Author Topic: Testing RF connectors and cables  (Read 16692 times)

0 Members and 1 Guest are viewing this topic.

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Testing RF connectors and cables
« on: August 24, 2019, 10:43:54 pm »
I was testing a bunch of SMA-F to N-F bulkhead adapters from ebay yesterday.  The neat thing about the SD-24 is you can test such things as fast as you can make the connections.  I was putting a very good Inmet N-M load on the N-F side.  The adapters came from wifi_expert.

Bear in mind In the time domain I'm testing to 20 GHz, but in the frequency domain I can only sweep to 2.9
GHz.  Notice that the TDR lets you see the SMA connection and the N connection separately.

TDR using a Tek 11801 w/ SD-24 20 GHz head

   Top is unmarked, but thought to be Suhner
   Middle is Astrolab
   Bottom is Chinese

819090-0

Next is a thru cal on my 8560A from 500 KHz to 500 MHz at 1 dB/div using an Astrolab part

819096-1

Chinese with same settings and using the Astrolab as the cal reference.

819102-2

Next is a thru cal on my 8560A from 500 KHz to 2.9 GHz at 1 dB/div using a Suhner (?) part

819114-3

Chinese with same settings using the Suhner as reference cal

819108-4

I'm building an RF switch deck to automate a full annual cal run of my bench.  I'll be using AR488 and Arduino Mega 2560s to control ST6P Radiall SMA relays for RF  and 44421A relay cards from a 3497A for DC, AC & resistance measurements. 

The Chinese SMA-F to N-F bulkhead adapters won't cut it for the cal system, but I'm also building an antenna and radio switch deck using another pair of the Radial relays.  Except for playing with an SDR and an abandoned Direct TV dish, I won't be doing anything above the 70 cm band.  I might use better grade adapters for those, but for HF/VHF/UHF the Chinese are just fine.

For those for whom an 11801 is out of reach, Leo Bodnar will provide his excellent BNC <40 ps square wave generator with a 1 MHz instead of 10 MHz period.  Naturally that limits you to the BW of your scope, but  it will test any connectors and cables you use with your scope to to full BW.

Just as you can do TDR on modern VNAs, you can do VNA by TDR with a DSO and computer.  I plan on dong some examples and writing software to do that. 

But I'm a severe ADD case, so no idea when that will happen. I've got a 7x12 Chinese mini-lathe awaiting further attention for making high precision RF stuff.  But first I have to rebuild the lathe to work to those tolerances.  That's a *lot* of work.  The machines are quite impressive, but it's really just a complete kit of parts which has been assembled to make sure they didn't forget something.  So for serious work they have to be completely disassembled, cleaned, critical parts replaced and then scraped by hand into alignment.  The hard part is learning to measure things accurately enough.
« Last Edit: August 24, 2019, 11:11:20 pm by rhb »
 
The following users thanked this post: bitseeker, Neomys Sapiens

Offline radiolistener

  • Super Contributor
  • ***
  • Posts: 3282
  • Country: ua
Re: Testing RF connectors and cables
« Reply #1 on: August 24, 2019, 11:26:41 pm »
Usually Chinese connectors works good when it's new, but after some time it is oxidized and it leads to weak connection. I'm cleaning it with ethanol, after that it works better.
 

Offline bitseeker

  • Super Contributor
  • ***
  • Posts: 9057
  • Country: us
  • Lots of engineer-tweakable parts inside!
Re: Testing RF connectors and cables
« Reply #2 on: August 24, 2019, 11:31:23 pm »
I've had some cheap BNC tee connectors fall apart. :palm:

As Dave says, "Built down to a price."
TEA is the way. | TEA Time channel
 

Offline joeqsmith

  • Super Contributor
  • ***
  • Posts: 11632
  • Country: us
Re: Testing RF connectors and cables
« Reply #3 on: August 25, 2019, 12:38:43 am »
I've had some cheap BNC tee connectors fall apart. :palm:

As Dave says, "Built down to a price."

Same.  I have had the lock fall off of the male ones.

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: Testing RF connectors and cables
« Reply #4 on: August 25, 2019, 01:05:31 am »
BNCs are particularly problematic.  I'm finding that it's easy to induce large reflections just by pressing on the male part.  The serious problem is that it doesn't always go away when you let go.

After my post to the FY600 thread I did some more testing and do *not* like the results.  I'm not sure what is going on yet.  I suspect it is just crappy BNCs, but I've got to find some that don't exhibit the problem to be sure it's not just inherent in the design of the BNC.

FWIW the little "w" at the end of the reflections in the TDR trace is the Inmet load.  On the Suhner and Astrolab adapters, there is a short flat spot separating the SMA and N connections where the adapter is purely resistive, but the Chinese is capacitive.   The level of detail you can see in the plots is amazing.  I'm tempted to cut a Chinese adapter in half lengthwise to compare it to the TDR trace.
 

Offline radiolistener

  • Super Contributor
  • ***
  • Posts: 3282
  • Country: ua
Re: Testing RF connectors and cables
« Reply #5 on: August 25, 2019, 01:09:26 am »
I also tried Chinese BNC connectors, but they almost unusable. They have very weak connection. So, now I'm using mainly SMA connectors with SMA-to-BNC adapters.
 

Offline xrunner

  • Super Contributor
  • ***
  • Posts: 7498
  • Country: us
  • hp>Agilent>Keysight>???
Re: Testing RF connectors and cables
« Reply #6 on: August 25, 2019, 01:24:55 am »
I've heard hams say over the years that putting an adapter in to extend your coax run was a bad idea. They'd say that it would cause 0.5 dB or 1 dB loss in the HF bands. Where they got the number I had no idea. So a few weeks ago I decided to test their claims. I put 7 various adapters (not just one) in series and ran some tests.

As you can see from the charts it was all an old wives tale.  :-DD Even in the 2 meter band it's not a big deal.

Now I have the data to call them out.
I told my friends I could teach them to be funny, but they all just laughed at me.
 

Offline 0culus

  • Super Contributor
  • ***
  • Posts: 3032
  • Country: us
  • Electronics, RF, and TEA Hobbyist
Re: Testing RF connectors and cables
« Reply #7 on: August 25, 2019, 02:09:25 am »
At those frequencies I'd be very surprised if it made any noticeable difference in operation. UHF would probably be a different story though.
 

Offline TheSteve

  • Supporter
  • ****
  • Posts: 3743
  • Country: ca
  • Living the Dream
Re: Testing RF connectors and cables
« Reply #8 on: August 25, 2019, 02:19:49 am »
Think of old school PL-259's with an SO-239 barrel between them, crappy crimps to the coax or terrible soldering, poorly installed and not super tight, been outside in the weather for years without proper sealing. Then it isn't too hard to lose half a dB even at HF frequencies.

rhb - I like the TDR pics, do you have a nice way to translate the images to return loss in dB?
VE7FM
 

Offline 0culus

  • Super Contributor
  • ***
  • Posts: 3032
  • Country: us
  • Electronics, RF, and TEA Hobbyist
Re: Testing RF connectors and cables
« Reply #9 on: August 25, 2019, 02:53:55 am »
Think of old school PL-259's with an SO-239 barrel between them, crappy crimps to the coax or terrible soldering, poorly installed and not super tight, been outside in the weather for years without proper sealing. Then it isn't too hard to lose half a dB even at HF frequencies.

rhb - I like the TDR pics, do you have a nice way to translate the images to return loss in dB?

Fair enough...I'm spoiled I guess.  :-DD
 

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: Testing RF connectors and cables
« Reply #10 on: August 25, 2019, 05:54:28 am »
Give me digits and it's easy.  The hard part is getting digits.

I *think* I know how to do it by inspection.  But I haven't  convinced myself I've got it right yet.

It's a couple of morning sessions of serious math,   Not hard, but I've got to be absolutely running at 100%.  I've just not been up to that level of late because of the vicissitudes of getting older.  Time eventually gets us all.

PL-259s are an abomination.
 

Offline bitseeker

  • Super Contributor
  • ***
  • Posts: 9057
  • Country: us
  • Lots of engineer-tweakable parts inside!
Re: Testing RF connectors and cables
« Reply #11 on: August 26, 2019, 12:03:57 am »
Shhh. Don’t say that stuff too loudly.  :-DD
TEA is the way. | TEA Time channel
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 16548
  • Country: us
  • DavidH
Re: Testing RF connectors and cables
« Reply #12 on: August 26, 2019, 04:05:19 am »
BNCs are particularly problematic.  I'm finding that it's easy to induce large reflections just by pressing on the male part.  The serious problem is that it doesn't always go away when you let go.

After my post to the FY600 thread I did some more testing and do *not* like the results.  I'm not sure what is going on yet.  I suspect it is just crappy BNCs, but I've got to find some that don't exhibit the problem to be sure it's not just inherent in the design of the BNC.

BNC quality varies considerably but bayonet style connectors all wear more quickly than threaded connectors.  TNCs are amazingly better than BNCs but unfortunately never really caught on for use in test instruments and would not be suitable for oscilloscope probes anyway.

Some companies did make clamp and thread style BNC male connectors for use with standard female BNC connectors on test instruments but they are cumbersome.  They were used where a more secure mechanical connection or more reliable long term high frequency performance was required.  Some active probes used them.

For smaller connectors which are not SMA, I really like SMB as an easy to use push-on connector but they never caught on either.

Quote
FWIW the little "w" at the end of the reflections in the TDR trace is the Inmet load.  On the Suhner and Astrolab adapters, there is a short flat spot separating the SMA and N connections where the adapter is purely resistive, but the Chinese is capacitive.   The level of detail you can see in the plots is amazing.  I'm tempted to cut a Chinese adapter in half lengthwise to compare it to the TDR trace.

VNAs provide more electrical detail but TDR is easier to relate to physical construction and faster.  You cannot do it with an intact connector but a trick for use with TDR on exposed circuits is to run a pencil along the circuit to relate time to position.
« Last Edit: August 26, 2019, 04:09:09 am by David Hess »
 
The following users thanked this post: Neomys Sapiens

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: Testing RF connectors and cables
« Reply #13 on: August 26, 2019, 03:47:05 pm »

VNAs provide more electrical detail but TDR is easier to relate to physical construction and faster.  You cannot do it with an intact connector but a trick for use with TDR on exposed circuits is to run a pencil along the circuit to relate time to position.

A minor quibble, both have the same information content assuming appropriate acquisition parameters.  Depending upon the task one may be easier to interpret than the other.

In seismic processing one goes between time and frequency so often that most experienced people know what the other domain looks like by inspection.  Knowing the major transform pairs in the pictorial dictionary in Bracewell is very useful.
 

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: Testing RF connectors and cables
« Reply #14 on: August 26, 2019, 04:55:18 pm »
First some fooling around with the nifty 0.1-3000 MHz RF bridges sold by 60dB.com.  First the setup.  I've got an SD-24 feeding a ~20 ps rise time square wave to the INP terminal.

820374-0

The DUT and REF terminals are connected to the same 20 GHz SD-26 head (an SD-24 w/o the pulser).   The OUT is connected to another SD-26.  The DUT and REF ports match very closely at 118 mV.  The OUT port pulse is around 900 uV.   So if I haven't botched my math, the OUT isolation is -42 dB from the REF port.  Note that I have not time aligned the OUT port head with the DUT & REF port head.  I should have thought to do that.

820380-1

Finally a sweep on my 8560A from 1-2900 MHz at 2 dB/div

820386-2
« Last Edit: August 26, 2019, 05:00:30 pm by rhb »
 

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23018
  • Country: gb
Re: Testing RF connectors and cables
« Reply #15 on: August 26, 2019, 05:05:09 pm »
I've heard hams say over the years that putting an adapter in to extend your coax run was a bad idea. They'd say that it would cause 0.5 dB or 1 dB loss in the HF bands. Where they got the number I had no idea. So a few weeks ago I decided to test their claims. I put 7 various adapters (not just one) in series and ran some tests.

As you can see from the charts it was all an old wives tale.  :-DD Even in the 2 meter band it's not a big deal.

Now I have the data to call them out.

The PL259 HF insertion loss myth probably came about from the average ham's inability to put a PL259 together without fucking it up plus the fact that nearly all PL259 connectors, apart from the really high end Amphenol ones are complete garbage, plus they probably used rotten mouldy water soaked coax they got from a hamfest as a "bargain".

Literally yesterday I saw someone eyeing up a roll of coax which had rodent nibblings on it  :palm: ... couple of hundred watts up that and I bet the new rat piss dielectric smelled marvelous.

At HF / VHF I am measuring around 0.3dB loss worst case per 10m on milspec RG58U (foam dielectric + really thick braid) with Amphenol clamp on BNCs. Above VHF, I do not do. Yet.
« Last Edit: August 26, 2019, 05:08:20 pm by bd139 »
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 16548
  • Country: us
  • DavidH
Re: Testing RF connectors and cables
« Reply #16 on: August 26, 2019, 05:11:53 pm »
VNAs provide more electrical detail but TDR is easier to relate to physical construction and faster.  You cannot do it with an intact connector but a trick for use with TDR on exposed circuits is to run a pencil along the circuit to relate time to position.

A minor quibble, both have the same information content assuming appropriate acquisition parameters.  Depending upon the task one may be easier to interpret than the other.

A few years ago one of the RF or electronic trade magazines did a comparison of state of the art TDR and VNA instruments.  VNAs had an edge in bandwidth, dynamic range, and accuracy but TDRs were faster to use.  The instruments they reviewed could display results in either format.  HP made some sampling based instruments which combined oscilloscope, VNA, and TDR capabilities but they remained a specialized instrument.
 

Offline xrunner

  • Super Contributor
  • ***
  • Posts: 7498
  • Country: us
  • hp>Agilent>Keysight>???
Re: Testing RF connectors and cables
« Reply #17 on: August 26, 2019, 05:29:36 pm »
Think of old school PL-259's with an SO-239 barrel between them, crappy crimps to the coax or terrible soldering, poorly installed and not super tight, been outside in the weather for years without proper sealing. Then it isn't too hard to lose half a dB even at HF frequencies.

The PL259 HF insertion loss myth probably came about from the average ham's inability to put a PL259 together without fucking it up plus the fact that nearly all PL259 connectors, apart from the really high end Amphenol ones are complete garbage, plus they probably used rotten mouldy water soaked coax they got from a hamfest as a "bargain".

Literally yesterday I saw someone eyeing up a roll of coax which had rodent nibblings on it  :palm: ... couple of hundred watts up that and I bet the new rat piss dielectric smelled marvelous.

At HF / VHF I am measuring around 0.3dB loss worst case per 10m on milspec RG58U (foam dielectric + really thick braid) with Amphenol clamp on BNCs. Above VHF, I do not do. Yet.

Absolutely, the quality and condition of the adapters makes a big difference. The next test would be to do it again with some crappy and old adapters. Unfortunately (for the testing not for myself), I only have good stuff on hand.  8)

However, I do know some hams whereby I could possibly obtain the needed crappy connectors.  :-DD



In case anyone's interested, the adapters used in series -

BNC-F | N-M--> N-F| N-F--> N-M | U-F--> U-M | U-M--> U-F | N-M--> N-F | N-F--> N-M | BNC-F
I told my friends I could teach them to be funny, but they all just laughed at me.
 

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: Testing RF connectors and cables
« Reply #18 on: August 26, 2019, 05:57:04 pm »
Here's TDR on an Anritsu SOL N-M calibrator. 

Top image is a comparison of the Inmet load I used in the first post on top and the Anritsu load on the bottom.  The load is the hardest part of a cal kit to get right.  There really is a difference.

820422-0

Next is the rise time of the open

820410-1

And the fall time of the short

820416-2

One of the really nice things about TDR is you  can see the signal as it passes through the transmission line.  So you see the SMA-M cable and the SMA-F to N-F connector mirror reversed either side of the reflections from the open and short.

FWIW I weaseled my statement about information equivalence with "appropriate acquisition parameters".  My point simply being that a *properly* done Fourier transform will get the same result in the other domain.  For a wide range of reasons the engineering of a TDR and VNA would be rather different and I'd be surprised if there was a one to one match in what any comparison of a TDR and VNA showed.

As for connectors in coax runs, I think it's really more a comment on connector aging which has been subjected to the "telephone game".  I'd like to note that when I redid the mast connections on a friend's sailboat I coated the connectors with silicone grease before applying  a couple of layers of heat shrink.  The boat's been sold at least twice since then and last I heard was headed for Tahiti.  I'd love to know how the connections held up as it's been over 15 years.

I'll try to do some TDR work with a DSO and one of Leo's 1 MHz square wave pulsers here shortly, followed then by VNA via pulser and DSO.  Admittedly, the appearance of the nanoVNA makes that less important, but I've got an Owon XDS2102A which collects 20 Mpts of 12 bit data.  So with a bit of slicing and dicing I should be able to get  very good dynamic range.  But I got a shiny new toy on Thursday, so I've got to play with it a bit ;-)

Have Fun!
Reg
« Last Edit: August 26, 2019, 05:59:48 pm by rhb »
 
The following users thanked this post: Neomys Sapiens

Offline TheUnnamedNewbie

  • Super Contributor
  • ***
  • Posts: 1208
  • Country: 00
  • mmwave RFIC/antenna designer
Re: Testing RF connectors and cables
« Reply #19 on: August 27, 2019, 04:49:20 am »

FWIW I weaseled my statement about information equivalence with "appropriate acquisition parameters".  My point simply being that a *properly* done Fourier transform will get the same result in the other domain.  For a wide range of reasons the engineering of a TDR and VNA would be rather different and I'd be surprised if there was a one to one match in what any comparison of a TDR and VNA showed.


This is actually something certain VNAs use a lot. If you are doing 'simple' one-port measurements, you can sometimes get away with almost no cal by doing a time-domain de-embed. What it will do is simply convert the f-dom trace to t-dom, chop off the first bit (as that would be your DUT connector you want to de-embed) and then go back to the f-dom.

Pretty cool stuff. Would be neat to see some things on cables, but I think that is hard to see with a TDR? I've done some measurements of cheap coax at the lab, and it was not pretty... But don't have any real results saved. I should re-do that some time.
The best part about magic is when it stops being magic and becomes science instead

"There was no road, but the people walked on it, and the road came to be, and the people followed it, for the road took the path of least resistance"
 

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: Testing RF connectors and cables
« Reply #20 on: August 28, 2019, 12:27:05 am »
This is an entirely random series of experiments.  A lot of it is driven by what I have on hand or some task such as checking the quality of the SMA-F to N-F adapters in the first post.

I did some tests with one of Leo Bodnar's pulsers and a 200 MHz DSO which I'll post eventually.  It's not as impressive as having 20 GHz BW, but still useful.  Though the appearance of the nanoVNA may well change things.  I've ordered one.  At $50 delivered with tax it was impossible not to try one out.
 

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23018
  • Country: gb
Re: Testing RF connectors and cables
« Reply #21 on: August 29, 2019, 11:30:08 pm »
Random experiment I did this evening. I ran out of continuous lengths of RG58U greater than 10m and I needed a 20m chunk so I figured I'd try something stupid.

I went "full CB troll", stripped 2 inches off each end and rammed a choc block in the middle of the two 10m lengths, wrapped it up in cheap insulation tape and chucked it on my RigExpert analyser.

VSWR was flat to 22MHz and at 55MHz it was still 1.07 (~29dB RL) compared to a baseline measurement of 1.0 on a properly terminated patch of the same length.

Ergo YMMV, at least if you don't push into VHF.

And no I'm not leaving it like that. I've got a 50m roll turning up on Monday :D

I shall leave the real measurements to those who can afford car priced VNA's :D
« Last Edit: August 29, 2019, 11:32:53 pm by bd139 »
 
The following users thanked this post: xrunner

Offline rhbTopic starter

  • Super Contributor
  • ***
  • Posts: 3476
  • Country: us
Re: Testing RF connectors and cables
« Reply #22 on: August 31, 2019, 03:20:04 am »
You need to start following the nanoVNA as it develops.  I got one today and am about to start beating on it.  But initial take is it's pretty cool even if it is a bit rough around the edges.  But for $50 delivered in 3 days it's quite impressive.

I should note that the SMA PCB connector snapped all 4 pins when I was tightening the short to do a cal using a Suhner torque wrench.  So if you get one, repackage it with more robust connections.

Have Fun!
Reg
 

Offline OwO

  • Super Contributor
  • ***
  • Posts: 1250
  • Country: cn
  • RF Engineer.
Re: Testing RF connectors and cables
« Reply #23 on: August 31, 2019, 05:13:31 am »
just got info that nanoVNA V2 will have frequency coverage to 2.5GHz and retail for the same price
Email: OwOwOwOwO123@outlook.com
 

Offline bd139

  • Super Contributor
  • ***
  • Posts: 23018
  • Country: gb
Re: Testing RF connectors and cables
« Reply #24 on: August 31, 2019, 09:06:57 am »
Interesting. I looked at some of the "amateur radio grade" VNAs on the market a while back and was unimpressed for the price. microVNA looks interesting though.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf