On Tue, 24 Jun 2008 16:39:29 -0500, "amdx" <(E-Mail Removed)> wrote:
> I don't have any Ghz equipment
That's what eBay is for. Most of this junk came from eBay:
<http://802.11junk.com/jeffl/pics/home/slides/BL-shop5.html>
Note the pile of HP 8620 sweep generators and plugins. If you want
one that works, buy three. Fix the best, use the worst for parts, and
save the middle one for when the best one blows up.
<http://802.11junk.com/jeffl/pics/home/slides/lab.html>
The 2 gray boxes on the left are Wiltron 610(?) sweep generators and
plugins. Good stuff but you gotta know how to fix them.
>so when I build an antenna I don't really
>know what
>I hav other than it works.
Use an antenna modeling program first. I suggest 4NEC2.
<http://home.ict.nl/~arivoors/>
The real antenna will never work better than the theoretical model.
Compare test results with the model and you'll have a clue what to do
to fix the antenna.
>I've compared signal strengths on the computers
>wireless
>utility so I know which is better, but not what makes an antenna match
>better.
Not good enough. You need to sweep the VWSR over the frequency range
to get it right. Never mind gain. If you have the matching nailed,
the gain will be close to theoretical with no inconvenient bumps in
the gain curve across the freq range.
Incidentally, if you want to build your own, I suggest you look at
AMOS or Franklin antennas. Same URL as the VWSR meter:
<http://pe2er.nl/wifisector/>
There are others but I'm too lazy to dig them out.
>When you mention balance cap, do you mean the cu foil that you trim or
>add.
Yep, that's it. Microwaves use very small value cazapitors.
>If so? I think you adjust this with a 50 ohm dummy load, to zero the output?
Yep. Good luck finding a dummy load that's exactly 50 ohms at 2.4GHz.
>I don't really see this? I think we're adjusting the 49.9 ohm resistors
>capacitance to ground. Tell me what you can, I'll have to study it more.
Not now. It's easier to build it and see what it does, than to
explain it. Besides, I'm not sure I can explain how it works and will
need to hit the books.
If you have time, just build a lower frequency version of the bridge.
Use whatever resistors you can. Scale the capacitor values. Almost
any decent diode will work. Use it a 50 or 150 MHz where you should
be able to scrounge some usable test equipment.
> Are you recommending a case on the head unit?
Yep. The ones in the photos are fine. Just make sure that things do
not move around, which was my problem. Also, you will have the null
change when you install the covers, so be prepared to adjust the null
cap through a small hole.
> PCB material ok?
See my comments about FR4/G10 PCB material. Use Polysulfone if
possible. However, G10/FR4 will work for a first attempt.
>> I didn't use a meter. The output went to a scope.
>
>You monitor the dc level with your scope? Never mind, I see that is a pulse
>coming
>out of the head unit.
You can integrate the DC output if you don't like pulses. That's what
the peak detector does. I don't use a Wi-Fi radio as a signal
generator so I don't have the pulse problem.
>> If you can find scrap polysulfone PCB material, use it instead of
>> G10/FR4. Less lossy.
>
> I have some Rogers Duroid .031", 5880 I think
er=2.2. Low loss glass and teflon composite dielectric. Loss tangent
of 0.0009 at 10Ghz. Probably overkill but I would certainly use it if
you have it. That's better than polysulfone. However, I would still
use G10/FR4 for the first attempt, which is by definition, doomed to
failure. Learn By Destroying(tm).
>> If using a Wi-Fi router for a signal source, you'll have problems
>> seeing anything on the meter without the peak detector circuit.
>
> Interesting, why do you think the author says it (the peak hold switch) is
>not needed?
>I think we're talking about the same thing!
Dunno. Probably because he's also not using a pulsing wi-fi router as
a signal source. Otherwise, his meter is fed by a rectifier and
integrator which gives tolerable output. I'm just guessing.
>Did you use the BAT 62-03W diode? If not what did you use? Some one
>recommended the BAT 62-02W it's smaller so should be ok, and I can get it.
I have no idea what I used. I have a mess of unlabelled diodes of
undetermined origin, that seem to be suitable for microwave
applications. No logo, no number, no code, not even a color dot.
Probably production rejects, but they're good enough for what I'm
doing. Sorry, no recommendation for substitutes other than it be a
shottky diode with low cazapitance and a sufficiently high reverse
voltage to prevent vaporization.
<http://www.infineon.com/dgdl/bat62series.pdf?folderId=db3a304314dca389011518104 e5d0df2&fileId=db3a304314dca3890115181141bf0df3>
About $5/ea and you'll probably have to buy a bunch to get over the
minimum. The 62-02W is very tiny. I hope you have the proper
handling tools (toothpick, superglue, and paper clip around the
soldering iron tip).
If you need a substitute, I can dig something out of Digikey, Mouser,
and accomplises. You can possibly also bug the distributors for
"samples".
--
# Jeff Liebermann 150 Felker St #D Santa Cruz CA 95060
# 831-336-2558
(E-Mail Removed)
#
http://802.11junk.com (E-Mail Removed)
#
http://www.LearnByDestroying.com AE6KS
Similar Threads
Linear Mode
