Author Topic: Testing RF connectors and cables  (Read 7655 times)

0 Members and 1 Guest are viewing this topic.

Offline orin

  • Frequent Contributor
  • **
  • Posts: 444
  • Country: us
Re: Testing RF connectors and cables
« Reply #75 on: October 24, 2019, 06:01:50 am »

The  capacitive discontinuity at the cable/connector interface is frequencies above 12 GHz.  So the  Chinese APC-7/N-F adapters I received are certainly acceptable up to 3 or 6  GHz.  Which is not to say the ones @orin received are.


I could always send them to you along with the F-F adapter that measured even worse.  My curiosity far out-weighs the cost of a priority mail box.

I have an N connector gauge arriving Monday, so they'll be available sometime next week after I get around to checking them.

Orin.
 

Offline rhb

  • Super Contributor
  • ***
  • Posts: 2815
  • Country: us
Re: Testing RF connectors and cables
« Reply #76 on: October 24, 2019, 02:21:32 pm »

The  capacitive discontinuity at the cable/connector interface is frequencies above 12 GHz.  So the  Chinese APC-7/N-F adapters I received are certainly acceptable up to 3 or 6  GHz.  Which is not to say the ones @orin received are.


I could always send them to you along with the F-F adapter that measured even worse.  My curiosity far out-weighs the cost of a priority mail box.

I have an N connector gauge arriving Monday, so they'll be available sometime next week after I get around to checking them.

Orin.

Mine also.  I'll be glad to split the shipping cost and pay for the return postage.  I'm really looking forward to getting my 8753B/85046A set up so I can do S11 and S21 in both the time and frequency domains.

Once I can get digits out of the instruments I'm going to write a FOSS  TDR/VNA program.
 

Offline rhb

  • Super Contributor
  • ***
  • Posts: 2815
  • Country: us
Re: Testing RF connectors and cables
« Reply #77 on: October 26, 2019, 11:54:23 pm »
I got an APC-7 cable today and an HP APC-7 O/S. My APC-7 loads are still in transit.  Here is TDR using the APC-7 cable and one of the $16.50 APC-7/N-F adapters.

I'm now using an SMA-M/APC-7 adapter and the APC-7 cable.  This gives me >11 GHz BW to the end of the cable and back.  I have 26 ps at the unterminated 3.5 mm connector and 31 ps at the cable end.

The first image shows the unterminated APC-7 at the left, next are the Chinese APC-7/N-F and  N-M O/S followed by the Anritsu O/S.  The electrical length of the APC-7/N-F is 172 ps.  The rising trace with slightly greater overshoot than the partially overlying highlighted trace is the APC-7/N-F unterminated.

861796-0

The second image shows the delay difference between the Chinese and Anritsu O/S reference planes which is 40 ps.  The reference plane for the Chinese O/S is the end of the N-F center conductor which I measure with calipers to be ~4 mm from the end of the center conductor.  The reference plane for the Anritsu is the face of the N-F outer shell.

Other than the reference plane location I don't see a dramatic difference between the  $18 Chinese O/S and the Anritsu which sells for 10x the price.  However, the load is another matter.  But I'll leave that for later when I have my APC-7 HP 909A loads.

861800-1


Have Fun!
Reg
« Last Edit: October 27, 2019, 12:00:57 am by rhb »
 

Offline orin

  • Frequent Contributor
  • **
  • Posts: 444
  • Country: us
Re: Testing RF connectors and cables
« Reply #78 on: October 29, 2019, 07:13:00 pm »
The N connector gage arrived and I took the Chinese adapters home to measure...

The kit came with a .197/.223 master gage, so measurements are relative to it.

One adapter N-F measured +0.007 and the other +0.009.  So 0.002" difference in location of the F contact.  These measurements are repeatable so I'm confident in the 0.002" difference.

The uncertainty is 0.0024 according to the Maury datasheet, so it's possible that one protrudes too far... I'll have to see how the HP adapters measure in comparison.
 

Offline rhb

  • Super Contributor
  • ***
  • Posts: 2815
  • Country: us
Re: Testing RF connectors and cables
« Reply #79 on: October 29, 2019, 11:12:04 pm »
Would you post some pictures of the gauge?  I've got several lathes, so making one is not a big issue.  And far cheaper than buying them.  Especially as I have several  0.0001" indicators, so all I'd need to do is make a piece to clamp on the stem.

I'd like to see how they did the dial indicator to RF connector interface.

Reg
 

Offline orin

  • Frequent Contributor
  • **
  • Posts: 444
  • Country: us
Re: Testing RF connectors and cables
« Reply #80 on: October 30, 2019, 12:09:38 am »
Would you post some pictures of the gauge?  I've got several lathes, so making one is not a big issue.  And far cheaper than buying them.  Especially as I have several  0.0001" indicators, so all I'd need to do is make a piece to clamp on the stem.

I'd like to see how they did the dial indicator to RF connector interface.

Reg

Page 22 here: https://www.maurymw.com/pdf/datasheets/2Y-051.pdf or ebay #254392757267 or #143315065493 will give you an idea of what they look like.  The master gage (Maury Microwave's spelling) is the knurled piece.  Mine has a .223" hole and .197" protrusion.  I'd have preferred .207 for both... Shouldn't be too hard to make myself one on the mini lathe though.

The end of the indicator shaft is bored (I assume) to accept a male pin though no doubt you could machine an extension for a regular indicator.

There is one piece stays on the dial gauge, then there are two other pieces that you have to swap for M/F connectors.  Small set screws hold the pieces to the indicator.

I'll try to take pics of mine tonight.

 

Offline orin

  • Frequent Contributor
  • **
  • Posts: 444
  • Country: us
Re: Testing RF connectors and cables
« Reply #81 on: October 30, 2019, 10:52:46 pm »
Pictures attached.

From the left - adapter for measuring F connectors, adapter for measuring M connectors, master gage (protrusion mimics the N F 'pin', the hole the shoulder of a male pin).

863990-0

Basically, you add the adapter to the indicator, push the master gage on and zero the indicator.  Then push the connector of interest on and read the pin position relative to the master gage.

I also took one of the F-F adapters that measured worse on the 8753C home.  It showed 0.000 and -0.0015 relative to the master gage, so the F pin is recessed by 0.010" from the N standard.  This is likely to avoid problems mating with low quality connectors... but the return loss isn't great when you get to 3GHz.
« Last Edit: October 31, 2019, 12:54:33 am by orin »
 

Offline rhb

  • Super Contributor
  • ***
  • Posts: 2815
  • Country: us
Re: Testing RF connectors and cables
« Reply #82 on: October 31, 2019, 05:37:59 pm »
Thanks for the pictures.  Those should be easy to make on the mini-lathe.  A gauge for checking the diameter of the male pins is more difficult.   A boring bar small enough is *very* fragile.  It might be easier to modify the tips of a caliper to allow reaching into a connector shell.

One can buy precision tapered hole gauges for a few bucks to check the female sockets.  So that is easy.

My HP 909A APC-7 loads came yesterday, so I'll try to set up the 8753B/85046A and learn to use it.  Once I  can get digits out of the 11801 and 8753B I'll be able to directly compare TDR and VNA measurements of the same devices via Fourier transform.

I've gotten quite interested in the possibility of measuring and correcting for the imperfections of cal kits by solving an inverse problem.  Solving an analogous problem of removing multiple  reflections in the water column is routine in seismic work.

Also while it is routine to acquire impulse responses for the instruments and source, typically those can't be found when reprocessing old data.  So the art of estimating those from the data is highly refined.
 

Offline rhb

  • Super Contributor
  • ***
  • Posts: 2815
  • Country: us
Re: Testing RF connectors and cables
« Reply #83 on: November 03, 2019, 05:29:00 pm »
I did a quick comparison of various SMA-M jumpers to answer a question the the TekScopes groups.io list.

First  some short high quality jumpers and a Chinese ebay jumper.  Top trace is the bent solid tube then the black 90 degree and the Chinese solder filled braid at the bottom.  In particular note the amplitude of the connection to the SD-24.  Note that the connector on the Chinese jumper is better than the black jumper, but not as good as the salvage jumper.

865818-0

Then a pair of longer RG402 soldered braid cables.  The blue plastic covered jumper is the bottom trace in both the reflected and the thru pairs.  While not shown, the rise time on both the thrus is ~31 ps, so slightly slower than the rise time of the SD-24 pulser and the SD-26 used for the thru.  The SD-24 is ~29 ps for the upper channel and ~25 ps for the lower channel and the inputs open.
865822-1

Here's what they all look like.

865814-2

Edit:  It turns out I had the speed of the channels backwards.  Here are photos showing the heads with the input connectors open.  The first is the full step response.  The 2nd is just the reflection from the open 3.5 mm connectors.  The upper input is highlighted.  The time between the steps is 2x how long it takes for the signal go go from the sampling point to the connector.
« Last Edit: November 03, 2019, 06:41:27 pm by rhb »
 

Offline rhb

  • Super Contributor
  • ***
  • Posts: 2815
  • Country: us
Re: Testing RF connectors and cables
« Reply #84 on: April 12, 2020, 06:51:41 pm »
 i finally got some APC-7/SMA-F connectors from China.  One has problems with the threaded sleeve not extending.  Both appear to have been assembled using something similar to Loctite, so I'm not able to fix the sleeve issue.

For comparison I have some new Weinschel APC-7/SMA-F adapters.

968446-0

from the bottom:

Chinese RG402 SMA-M to SMA-M open

with Wienschel adapter open

with HP 909A APC-7 50 ohm load.


The vertical lines on the lower two traces show the plane of the open. Obviously the adapter is better than the cable termination.

The Chinese APC-7/SMA-F are clearly not as good.  Though they look as if they would be usable on an 8753x up to 6 GHz.

968450-1

The traces are as the Weinschel case, just different APC-7/SMA-F adapter.

I got 4 Weinschels for $100 and 2 of the Chinese for $30.

It's a bit harder to grasp what is going on with the Chinese stuff as a capacitive bump appears when the load is connected which is not there when it is open.  The fingers on the center conductor stick out farther than on the 909A loads or the Weinschel adapters.

Once I get all the cabling for my 8510C I'll update the TDR displays with VNA sweeps.

Edit:  I just noticed that the photo for the Chinese adapter has the bottom trace at 10 mV/div.  The other traces are 20 mV/div.  The capacitive reflection is about 0.04 for both the SMA-M on the RG402 cable and the -0.04 for the adapter.  Not great, but usable below 6 GHz.

I was able to disassemble the adapter with the non-working sleeve and fix it.  They appear to have been epoxied.  So a hot air gun should simplify taking them apart.

These are obviously not a good deal relative to the Weinschels, but you might have to wait a long time to get the sort of deal I found.

Addendum:

I just measured the rise time going through the two Chinese AP-7/SMA-F adapters and two Weinschels and both have a thru rise time of 35 ps.  With a single20"  RG402 jumper I measure 30 ps.

Reg
« Last Edit: April 12, 2020, 11:42:44 pm by rhb »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf