

There are too many unpredictable variables. If your airplane tilts in
just the slightest, you'll loose the signal. All you can really do is
try it out and see how far it works.

Adam Chapman
04-02-08, 10:55 AM
On Apr 2, 4:37*pm, DTC <m...@nothingtoseehere.zzx> wrote:
> Adam Chapman wrote:
> > My problem is that I don't know how to calculate the achievable
> > transmission range from power and frequency. Is there a rule of thumb
> > calculation for finding the transmission range?
>
> There are too many unpredictable variables. If your airplane tilts in
> just the slightest, you'll loose the signal. All you can really do is
> try it out and see how far it works.

even if i have an omni-directional antenna?

LR
04-02-08, 01:16 PM
Adam Chapman wrote:
> Hi,
>
> I've ordered a wireless webcam that will be out onboard an autonomous
> aircraft for a university project.
>
> The camera has a "2dBi Diversity Antenna" and transmits video over an
> IP protocol in the 2.4-2.4835 GHz range.
>
> Now the aircraft needs to operate in a 500x500m area, although I
> expect that my ground station computer will be placed some distance
> away from the operating area.
>
> My problem is that I don't know how to calculate the achievable
> transmission range from power and frequency. Is there a rule of thumb
> calculation for finding the transmission range?

Communication is a 2 way operation and you would need to know the Rx
sensitivity, Tx o/p, Antenna gain... of both the devices to get even a
vague idea of the Link distance that would be achieved by 2 staic devices.
If one of the devices is moving in 3 dimensions that would make it
extremely difficult to calculate. You say it would operate in an area
500x500 metres...not true as it would be 500x500xheight metres.
If you were
>
> UK regulations say that I cant have a power above 100mW- does anybody
> know what range is achievable with this power in the 2.4 GHz band?
>
> Any help is heartfuly appreciated.
>
> Adam

LR
04-02-08, 01:37 PM
LR wrote:
> Adam Chapman wrote:
>> Hi,
>>
>> I've ordered a wireless webcam that will be out onboard an autonomous
>> aircraft for a university project.
>>
>> The camera has a "2dBi Diversity Antenna" and transmits video over an
>> IP protocol in the 2.4-2.4835 GHz range.
>>
>> Now the aircraft needs to operate in a 500x500m area, although I
>> expect that my ground station computer will be placed some distance
>> away from the operating area.
>>
>> My problem is that I don't know how to calculate the achievable
>> transmission range from power and frequency. Is there a rule of thumb
>> calculation for finding the transmission range?
>
> Communication is a 2 way operation and you would need to know the Rx
> sensitivity, Tx o/p, Antenna gain... of both the devices to get even a
> vague idea of the Link distance that would be achieved by 2 staic devices.
> If one of the devices is moving in 3 dimensions that would make it
> extremely difficult to calculate. You say it would operate in an area
> 500x500 metres...not true as it would be 500x500xheight metres.

Bleeding cats..grrh.
How do you intend to keep the Antennas in the same plane.Depending on
the gain of the omni on the ground unit you may find that if your
aircraft goes to high the camera's antenna may not be in the beamwidth
of the groundstaions antenna.
http://my.athenet.net/~multiplx/cgi-bin/omni.main.cgi

You could perhaps get a rough idea of the max distance of a static link
using a link calculator:-
http://www.wifiextreme.com.au/index.php?main_page=page_5
However adding movement in will just just reduce the workable distance.
You will also have to try and factor in the Data transfer rate you need
to use for an acceptable video quality. The higher the rate the less
distance.

>>
>> UK regulations say that I cant have a power above 100mW- does anybody
>> know what range is achievable with this power in the 2.4 GHz band?
>>
>> Any help is heartfuly appreciated.
>>
>> Adam

P.Schuman
04-02-08, 03:33 PM
Adam Chapman wrote:
> On Apr 2, 4:37 pm, DTC <m...@nothingtoseehere.zzx> wrote:
>> Adam Chapman wrote:
>>> My problem is that I don't know how to calculate the achievable
>>> transmission range from power and frequency. Is there a rule of
>>> thumb calculation for finding the transmission range?
>>
>> There are too many unpredictable variables. If your airplane tilts in
>> just the slightest, you'll loose the signal. All you can really do is
>> try it out and see how far it works.
>
> even if i have an omni-directional antenna?

remember that any omni antenna readiates as a donut sitting over the
antenna.
So - just as an example - imagine if the plane was directly overhead,
you might be able to see it, but not receive the video since your receiving
"donut"
and the plane's transmitting "donut" would not intersect.

As you can see - the name of the game is for both plane + ground station
to have intersecting donuts....
As the plane flies away, you may actually receive a better signal
as the donut beamwidth pattern increases with distance - but strength
decreases -
SO - ??

Jeff Liebermann
04-02-08, 05:54 PM
On Wed, 2 Apr 2008 08:00:06 -0700 (PDT), Adam Chapman
<adam.chapman@student.manchester.ac.uk> wrote:

>I've ordered a wireless webcam that will be out onboard an autonomous
>aircraft for a university project.

I've done that. You're about to have a battery problem.

>The camera has a "2dBi Diversity Antenna" and transmits video over an
>IP protocol in the 2.4-2.4835 GHz range.

Duz this camera have a manufacturer and a model number? A URL with
the specifications would be nice.

2dBi is a simple monopole antenna. It has a radiation pattern that
looks like a donut. There's a big null when in line with the antenna,
such as when the aircraft is flying directly overhead.

>Now the aircraft needs to operate in a 500x500m area, although I
>expect that my ground station computer will be placed some distance
>away from the operating area.

I have this thing about numbers. How far is "some distance"? How
high does the air thing have to fly? What's the MAXIMUM distance you
expect to see an image?

>My problem is that I don't know how to calculate the achievable
>transmission range from power and frequency.

Neither do I. You've supplied exactly one non-ambiguous number (the
2dBi antenna gain). What's missing is everything else that's on the
specification sheet.

>Is there a rule of thumb
>calculation for finding the transmission range?

If you can see your thumb, you can communicate.

Yes, there are calculations possible. They're actually fairly easy.
See example at:
<http://wireless.wikia.com/wiki/Wi-Fi#Link_Calculations>
However, you're doing video which requires some tweaks to the
calculations (depending on bandwidth and modulation method). I can't
help without you supplying some numbers.

>UK regulations say that I cant have a power above 100mW- does anybody
>know what range is achievable with this power in the 2.4 GHz band?

Sorry, no answer without knowing the antenna gains, receiver
sensitivity, SNR required, and modulation method.

>Any help is heartfuly appreciated.

Take your video xmitter and receiver and connect to a portable TV.
Start walking. When the picture starts to look horrible, stop. Record
the distance. Let's say it's 100 meters.

If you increase the antenna gain by 6dB, you will get twice the range
or 200 meters. 12dB increase in gain will give you 4 times the range,
or 400 meters. 18dB gain will yield 8 times the range. Ad nasusium.
The problem is that every time you increase the gain, the beamwidth
and pattern of the antenna gets narrower and narrower. The airplane
mounted video camera I helped throw together had about a 1000 meter
range. However, we used a 24dBi dish antenna on the ground, that
tracked the aircraft (both automagically and manually).

--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Jeff Liebermann
04-02-08, 06:13 PM
On Wed, 02 Apr 2008 15:37:48 GMT, DTC <me@nothingtoseehere.zzx> wrote:

>If your airplane tilts in
>just the slightest, you'll loose the signal.

Nope. A 2dBi donut pattern isn't the same as a hemispherical pattern,
but close enough. When we were tinkering with model airplane video,
the best antenna was a 3cm coaxial antenna (1/4 wave driven element,
1/4 wave coax sleeve) made out of a piece of RG-175, dangling roughly
downward. The only times when the signal disappeared were when the
plane flew overhead and when the body blocked the signal (i.e.
inverted flying). The null was rather sharp and the image returned
quickly. It's not as bad as it would seem. Some of the other model
airplane pilots were using patch antenna, which had less drag, and
were only useful in level flight.

What was really cool was flying the plane with the camera in the
cockpit and watching through video goggles. 640x480 for each eye:
<http://www.vuzix.com/iwear/products_vr920.html>
Nice for watching 3D TV when not flying. Say goodby to about $350 for
those.

Incidentally, almost all the ground based video receivers had some
manner of directional antenna attached. Ours was overkill with a
24dBi dish (because I had one handy). Others were using everything
from cantennas to big panels. There were a few high gain omnis, but
they didn't work too well.

>All you can really do is
>try it out and see how far it works.

Yep. I kinda prefer calculations first, but a sanity check is always
a good idea.

Thought 2.4GHz was kinda crowded? Futaba 2.4GHz radio control.
<http://2.4gigahertz.com/>
The wireless video and these should interfere nicely.

Ugh. Back to working on my taxes...


--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

msg
04-02-08, 09:33 PM
Jeff Liebermann wrote:

<snip>

> The airplane mounted video camera I helped throw together had about
> a 1000 meter range. However, we used a 24dBi dish antenna on the
> ground, that tracked the aircraft (both automagically and manually).

Did you build a nutating feed for the ground station, or perhaps a
phased array with electronic nutation?

Michael

Jeff Liebermann
04-03-08, 01:43 AM
On Wed, 02 Apr 2008 20:33:34 -0600, msg <msg@_cybertheque.org_> wrote:

>Jeff Liebermann wrote:
>
><snip>
>
>> The airplane mounted video camera I helped throw together had about
>> a 1000 meter range. However, we used a 24dBi dish antenna on the
>> ground, that tracked the aircraft (both automagically and manually).

>Did you build a nutating feed for the ground station, or perhaps a
>phased array with electronic nutation?
>Michael

Nope. There were numerous models of tracking antenna. That was my
job. The problem was that the spread spectrum signal does not have a
carrier, which made it difficult to throw together an easy tracking
circuit. The ability of the airplane to fly close and perpendicular
to the antenna made tracking even more difficult. Worse, the 24dBi
dish has a -3dB beamwidth of about 7 degrees. What seemed like an
easy project turned into a real mess.

One scheme that came really close to working required 4 additional
receivers. The center feed mount was extended beyond the feed and
made from an RF absorber (i.e. PVC pipe fill of water). 4 antennas
were arranged in 4 quadrants on each side of the center feed pipe
extension. With the signal source dead ahead, the signal levels at
all 4 receivers is (allegedly) identical. If the antenna were offset
in one direction, the center pipe casts an RF "shadow" on the one of
the 4 antennas. A differential amplifier runs a gimbal mounted pair
of motors to correct the direction.

This worked well fairly for tracking under ideal conditions, but had
acquisition problems that drove me into overtime and later panic.
There were just too many reflections that screwed up the direction. If
I had a deep solid dish (large f/D ratio), instead of a barbeque grill
dish, the reflections would have been minimized and it might have
worked better. However, once the direction antennas were hit by a
reflection, the motor would swerve the antenna radically, losing lock,
and not easily recovering.

The radio link was via 802.11b wi-fi so unlike the ATV link, there was
a transmitter involved at the dish. When it transmitted, the
directional receivers were instantly overloaded. I worked around this
problem by temporarily disabling the servos in transmit.

I eventually gave up on the RF approach.

There were several other attempts to build a tracker. I finally threw
together an optical system that worked. It was similar to the RF
tracker, but was immune to all but the most disgusting reflections.
The aircraft carried several green LED's that pulsed at about 100 Hz.
The 4 antennas were replaced by 4 security camera lenses and photo
transistors. The center pipe was replaced by a fiberglass tube, which
blocks light, but passes RF.

This system worked much better, especially at night. It crapped out
when pointed into the sun, when some dingbat shines his flashlight on
the antenna, and when I took a flash photo. Range was limited to
about 300 meters. I never did nail down the servo loop damping
factor, so it tended to either crawl across the traverse, or twitch
badly as it moved. Neither seemed to bother the RF data link, but it
sure made everyone around the antenna nervous. Like the RF solution,
it worked for tracking and sucked for acquisition.

I had thought of using a nutating (conical scan) feed (straight out of
the WWII SCR-584 radar):
<http://www.carnarvonspace.com/wiki/index.php?title=SCR-584_Technical_Description>
However, I didn't have time to machine the required components and the
corresponding control system. When I was much younger, there were
tons of those feeds available in the WWII surplus stores, but those
are long gone today.

I have some other ideas on how to do a wi-fi tracking antenna.
However, I keep seeing high skool and college project proposals that
involve the construction of such systems. I don't wanna ruin it for
the students.


--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Adam Chapman
04-03-08, 10:12 AM
Thanks to all of you, I replied to this particular post because it had
the most questions for me to answer.
I do hope that Im not appearing to be lazy here, I've never had to
work in wireless systems as I am an aerodynamisist recently turned
image processing geek. I was expecting my question to have an easy
answer but as usual here wasn't so sorry and thank you for taking your
own time to help me.

On Apr 2, 11:54 pm, Jeff Liebermann <je...@cruzio.com> wrote:
> On Wed, 2 Apr 2008 08:00:06 -0700 (PDT), Adam Chapman
>
> <adam.chap...@student.manchester.ac.uk> wrote:
> >I've ordered a wireless webcam that will be out onboard an autonomous
> >aircraft for a university project.
>
> I've done that. You're about to have a battery problem.
>
> >The camera has a "2dBi Diversity Antenna" and transmits video over an
> >IP protocol in the 2.4-2.4835 GHz range.
>
> Duz this camera have a manufacturer and a model number? A URL with
> the specifications would be nice.
>

The camera is a Trendnet IP-TV301W model, specification at
http://trendnet.com/products/proddetail.asp?prod=110_TV-IP301W&cat=48.
I don't physically have the camera yet because the university is still
processing the account, which in my experience can take a very long
time. If it is possible to estimate transmission ranges with certain
equipment, I would have an easier job justifying the expense to my
department.

> 2dBi is a simple monopole antenna. It has a radiation pattern that
> looks like a donut. There's a big null when in line with the antenna,
> such as when the aircraft is flying directly overhead.
>

The 2dBi antenna can be replaced, if there is a better type then I am
happy to use it. I would prefer an antenna that emits as spherical a
pattern as possible, otherwise we will have to limit aircraft
maneuevres.



> >Now the aircraft needs to operate in a 500x500m area, although I
> >expect that my ground station computer will be placed some distance
> >away from the operating area.
>
> I have this thing about numbers. How far is "some distance"? How
> high does the air thing have to fly? What's the MAXIMUM distance you
> expect to see an image?
>

I have just recieved more information from the competition organisers,
who tell me that all flying will be within 500m of the launch site and
flying is not permitted above 400ft (121.92 m) This gives the max.
Euclidean distance as 514.65m, although i expect we will be measuring
that distance on the aircraft with GPS, and with the civilian accuracy
limitations on GPS (~ +- 12m), I would suggest a target transmission
distance of 550m.

Due to safety rules we will never lose sight of the aircraft so we can
expect no obstacles between transmitter and reciever.

I've been slowly writing this reply all day, and thinking a lot as i
do so. A helical antenna looks the best as it is the most isotropic,
although it is also pretty big for our aircraft. However even that has
+-45 degree 3dB lobe, so i might have to have the aircraft bank more
when closer to the centre of its circular mission area.

The link calculator link from LR told me that if i put an antenna with
a 7dBi gain (like the one at _) on the UAV and at the ground station,
i will have a range of over 7km with a 100mW transmitter, which I
assume can be achieved using an in-line amplifier. Does this
calculator use the range in the direction of the strongest energy
emisiion from the antenna?

I guess that more than on antenna on the same aircraft with different
orientations would interfere with each others signals?

Adam

Adam Chapman
04-03-08, 10:15 AM
Thanks to all of you, I replied to this particular post because it had
the most questions for me to answer.
I do hope that Im not appearing to be lazy here, I've never had to
work in wireless systems as I am an aerodynamisist recently turned
image processing geek. I was expecting my question to have an easy
answer but as usual here wasn't so sorry and thank you for taking your
own time to help me.

On Apr 2, 11:54 pm, Jeff Liebermann <je...@cruzio.com> wrote:
> On Wed, 2 Apr 2008 08:00:06 -0700 (PDT), Adam Chapman
>
> <adam.chap...@student.manchester.ac.uk> wrote:
> >I've ordered a wireless webcam that will be out onboard an autonomous
> >aircraft for a university project.
>
> I've done that. You're about to have a battery problem.
>
> >The camera has a "2dBi Diversity Antenna" and transmits video over an
> >IP protocol in the 2.4-2.4835 GHz range.
>
> Duz this camera have a manufacturer and a model number? A URL with
> the specifications would be nice.
>

The camera is a Trendnet IP-TV301W model, specification at
http://trendnet.com/products/proddetail.asp?prod=110_TV-IP301W&cat=48.
I don't physically have the camera yet because the university is still
processing the account, which in my experience can take a very long
time. If it is possible to estimate transmission ranges with certain
equipment, I would have an easier job justifying the expense to my
department.

> 2dBi is a simple monopole antenna. It has a radiation pattern that
> looks like a donut. There's a big null when in line with the antenna,
> such as when the aircraft is flying directly overhead.
>

The 2dBi antenna can be replaced, if there is a better type then I am
happy to use it. I would prefer an antenna that emits as spherical a
pattern as possible, otherwise we will have to limit aircraft
maneuevres.



> >Now the aircraft needs to operate in a 500x500m area, although I
> >expect that my ground station computer will be placed some distance
> >away from the operating area.
>
> I have this thing about numbers. How far is "some distance"? How
> high does the air thing have to fly? What's the MAXIMUM distance you
> expect to see an image?
>

I have just recieved more information from the competition organisers,
who tell me that all flying will be within 500m of the launch site and
flying is not permitted above 400ft (121.92 m) This gives the max.
Euclidean distance as 514.65m, although i expect we will be measuring
that distance on the aircraft with GPS, and with the civilian accuracy
limitations on GPS (~ +- 12m), I would suggest a target transmission
distance of 550m.

Due to safety rules we will never lose sight of the aircraft so we can
expect no obstacles between transmitter and reciever.

I've been slowly writing this reply all day, and thinking a lot as i
do so. A helical antenna looks the best as it is the most isotropic,
although it is also pretty big for our aircraft. However even that has
+-45 degree 3dB lobe, so i might have to have the aircraft bank more
when closer to the centre of its circular mission area.

The link calculator link from LR told me that if i put an antenna with
a 7dBi gain (like the one at _) on the UAV and at the ground station,
i will have a range of over 7km with a 100mW transmitter, which I
assume can be achieved using an in-line amplifier. Does this
calculator use the range in the direction of the strongest energy
emisiion from the antenna?

I guess that more than on antenna on the same aircraft with different
orientations would interfere with each others signals?

Adam

msg
04-03-08, 11:41 AM
Adam Chapman wrote:

> A helical antenna looks the best as it is the most isotropic,
> although it is also pretty big for our aircraft. However even that has
> +-45 degree 3dB lobe, so i might have to have the aircraft bank more
> when closer to the centre of its circular mission area.

How about taking a lesson from radiosonde antennas; I don't know about
current models, but for a very long time the antenna at 1680 MHz was
a 1/4 wave stub mounted at the center of an inverted conical reflector
(about 60 degrees). When suspended from the balloon, the antenna points
downward and illuminates the ground with a pattern formed from that
of the stub combined with the reflector, giving horizon-wide coverage
with no nulls for high inclination either.

Also, you could mount two such antennas on the fuselage, at 180 degrees
top and bottom and use a splitter to drive them so that there is
coverage during maneuvers.

Michael

msg
04-03-08, 11:43 AM
Jeff Liebermann wrote:

<snip>

Sounds like a heck of a project you did ;)

>
> I had thought of using a nutating (conical scan) feed (straight out of
> the WWII SCR-584 radar):
> <http://www.carnarvonspace.com/wiki/index.php?title=SCR-584_Technical_Description>
> However, I didn't have time to machine the required components and the
> corresponding control system. When I was much younger, there were
> tons of those feeds available in the WWII surplus stores, but those
> are long gone today.

In days gone by, I had feeds from the MPQ-10A mortar tracking radar as well
as the GMD-1 ground station.

Michael

Jeff Liebermann
04-03-08, 12:02 PM
On Thu, 3 Apr 2008 08:12:00 -0700 (PDT), Adam Chapman
<adam.chapman@student.manchester.ac.uk> wrote:

>I do hope that Im not appearing to be lazy here, I've never had to
>work in wireless systems as I am an aerodynamisist recently turned
>image processing geek.

If this is a skool project, you are expected to do your own research
and calculations. You don't learn much by having us doing it for you.
If you have a *SPECIFIC* problem with some aspect of the project, feel
free to ask. However, the basic design in implimentation should be
all your work.

>I was expecting my question to have an easy
>answer but as usual here wasn't so sorry and thank you for taking your
>own time to help me.

Welcome to engineering at its worst. The URL I listed:
<http://wireless.wikia.com/wiki/Wi-Fi#Link_Calculations>
shows a simplified wireless link calculation. You are apparently
using 802.11b/g instead of VSB modulation for video (and data) so the
examples should work. Use this as the basis for your calcs. However,
please note that the resultant numbers are the *BEST* case
calculations. Real world situations only make things worse. For
example, the 2dBi gain of the antenna is actually the MAXIMUM gain of
the antenna. If you are off axis from the antenna pattern, the gain
(and range) will be less. Receiver sensitivity is never as good as
they claim. This should help:
<http://802.11junk.com/jeffl/rx-sens/receiver%20sensitivity.htm>
You will need to specify the other end of your radio link (equipment,
coax, antenna) before I can help with the numbers. It's really quite
simple.

>The camera is a Trendnet IP-TV301W model, specification at
>http://trendnet.com/products/proddetail.asp?prod=110_TV-IP301W&cat=48.

OK. 802.11b/g. Draws 7 watts (at 12V) and weighs 0.5kg. Do you
have any idea of how big a battery will need to be to supply 7 watts
of continuous power? 4ea LIPO cell phone batteries will do the trick,
but you'll only have perhaps 15 minutes of operating time. To lift
all this, the plane doesn't need to be huge, but it certainly won't be
a styrofoam fly weight.

>I don't physically have the camera yet because the university is still
>processing the account, which in my experience can take a very long
>time.

The local university owes me some money from a project that is now 22
years overdue. Good luck.

>If it is possible to estimate transmission ranges with certain
>equipment, I would have an easier job justifying the expense to my
>department.

I don't see the connection between range and funding. I suspect
you'll do better with a feasability and finance study, than with a
performance estimate.

>The 2dBi antenna can be replaced, if there is a better type then I am
>happy to use it. I would prefer an antenna that emits as spherical a
>pattern as possible, otherwise we will have to limit aircraft
>maneuevres.

It really depends on what material you make the airplane from. If RF
transparent, a simple monopole (as supplied) will work. When
inverted, the signal will go through the aircraft. However, if you
make the airplane from reflective or absorptive materials, you'll need
multiple antennas. One 1/4 wave monopole on top and one on the bottom
should be sufficient. With a small ground plane under the antenna,
the pattern should be mostly cardioid shaped. If you have a wind
resistance problem, you can use a small ceramic patch antenna instead.
It is possible to model the antenna pattern using (free) antenna
modeling tools:
<http://home.ict.nl/~arivoors/>
The sample files include a 747 aircraft (which will need to be
scaled). This is not exactly a trivial exercise, a potential time
burning diversion, and is probably overkill for what you're
attempting. It's easy enough to construct a model, install the
camera, and just play with the position of the antenna until it's a
workable compromise.

Note that installing a bigger antenna on the airplane is not what you
want. Bigger antennas imply more gain. You want hemispherical
antenna pattern which requires a smaller or simpler antenna. Here's
an approximation of an isotropic antenna:
<http://802.11junk.com/jeffl/antennas/isotropic/index.html>
I think this is overkill and probably not practical on an airplane. If
you're going to do work with the antenna, do it on the ground station.

>I have just recieved more information from the competition organisers,
>who tell me that all flying will be within 500m of the launch site and
>flying is not permitted above 400ft (121.92 m) This gives the max.
>Euclidean distance as 514.65m, although i expect we will be measuring
>that distance on the aircraft with GPS, and with the civilian accuracy
>limitations on GPS (~ +- 12m), I would suggest a target transmission
>distance of 550m.
>
>Due to safety rules we will never lose sight of the aircraft so we can
>expect no obstacles between transmitter and reciever.

If this is competition flying, it's also likely (but not guaranteed)
that you'll also get minimal 2.4GHz interference. It's traditional to
confiscate the controllers during the competition to prevent sabotage.
That works because RC has dedicated frequencies. That's NOT the case
with 2.4GHz, where there's a chance that some lunatic, with a portable
wireless router, can crash your flying machine. You can also lose
control when the aircraft flys over a coffee shop wireless hot spot.
(I'm assuming that you'll be controlling on 2.4GHz). I sure hope this
is going to be run in the middle of nowhere. Testing it at skool,
where there's lots of 2.4GHz interference, is going to be ummm...
interesting.

>I've been slowly writing this reply all day, and thinking a lot as i
>do so.

At the beginning of any project, leaps of faith, wild ideas, the
shooting from the hip, are standard procedure. Once the ideas are
ossified, then you can be more studious and careful.

>A helical antenna looks the best as it is the most isotropic,

Too big, too much drag, too critical, poor gain to volume ratio. If
you're going to go to that level of complexity, think about two
ceramic patch antennas with a -3dB Wilkinson combiner. Two back to
back cartoid patterns equal roughly a hemisphere.

>although it is also pretty big for our aircraft. However even that has
>+-45 degree 3dB lobe, so i might have to have the aircraft bank more
>when closer to the centre of its circular mission area.

Translation. If you have a hole in the antenna pattern, you're going
to lose control. A helix also has no gain in the reverse direction,
so you'll need two. Hint: Look at what's on a real airplane or
guided missile.

>The link calculator link from LR told me that if i put an antenna with
>a 7dBi gain (like the one at _) on the UAV and at the ground station,
>i will have a range of over 7km with a 100mW transmitter, which I
>assume can be achieved using an in-line amplifier. Does this
>calculator use the range in the direction of the strongest energy
>emisiion from the antenna?

I'm still missing some numbers. See the example calcs on the URL I
mentioned. Copy the table of numbers and plug in YOUR numbers.
Remember, those are the BEST case calculations. Reality only makes it
worse.

>I guess that more than on antenna on the same aircraft with different
>orientations would interfere with each others signals?

If the two antennas cannot "see" each other, they will not mangle the
combined pattern. What other signals?

Good luck. Sounds like an interesting project. Watch your weight and
power consumption.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Jeff Liebermann
04-03-08, 12:07 PM
On Thu, 03 Apr 2008 10:41:08 -0600, msg <msg@_cybertheque.org_> wrote:

>Adam Chapman wrote:
>
>> A helical antenna looks the best as it is the most isotropic,
>> although it is also pretty big for our aircraft. However even that has
>> +-45 degree 3dB lobe, so i might have to have the aircraft bank more
>> when closer to the centre of its circular mission area.
>
>How about taking a lesson from radiosonde antennas; I don't know about
>current models, but for a very long time the antenna at 1680 MHz was
>a 1/4 wave stub mounted at the center of an inverted conical reflector
>(about 60 degrees). When suspended from the balloon, the antenna points
>downward and illuminates the ground with a pattern formed from that
>of the stub combined with the reflector, giving horizon-wide coverage
>with no nulls for high inclination either.

Radiosondes have two basic assumptions. The transmitter is always
above the receiver and that antenna is always roughly vertically
oriented. Those assumptions don't work well with an RC airplane,
which can easily be inverted, at odd angles, and at the same elevation
as the controller. Incidentally, the usual RC screwup is losing
control on landing, where the ground (small hills, cars, buildings)
block the signal and the radio link loses control.

>Also, you could mount two such antennas on the fuselage, at 180 degrees
>top and bottom and use a splitter to drive them so that there is
>coverage during maneuvers.

Yep. That's what I'm thinking. Two cardioid patch antennas make a
fair hemispherical pattern. The receive sensitivity won't be affected
much by the splitter, but the xmit power will be equally divided
between the two antennas. Link calcs in each direction will need to
be independently calculated.

--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Jeff Liebermann
04-03-08, 12:26 PM
On Thu, 03 Apr 2008 10:43:07 -0600, msg <msg@_cybertheque.org_> wrote:

>Jeff Liebermann wrote:
>
><snip>
>
>Sounds like a heck of a project you did ;)

Not "my" project. I originally got involved in bailing out a group of
arrogant engineers by cleaning up the antenna design. That grew
rapidly into cleaning up the control system, data link, and tracking
mechanism. The basic design was already done when I arrived. I just
made it all work.

One of the bad habits found in engineering is stopping when there is a
problem. These guys were running in circles around the problems
without actually attacking the problem. One had spent about a month
doing a computer simulation of a fundamentally flawed control system.
So, I got to jump in with both feet, turn the muddy waters into
quicksand, and kick those involved into action. Once the sacred cows
were slaughtered and sacrificed, and those involved were willing to
question their own assumptions, progress was rapid. I wasn't the only
one spending sleepless nights on this project.

>In days gone by, I had feeds from the MPQ-10A mortar tracking radar as well
>as the GMD-1 ground station.

2.7GHz, as I recall. Nice:
<http://www.primeportal.net/artillery/don/mpq10a_walk_1.htm>
It should work well at 2.4GHz for a Wi-Fi "shoot out". Got any more
sitting around?


--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

DTC
04-03-08, 04:57 PM
Jeff Liebermann wrote:
> Radiosondes have two basic assumptions. The transmitter is always
> above the receiver and that antenna is always roughly vertically
> oriented.

Not so. The weather reporting radiosonde used in the 60s and 70s
aways drifted off to the horizon, driven by the prevailing wind.

The radio receiver was something like a ten foot diameter dish
under a plastic radome that tracked the radiosonde.

As i recall, they had a single tube transmitter that was frequency
modulated by the thermometer and air pressure was sensed by a
bellows that was mechanically connected to a rheostat that pulse
proportional modulated the signal. All this is going back over
thirty years ago when I played with the project.

Direction was resolved by the rotation coordinates of the dish.

Adam Chapman
04-03-08, 06:13 PM
On Apr 3, 6:02*pm, Jeff Liebermann <je...@cruzio.com> wrote:
> On Thu, 3 Apr 2008 08:12:00 -0700 (PDT), Adam Chapman
>
> <adam.chap...@student.manchester.ac.uk> wrote:
> >I do hope that Im not appearing to be lazy here, I've never had to
> >work in wireless systems as I am an aerodynamisist recently turned
> >image processing geek.
>
> If this is a skool project, you are expected to do your own research
> and calculations. *You don't learn much by having us doing it for you.
> If you have a *SPECIFIC* problem with some aspect of the project, feel
> free to ask. *However, the basic design in implimentation should be
> all your work.
>
> >I was expecting my question to have an easy
> >answer but as usual here wasn't so sorry and thank you for taking your
> >own time to help me.
>
> Welcome to engineering at its worst. *The URL I listed:
> <http://wireless.wikia.com/wiki/Wi-Fi#Link_Calculations>
> shows a simplified wireless link calculation. *You are apparently
> using 802.11b/g instead of VSB modulation for video (and data) so the
> examples should work. *Use this as the basis for your calcs. *However,
> please note that the resultant numbers are the *BEST* case
> calculations. *Real world situations only make things worse. *For
> example, the 2dBi gain of the antenna is actually the MAXIMUM gain of
> the antenna. *If you are off axis from the antenna pattern, the gain
> (and range) will be less. *Receiver sensitivity is never as good as
> they claim. *This should help:
> <http://802.11junk.com/jeffl/rx-sens/receiver%20sensitivity.htm>
> You will need to specify the other end of your radio link (equipment,
> coax, antenna) before I can help with the numbers. *It's really quite
> simple.
>
> >The camera is a Trendnet IP-TV301W model, specification at
> >http://trendnet.com/products/proddetail.asp?prod=110_TV-IP301W&cat=48..
>
> OK. *802.11b/g. * Draws 7 watts (at 12V) and weighs 0.5kg. *Do you
> have any idea of how big a battery will need to be to supply 7 watts
> of continuous power? *4ea LIPO cell phone batteries will do the trick,
> but you'll only have perhaps 15 minutes of operating time. *To lift
> all this, the plane doesn't need to be huge, but it certainly won't be
> a styrofoam fly weight.
>
> >I don't physically have the camera yet because the university is still
> >processing the account, which in my experience can take a very long
> >time.
>
> The local university owes me some money from a project that is now 22
> years overdue. *Good luck.
>
> >If it is possible to estimate transmission ranges with certain
> >equipment, I would have an easier job justifying the expense to my
> >department.
>
> I don't see the connection between range and funding. *I suspect
> you'll do better with a feasability and finance study, than with a
> performance estimate.
>
> >The 2dBi antenna can be replaced, if there is a better type then I am
> >happy to use it. I would prefer an antenna that emits as spherical a
> >pattern as possible, otherwise we will have to limit aircraft
> >maneuevres.
>
> It really depends on what material you make the airplane from. *If RF
> transparent, a simple monopole (as supplied) will work. *When
> inverted, the signal will go through the aircraft. *However, if you
> make the airplane from reflective or absorptive materials, you'll need
> multiple antennas. *One 1/4 wave monopole on top and one on the bottom
> should be sufficient. *With a small ground plane under the antenna,
> the pattern should be mostly cardioid shaped. *If you have a wind
> resistance problem, you can use a small ceramic patch antenna instead.
> It is possible to model the antenna pattern using (free) antenna
> modeling tools:
> <http://home.ict.nl/~arivoors/>
> The sample files include a 747 aircraft (which will need to be
> scaled). *This is not exactly a trivial exercise, a potential time
> burning diversion, and is probably overkill for what you're
> attempting. *It's easy enough to construct a model, install the
> camera, and just play with the position of the antenna until it's a
> workable compromise. *
>
> Note that installing a bigger antenna on the airplane is not what you
> want. *Bigger antennas imply more gain. *You want hemispherical
> antenna pattern which requires a smaller or simpler antenna. *Here's
> an approximation of an isotropic antenna:
> <http://802.11junk.com/jeffl/antennas/isotropic/index.html>
> I think this is overkill and probably not practical on an airplane. If
> you're going to do work with the antenna, do it on the ground station.
>
> >I have just recieved more information from the competition organisers,
> >who tell me that all flying will be within 500m of the launch site and
> >flying is not permitted above 400ft (121.92 m) This gives the max.
> >Euclidean distance as 514.65m, although i expect we will be measuring
> >that distance on the aircraft with GPS, and with the civilian accuracy
> >limitations on GPS (~ +- 12m), I would suggest a target transmission
> >distance of 550m.
>
> >Due to safety rules we will never lose sight of the aircraft so we can
> >expect no obstacles between transmitter and reciever.
>
> If this is competition flying, it's also likely (but not guaranteed)
> that you'll also get minimal 2.4GHz interference. *It's traditional to
> confiscate the controllers during the competition to prevent sabotage.
> That works because RC has dedicated frequencies. *That's NOT the case
> with 2.4GHz, where there's a chance that some lunatic, with a portable
> wireless router, can crash your flying machine. *You can also lose
> control when the aircraft flys over a coffee shop wireless hot spot.
> (I'm assuming that you'll be controlling on 2.4GHz). *I sure hope this
> is going to be run in the middle of nowhere. *Testing it at skool,
> where there's lots of 2.4GHz interference, is going to be ummm...
> interesting.
>
> >I've been slowly writing this reply all day, and thinking a lot as i
> >do so.
>
> At the beginning of any project, leaps of faith, wild ideas, the
> shooting from the hip, are standard procedure. *Once the ideas are
> ossified, then you can be more studious and careful.
>
> >A helical antenna looks the best as it is the most isotropic,
>
> Too big, too much drag, too critical, poor gain to volume ratio. *If
> you're going to go to that level of complexity, think about two
> ceramic patch antennas with a -3dB Wilkinson combiner. *Two back to
> back cartoid patterns equal roughly a hemisphere.
>
> >although it is also pretty big for our aircraft. However even that has
> >+-45 degree 3dB lobe, so i might have to have the aircraft bank more
> >when closer to the centre of its circular mission area.
>
> Translation. *If you have a hole in the antenna pattern, you're going
> to lose control. *A helix also has no gain in the reverse direction,
> so you'll need two. *Hint: *Look at what's on a real airplane or
> guided missile.
>
> >The link calculator link from LR told me that if i put an antenna with
> >a 7dBi gain (like the one at _) on the UAV and at the ground station,
> >i will have a range of over 7km with a 100mW transmitter, which I
> >assume can be achieved using an in-line amplifier. Does this
> >calculator use the range in the direction of the strongest energy
> >emisiion from the antenna?
>
> I'm still missing some numbers. *See the example calcs on the URL I
> mentioned. *Copy the table of numbers and plug in YOUR numbers.
> Remember, those are the BEST case calculations. *Reality only makes it
> worse.
>
> >I guess that more than on antenna on the same aircraft with different
> >orientations would interfere with each others signals?
>
> If the two antennas cannot "see" each other, they will not mangle the
> combined pattern. *What other signals?
>
> Good luck. *Sounds like an interesting project. *Watch your weight and
> power consumption. *
> --
> Jeff Liebermann * * je...@cruzio.com
> 150 Felker St #D * *http://www.LearnByDestroying.com
> Santa Cruz CA 95060http://802.11junk.com
> Skype: JeffLiebermann * * AE6KS * *831-336-2558

I have found a commercially available patch antenna, here at
http://www.cisco.com/en/US/docs/wireless/antenna/installation/guide/ant2485.html
and a groundstation antenna at http://www.radiolabs.com/products/wireless/waverv2.php
.
Now the cisco site I linked for the patch antenna does not say what
the Tx power is, althoug at a another site (http://www.cisco.com/en/US/
prod/collateral/wireless/ps7183/ps469/
product_data_sheet09186a008008883b.html) I read that the Cisco Aironet
2.4 GHz Bridge transmitter power is 20 dBm. I cant seem to find a
defintion for a bridge transmitter though, so im not sure if i know
the Tx power of the patch antenna or not. Does 20 dBm seem reasonable
for this antenna?


If the patch antenna power is 20dBm (which is the legal maximum in the
UK), then over a 550m range the SOM should be just below 28 dB
according to my calculations. Although I was aiming for 28 or above
the PER does not appear to be too bad.

Thanks again
Adam

msg
04-03-08, 09:34 PM
DTC wrote:

> Jeff Liebermann wrote:
>
>> Radiosondes have two basic assumptions. The transmitter is always
>> above the receiver and that antenna is always roughly vertically
>> oriented.
>
> Not so. The weather reporting radiosonde used in the 60s and 70s
> aways drifted off to the horizon, driven by the prevailing wind.

Indeed, the antenna was vertical, but the pattern permitted
horizon to horizon coverage but without nulls overhead.

>
> The radio receiver was something like a ten foot diameter dish
> under a plastic radome that tracked the radiosonde.

The GMD-1 dish often was exposed and not in a radome.

>
> As i recall, they had a single tube transmitter that was frequency
> modulated by the thermometer and air pressure was sensed by a
> bellows that was mechanically connected to a rheostat that pulse
> proportional modulated the signal.

There was a baroswitch which selected the sensor to read based on
increments in altitude; the modulation was PRR. The oscillator
was a pencil triode in a cavity resonating at 1680 MHz and tuned
with a screw probe.

Michael

Jeff Liebermann
04-04-08, 10:17 AM
On Thu, 3 Apr 2008 16:13:32 -0700 (PDT), Adam Chapman
<adam.chapman@student.manchester.ac.uk> wrote:

>I read that the Cisco Aironet
>2.4 GHz Bridge transmitter power is 20 dBm. I cant seem to find a
>defintion for a bridge transmitter though, so im not sure if i know
>the Tx power of the patch antenna or not. Does 20 dBm seem reasonable
>for this antenna?

The power rating of an antenna is the point where the antenna can be
safely operated without melting, arcing, or self-destructing. There's
no way an antenna is going to do any of these at 20dBm (100
milliwatts).

>If the patch antenna power is 20dBm (which is the legal maximum in the
>UK), then over a 550m range the SOM should be just below 28 dB
>according to my calculations. Although I was aiming for 28 or above
>the PER does not appear to be too bad.

Antenna power handling ability has nothing to do with range.

Any antenna that has the proper gain, size, and weight characteristics
will work. It's up to you to determine the requirements.

--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Jeff Liebermann
04-04-08, 10:29 AM
On Thu, 03 Apr 2008 16:57:55 -0500, DTC <me@nothingtoseehere.zzx>
wrote:

>Jeff Liebermann wrote:
>> Radiosondes have two basic assumptions. The transmitter is always
>> above the receiver and that antenna is always roughly vertically
>> oriented.
>
>Not so. The weather reporting radiosonde used in the 60s and 70s
>aways drifted off to the horizon, driven by the prevailing wind.

Sure. Assuming a flat earth model, the tracking antenna cannot be
expected to dip much below horizontal. My point is that there was no
need for any RF to be radiated by the radiosonde ABOVE the horizon or
in an upwards direction. You can't make that assumption with a model
airplane, which can fly upside down, and does require RF to be
radiated in the upward direction.

>The radio receiver was something like a ten foot diameter dish
>under a plastic radome that tracked the radiosonde.

I have one in the closet somewhere. I'll dig it out and take some
photos (later). It's amazingly crude and cheaply made.

>As i recall, they had a single tube transmitter that was frequency
>modulated by the thermometer and air pressure was sensed by a
>bellows that was mechanically connected to a rheostat that pulse
>proportional modulated the signal. All this is going back over
>thirty years ago when I played with the project.

Tube? Thermionic Valve? Whazzat? Dis is da space age. We don't do
no stinkin tubes.

>Direction was resolved by the rotation coordinates of the dish.


--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

msg
04-04-08, 02:35 PM
Jeff Liebermann wrote:

<snip>
>
> Tube? Thermionic Valve? Whazzat? Dis is da space age. We don't do
> no stinkin tubes.

Ground support computers for the first few Redstone/Mercury launches
included Bendix G-15 vacuum tube machines; most of the hardware
was in fact tube-based. IIRC, the Mercury capsule even contained
some tubes.

Michael

DTC
04-04-08, 06:12 PM
msg wrote:
> There was a baroswitch which selected the sensor to read based on
> increments in altitude; the modulation was PRR.

That's the phrase I was trying to remember, Pulse Repletion Rate.

P.Schuman
04-05-08, 10:49 AM
Adam Chapman wrote:>
> The camera is a Trendnet IP-TV301W model, specification at
> http://trendnet.com/products/proddetail.asp?prod=110_TV-IP301W&cat=48.

It's been interesting reading....
but was wondering why you went with a larger heavy camera
vs the tiny 2.4Ghz spycams you see advertised.

Do you only need a webcam for a visual view,
OR
is this part of the onboard autonomous navigation
and you need a "good" camera image to use for processing ?

P.Schuman
04-05-08, 10:55 AM
Adam Chapman wrote:
> Thanks to all of you, I replied to this particular post because it had
> the most questions for me to answer.
> I do hope that Im not appearing to be lazy here, I've never had to
> work in wireless systems as I am an aerodynamisist recently turned
> image processing geek. I was expecting my question to have an easy
> answer but as usual here wasn't so sorry and thank you for taking your
> own time to help me.
>
> On Apr 2, 11:54 pm, Jeff Liebermann <je...@cruzio.com> wrote:
>> On Wed, 2 Apr 2008 08:00:06 -0700 (PDT), Adam Chapman
>>
>> <adam.chap...@student.manchester.ac.uk> wrote:
>>> I've ordered a wireless webcam that will be out onboard an
>>> autonomous aircraft for a university project.
>>
>> I've done that. You're about to have a battery problem.
>>
>>> The camera has a "2dBi Diversity Antenna" and transmits video over
>>> an IP protocol in the 2.4-2.4835 GHz range.
>>
>> Duz this camera have a manufacturer and a model number? A URL with
>> the specifications would be nice.
>>
>
> The camera is a Trendnet IP-TV301W model, specification at
> http://trendnet.com/products/proddetail.asp?prod=110_TV-IP301W&cat=48.
> I don't physically have the camera yet because the university is still
> processing the account, which in my experience can take a very long
> time. If it is possible to estimate transmission ranges with certain
> equipment, I would have an easier job justifying the expense to my
> department.
>
>> 2dBi is a simple monopole antenna. It has a radiation pattern that
>> looks like a donut. There's a big null when in line with the
>> antenna, such as when the aircraft is flying directly overhead.
>>
>
> The 2dBi antenna can be replaced, if there is a better type then I am
> happy to use it. I would prefer an antenna that emits as spherical a
> pattern as possible, otherwise we will have to limit aircraft
> maneuevres.
>
>
>
>>> Now the aircraft needs to operate in a 500x500m area, although I
>>> expect that my ground station computer will be placed some distance
>>> away from the operating area.
>>
>> I have this thing about numbers. How far is "some distance"? How
>> high does the air thing have to fly? What's the MAXIMUM distance you
>> expect to see an image?
>>
>
> I have just recieved more information from the competition organisers,
> who tell me that all flying will be within 500m of the launch site and
> flying is not permitted above 400ft (121.92 m) This gives the max.
> Euclidean distance as 514.65m, although i expect we will be measuring
> that distance on the aircraft with GPS, and with the civilian accuracy
> limitations on GPS (~ +- 12m), I would suggest a target transmission
> distance of 550m.
>
> Due to safety rules we will never lose sight of the aircraft so we can
> expect no obstacles between transmitter and reciever.
>
> I've been slowly writing this reply all day, and thinking a lot as i
> do so. A helical antenna looks the best as it is the most isotropic,
> although it is also pretty big for our aircraft. However even that has
> +-45 degree 3dB lobe, so i might have to have the aircraft bank more
> when closer to the centre of its circular mission area.
>
> The link calculator link from LR told me that if i put an antenna with
> a 7dBi gain (like the one at _) on the UAV and at the ground station,
> i will have a range of over 7km with a 100mW transmitter, which I
> assume can be achieved using an in-line amplifier. Does this
> calculator use the range in the direction of the strongest energy
> emisiion from the antenna?
>
> I guess that more than on antenna on the same aircraft with different
> orientations would interfere with each others signals?
>

how big is the plane ?
do you have a UK link for the project, or photos ?

Jeff Liebermann
04-05-08, 11:40 AM
On Sat, 5 Apr 2008 10:55:27 -0500, "P.Schuman"
<pschuman_no_spam_me@interserv.com> wrote:

>do you have a UK link for the project, or photos ?

Well, he attends the University of Manchester:
<http://www.manchester.ac.uk>
There might be something on the web site. I'm too lazy to search.

Duh. I just noticed that the title is for an "autonomous" aircraft.
That means no direct remote control with lots of computing horsepower
in the air to avoid obstacles (buildings) and to navigate. This is
gonna be a big and heavy aircraft or the computing is going to be done
on the ground via a wireless link. (Note: I'm guessing).

As I previously mentioned, it's considered good form for students to
do their own homework and research. Most universities have adequate
resources for answering basic path loss and antenna questions.

Hmmm... this is probably not the same project, but the university does
have some interesting projects:
<http://personalpages.manchester.ac.uk/staff/Dominic.Diston/MyTeaching/MACE31501/Aircraft%20Group%20Design%202007%20v2.1.pdf>



--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

P.Schuman
04-05-08, 12:00 PM
"Jeff Liebermann" <jeffl@cruzio.com> wrote in message
news:63afv35jtmah1u13vkuj3o0bp2p6ogv8eh@4ax.com...
> On Sat, 5 Apr 2008 10:55:27 -0500, "P.Schuman"
> <pschuman_no_spam_me@interserv.com> wrote:
>
>>do you have a UK link for the project, or photos ?
>
> Well, he attends the University of Manchester:
> <http://www.manchester.ac.uk>
> There might be something on the web site. I'm too lazy to search.
>
> Duh. I just noticed that the title is for an "autonomous" aircraft.
> That means no direct remote control with lots of computing horsepower
> in the air to avoid obstacles (buildings) and to navigate. This is
> gonna be a big and heavy aircraft or the computing is going to be done
> on the ground via a wireless link. (Note: I'm guessing).
>
> As I previously mentioned, it's considered good form for students to
> do their own homework and research. Most universities have adequate
> resources for answering basic path loss and antenna questions.
>
> Hmmm... this is probably not the same project, but the university does
> have some interesting projects:
> <http://personalpages.manchester.ac.uk/staff/Dominic.Diston/MyTeaching/MACE31501/Aircraft%20Group%20Design%202007%20v2.1.pdf>
>
>
I've seen on either TLC or Discovery the DARPA competition for the land
based vehicles.
Pretty amazing.....
http://en.wikipedia.org/wiki/DARPA_Grand_Challenge

Adam Chapman
04-05-08, 03:19 PM
On Apr 5, 4:49*pm, "P.Schuman" <pschuman_no_spam...@interserv.com>
wrote:
> Adam Chapman wrote:>
> > The camera is a Trendnet IP-TV301W model, specification at
> >http://trendnet.com/products/proddetail.asp?prod=110_TV-IP301W&cat=48..
>
> It's been interesting reading....
> but was wondering why you went with a larger heavy camera
> vs the tiny 2.4Ghz spycams you see advertised.
>
> Do you only need a webcam for a visual view,
> OR
> is this part of the onboard autonomous navigation
> and you need a "good" camera image to use for processing ?

I wanted a digital image to be processed in Matlab, which can aquire
images over an IP protocol, and is excellent software for image
processing because it is built around matrix algebra.
Also I wanted a camera that output images in digital 'pixel' format
because converiing a frame from line-tracing cameras into digital
format is very slow. In another project I am working on for the
(http://www.challenge.mod.uk/) we tried using an analogue camera on an
RC helicopter and the quality was awful- particularly we had problems
with colour synchronisation and ghosting.
The aircraft in this project is for the ESO-UAS competition (http://
www.hertfordshire-connect.com/default.asp?ContentID=1234) where we are
legally allowed a UAV weighing up to 20kg! We can't make it this heavy
though because the wingspan would be ridiculous.
I wanted to put a mini-itx computer onboard to do the processing but
the weight (~2kg) was not justified.
Basically I chose a camera that had a good digital quality and allowed
tuning of parameters like frame rate, exposure time etc. I also liked
the fact that it had a serial output, which could perhaps be used on a
future aircraft to drive servos, allowing a pan/tilt system or even
the aircraft flight control system!
I'm only working towards an image recognition capability at the moment
so that's why I chose a good quality camera. I can probably take the
casing off to save some weight too. Next year I would like to develop
a visual gudance capability, which would be really cool and far more
accurate than GPS.


I think I have my antenna setup sorted- although I have another
question: If I use a 2-way splitter on the aircraft to feed the 2
antennas, will there be any problems with recieving a signal from the
ground station? What I mean is; is a splitter a bi-directioinal
component or does it only allow a signal to travel in one direction?

Adam

Adam Chapman
04-05-08, 03:29 PM
On Apr 5, 9:19*pm, Adam Chapman
<adam.chap...@student.manchester.ac.uk> wrote:
> On Apr 5, 4:49*pm, "P.Schuman" <pschuman_no_spam...@interserv.com>
> wrote:
>
> > Adam Chapman wrote:>
> > > The camera is a Trendnet IP-TV301W model, specification at
> > >http://trendnet.com/products/proddetail.asp?prod=110_TV-IP301W&cat=48.
>
> > It's been interesting reading....
> > but was wondering why you went with a larger heavy camera
> > vs the tiny 2.4Ghz spycams you see advertised.
>
> > Do you only need a webcam for a visual view,
> > OR
> > is this part of the onboard autonomous navigation
> > and you need a "good" camera image to use for processing ?
>
> I wanted a digital image to be processed in Matlab, which can aquire
> images over an IP protocol, and is excellent software for image
> processing because it is built around matrix algebra.
> Also I wanted a camera that output images in digital 'pixel' format
> because converiing a frame from line-tracing cameras into digital
> format is very slow. In another project I am working on for the
> (http://www.challenge.mod.uk/) we tried using an analogue camera on an
> RC helicopter and the quality was awful- particularly we had problems
> with colour synchronisation and ghosting.
> The aircraft in this project is for the ESO-UAS competition (http://www.hertfordshire-connect.com/default.asp?ContentID=1234) where we are
> legally allowed a UAV weighing up to 20kg! We can't make it this heavy
> though because the wingspan would be ridiculous.
> I wanted to put a mini-itx computer onboard to do the processing but
> the weight (~2kg) was not justified.
> Basically I chose a camera that had a good digital quality and allowed
> tuning of parameters like frame rate, exposure time etc. I also liked
> the fact that it had a serial output, which could perhaps be used on a
> future aircraft to drive servos, allowing a pan/tilt system or even
> the aircraft flight control system!
> I'm only working towards an image recognition capability at the moment
> so that's why I chose a good quality camera. I can probably take the
> casing off to save some weight too. Next year I would like to develop
> a visual gudance capability, which would be really cool and far more
> accurate than GPS.
>
> I think I have my antenna setup sorted- although I have another
> question: If I use a 2-way splitter on the aircraft to feed the 2
> antennas, will there be any problems with recieving a signal from the
> ground station? What I mean is; is a splitter a bi-directioinal
> component or does it only allow a signal to travel in one direction?
>
> Adam

I think our aircraft mass is 12kg, 6m wingspan. We won't release
photos until the competition is held in June/July so other teams can't
steal our ideas. I'd be happy to fill you all in with info then
though.
I am also working on the air-to-air refuelling project but it's
nothing to do with the aircraft this thread is about.

Adam

P.Schuman
04-05-08, 05:02 PM
here is the link for the Site -
http://www.hertfordshire-connect.com/default.asp?ContentID=1234
Rules -
http://www.hertfordshire-connect.com/default.asp?ContentID=1239

Adam Chapman
04-05-08, 06:36 PM
On Apr 5, 11:02*pm, "P.Schuman" <pschuman_no_spam...@interserv.com>
wrote:
> here is the link for the Site -http://www.hertfordshire-connect.com/default.asp?ContentID=1234
> Rules -http://www.hertfordshire-connect.com/default.asp?ContentID=1239

It's a new competition this year but is planned to be an annual event
in the future. Hopefully lots of new and exciting things to come!

Jeff Liebermann
04-06-08, 11:24 AM
On Sat, 05 Apr 2008 17:00:32 GMT, "P.Schuman"
<pschuman_no_spam_me@interserv.com> wrote:

>I've seen on either TLC or Discovery the DARPA competition for the land
>based vehicles.
>Pretty amazing.....
>http://en.wikipedia.org/wiki/DARPA_Grand_Challenge

Just add wings? A flying Hummer? There are fewer obstacles in the
air than on the ground. They're also easier to see. However, weight
doesn't seem to be an issue with the ground vehicles, while it's
everything in the air. Also, the not so minor problem of hands off
approach and landing. Doing all the computing on the ground (with a
connecting wireless radio link) will certainly help the weight load,
but there are still the sensors and datacomm hardware that need to be
in the airplane. If my guessing is correct, this sounds like a real
challenge.


--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Adam Chapman
04-06-08, 11:40 AM
On Apr 6, 5:24*pm, Jeff Liebermann <je...@cruzio.com> wrote:
> On Sat, 05 Apr 2008 17:00:32 GMT, "P.Schuman"
>
> <pschuman_no_spam...@interserv.com> wrote:
> >I've seen on either TLC or Discovery the DARPA competition for the land
> >based vehicles.
> >Pretty amazing.....
> >http://en.wikipedia.org/wiki/DARPA_Grand_Challenge
>
> Just add wings? *A flying Hummer? *There are fewer obstacles in the
> air than on the ground. *They're also easier to see. *However, weight
> doesn't seem to be an issue with the ground vehicles, while it's
> everything in the air. *Also, the not so minor problem of hands off
> approach and landing. *Doing all the computing on the ground (with a
> connecting wireless radio link) will certainly help the weight load,
> but there are still the sensors and datacomm hardware that need to be
> in the airplane. *If my guessing is correct, this sounds like a real
> challenge.
>
> --
> Jeff Liebermann * * je...@cruzio.com
> 150 Felker St #D * *http://www.LearnByDestroying.com
> Santa Cruz CA 95060http://802.11junk.com
> Skype: JeffLiebermann * * AE6KS * *831-336-2558

indeed!

Jeff Liebermann
04-06-08, 12:26 PM
On Sat, 5 Apr 2008 13:19:37 -0700 (PDT), Adam Chapman
<adam.chapman@student.manchester.ac.uk> wrote:

>I wanted a digital image to be processed in Matlab, which can aquire
>images over an IP protocol, and is excellent software for image
>processing because it is built around matrix algebra.

OK, so you're doing your computing in the airplane, not on the ground.

>Also I wanted a camera that output images in digital 'pixel' format
>because converiing a frame from line-tracing cameras into digital
>format is very slow.

There are many different types of cameras available. Most (not all)
have a "raw" format which outputs a bit mapped image. This is the
fastest mode because it involves no post processing (compression,
image enhancement, image stabilization, or color correction). You
won't find this mode in the cheap cameras, which output JPG's
directly. However, since you have a flying Matlab computah, you
probably have the horsepower to do your own rasterization in the
computah, which means you can use a cheap USB camera.

>In another project I am working on for the
>(http://www.challenge.mod.uk/) we tried using an analogue camera on an
>RC helicopter and the quality was awful- particularly we had problems
>with colour synchronisation and ghosting.

There's only one way that I know to get ghosting. That's if you use
NTSC/PAL video, and you have reflections on the interconnect cables
due to mismatched impedance terminations. Color smear is caused by
bad group delay on the same interconnect cables, where the different
frequency components of the video signal, have different (or changing)
phase shifts.

However, you're correct about analog cameras not being the right
choice. NSTC video only needs about 320x240 resolution to be useable.
No need for more dots for TV. However, if you're going to split
pixels with Matlab, you need all the dots on the screen as possible. I
suggest you look into using a hacked digital camera with a really big
CCD sensor.

>The aircraft in this project is for the ESO-UAS competition (http://
>www.hertfordshire-connect.com/default.asp?ContentID=1234) where we are
>legally allowed a UAV weighing up to 20kg! We can't make it this heavy
>though because the wingspan would be ridiculous.

"Teams will be required to demonstrate four aspects of autonomous
flight, including image recognition, a pylon race against time,
aspects of navigation and being able to land the unmanned aircraft at
a designated spot."

Ouch. That's not going to be easy. I would be tempted to suggest big
and slow, so as to give the processors more time to make decisions.

>I wanted to put a mini-itx computer onboard to do the processing but
>the weight (~2kg) was not justified.

It's not just the weight of the board and cooling system, it's also
the weight of the power system. How many minutes/hours flight
duration is expected? I couldn't find anything in the terse press
release and the almost blank "Preliminary Design Specifications"?

>Basically I chose a camera that had a good digital quality and allowed
>tuning of parameters like frame rate, exposure time etc. I also liked
>the fact that it had a serial output, which could perhaps be used on a
>future aircraft to drive servos, allowing a pan/tilt system or even
>the aircraft flight control system!

Good logic but I wouldn't have done it quite the same way. If you're
going to have a mess of independent sensors, that all communicated to
a mess of dedicated computahs, each with their own particular
function, then the camera has to fit into the network communications
topology. If you're using TCP/IP for onboard communications, you
might want to look into an IP network camera.

>I'm only working towards an image recognition capability at the moment
>so that's why I chose a good quality camera. I can probably take the
>casing off to save some weight too. Next year I would like to develop
>a visual gudance capability, which would be really cool and far more
>accurate than GPS.

The ground based autonomous vehicles have the advantage of really only
needing image recognition in 2D. In the air, you have to recognize
and locate objects in 3D, which literally adds a new dimension to the
problem.

I guessed that you were going to be flying this year. If you're just
building parts of the puzzle on the ground this year, there's no
reason to economize on weight and power at this point. Might was well
go with your selected camera.

>I think I have my antenna setup sorted- although I have another
>question: If I use a 2-way splitter on the aircraft to feed the 2
>antennas, will there be any problems with recieving a signal from the
>ground station? What I mean is; is a splitter a bi-directioinal
>component or does it only allow a signal to travel in one direction?

Excellent question and good thinking. A splitter (also known as a
Wilkinson power divider) divides the signal in two. It's totally
passive and bi-directional. At 2.4GHz, it's just two 1/4 wave strip
lines and a 100 ohm resistor. See diagram at:
<http://www.microwaves101.com/encyclopedia/Wilkinson_splitters.cfm#twoport>
<http://www.qsl.net/yu1aw/2G4spliter2.gif>
<http://www.qsl.net/yu1aw/2G4spliterN.gif>
Let's assume that there's no internal losses for now. A 2.4GHz signal
applied to the input/output port, will be divided equally between the
two other ports. That means that each antenna will get half the RF
power or 3dB. 3dB is NOT a large drop in signal level. For example,
the rule of thumb for range versus power is a 6dB drop in power equals
a 50% drop in range. 3dB loss is equal to about 0.7 times the range.
If possible, simply increase the transmit power 3dB (double the
milliwatts) and you'll be exactly the same as before.

The situation is also good in receive. There is no loss between
either antenna port and the input/output port. Every dB received by
either antenna is delivered directly to the input/output port.

There is also considerably isolation (10-30dB) between antenna ports.
This is a huge help in keeping the antenna termination impedance near
the desired 50 ohms, thus keeping the VSWR and losses low. If you can
tolerate minimal isolation, the 100 ohm resistor can be eliminated.

However, there are some potential problems. RF radiating from both
antennas simultaneously will cancel and add at various points. This
will create nulls and peaks in the antenna pattern. The basic idea is
that each antenna should not "see" each other. More simply, there
should be no position, where you can see both antennas at the same
time. That's not going to be easy with the donut shaped pattern of a
common vertical 1/4 wave antenna. The patterns will overlap and
possibly cancel. However, it is possible with cardioid patterns, such
as patch or panel antennas.

There's also the question of where to mount the antennas. On top and
on the belly are the obvious locations. However, I think you'll get
better hemispherical coverage with one antenna mounted in the nose,
pointing forward, and another out the tail pointing aft. This allows
the body to provide some shielding between antennas. There may be
some signal loss dead ahead, which will be critical during landing.
However, that will also be the point of closest approach so distances
will be minimal.

Since you're not flying this year, I suggest you start with just one
antenna. Play with it in a mockup or model and see what it does. Hang
the model from a string in a tree, and measure the received signal
strength. That will determine if two antenna are necessary.

Also, I suggest you consider separating the video, data, and control
links. Control can be done at much lower frequencies (27, 72MHz)
using conventional RC hardware. This has the advantage of being
fairly orientation insensitive. I don't think you have the 900MHz
band in UK, so that's out. 2.4GHz is it for data and video, but I
suspect that the video will saturate the link leaving no time slots
for data. If there's some other frequencies available in UK for
video, I would use it.

Gotta run... good luck.

--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

LR
04-07-08, 12:17 PM
Jeff Liebermann wrote:
> On Sat, 5 Apr 2008 13:19:37 -0700 (PDT), Adam Chapman
> <adam.chapman@student.manchester.ac.uk> wrote:
>
>> I wanted a digital image to be processed in Matlab, which can aquire
>> images over an IP protocol, and is excellent software for image
>> processing because it is built around matrix algebra.
>
> OK, so you're doing your computing in the airplane, not on the ground.
>
>> Also I wanted a camera that output images in digital 'pixel' format
>> because converiing a frame from line-tracing cameras into digital
>> format is very slow.
>
> There are many different types of cameras available. Most (not all)
> have a "raw" format which outputs a bit mapped image. This is the
> fastest mode because it involves no post processing (compression,
> image enhancement, image stabilization, or color correction). You
> won't find this mode in the cheap cameras, which output JPG's
> directly. However, since you have a flying Matlab computah, you
> probably have the horsepower to do your own rasterization in the
> computah, which means you can use a cheap USB camera.
>
>> In another project I am working on for the
>> (http://www.challenge.mod.uk/) we tried using an analogue camera on an
>> RC helicopter and the quality was awful- particularly we had problems
>> with colour synchronisation and ghosting.
>
> There's only one way that I know to get ghosting. That's if you use
> NTSC/PAL video, and you have reflections on the interconnect cables
> due to mismatched impedance terminations. Color smear is caused by
> bad group delay on the same interconnect cables, where the different
> frequency components of the video signal, have different (or changing)
> phase shifts.
>
> However, you're correct about analog cameras not being the right
> choice. NSTC video only needs about 320x240 resolution to be useable.
> No need for more dots for TV. However, if you're going to split
> pixels with Matlab, you need all the dots on the screen as possible. I
> suggest you look into using a hacked digital camera with a really big
> CCD sensor.
>
>> The aircraft in this project is for the ESO-UAS competition (http://
>> www.hertfordshire-connect.com/default.asp?ContentID=1234) where we are
>> legally allowed a UAV weighing up to 20kg! We can't make it this heavy
>> though because the wingspan would be ridiculous.
>
> "Teams will be required to demonstrate four aspects of autonomous
> flight, including image recognition, a pylon race against time,
> aspects of navigation and being able to land the unmanned aircraft at
> a designated spot."
>
> Ouch. That's not going to be easy. I would be tempted to suggest big
> and slow, so as to give the processors more time to make decisions.
>
>> I wanted to put a mini-itx computer onboard to do the processing but
>> the weight (~2kg) was not justified.
>
> It's not just the weight of the board and cooling system, it's also
> the weight of the power system. How many minutes/hours flight
> duration is expected? I couldn't find anything in the terse press
> release and the almost blank "Preliminary Design Specifications"?
>
>> Basically I chose a camera that had a good digital quality and allowed
>> tuning of parameters like frame rate, exposure time etc. I also liked
>> the fact that it had a serial output, which could perhaps be used on a
>> future aircraft to drive servos, allowing a pan/tilt system or even
>> the aircraft flight control system!
>
> Good logic but I wouldn't have done it quite the same way. If you're
> going to have a mess of independent sensors, that all communicated to
> a mess of dedicated computahs, each with their own particular
> function, then the camera has to fit into the network communications
> topology. If you're using TCP/IP for onboard communications, you
> might want to look into an IP network camera.
>
>> I'm only working towards an image recognition capability at the moment
>> so that's why I chose a good quality camera. I can probably take the
>> casing off to save some weight too. Next year I would like to develop
>> a visual gudance capability, which would be really cool and far more
>> accurate than GPS.
>
> The ground based autonomous vehicles have the advantage of really only
> needing image recognition in 2D. In the air, you have to recognize
> and locate objects in 3D, which literally adds a new dimension to the
> problem.
>
> I guessed that you were going to be flying this year. If you're just
> building parts of the puzzle on the ground this year, there's no
> reason to economize on weight and power at this point. Might was well
> go with your selected camera.
>
>> I think I have my antenna setup sorted- although I have another
>> question: If I use a 2-way splitter on the aircraft to feed the 2
>> antennas, will there be any problems with recieving a signal from the
>> ground station? What I mean is; is a splitter a bi-directioinal
>> component or does it only allow a signal to travel in one direction?
>
> Excellent question and good thinking. A splitter (also known as a
> Wilkinson power divider) divides the signal in two. It's totally
> passive and bi-directional. At 2.4GHz, it's just two 1/4 wave strip
> lines and a 100 ohm resistor. See diagram at:
> <http://www.microwaves101.com/encyclopedia/Wilkinson_splitters.cfm#twoport>
> <http://www.qsl.net/yu1aw/2G4spliter2.gif>
> <http://www.qsl.net/yu1aw/2G4spliterN.gif>
> Let's assume that there's no internal losses for now. A 2.4GHz signal
> applied to the input/output port, will be divided equally between the
> two other ports. That means that each antenna will get half the RF
> power or 3dB. 3dB is NOT a large drop in signal level. For example,
> the rule of thumb for range versus power is a 6dB drop in power equals
> a 50% drop in range. 3dB loss is equal to about 0.7 times the range.
> If possible, simply increase the transmit power 3dB (double the
> milliwatts) and you'll be exactly the same as before.
>
> The situation is also good in receive. There is no loss between
> either antenna port and the input/output port. Every dB received by
> either antenna is delivered directly to the input/output port.
>
> There is also considerably isolation (10-30dB) between antenna ports.
> This is a huge help in keeping the antenna termination impedance near
> the desired 50 ohms, thus keeping the VSWR and losses low. If you can
> tolerate minimal isolation, the 100 ohm resistor can be eliminated.
>
> However, there are some potential problems. RF radiating from both
> antennas simultaneously will cancel and add at various points. This
> will create nulls and peaks in the antenna pattern. The basic idea is
> that each antenna should not "see" each other. More simply, there
> should be no position, where you can see both antennas at the same
> time. That's not going to be easy with the donut shaped pattern of a
> common vertical 1/4 wave antenna. The patterns will overlap and
> possibly cancel. However, it is possible with cardioid patterns, such
> as patch or panel antennas.
>
> There's also the question of where to mount the antennas. On top and
> on the belly are the obvious locations. However, I think you'll get
> better hemispherical coverage with one antenna mounted in the nose,
> pointing forward, and another out the tail pointing aft. This allows
> the body to provide some shielding between antennas. There may be
> some signal loss dead ahead, which will be critical during landing.
> However, that will also be the point of closest approach so distances
> will be minimal.
>
> Since you're not flying this year, I suggest you start with just one
> antenna. Play with it in a mockup or model and see what it does. Hang
> the model from a string in a tree, and measure the received signal
> strength. That will determine if two antenna are necessary.
>
> Also, I suggest you consider separating the video, data, and control
> links. Control can be done at much lower frequencies (27, 72MHz)
> using conventional RC hardware. This has the advantage of being
> fairly orientation insensitive. I don't think you have the 900MHz
> band in UK, so that's out. 2.4GHz is it for data and video, but I
> suspect that the video will saturate the link leaving no time slots
> for data. If there's some other frequencies available in UK for
> video, I would use it.
>
> Gotta run... good luck.
>
It was interesting to note that the modern 2.4GHz rc controllers
transmit on 2 channels and have 2 receivers, the antennas of which are
dipoles and are usually placed perpendicular to each other.

http://www.spektrumrc.com/ProdInfo/Files/SPM2710_DX7_Manual.pdf

One of the teams involved in the "MOD Challenge" did a bit of work on
imaging and the problems they encountered, worth a read.
Imaging through to motion deblur.
http://www.barnardmicrosystems.com/L4E_imaging.htm

Adam Chapman
04-08-08, 05:49 PM
On Apr 7, 6:17*pm, LR <l...@privacy.net> wrote:
> Jeff Liebermann wrote:
> > On Sat, 5 Apr 2008 13:19:37 -0700 (PDT), Adam Chapman
> > <adam.chap...@student.manchester.ac.uk> wrote:
>
> >> I wanted a digital image to be processed in Matlab, which can aquire
> >> images over an IP protocol, and is excellent software for image
> >> processing because it is built around matrix algebra.
>
> > OK, so you're doing your computing in the airplane, not on the ground.
>
> >> Also I wanted a camera that output images in digital 'pixel' format
> >> because converiing a frame from line-tracing cameras into digital
> >> format is very slow.
>
> > There are many different types of cameras available. *Most (not all)
> > have a "raw" format which outputs a bit mapped image. *This is the
> > fastest mode because it involves no post processing (compression,
> > image enhancement, image stabilization, or color correction). *You
> > won't find this mode in the cheap cameras, which output JPG's
> > directly. *However, since you have a flying Matlab computah, you
> > probably have the horsepower to do your own rasterization in the
> > computah, which means you can use a cheap USB camera.
>
> >> In another project I am working on for the
> >> (http://www.challenge.mod.uk/) we tried using an analogue camera on an
> >> RC helicopter and the quality was awful- particularly we had problems
> >> with colour synchronisation and ghosting.
>
> > There's only one way that I know to get ghosting. *That's if you use
> > NTSC/PAL video, and you have reflections on the interconnect cables
> > due to mismatched impedance terminations. *Color smear is caused by
> > bad group delay on the same interconnect cables, where the different
> > frequency components of the video signal, have different (or changing)
> > phase shifts.
>
> > However, you're correct about analog cameras not being the right
> > choice. *NSTC video only needs about 320x240 resolution to be useable.
> > No need for more dots for TV. *However, if you're going to split
> > pixels with Matlab, you need all the dots on the screen as possible. I
> > suggest you look into using a hacked digital camera with a really big
> > CCD sensor.
>
> >> The aircraft in this project is for the ESO-UAS competition (http://
> >>www.hertfordshire-connect.com/default.asp?ContentID=1234) where we are
> >> legally allowed a UAV weighing up to 20kg! We can't make it this heavy
> >> though because the wingspan would be ridiculous.
>
> > "Teams will be required to demonstrate four aspects of autonomous
> > flight, including image recognition, a pylon race against time,
> > aspects of navigation and being able to land the unmanned aircraft at
> > a designated spot."
>
> > Ouch. *That's not going to be easy. *I would be tempted to suggest big
> > and slow, so as to give the processors more time to make decisions.
>
> >> I wanted to put a mini-itx computer onboard to do the processing but
> >> the weight (~2kg) was not justified.
>
> > It's not just the weight of the board and cooling system, it's also
> > the weight of the power system. *How many minutes/hours flight
> > duration is expected? *I couldn't find anything in the terse press
> > release and the almost blank "Preliminary Design Specifications"?
>
> >> Basically I chose a camera that had a good digital quality and allowed
> >> tuning of parameters like frame rate, exposure time etc. I also liked
> >> the fact that it had a serial output, which could perhaps be used on a
> >> future aircraft to drive servos, allowing a pan/tilt system or even
> >> the aircraft flight control system!
>
> > Good logic but I wouldn't have done it quite the same way. *If you're
> > going to have a mess of independent sensors, that all communicated to
> > a mess of dedicated computahs, each with their own particular
> > function, then the camera has to fit into the network communications
> > topology. *If you're using TCP/IP for onboard communications, you
> > might want to look into an IP network camera.
>
> >> I'm only working towards an image recognition capability at the moment
> >> so that's why I chose a good quality camera. I can probably take the
> >> casing off to save some weight too. Next year I would like to develop
> >> a visual gudance capability, which would be really cool and far more
> >> accurate than GPS.
>
> > The ground based autonomous vehicles have the advantage of really only
> > needing image recognition in 2D. *In the air, you have to recognize
> > and locate objects in 3D, which literally adds a new dimension to the
> > problem.
>
> > I guessed that you were going to be flying this year. *If you're just
> > building parts of the puzzle on the ground this year, there's no
> > reason to economize on weight and power at this point. *Might was well
> > go with your selected camera.
>
> >> I think I have my antenna setup sorted- although I have another
> >> question: If I use a 2-way splitter on the aircraft to feed the 2
> >> antennas, will there be any problems with recieving a signal from the
> >> ground station? What I mean is; is a splitter a bi-directioinal
> >> component or does it only allow a signal to travel in one direction?
>
> > Excellent question and good thinking. *A splitter (also known as a
> > Wilkinson power divider) divides the signal in two. *It's totally
> > passive and bi-directional. *At 2.4GHz, it's just two 1/4 wave strip
> > lines and a 100 ohm resistor. *See diagram at:
> > <http://www.microwaves101.com/encyclopedia/Wilkinson_splitters.cfm#two....>
> > <http://www.qsl.net/yu1aw/2G4spliter2.gif>
> > <http://www.qsl.net/yu1aw/2G4spliterN.gif>
> > Let's assume that there's no internal losses for now. *A 2.4GHz signal
> > applied to the input/output port, will be divided equally between the
> > two other ports. *That means that each antenna will get half the RF
> > power or 3dB. *3dB is NOT a large drop in signal level. *For example,
> > the rule of thumb for range versus power is a 6dB drop in power equals
> > a 50% drop in range. *3dB loss is equal to about 0.7 times the range.
> > If possible, simply increase the transmit power 3dB (double the
> > milliwatts) and you'll be exactly the same as before.
>
> > The situation is also good in receive. *There is no loss between
> > either antenna port and the input/output port. *Every dB received by
> > either antenna is delivered directly to the input/output port.
>
> > There is also considerably isolation (10-30dB) between antenna ports.
> > This is a huge help in keeping the antenna termination impedance near
> > the desired 50 ohms, thus keeping the VSWR and losses low. *If you can
> > tolerate minimal isolation, the 100 ohm resistor can be eliminated.
>
> > However, there are some potential problems. *RF radiating from both
> > antennas simultaneously will cancel and add at various points. *This
> > will create nulls and peaks in the antenna pattern. *The basic idea is
> > that each antenna should not "see" each other. *More simply, there
> > should be no position, where you can see both antennas at the same
> > time. *That's not going to be easy with the donut shaped pattern of a
> > common vertical 1/4 wave antenna. *The patterns will overlap and
> > possibly cancel. *However, it is possible with cardioid patterns, such
> > as patch or panel antennas.
>
> > There's also the question of where to mount the antennas. *On top and
> > on the belly are the obvious locations. *However, I think you'll get
> > better hemispherical coverage with one antenna mounted in the nose,
> > pointing forward, and another out the tail pointing aft. *This allows
> > the body to provide some shielding between antennas. *There may be
> > some signal loss dead ahead, which will be critical during landing.
> > However, that will also be the point of closest approach so distances
> > will be minimal.
>
> > Since you're not flying this year, I suggest you start with just one
> > antenna. *Play with it in a mockup or model and see what it does. Hang
> > the model from a string in a tree, and measure the received signal
> > strength. *That will determine if two antenna are necessary.
>
> > Also, I suggest you consider separating the video, data, and control
> > links. *Control can be done at much lower frequencies (27, 72MHz)
> > using conventional RC hardware. *This has the advantage of being
> > fairly orientation insensitive. *I don't think you have the 900MHz
> > band in UK, so that's out. *2.4GHz is it for data and video, but I
> > suspect that the video will saturate the link leaving no time slots
> > for data. *If there's some other frequencies available in UK for
> > video, I would use it.
>
> > Gotta run... good luck.
>
> It was interesting to note that the modern 2.4GHz rc controllers
> transmit on 2 channels and have 2 receivers, the antennas of which are
> dipoles and are usually placed perpendicular to each other.
>
> http://www.spektrumrc.com/ProdInfo/Files/SPM2710_DX7_Manual.pdf
>
> One of the teams involved in the "MOD Challenge" did a bit of work on
> imaging and the problems they encountered, worth a read.
> Imaging through to motion deblur.http://www.barnardmicrosystems.com/L4E_imaging.htm- Hide quoted text -
>
> - Show quoted text -

Thanks everyone for all your help, especially Jeff. I will acknowledge
you in my final report if thats ok with you

Adam Chapman
04-25-08, 08:18 AM
> I've done that. You're about to have a battery problem.

Do you think an 11.1V battery would do the trick? Im looking at this
one
http://www.flightpower.co.uk/index.asp?page=Products/index.asp^currency=GBP
At the cost of another 40 grams i could use a 14.8V battery and put
some resistors in.

Jeff Liebermann
04-25-08, 10:56 AM
On Fri, 25 Apr 2008 06:18:16 -0700 (PDT), Adam Chapman
<adam.chapman@student.manchester.ac.uk> wrote:

>> I've done that. You're about to have a battery problem.
>
>Do you think an 11.1V battery would do the trick? Im looking at this
>one
>http://www.flightpower.co.uk/index.asp?page=Products/index.asp^currency=GBP
>At the cost of another 40 grams i could use a 14.8V battery and put
>some resistors in.

Which battery were you thinking of? I'm not going to pick one for you
as there are quite a few on the site. Is this going to also power the
servos, receiver, propellers, etc or just the camera? You do the
numbers, and I'll check the results. I'm not doing your homework for
you.

The specs say that the camera at:
<http://trendnet.com/products/proddetail.asp?prod=110_TV-IP301W&cat=48>
show that it has a 12v 1.5A wall wart. That's not what it draws in
power, but that's what the power supply delivers. Your first step is
to *MEASURE* what it really draws. Don't be suprised if you get wide
variations depending on what the camera is doing, whether the IR
illuminator is running, frame rate settings, etc. Use and adjustable
voltage power supply and see what voltage range the camera can
tolerate. My guess(tm) is that the camera will run down to perhaps
8VDC or maybe even lower if it has an efficient switching regulator
inside. Measure the current at various points and calculate the
efficiency. I wouldn't be suprised if the camera can run on a 2 cell
battery pack (7.4VDC).

Once you know how much energy (i.e. watts/time) is required. LIPO
batteries have a maximum current draw spec which should not be
exceeded or you kill the battery. Watch out for so called "burst"
current ratings on the batteries. It's the battery that will burst if
you go over. Be sure to allow for temperature variations as it has a
huge effect on battery capacity. I'm not sure what percentage of the
capacity of the battery you can discharge down to, but if you go too
far, you also kill the battery. You'll need to know this number. How
long were you planning on draining the battery, which really means how
long is this thing suppose to be in the air? If it powers just the
camera, you can turn it off for the takeoff and landing, thus saving
energy.

Anyway, do the numbers. No Mathcad model needed for this one. Just
some measurements and arithmetic.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

msg
04-25-08, 07:54 PM
Jeff Liebermann wrote:

<snip>

> Anyway, do the numbers. No Mathcad model needed for this one. Just
> some measurements and arithmetic.

Hey Jeff,

Do you have the Mathcad 4.0 "Signal Processing Function Pack" or the
smaller "EE Electronic Handbook"? From the ad literature (way back)
the filter design tools appear to provide a level of interactivity
beyond the average (probably much better than the old tools that I
use).

Michael

Jeff Liebermann
04-26-08, 12:01 AM
On Fri, 25 Apr 2008 19:54:11 -0500, msg <msg@_cybertheque.org_> wrote:

>Do you have the Mathcad 4.0 "Signal Processing Function Pack" or the
>smaller "EE Electronic Handbook"?

Nope. I think you mean the Signal Processing Extension Pack. I'm
using a "borrowed" copy of Mathcad 3.1. A former employer has the
real copy and some of the add ons. We're not currently on good terms
so I have no access to the nifty software or test equipment.

Mathcad 4.0???? The current version is 14.0. Only $1200 for Mathcad
and $345 for the Sig Proc Exten Pack. Ouch.
<http://store.ptc.com/DRHM/servlet/ControllerServlet?Action=DisplayProductDetailsPage&SiteID=ptc&Locale=en_US&Env=BASE&productID=70040200>

>From the ad literature (way back)
>the filter design tools appear to provide a level of interactivity
>beyond the average (probably much better than the old tools that I
>use).

Dunno. It's been a while since I did anything with Mathcad and don't
recall the details. I'm also not very good with using it.

>Michael
--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831-336-2558 jeffl@comix.santa-cruz.ca.us
# http://802.11junk.com jeffl@cruzio.com
# http://www.LearnByDestroying.com AE6KS

Adam Chapman
04-27-08, 08:45 AM
On Apr 25, 4:56*pm, Jeff Liebermann <je...@cruzio.com> wrote:
> On Fri, 25 Apr 2008 06:18:16 -0700 (PDT), Adam Chapman
>
> <adam.chap...@student.manchester.ac.uk> wrote:
> >> I've done that. *You're about to have a battery problem.
>
> >Do you think an 11.1V battery would do the trick? Im looking at this
> >one
> >http://www.flightpower.co.uk/index.asp?page=Products/index.asp^currency=GBP
> >At the cost of another 40 grams i could use a 14.8V battery and put
> >some resistors in.
>
> Which battery were you thinking of? *I'm not going to pick one for you
> as there are quite a few on the site. *Is this going to also power the
> servos, receiver, propellers, etc or just the camera? *You do the
> numbers, and I'll check the results. *I'm not doing your homework for
> you.
>
> The specs say that the camera at:
> <http://trendnet.com/products/proddetail.asp?prod=110_TV-IP301W&cat=48>
> show that it has a 12v 1.5A wall wart. *That's not what it draws in
> power, but that's what the power supply delivers. *Your first step is
> to *MEASURE* what it really draws. *Don't be suprised if you get wide
> variations depending on what the camera is doing, whether the IR
> illuminator is running, frame rate settings, etc. *Use and adjustable
> voltage power supply and see what voltage range the camera can
> tolerate. *My guess(tm) is that the camera will run down to perhaps
> 8VDC or maybe even lower if it has an efficient switching regulator
> inside. *Measure the current at various points and calculate the
> efficiency. *I wouldn't be suprised if the camera can run on a 2 cell
> battery pack (7.4VDC).
>
> Once you know how much energy (i.e. watts/time) is required. *LIPO
> batteries have a maximum current draw spec which should not be
> exceeded or you kill the battery. *Watch out for so called "burst"
> current ratings on the batteries. *It's the battery that will burst if
> you go over. *Be sure to allow for temperature variations as it has a
> huge effect on battery capacity. *I'm not sure what percentage of the
> capacity of the battery you can discharge down to, but if you go too
> far, you also kill the battery. *You'll need to know this number. *How
> long were you planning on draining the battery, which really means how
> long is this thing suppose to be in the air? *If it powers just the
> camera, you can turn it off for the takeoff and landing, thus saving
> energy.
>
> Anyway, do the numbers. *No Mathcad model needed for this one. *Just
> some measurements and arithmetic.
> --
> Jeff Liebermann * * je...@cruzio.com
> 150 Felker St #D * *http://www.LearnByDestroying.com
> Santa Cruz CA 95060http://802.11junk.com
> Skype: JeffLiebermann * * AE6KS * *831-336-2558

Sorry i didnt realise the link was for the list rather than the
specific battery i chose. This battery will only be powering the
camera and onboard wireless stuff. I just wondered if 11.1 volts would
be enough. Your view was similar to mine, that the camera has a
tolerance. I should be able to get 13 minutes out of the camera, and
if we run out of power for the camera in flight, we can just land and
plug a spare in.

Thanks
Adam