Advantest R3132 Spectrum Analyzer repair information

[Fraser]

There is already  thread discussing the repair of the Advantest R3131 spectrum analyzer and that also contains R3132 discussions.

I thought it might be useful to start a new thread for the R3132 model so that information on its repair may easily be found with searches.

I am uploading information that I have authored whilst repairing several of these analyzers in 2013. The information relates to the 2nd Local Oscillator as this drifts off tune with age. It is a very common fault that causes a self-calibration error.

I hope this information is of interest and help to others.

I am a member of the Yahoo Advantest Instruments Group and have uloaded these files to that Groups file area already.

Fraser

[Fraser]

R3132 RF module block diagram

[Fraser]

R3132 2nd Local Oscillator pseudo schematics (sketches really)

There are two versions of the 2nd Local Oscillator. The early version is slightly different and uses different supply voltages to those of the later version.

Fraser

[Fraser]

The 2nd Local Oscillator alignment guide that I created from my experience with the later Type 2 RF modules. It must be modified when used on the early models as the supply voltage is incorrect for the early type 1 module. The alignment routine is somewhat unorthodox but I had five RF modules to repair and it was the most efficient way to align them. I had a known good RF module that provided the information I needed to produce these alignment process.

The file is a tiny bit over Dave's file size limit so I have had to compress it in a Zip file.

Fraser

[Fraser]

After discussion on the Yahoo Advantest Instruments forum, I quickly wrote some comment on the earlier type of 2nd L.O and some thoughts on another alignment method that is more scientific. Here it is in all its rough glory 

Fraser

[KarelP]

I did a complete reverse engineering job on the PLL circuit of the 2e local oscillator and made a 8 page document containing pictures with the circuit and measurement information. It is in PDF format.  There is little information on this specrum analyser I think it might be benificial to a lot of people. I included the circuit but there is more information in the 8 page document but the file is to big to uplode

[KarelP]

I found a way to downsize my document to send it as an attachment. The circuit in the document is a bit small now but i will include a full size circuit as a separate file.
Help and comments welcome 

[KarelP]

I completed the 2e Local oscillator circuit further. The 2e page of the circuit now also contains the VCO and the Local power supply circuit for the PLL and VCO.
I’m still missing  information on some of the components. Douse anyone know, e.g.  the type of the Varicap diodes in the VCO some IC’s and transistors are also still a mystery. ?
I’m still waiting for the MMIC’s that I ordered but I’m not so sure of my first failure analysis anymore.
I have following problem: The spectrum analyzer works fine for a couple of minutes but when it warms up the level of the 2e LO fades away and the PLL get’s Out-Of-Lock.  The VCO and the first gain block (ERA-5) are under suspicion. However when I test the unit open on the table with only the 11.5V connected to the VCO then the VCO output level is stable. The power supplies are stable, no problem there. The VCO output is measured with a power meter connected to the SMA connector.
Anyone, any ideas? Help is very welcome

[KarelP]

Problem(s) solved:
The trick, I placed a small piece of copper tape on the dielectric trim and the frequency instantly jumped up.
I also mannaged to get the tracking generator back on track in a similar manner. See document in attchment.

[sbvr4]

Just stumbled upon this post. KarelP, this is awesome. Thank you very much. 

I believe your mystery C1J chip is a NEC UPC2713T. I've attached the datasheet for it. 

sbvr4

[KarelP]

Thanks for datasheet sbvr4 the markings on the chip are correct and also the function is correct.  I think you nailed it . Also thank you for the reply I was under the impression that nobody was reading my posts and that I used the wrong media in my attempt to help fellow Advantest owners. I’m convinced that the problem of VCO frequency drift is happening  :-+in all their products using this type of VCO’s and PCB material

[sbvr4]

No problem. I have two different Advantest units that I am in the process of repairing, which would not be possible without the tremendous contribution both you and Fraser have made regarding these SAs. For what it is worth, I've found the R3265A service manual (link below) to be a huge help as well with clarification of diagnoses. The manual precedes Rohde and Schwarz's acquisition, and is from the golden age of component-level diagnoses. It is split up into 3 parts and includes very detailed block diagrams, troubleshooting techniques and schematics. The theory of operation is fundamentally the same despite the architecture being largely different.

http://www.ko4bb.com/getsimple/index.php?id=manuals&dir=Advantest

[RF_STUDENT]

Thank you for the links to the  Advantest R3265A Service manuals and the work you guys have put in. I just purchased a R3265A on Ebay for $700 which I thought was a good deal.

[fozzyvis]

Also thank you for the reply I was under the impression that nobody was reading my posts and that I used the wrong media in my attempt to help fellow Advantest owners. I’m convinced that the problem of VCO frequency drift is happening  :-+in all their products using this type of VCO’s and PCB material

Karel, be assured that your posts (and those) of Fraser are very valuable for others. I have a r3132 with similar issues that (once I have more spare time) I need to look into. For sure the information from both of you will be very valuable...

Greetings from Ghent

[ARF]

I'll update the topic.
Recently I've got the R3132 and it doesn't boot. The screen does not initialize.
Did a quick look:
- supply voltages are OK.
- the capacitors on the processor board are OK.
- battery was replaced (it should not affect the booting)
- I tried to connect another CF card with firmware.
- I tried to connect an external display - the problem is definitely in the processor board.
- LEDs on the processor board light up. Their meaning is unknown.

[Gianni]

I revive this topic that I found very useful for me, hoping that someone is still there to give me some help.
I got for free an Advantest R3132 with TG. Replaced the battery, it "seems" to work but really doesn't, because the trace (and the TG) is shifted about 20MHz toward left and the signal is 5dBm lower than real and noisy. The frequency is not stable and the trace drifts slowly toward left.
The manual self-test (SHIFT/1) says that RF does not pass. I already read the documents above and I am going to do the first test, but I would like to have some help. I am not definitely a RF expert.
I understand that I must do what it follows:
•   remove the RF group from the housing (done);
•   remove the lid from the RF group, paying attention to the rubber details that risk to fall when you re-move the lid itself (done);
•   place the RF group on the table, with no connection (done);
•   identify the LO power supply point and solder a small piece of wire to it (identified - see the arrow in the photo - is that point OK?);
•   identify a ground point and solder a small piece of wire to it;
•   determine the power supply voltage; in my case, there is not the big 39 ohm resistor, so I need to give 6V power supply, not 11;
•   connect a suitable 50 ohm spectrum analyzer to the 2nd Local Oscillator monitoring port; I didn’t have such SA, so I am going to use a RACAL DANA 2101 frequency meter (up to 22 GHz) - what do you think?;
•   give power supply (6V);
and then proceed as described in the above documents. Did I understand correctly?

Then I have other questions about the VCO centering, but perhaps it is better to proceed by steps.
Thanks for any help…

[Fraser]

All sounds right to me. I have used a frequency counter to setup the VCO dielectric trimmer but that adjustment is tricky at the best of times. The amplitude of the oscillators output varies greatly whilst trying to set the centre frequency. This is why another spectrum Analyzer makes the job a lot easier. With an SA you can choose the best location of the trimmer for both correct frequency and highest output amplitude.

Be warned, these VCO’s can be so far out of tune due to ageing that the only way to get the correct centre frequency is to remove the dielectric trimmer arm completely and then adding small pieces of copper foil to the PCB traces where the arm used to sweep. I have seen SMD 0 Ohm ‘resistors’ used in place of copper foil so they may be easier to try. Size 0603 and 0804 are suitable sizes of zero ohm ‘resistor’. If you are very lucky  just removing the dielectric trimmer arm will bring the VCO close enough to the desired frequency and no soldering of SMD ‘trimming’ resistors will be required.

Fraser

[Gianni]

Wow, thanks for replying, it's an honor to talk with you!

In the meantime, following your documents, I could fix the SA and the TG. I could not check the VCO behavior because my R3132 is of an older type, and there is not the passage for the "blue wire", but just an internal test point. In the photos you can see what I did. In the case of the TG, there is also a small fragment of wire in the bubble of tin. Now everything seems to work rather well: both the SA and the TG passes the tests and the self-calibrations and appear to be precise. My R3132 has in addition the "high stability oscillator" option n.20. A nice instrument.

Thank you very much for your work, without it I could use my R3132 just as door stop.

Gianni from Florence, Italy

[Gianni]

Hi,

here you can find a document that resumes my experience: http://www.k100.biz/zParade/Advantest_R3132.pdf

It will be a chapter in the HP book, here you can find the latest draft: http://www.k100.biz/zParade/HP_book_draft_9sep.pdf

[KE5FX]

Hi,

here you can find a document that resumes my experience: http://www.k100.biz/zParade/Advantest_R3132.pdf

It will be a chapter in the HP book, here you can find the latest draft: http://www.k100.biz/zParade/HP_book_draft_9sep.pdf

Great to see you're doing coffee-table books for HP and Advantest.   Looking good!

(OT, but while you're working on those 8640Bs, consider adding a small light bulb or two just above the clear plastic meter housing before you close them back up.  When you see how good the illuminated scale looks, you'll wonder why HP didn't do that themselves.)

[Fraser]

I have enjoyed reading the HP book draft 

I was pleased to see a section on the HP calculators that included one that I have... the HP-41C, magnetic card reader and some programming cards. The HP-41C was part of an AGA 7xx series thermal camera kit that I bought. Sadly I am not a collector of HP calculators so that unit will likely end up on eBay. It was interesting to read about the card reader mechanism and the HP-41C though 

Fraser

[smaultre]

Hello i have the same problem with 3172. I couldn't reach 3840MHz on any points of VCO tuning Voltages.
When i close the cover any reached generations stops. I only get generation on nearby harmonic that's caused frequency shifting. Is there is any way to test varicaps?
I test all the +\- voltage Op amps rails 200,40,20MHz are presents. Its simply stop generating.

[smaultre]

200,40,20MHz seems Ok..

The diagnostic message says that VATT_H and VCOTUNE are not ok

[smaultre]

Table of measurements states (points on photo) :

     "locked"    unlocked
1)  10,55V  -3,3V
2)  0,3mV    0,3mV
3) -10,06V  -10,06V
4)  1,3mV    60mV
5)  4,9V       4,9V
6)  11,29V   11,29V
7)  11,48V   11,48V

[EE54]

I though I would pitch in as well.
I'm repairing an Advantest R3132N SA (same as 3132 but 75 Ohm input) with tracking gen. Unit does not detect input and fail self test.
Took the unit apart am I tried to tune the 2nd LO according to the instruction. But I foolishly supplied 11.3V instead of 6V (my 2nd LO is the 1st revision). I measured 3.78 GHz at about -30 to -40 dBm.
Gonna try again with the CORRECT voltage of 6V this time.
Man I suck at repairing stuff.