Author Topic: Advantest R3131A Spectrum Analyzer 9KHz-3GHz  (Read 90035 times)

0 Members and 1 Guest are viewing this topic.

Offline jc81382

  • Regular Contributor
  • *
  • Posts: 68
  • Country: us
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #75 on: February 03, 2016, 11:16:31 pm »
nctnico,

Thanks for the idea. I just tried it and all it did was reset the unit, but worth a try.

I decided to take out, what I assume to be the RF Board, to inspect this "dielectric switch" that I keep seeing in the pictures.

Now I am not a RF or Electronics technician, I graduated from a tech school for Electronics, I work at a Calibration Lab and I like to tinker, so a lot of this is over my head. But if anyone can explain to my why the position of these switch's seems to be different on every unit and how adjusting them may or may not help my situation that would be greatly appreciated.

Attached is a pic of the position I found my unit it.

Thanks
 

Online T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 15926
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #76 on: February 04, 2016, 12:14:28 am »
A structure like that would be for fine tuning a filter's frequency, bandpass, matching, etc.  It might compensate for tolerances in the PCB fabrication, or variations in device impedance.

I think you'd want to run a full E&M sim on it, just to be sure.  The range of effects it can have is quite broad (its position varies by electrical angle along the transmission line structures, as well as contributing loading impedance with those isolated chits, which are at some distance as well, so will affect the response in a complicated way again).

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Online nctnico

  • Super Contributor
  • ***
  • Posts: 19980
  • Country: nl
    • NCT Developments
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #77 on: February 04, 2016, 12:32:14 am »
The calibration signal being absent can have any number of reasons. The only way to find out is by determining where the calibration signal is generated and follow it from there. The structure on the picture is not a switch but some kind of adjustment so don't touch that.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline jc81382

  • Regular Contributor
  • *
  • Posts: 68
  • Country: us
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #78 on: February 04, 2016, 12:41:36 am »
Thanks guys,

I noticed that the "cal out" has a 30 mHz -20db printed on the housing. I assume that is the standard output.

Maybe I will take the unit to work and try to supply 30 mHz -20 db into the RF in and run the "auto cal" program. I assume that as long as the input sees that input, it should adjust/offset. I know the RF in works because although it failed calibration when we attempted it at work it was getting readings. So maybe it is just the cal out that is bad.

I can also see if I get a reading on another Spec A from the cal out when I initiate the auto cal.

I think that will be the extent of my abilities.

Thanks again, if anyone has any other suggestions it is greatly appreciated.
 

Online nctnico

  • Super Contributor
  • ***
  • Posts: 19980
  • Country: nl
    • NCT Developments
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #79 on: February 04, 2016, 12:44:59 am »
Wait a minute... you have a model with a cal out connector! That should be connected to the input when doing a self calibration!
In the R3131A that connection is made internally but in your spectrum analyser you'll have to make an external connection.
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline jc81382

  • Regular Contributor
  • *
  • Posts: 68
  • Country: us
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #80 on: February 04, 2016, 01:22:28 am »
yeah, I know... Unless i have a bad cable, its as if the cal out is not putting out a signal or the RF in is not sensing it.

I assume the RF in works because we were able to get it to read a signal when fead form an external source when we attempted calibration (not the internal calibration but they type provided by a Calibration Lab), it was just OOT.

So that leaves me to assume the issue is with the cal out.

When i run the self cal I am still getting a RF Fail. So does the system know that there is an issue with the RF board? I have no idea.
 

Online Fraser

  • Super Contributor
  • ***
  • Posts: 9967
  • Country: gb
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #81 on: February 04, 2016, 11:41:32 am »
Some comment on this........

First check that the calibration signal is present at the output port. It should be the frequency and amplitude stated in the spec. The calibration signal is unlikely to fail on these units. It should still be rested though.

The PCB Dielectric trimmer is part of the 2nd local oscillator VCO tuned circuit, and sets the correct centre frequency around which the Local Oscillator may be tuned by the phase locked loop. There are a pair of varicap diodes under control of the LO PLL that adjust the frequency as required whilst the SA is operating.

When the SA starts its self calibration and test routine it checks that it can see the Calibration signal and that the levels are correct. If The LO has drifted too far off its centre frequency, the PLL cannot bring it into lock and so the Calibration signal is not seen on the correct frequency or correct level.

The SA will often still work but as its LO is not working correctly, accuracy will be off.

Correction of this situation involves adjustment of the dielectric trimmer to bring the LO centre frequency back to spec and enabling the PLL to drive it over the correct operating range. Adjustment is done as detailed in my guide which I believe I posted in this thread ?

The adjustment of the dielectric trimmer is very sensitive and it is important to set the centre frequency accurately using an accurate SA or frequency counter connected to the LO monitoring port. It is not recommended that the dielectric trimmer be adjusted without monitoring the frequency as it does not operate in a predictable manner when moved.

I just took a quick look at this thread and did not see my 2nd LO dielectric trimmer adjustment guide. If it's not on the forum I will upload it.

Fraser
« Last Edit: February 04, 2016, 12:54:47 pm by Fraser »
Cogito, ergo sum
 

Online Fraser

  • Super Contributor
  • ***
  • Posts: 9967
  • Country: gb
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #82 on: February 04, 2016, 11:55:07 am »
Just remembered, I uploaded my 2nd LO adjustment guide to the Advantestinstrument Yahoo Group. You can download it from there. Its over 1MB so cannot be attached here.

https://groups.yahoo.com/neo/groups/advantestinstrument/files/-%3DAdvantest%20-%20R3132%3D-/

UPDATE & WARNING

Please read all of this thread. It has become apparent that two versions of R3132 RF module were made. My newer modules use a dropper resistor in series with the 2nd LO MMIC and buffer MMIC supply rails. Earlier versions of the RF Module do not contain this dropper resistor ad use a lower supply voltage to the MMIC's. This thread contains details of teh lower voltage level measured by a fellow forum member.

Fraser
« Last Edit: August 28, 2016, 01:14:46 am by Fraser »
Cogito, ergo sum
 

Offline jc81382

  • Regular Contributor
  • *
  • Posts: 68
  • Country: us
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #83 on: February 04, 2016, 12:21:26 pm »
Fraser,

Thank you so much for the info and for the link. I wasn't able to see it on the Yahoo group, at first, because I wasn't a member of the group originally.

I will ask around at work to see if I can get my hands on some of the necessary tools/equipment to attempt the procedure called out in your adjustment guide. Can you confirm what I should have? I assume another Spec A and a Power Supply?

I am usually willing to try anything, especially since its not really working now I almost have nothing to lose.
 

Online Fraser

  • Super Contributor
  • ***
  • Posts: 9967
  • Country: gb
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #84 on: February 04, 2016, 12:48:15 pm »
You will need a power supply, multimeter and some means to measure the 3805 MHz that comes out of the Local Oscillator monitoring port (SMA socket located on the side of the module).

The power supply should be of good quality and provide an accurate 11.3V for the module. The multimeter can be any digital unit capable of reading the approx 10.3V Varicap diode bias voltage.

Fraser
« Last Edit: February 04, 2016, 12:50:58 pm by Fraser »
Cogito, ergo sum
 

Offline l0wside

  • Regular Contributor
  • *
  • Posts: 118
  • Country: de
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #85 on: February 04, 2016, 12:52:33 pm »
An R3131 is on the way to me, and before I destroy it the very first day, I´d rather ask the experts here before.

The R3131 user´s manual says about the input:
"N-type input connector 50Ohms. Analyzer input connector: Frequency range is 9 kHz to 3 GHz; the maximum input level is +20 dBm or ±50 VDC max"

The last part (in bold) confuses me. At a SA training at Agilent, the trainer had the complete group repeat the phrase "never ever apply DC to an SA input". The R3131 manual suggests otherwise.
What should I believe?

Max
 

Offline jc81382

  • Regular Contributor
  • *
  • Posts: 68
  • Country: us
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #86 on: February 04, 2016, 12:57:54 pm »
Fraser,

Excellent, thank you again.

I have our Electronic Team Leader/Main RF Tech reviewing your procedure make sure I don't miss anything and to give me any pointers.

I will keep you guys posted.

 

Online Fraser

  • Super Contributor
  • ***
  • Posts: 9967
  • Country: gb
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #87 on: February 04, 2016, 01:21:49 pm »
@IOwside

As a spectrum analyser is a test equipment for AC signals you would not normally inject DC into its input.....BUT an AC signal can be sat on a DC bias voltage.

The Advantest SA contains a DC blocking capacitor rated at 50V DC. Hence the specification on the front panel. I also use a higher rated DC blocking adapter when I know DC bias is a possibility. It's just good practice, but the SA will cope with a DC bias up to 50V without damage.

Static discharge is also the enemy of Spectrum Analyzers. Be vet careful if considering connection an SA to a non DC shorted antenna such as a discone. Static discharge can damage your SA's first mixer when running with no input attenuation applied.

Another way to fry your SA input is by injecting a spectrum of signals whose total energy level overloads the SA's input specification. This is especially important if working with spread spectrum or noise generators. The total energy present is not just that of one signal 'spike' , it is the total energy of all signals within the pass-band.

Fraser
Cogito, ergo sum
 

Offline l0wside

  • Regular Contributor
  • *
  • Posts: 118
  • Country: de
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #88 on: February 04, 2016, 01:41:11 pm »
Fraser,

thank you so much. I am aware that measuring DC with an SA does not make any sense, but my application (EMC precompliance on devices with a 30V DC supply) seems to match the R3131´s possibilities quite well (and fits your description "AC sat on a DC bias voltage").

I have ordered a DC blocker anyway (set me back under 7 EUR, direct from China) and will put it in the drawer, just in case.

Max
 

Offline jc81382

  • Regular Contributor
  • *
  • Posts: 68
  • Country: us
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #89 on: February 05, 2016, 03:03:15 pm »
Fraser,

I was just reviewing your procedure with our QA Manager who uses to repair SA's, hoping to get additional help and he noticed something, that i saw before but didn't may much attention to.

If you look closely at the picture in your procedure and a picture I took when I opened my unit up there seems to be an additional trace or component that was added to the circuit. I have attached pictures and circled where this is located.

I also noticed originally that when i went to open the case the first time there were 2 sets of marks on the housing screws (one black and 1 red... all of the other marks are red so I assume that there was a repair done on the unit and they used black marker to mark to position of the screws).

Do you have any idea what this additional trace might be for? Should I consider removing it before temptation your procedure?

Please let me know your thoughts.
 

Online Fraser

  • Super Contributor
  • ***
  • Posts: 9967
  • Country: gb
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #90 on: February 05, 2016, 04:20:43 pm »
You appear to have an earlier version of the module. Fortunately I have also repaired that exact same version as I worked on five of these SA's, all with the same problem. The components added across the varicap diodes are a modification carried out due to issues with tuning the 2nd local oscillator.

My procedure is still applicable and you should not make any changes to the unit.

Best Wishes

Fraser
Cogito, ergo sum
 

Offline jc81382

  • Regular Contributor
  • *
  • Posts: 68
  • Country: us
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #91 on: February 06, 2016, 12:22:32 am »
Fraiser,

As always thank you for the input...

So I was unable to get my hands on another Spec A so I opted for a Frequency Counter, knowing that it would not let me make all of the measurements I needed but I was hopeful I would be able to dial in on the 3.805 Ghz Frequency.

I started off by checking the frequency coming off the Cal Out and it was 29.997...

Then I took everything apart and took my first measurement on the SMA Port and got 3.742...

I the adjusted the pad and got it as close as I could to 3.805 and ended up with 3.797...

I figured since this was as good as I could get i would put it back together and give it a try and it still failed RF Self Test and Auto Cal (as you can see by the red warning in the pic).

At this point my cal out is giving me almost exactly 30 MHz though.

Anything I am missing?
 

Online Fraser

  • Super Contributor
  • ***
  • Posts: 9967
  • Country: gb
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #92 on: February 06, 2016, 12:25:10 pm »
Yes there are two more checks to carry out

1. I monitored the 2nd LO frequency when it was back in the SA to ensure that the PLL brought it onto the correct operating frequency. This is not an easy test however as access to the side mounted monitoring port is obscured by the chassis rails. I solved this by placing as this copper wire into the SMA socket centre pin to act as an aerial. Make sure it approx 0.5 inches long and apart from the connection with the centre pin, insulated from the chassis.
I them used an EMC probe to sniff the 2nd LO signal frequency.

2. I used a meter to measure the 2nd LO PLL control voltage at the detailed feed through point. It should be very close to 0V.

In your case your frequency counter will not be sensitive enough to pick up the signal from my suggested tiny leakage antenna in the monitoring port. If you can use a longer wire and direct connection to the counter input you may get enough signal to count. The monitoring port is fed from a directional coupler pick-off so a longer wire should not cause you any problems and this is not accurate level measurement test anyway.

Test 2 just requires you to place a long test wire on the test point and feed it to the rear of the chassis to Connect to your meter.

Did you measure the voltage on the Varicap diode drive point when you powered the module ? It should have been around 10.33V wrt chassis. Anything else would be evidence of an issue and would make alignment using my simple method void.

If the 2nd L.O is working fine and all checks prove it is on frequency and the PLL is happy, you have another problem in the signal path that we will need to diagnose. 1st Mixer damage is one possibility but hopefully not the case.

Fraser
« Last Edit: February 06, 2016, 02:48:40 pm by Fraser »
Cogito, ergo sum
 

Offline jc81382

  • Regular Contributor
  • *
  • Posts: 68
  • Country: us
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #93 on: February 06, 2016, 02:26:52 pm »
I did not get to check the voltage when it was back in the SA yet, but the voltage at the varicap ranged from 10.3 to 10.4. I checked it multiple times while adjusting the freq.
 

Online Fraser

  • Super Contributor
  • ***
  • Posts: 9967
  • Country: gb
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #94 on: February 06, 2016, 02:50:51 pm »
Varicap voltage looks good. Just check if it is close to 0V on the PLL control line when in the chassis and working. If so, you should have PLL lock.
Cogito, ergo sum
 

Offline jc81382

  • Regular Contributor
  • *
  • Posts: 68
  • Country: us
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #95 on: February 06, 2016, 03:30:55 pm »
Fraser,

Do any of the pictures say where to connect tot make the " Connect a DC voltage meter between the PLL control voltage monitor wire and chassis"?

I have the cover back on and am about to put it back in the SA.
 

Offline jc81382

  • Regular Contributor
  • *
  • Posts: 68
  • Country: us
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #96 on: February 06, 2016, 03:33:33 pm »
Fraser,

Never mind, I see that you saw a few steps before.

It looks like the other side of the blue wire that was feeding the power is the adjustment stage. Sorry
 

Offline jc81382

  • Regular Contributor
  • *
  • Posts: 68
  • Country: us
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #97 on: February 06, 2016, 03:54:09 pm »
Fraiser,

I don't think i'm taking the reading the right place.

The only passthrough that I could see with a blue wired was on the opposite end of the of the board, sandwiched between the top and bottom boards in the "gap area" (for lack of better works. Plus it was on the opposite end of the board  from the Cal Out BNC. When i tried to take a measurement here I got 0V, like absolute 0V like there was no voltage there so I assume I was in the wrong place.

Do you see in my pictures where I should be getting the measurement from? I assume I should be grounding to the chassis?

Thanks
 

Offline jc81382

  • Regular Contributor
  • *
  • Posts: 68
  • Country: us
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #98 on: February 06, 2016, 04:00:03 pm »
Sorry forgot to attache the pics
 

Online Fraser

  • Super Contributor
  • ***
  • Posts: 9967
  • Country: gb
Re: Advantest R3131A Spectrum Analyzer 9KHz-3GHz
« Reply #99 on: February 06, 2016, 04:54:31 pm »
Hi

Take a look at the last picture in my alignment guide. The feedthru is right behind the Calibration socket and has a blue wire coming off of it. Exactly 0V wrt chassis would be unusual but not impossible.

The final confirmation would be to measure the frequency on the 2nd LO monitoring port whilst the module is back in the chassis.

Fraser
Cogito, ergo sum
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf