Attn: Jeff Lieberman - Network Analyzer Info

Hi Jeff,

Better late than never.

I did some tests on an HP Network Analyzer regarding the multiple
adapters subject.

The results were kind of interesting. Here's a zip with some not-so-high
quality pix.

http://users.adelphia.net/~thisnthat/adapters.zip

I did this at 450 mHz and at 2.4 gHz to illustrate the difference in
different frequency ranges.

There is a pic there showing the adapter 'rig'. It's a mix of 'N', 'BNC',
and some 'TNC' connectors. There's around 25 connections in this chain.

The names of the files describe the measurements. I never did buy into
the whole .25db per connection loss. The vertical scale on all pix is 1dB
and the freq. span is 500 mHz....50 mHz/div.

At 450 mHz the total loss is only .16 dB..... a far cry from the
estimated 6 ish db of loss if you go by .25/connection. Much more loss
though at 2.45 gHz center, but still only 2.1 dB @ 2450 mHz. The curve is
much more rough though, probably a lot to do with the adapters rapid
change in diameter.

I also included SWR measurements for 450 & 2.4 as well.

(Back to work.)

Regards,

DanS

DanS <(E-Mail Removed)> hath wroth:

>Better late than never.

Neato. Much thanks.

>I did some tests on an HP Network Analyzer regarding the multiple
>adapters subject.
>
>The results were kind of interesting. Here's a zip with some not-so-high
>quality pix.
>
>http://users.adelphia.net/~thisnthat/adapters.zip

I did some image processing and resizing for easier viewing.
| http://www.80211junk.com/jeffl/anten...oss/index.html

>I did this at 450 mHz and at 2.4 gHz to illustrate the difference in
>different frequency ranges.
>
>There is a pic there showing the adapter 'rig'. It's a mix of 'N', 'BNC',
>and some 'TNC' connectors. There's around 25 connections in this chain.

I think it's spelled "kludge".
http://en.wikipedia.org/wiki/Kludge

>The names of the files describe the measurements. I never did buy into
>the whole .25db per connection loss. The vertical scale on all pix is 1dB
>and the freq. span is 500 mHz....50 mHz/div.
>
>At 450 mHz the total loss is only .16 dB..... a far cry from the
>estimated 6 ish db of loss if you go by .25/connection. Much more loss
>though at 2.45 gHz center, but still only 2.1 dB @ 2450 mHz. The curve is
>much more rough though, probably a lot to do with the adapters rapid
>change in diameter.

My guess(tm) is that the frequency response is more a problem with the
BNC connectors used. They just don't make a good fit. The N and TNC
are much better. Also, dissimilar metals (cadmium, silver, gold,
nickel) plating might cause a slight increase in loss.

Using your results at 2.4GHz, that's:
2.1dB / 23 connector junctions = 0.09dB per connector pair.

Looking at the photo, my guess is about 18" of connectors.
LMR-240 is good for 12.9dB/100ft. 18" would have a loss of 0.194dB.
Even RG-174 would be only about 0.9dB loss. So much for my statement
that it's the same loss as an equivalent piece of coax cable. The
connector string are worse.

>I also included SWR measurements for 450 & 2.4 as well.
>(Back to work.)

Thanks again.
--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

On Wed, 12 Jul 2006 09:01:32 -0700, Jeff Liebermann
<(E-Mail Removed)> wrote in
<(E-Mail Removed)>:

>My guess(tm) ...

Wouldn't (sm) would be more appropriate?

--
Best regards, FAQ for Wireless Internet: <http://Wireless.wikia.com>
John Navas FAQ for Wi-Fi: <http://wireless.wikia.com/wiki/Wi-Fi>
Wi-Fi How To: <http://wireless.wikia.com/wiki/Wi-Fi_How_To>
Fixes to Wi-Fi Problems: <http://wireless.wikia.com/wiki/Wi-Fi_Fixes>

John Navas <(E-Mail Removed)> hath wroth:

>On Wed, 12 Jul 2006 09:01:32 -0700, Jeff Liebermann
><(E-Mail Removed)> wrote in
><(E-Mail Removed)>:
>
>>My guess(tm) ...
>
>Wouldn't (sm) would be more appropriate?

It's actually an inside joke. One of my former clients was looking
for a suitable acronym for their for-pay support organization. I
suggested GUESS (Global User Extended Services and Support). Although
management failed to appreciate my suggestion, the acronym stuck and
was commonly used until almost all support was outsourced to India.
Since then, I've used the term to describe my highly intuitive and
seat of the pants method of answering technical questions.

However, I think you're right. It really is a service mark, not a
trademark. The problem is that most users wouldn't understand the
(sm) tag. Methinks I'll continue to mislabel it as (tm) to avoid
having to answer "what's the (sm) mean?".


--
Jeff Liebermann (E-Mail Removed)
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558

Jeff Liebermann <(E-Mail Removed)> wrote in
news:(E-Mail Removed):

> DanS <(E-Mail Removed)> hath wroth:
>
>>Better late than never.
>
> Neato. Much thanks.
>
>>I did some tests on an HP Network Analyzer regarding the multiple
>>adapters subject.
>>
>>The results were kind of interesting. Here's a zip with some
>>not-so-high quality pix.
>>
>>http://users.adelphia.net/~thisnthat/adapters.zip
>
> I did some image processing and resizing for easier viewing.
>| http://www.80211junk.com/jeffl/anten...oss/index.html
>
>>I did this at 450 mHz and at 2.4 gHz to illustrate the difference in
>>different frequency ranges.
>>
>>There is a pic there showing the adapter 'rig'. It's a mix of 'N',
>>'BNC', and some 'TNC' connectors. There's around 25 connections in
>>this chain.
>
> I think it's spelled "kludge".
> http://en.wikipedia.org/wiki/Kludge
>
>>The names of the files describe the measurements. I never did buy into
>>the whole .25db per connection loss. The vertical scale on all pix is
>>1dB and the freq. span is 500 mHz....50 mHz/div.
>>
>>At 450 mHz the total loss is only .16 dB..... a far cry from the
>>estimated 6 ish db of loss if you go by .25/connection. Much more loss
>>though at 2.45 gHz center, but still only 2.1 dB @ 2450 mHz. The curve
>>is much more rough though, probably a lot to do with the adapters
>>rapid change in diameter.
>
> My guess(tm) is that the frequency response is more a problem with the
> BNC connectors used. They just don't make a good fit. The N and TNC
> are much better. Also, dissimilar metals (cadmium, silver, gold,
> nickel) plating might cause a slight increase in loss.
>
> Using your results at 2.4GHz, that's:
> 2.1dB / 23 connector junctions = 0.09dB per connector pair.
>
> Looking at the photo, my guess is about 18" of connectors.
> LMR-240 is good for 12.9dB/100ft. 18" would have a loss of 0.194dB.
> Even RG-174 would be only about 0.9dB loss. So much for my statement
> that it's the same loss as an equivalent piece of coax cable. The
> connector string are worse.
>
>>I also included SWR measurements for 450 & 2.4 as well.
>>(Back to work.)
>
> Thanks again.

I don't know if the BNC _design_ has anything to do with it. They are the
same as TNC, except quick-release and not threaded. The chain had 2 BNC
MALE barrels in it that are quite old silver plated ones.

I was actually a little suprised about the 2.4 gHz performance. I didn't
think that there would be that much loss. (Although still far from the
..25 dB I was taught.)

I'll try to see if I can dig up a bunch of 'N' only barrels and re-check
it then with out the connector type change.

If I am able to, I'll be sure the pix are a little better. Those were
just quicky, non-tripod shots. It would be a good idea to to the picture
resolution back up and shoot from further back to minimize the CRT's
apparent curvature, although I've gotten much better shots with the only
difference being a tripod.

The final outcome though is really that connector connections do not
introduce nearly as much loss as expected, and having an adapter or 2 in-
line will most likely make no detectable difference. If it does, there
has to be something electrically wrong with it, like an errant braiding
hair or a bad crimp.

As a side-question, have you ever used Radio Mobile (a freeware prop
study tool) ?

Regards,

DanS

DanS <(E-Mail Removed)> wrote:
>The final outcome though is really that connector connections do not
>introduce nearly as much loss as expected, and having an adapter or 2 in-
>line will most likely make no detectable difference.

Wouldn't the fair test be a single piece of coax at least many
wavelengths long, with a single connector pair introduced into the
middle of it? I'd think that the reason you can have all those
connectors stuck together with so little loss is that they interfere
with each other because they are so close. For instance, a capacitive
section near an inductive section (if close enough together) might
look less lossy than either of them individually.

On Wed, 12 Jul 2006 12:27:53 -0500, DanS
<(E-Mail Removed)> wrote:

>I don't know if the BNC _design_ has anything to do with it. They are the
>same as TNC, except quick-release and not threaded. The chain had 2 BNC
>MALE barrels in it that are quite old silver plated ones.

True. However, the connector alignment is much better with a TNC than
with a BNC. The BNC tends to wiggle quite a bit. If the center pin
or shield fingers get misaligned, the loss and leakage goes up
somewhat.

>I was actually a little suprised about the 2.4 gHz performance. I didn't
>think that there would be that much loss. (Although still far from the
>.25 dB I was taught.)

Agreed. 2.1dB is a little higher than I would have expected for 18"
of connectors at 2.4Ghz. Possibly, one of the adapters has a bad
dielectric. Any phenolic insulators mixed in there (I doubt it but I
sometimes get a surprise). I have a few N connectors with mica filled
phenolic I save to show the non-believers.

>I'll try to see if I can dig up a bunch of 'N' only barrels and re-check
>it then with out the connector type change.

Sounds reasonable.

>If I am able to, I'll be sure the pix are a little better. Those were
>just quicky, non-tripod shots. It would be a good idea to to the picture
>resolution back up and shoot from further back to minimize the CRT's
>apparent curvature, although I've gotten much better shots with the only
>difference being a tripod.
>
>The final outcome though is really that connector connections do not
>introduce nearly as much loss as expected, and having an adapter or 2 in-
>line will most likely make no detectable difference. If it does, there
>has to be something electrically wrong with it, like an errant braiding
>hair or a bad crimp.

Yep. The number of the week is 0.1dB per connector pair.

>As a side-question, have you ever used Radio Mobile (a freeware prop
>study tool) ?

Oh yes. I've been abusing it for years.
http://www.cplus.org/rmw/english1.html
I carry the California west coast SRTM data around with me on a DVD.
Here's one I did of the local ham radio repeater:
http://www.LearnByDestroying.com/maps/k6bj3-3d.jpg
I can post more if you're interested (after I clear out some diskspace
so I don't go over my quota). I have a bunch I did at 2.4Ghz for
various WISP's, but I can't post those. Suffice to say that theory
and reality were fairly close agreement. I also like to do the
stereoscopic views with the red and blue glasses. I've also been
tinkering with using Google Earth for the map.

Hint: Use SRTM maps, not DTED or DEM.

The problem with Radio-Mobile is that the learning curve is a bit long
and the order and sequence of operations is far from intuitive. Of
course, like all good programs, the documentation sucks. See:
http://www.pizon.org/radio-mobile-tutorial/index.html
http://groups.yahoo.com/group/Radio_Mobile_Deluxe/
for tutorials and support.

--
# 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

Jeff Liebermann <(E-Mail Removed)> wrote in
news(E-Mail Removed):


>
> The problem with Radio-Mobile is that the learning curve is a bit long
> and the order and sequence of operations is far from intuitive.

You are right ! A lot of the terminology is 'odd'. No offense intended, but
I think it is because it was written by a ham guy !!

William P.N. Smith <(E-Mail Removed)> wrote in
news:(E-Mail Removed):

> DanS <(E-Mail Removed)> wrote:
>>The final outcome though is really that connector connections do not
>>introduce nearly as much loss as expected, and having an adapter or 2
in-
>>line will most likely make no detectable difference.
>
> Wouldn't the fair test be a single piece of coax at least many
> wavelengths long, with a single connector pair introduced into the
> middle of it? I'd think that the reason you can have all those
> connectors stuck together with so little loss is that they interfere
> with each other because they are so close. For instance, a capacitive
> section near an inductive section (if close enough together) might
> look less lossy than either of them individually.
>

I can do that, I'd like to see the results myself.

The supply of smaller length cables is limited though. We have little on-
hand, only 8" RG58 w/BNC's, pre-assembled crimps. Other than that, I
don't think I want to try the super-cheesy mini-RCA cables made of RG188
(?).

It will be next week. I did two half-days and am off tomorrow. (Sometimes
life is nice....rarely....but sometimes.)

Regards,

DanS

DanS <(E-Mail Removed)> wrote:
>William P.N. Smith <(E-Mail Removed)> wrote in
>> Wouldn't the fair test be a single piece of coax at least many
>> wavelengths long, with a single connector pair introduced into the
>> middle of it?

>I can do that, I'd like to see the results myself.
>
>The supply of smaller length cables is limited though. We have little on-
>hand, only 8" RG58 w/BNC's, pre-assembled crimps.

You don't want little short lengths, you want long lengths to swamp
out the impedance discontinuities. Take a 10-foot cable, measure it,
cut it in half, and put a male and female BNC on either half...

>It will be next week. I did two half-days and am off tomorrow. (Sometimes
>life is nice....rarely....but sometimes.)

Take your time, and remember how much you are getting paid for this.
8*)