Prev: WISP Routing into house WiFi network
Next: Wireless (Internet) Wiki (FAQ)
From: Jeff Liebermann on 20 Apr 2008 01:23
On Sat, 19 Apr 2008 12:10:08 -0700 (PDT), Adam Chapman
<adam.chapman(a)student.manchester.ac.uk> wrote:

>I am going to put 2 cisco 6dB patch antennas (http://www.cisco.com/en/
>US/docs/wireless/antenna/installation/guide/ant2460.html)

Good choice but pricey. Because you have a fiberglass aircraft, the
antennas can be inboard and do not need to be aerodynamic. To save
weight, you can remove the fiberglass radome from the antenna. Other
similar antennas to look at are:
<http://www.fab-corp.com/product.php?productid=1459&cat=255&page=1>
<http://www.hyperlinktech.com/familylist.aspx?id=147>
Keep the gain at around 8dBi or LOWER so that you have a fairly wide
(60 degree) antenna pattern. One antenna on each side of the
aircraft.

>back to
>back, and would like to stop them 'seeing' each other by using some RF
>absorbing material at http://www.kemtron.co.uk/arc.html, which I
>havent looked at in depth yet but expect to be lighter then
>aluminium.

No, not back to back. Put the antennas facing the outside of the
aircraft belly, pointed in opposite directions. The antenna pattern
sprays RF all in one direction, with very little going out the back of
the antenna. You won't need any shielding other than what is supplied
with the aluminum backing in the antennas. Also, mounting the
antennas against the side of the aircraft will prevent any reflections
from bouncing back into the pattern of the opposite antenna.

I'll assume that you want a downwards facing pattern. The -3dB
beamwidth of the antenna is 60 degrees. Therefore, the MINIMUM angle
between the two antennas is 60 degrees. I would go for somewhat more,
perhaps 90 degrees. The catch is that with this arrangement, in level
flight, you'll have no RF directed directly downward. If this is a
problem, I suggest you try using only one antenna, facing downward.
However, if you're doing acrobatics and need to communicate at any
attitude, you'll need more than 2 antennas.

Think of it in terms of illuminating with lights. Each 8dBi gain
antenna is a flood light, with a 60 degree wide beam. What gets lit
up, gets RF. What doesn't, is in a dead area. Lower gain antennas
have a wider beamwidth.

>The aircraft will be made of fibreglass so the shielding is needed.
>I'm using a wifi network so the frequency is 2.4 gigs.

Model airplane? Manned airplane? Experimental aircraft? Human
powered flight? Ornithopter?

>I dontthink the power-pass is a requirenment, although power-pass
>splitters seem to be lighter and cheaper. Would using a power pass
>enabled splitter actually cause any problems?

No benefit to using a DC power pass system. There are no active
devices in the antenna.

>I also need an amplifier in my system and this one (http://www.rf-
>links.com/AMP8_24.pdf) looks nice and small, although I dont know if
>it amplifies both ways, i.e. ampllifies a trnamitted signal AND a
>recieved signal. Can any of you guys tell by looking at it?

No. Amplifiers tend to be a problem. They also ost plenty and
weigh a bunch. I don't want to go into details. Search this news
group for "RF power amplifier" or bettery yet, search for "alligator".

I've done model airplane drone video and data systems. There was no
way to do it using static antennas at 2.4GHz. The drone had to be
tracked from the ground using a high gain dish antenna. The antenna
on the drone was an attempt at being omnidirectional with a
hemispherical pattern using just a 1/4 wave monopole antenna. That's
not much gain, which explains the 24dBi tracking dish antenna on the
ground. See my comments under:
<http://groups.google.com/group/alt.internet.wireless/browse_thread/thread/bf1b984da67cb8d2/>

What's the ground end of this video link going to look like?

--
Jeff Liebermann jeffl(a)cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
From: Adam Chapman on 20 Apr 2008 07:43
On Apr 20, 6:23 am, Jeff Liebermann <je...(a)cruzio.com> wrote:
> On Sat, 19 Apr 2008 12:10:08 -0700 (PDT), Adam Chapman
>
> <adam.chap...(a)student.manchester.ac.uk> wrote:
> >I am going to put 2 cisco 6dB patch antennas (http://www.cisco.com/en/
> >US/docs/wireless/antenna/installation/guide/ant2460.html)
>
> Good choice but pricey.  Because you have a fiberglass aircraft, the
> antennas can be inboard and do not need to be aerodynamic.  To save
> weight, you can remove the fiberglass radome from the antenna.  Other
> similar antennas to look at are:
> <http://www.fab-corp.com/product.php?productid=1459&cat=255&page=1>
> <http://www.hyperlinktech.com/familylist.aspx?id=147>
> Keep the gain at around 8dBi or LOWER so that you have a fairly wide
> (60 degree) antenna pattern.  One antenna on each side of the
> aircraft.
>
> >back to
> >back, and would like to stop them 'seeing' each other by using some RF
> >absorbing material athttp://www.kemtron.co.uk/arc.html, which I
> >havent looked at in depth yet but expect to be lighter then
> >aluminium.
>
> No, not back to back.  Put the antennas facing the outside of the
> aircraft belly, pointed in opposite directions.  The antenna pattern
> sprays RF all in one direction, with very little going out the back of
> the antenna.  You won't need any shielding other than what is supplied
> with the aluminum backing in the antennas.  Also, mounting the
> antennas against the side of the aircraft will prevent any reflections
> from bouncing back into the pattern of the opposite antenna.  
>
> I'll assume that you want a downwards facing pattern.  The -3dB
> beamwidth of the antenna is 60 degrees.  Therefore, the MINIMUM angle
> between the two antennas is 60 degrees.  I would go for somewhat more,
> perhaps 90 degrees.  The catch is that with this arrangement, in level
> flight, you'll have no RF directed directly downward.  If this is a
> problem, I suggest you try using only one antenna, facing downward.
> However, if you're doing acrobatics and need to communicate at any
> attitude, you'll need more than 2 antennas.  
>
> Think of it in terms of illuminating with lights.  Each 8dBi gain
> antenna is a flood light, with a 60 degree wide beam.  What gets lit
> up, gets RF.  What doesn't, is in a dead area.  Lower gain antennas
> have a wider beamwidth.
>
> >The aircraft will be made of fibreglass so the shielding is needed.
> >I'm using a wifi network so the frequency is 2.4 gigs.
>
> Model airplane?  Manned airplane?  Experimental aircraft?  Human
> powered flight?  Ornithopter?
>
> >I dontthink the power-pass is a requirenment, although power-pass
> >splitters seem to be lighter and cheaper. Would using a power pass
> >enabled splitter actually cause any problems?
>
> No benefit to using a DC power pass system.  There are no active
> devices in the antenna.
>
> >I also need an amplifier in my system and this one (http://www.rf-
> >links.com/AMP8_24.pdf) looks nice and small, although I dont know if
> >it amplifies both ways, i.e. ampllifies a trnamitted signal AND a
> >recieved signal. Can any of you guys tell by looking at it?
>
> No.  Amplifiers tend to be a problem.    They also ost plenty and
> weigh a bunch.  I don't want to go into details.  Search this news
> group for "RF power amplifier" or bettery yet, search for "alligator".
>
> I've done model airplane drone video and data systems.  There was no
> way to do it using static antennas at 2.4GHz.  The drone had to be
> tracked from the ground using a high gain dish antenna.  The antenna
> on the drone was an attempt at being omnidirectional with a
> hemispherical pattern using just a 1/4 wave monopole antenna.  That's
> not much gain, which explains the 24dBi tracking dish antenna on the
> ground.  See my comments under:
> <http://groups.google.com/group/alt.internet.wireless/browse_thread/th...>
>
> What's the ground end of this video link going to look like?
>
> --
> Jeff Liebermann     je...(a)cruzio.com
> 150 Felker St #D    http://www.LearnByDestroying.com
> Santa Cruz CA 95060http://802.11junk.com
> Skype: JeffLiebermann     AE6KS    831-336-2558

I want to put this: http://www.radiolabs.com/products/wireless/waverv2.php
on the ground, I was planning to be sneaky on the fly-off day. The
camera system uses an http protocol (not perfect i know, but the
easiest to interface with matlab for now).
It's for a model sized aircraft (about 3m wingspan). Flying range has
been specified as 500m, but also within visual range at all times for
safety. Im not sure how 'visible' the aircraft will be over 500m, so
we will probably be flying closer, and probably no higer than 100ft
too. Because the little laptop antenna i linked fits onto the screen
of the laptop, i was hoping to tilt the screen so the high gain region
points at the aircraft whilst a pretended to squint at the screen.

Because we will be flying pretty low and looking for targets on the
ground with the camera, which looks directly down and is fixed to the
aircraft belly, we don't really need a signal to be transmitted
downward. (we won't need to identify ourselves as targets
hopefully!).

I see what the alligator term represents, big mouth, little ears. A
linear amplifier amplifies a signal in ONE direction and not in the
other, but does it reduce the signal in its non-amplifying direction?
I guess I could use two alligators if they had big enough mouths.

Adam
From: Jeff Liebermann on 20 Apr 2008 12:23
On Sun, 20 Apr 2008 04:43:38 -0700 (PDT), Adam Chapman
<adam.chapman(a)student.manchester.ac.uk> wrote:

>I want to put this: http://www.radiolabs.com/products/wireless/waverv2.php
>on the ground,

That won't do much for you. Your airplane will be able to hear the
signals from the ground somewhat better, but your ground station will
not be able to do likewise. You would need such an amplifier at both
ends in order for it to be effective.

>I was planning to be sneaky on the fly-off day.

Cleverness and sneakiness are good things. However, I think the
judges are looking for good engineering.

>The
>camera system uses an http protocol (not perfect i know, but the
>easiest to interface with matlab for now).
>It's for a model sized aircraft (about 3m wingspan).

Duh. I forgot that you were the one building the autonomous model
airplane. I've been engaged in various email exchanges on the subject
of flying a camera in a much larger airplane, and was thinking along
those lines.

>Flying range has
>been specified as 500m, but also within visual range at all times for
>safety. Im not sure how 'visible' the aircraft will be over 500m, so
>we will probably be flying closer, and probably no higer than 100ft
>too.

Well, I can see you haven't bothered to do the math. I'll do a dry
run for you. See:
<http://wireless.wikia.com/wiki/Wi-Fi#Link_Calculations>
I'll assume that you're using the "Waverider" gizmo, which belches
200mw of RF (+23dBm) and a 3dB power divider with a 0.5dB port loss.
I'm not sure what frame rate you're expecting so I'll assume that you
can tolerate the slowest OFDM speed (6Mbits/sec) as the 802.11b speeds
are generally not that useful for streaming anything. As I recall,
you're using this camera:
<http://trendnet.com/products/proddetail.asp?prod=110_TV-IP301W&cat=48>
Unfortunately, the data sheet does not offer RF specifications, so
I'll conjur my own. That basic idea is to see at what range you can
maintain a 20dB fade margin.

From the air to the ground:
TX power +15dBm
TX coax loss -1dB (includes connector losses)
Divider loss -3.5dB (half the power less internal losses)
TX ant gain 6dBi
Distance unknown
RX ant gain 2dBi
RX coax loss 0dB (assumes good RF amp in Waverider thing).
RX sens -88dBm (at 6Mbits/sec)
Fade margin 20dB

Plugging the numbers into:
<http://www.terabeam.com/support/calculations/som.php>
I get 0.12 miles or about 200 meters. Not very good.

The other direction, from the ground to the air is a bit different.
The power divider does NOT split the signal coming in from the two
ports. Therefore, the loss is less.

From ground to air:
TX power +22dBm
TX coax loss -0.5dB (connector losses)
TX ant gain 2dBi
Distance unknown
RX ant gain 6dBi
Divider loss -0.5dB (internal divider losses)
RX coax loss -0.5dB (connector losses).
RX sens -88dBm (at 6Mbits/sec)
Fade margin 20dB

Plugging the numbers into the calculator, I get 0.4 miles or about 650
meters. Much better but barely adequate.

Worse is yet to come. The above guesswork is based totally on the
assumption that your ground and air antennas are oriented for maximum
gain. This is obviously not a good assumption as the aircraft be off
axis for the majority of possible orientations. Therefore the 6dBi
antenna gain may actually be considerably less when oriented off axis.
For example, if you lose 3dB of gain due to orientation errors, your
range will be 0.707 times as far. 6dB loss is good for 0.5 times the
range.

Someone is sure to suggest that slowing down the system from
6Mbits/sec to 1Mbit/sec will dramatically increase the range. That
doesn't work. Look at the receiver sensitivity chart at:
<http://wireless.wikia.com/wiki/Wi-Fi#Link_Calculations>
At 6Mbits/sec OFDM, the rx sensitivity is -88dBm. At 1Mbits/sec BPSK,
the rx sensitivity is -89dBm. For 1dB of additional gain, it's hardly
worth a 1/6th decrease in speed. That's also why I suggest you use
the slowed OFDM speed.

>Because the little laptop antenna i linked fits onto the screen
>of the laptop, i was hoping to tilt the screen so the high gain region
>points at the aircraft whilst a pretended to squint at the screen.

Get a tripod. Mount a fairly directional antenna on the tripod. You
don't need much gain, but you do need a wide beamwidth. Another 6 or
8dBi panel will do. Point it in the general direction of the test
range. Attach to your laptop. Put a photographers blind black cloth
over your head. Don't worry about looking weird.

>Because we will be flying pretty low and looking for targets on the
>ground with the camera, which looks directly down and is fixed to the
>aircraft belly, we don't really need a signal to be transmitted
>downward. (we won't need to identify ourselves as targets
>hopefully!).

Perfect. If you're not going to fly at much altitude, an
omnidirectional antenna (monopole) sticking out of the belly, will
work just fine. The radiation pattern is a horizontal donut. As long
as you maintain level flight, you're fine. Do aerobatics and you'll
have problems, but level flight up to about 30 degrees is just fine.

>I see what the alligator term represents, big mouth, little ears.

Yep. It's asymmetrical.

>A
>linear amplifier amplifies a signal in ONE direction and not in the
>other, but does it reduce the signal in its non-amplifying direction?

Most such RF amps have a receive amplifier. The problem is that it
doesn't improve the receive sensitivity in any useful way. The access
point (or laptop wireless card) already is running at the limit of
receiver sensitivity. Adding additional gain will only add additional
noise and reduce dynamic range of the receiver. What the receive
amplifier will do that's useful is compensate for the coax cable
losses between the amplifier and the receiver. However, for your
laptop installation, that's very little coax, very little loss, and
very little benifit.

>I guess I could use two alligators if they had big enough mouths.

Yep. Try to keep the tx power at both ends roughly the same. If
you're going to fly a power splitter, you'll need 3dB (twice) the TX
power output of the ground station in the airplane in order to
compensate for the splitter losses.

Good luck.

--
Jeff Liebermann jeffl(a)cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
From: Adam Chapman on 20 Apr 2008 15:23
On Apr 20, 5:23 pm, Jeff Liebermann <je...(a)cruzio.com> wrote:
> On Sun, 20 Apr 2008 04:43:38 -0700 (PDT), Adam Chapman
>
> <adam.chap...(a)student.manchester.ac.uk> wrote:
> >I want to put this:  http://www.radiolabs.com/products/wireless/waverv2..php
> >on the ground,
>
> That won't do much for you.  Your airplane will be able to hear the
> signals from the ground somewhat better, but your ground station will
> not be able to do likewise.  You would need such an amplifier at both
> ends in order for it to be effective.
>
> >I was planning to be sneaky on the fly-off day.
>
> Cleverness and sneakiness are good things.  However, I think the
> judges are looking for good engineering.
>
> >The
> >camera system uses an http protocol (not perfect i know, but the
> >easiest to interface with matlab for now).
> >It's for a model sized aircraft (about 3m wingspan).
>
> Duh.  I forgot that you were the one building the autonomous model
> airplane.  I've been engaged in various email exchanges on the subject
> of flying a camera in a much larger airplane, and was thinking along
> those lines.
>
> >Flying range has
> >been specified as 500m, but also within visual range at all times for
> >safety. Im not sure how 'visible' the aircraft will be over 500m, so
> >we will probably be flying closer, and probably no higer than 100ft
> >too.
>
> Well, I can see you haven't bothered to do the math.  I'll do a dry
> run for you.  See:
> <http://wireless.wikia.com/wiki/Wi-Fi#Link_Calculations>
> I'll assume that you're using the "Waverider" gizmo, which belches
> 200mw of RF (+23dBm) and a 3dB power divider with a 0.5dB port loss.
> I'm not sure what frame rate you're expecting so I'll assume that you
> can tolerate the slowest OFDM speed (6Mbits/sec) as the 802.11b speeds
> are generally not that useful for streaming anything.  As I recall,
> you're using this camera:
> <http://trendnet.com/products/proddetail.asp?prod=110_TV-IP301W&cat=48>
> Unfortunately, the data sheet does not offer RF specifications, so
> I'll conjur my own.  That basic idea is to see at what range you can
> maintain a 20dB fade margin.
>
> From the air to the ground:
>   TX power        +15dBm
>   TX coax loss     -1dB (includes connector losses)
>   Divider loss   -3.5dB (half the power less internal losses)
>   TX ant gain       6dBi
>   Distance       unknown
>   RX ant gain       2dBi
>   RX coax loss      0dB  (assumes good RF amp in Waverider thing).
>   RX sens         -88dBm (at 6Mbits/sec)
>   Fade margin      20dB
>
> Plugging the numbers into:
> <http://www.terabeam.com/support/calculations/som.php>
> I get 0.12 miles or about 200 meters.  Not very good.
>
> The other direction, from the ground to the air is a bit different.
> The power divider does NOT split the signal coming in from the two
> ports.  Therefore, the loss is less.
>
> From ground to air:
>   TX power        +22dBm
>   TX coax loss    -0.5dB (connector losses)
>   TX ant gain       2dBi
>   Distance       unknown
>   RX ant gain       6dBi
>   Divider loss   -0.5dB  (internal divider losses)
>   RX coax loss   -0.5dB  (connector losses).
>   RX sens         -88dBm (at 6Mbits/sec)
>   Fade margin      20dB
>
> Plugging the numbers into the calculator, I get 0.4 miles or about 650
> meters.  Much better but barely adequate.
>
> Worse is yet to come.  The above guesswork is based totally on the
> assumption that your ground and air antennas are oriented for maximum
> gain.  This is obviously not a good assumption as the aircraft be off
> axis for the majority of possible orientations.  Therefore the 6dBi
> antenna gain may actually be considerably less when oriented off axis.
> For example, if you lose 3dB of gain due to orientation errors, your
> range will be 0.707 times as far.  6dB loss is good for 0.5 times the
> range.
>
> Someone is sure to suggest that slowing down the system from
> 6Mbits/sec to 1Mbit/sec will dramatically increase the range.  That
> doesn't work.  Look at the receiver sensitivity chart at:
> <http://wireless.wikia.com/wiki/Wi-Fi#Link_Calculations>
> At 6Mbits/sec OFDM, the rx sensitivity is -88dBm.  At 1Mbits/sec BPSK,
> the rx sensitivity is -89dBm.  For 1dB of additional gain, it's hardly
> worth a 1/6th decrease in speed.  That's also why I suggest you use
> the slowed OFDM speed.
>
> >Because the little laptop antenna i linked fits onto the screen
> >of the laptop, i was hoping to tilt the screen so the high gain region
> >points at the aircraft whilst a pretended to squint at the screen.
>
> Get a tripod.  Mount a fairly directional antenna on the tripod.  You
> don't need much gain, but you do need a wide beamwidth.  Another 6 or
> 8dBi panel will do.  Point it in the general direction of the test
> range.  Attach to your laptop.  Put a photographers blind black cloth
> over your head.  Don't worry about looking weird.
>
> >Because we will be flying pretty low and looking for targets on the
> >ground with the camera, which looks directly down and is fixed to the
> >aircraft belly, we don't really need a signal to be transmitted
> >downward. (we won't need to identify ourselves as targets
> >hopefully!).
>
> Perfect.  If you're not going to fly at much altitude, an
> omnidirectional antenna (monopole) sticking out of the belly, will
> work just fine.  The radiation pattern is a horizontal donut.  As long
> as you maintain level flight, you're fine.  Do aerobatics and you'll
> have problems, but level flight up to about 30 degrees is just fine.
>
> >I see what the alligator term represents, big mouth, little ears.
>
> Yep.  It's asymmetrical.
>
> >A
> >linear amplifier amplifies a signal in ONE direction and not in the
> >other, but does it reduce the signal in its non-amplifying direction?
>
> Most such RF amps have a receive amplifier.  The problem is that it
> doesn't improve the receive sensitivity in any useful way.  The access
> point (or laptop wireless card) already is running at the limit of
> receiver sensitivity.  Adding additional gain will only add additional
> noise and reduce dynamic range of the receiver.  What the receive
> amplifier will do that's useful is compensate for the coax cable
> losses between the amplifier and the receiver.  However, for your
> laptop installation, that's very little coax, very little loss, and
> very little benifit.
>
> >I guess I could use two alligators if they had big enough mouths.
>
> Yep.  Try to keep the tx power at both ends roughly the same.  If
> you're going to fly a power splitter, you'll need 3dB (twice) the TX
> power output of the ground station in the airplane in order to
> compensate for the splitter losses.
>
> Good luck.
>
> --
> Jeff Liebermann     je...(a)cruzio.com
> 150 Felker St #D    http://www.LearnByDestroying.com
> Santa Cruz CA 95060http://802.11junk.com
> Skype: JeffLiebermann     AE6KS    831-336-2558

Thanks, I did do the maths before but admittedly I did it wrong. I
assumed that the Rx gain would be the same as the Tx power. Silly me!
First  |  Prev  | 
Pages: 1 2
Prev: WISP Routing into house WiFi network
Next: Wireless (Internet) Wiki (FAQ)