Repair of Wiltron-Anritsu vector network analyzer 37269A

[smal_edelweiss]

Dear all
recently I got a Wiltron-Anritsu vector network analyzer 37269A in my possession. Unfortunately at startup selftest fails with following messages:
4500 IF CAL FAIL
6008 PHS LCK FAIL D (for some moment I can also see the message 6024  PHS LCK FAIL E).
5110 RF PWR UNLEVELED
Is here anybody who has some general knowledge about this device or experience repairing this kind of device? Or have an electrical and electronics schematic?
Best regards
smal_edelweiss

[vaualbus]

You may check the internal 10Mhz reference that start to check voltages.
A teardown would be very good....
Best regard, ALberto

[vaualbus]

Look for the error from page 53 in the operator manual.
http://downloadfile.anritsu.com/Files/en-US/Manuals/Maintenance-Manual/10410-00153.pdf?f4739ea0f83b43ad1015d39273bcf8beccd11e5ac9d1d8ff2dbf45c4e0f70a1c354008d5d2cf4c42353056e697f2899f3b69186b2045930ba6719b7606af7784653ae2f25947974d5db48be287e073f0dc751cd57399d0e95a9a84ee6e6ed72786de63ab78f26e95bd03f5f158&10410-00153.pdf.

[smal_edelweiss]

Hello,
I have attached the oscilloscope on the back 10MHz reference out BNC connector and I can clearly see that 10MHz 1Vpp signal is present. I suppose that this signal is also present for other boards in the VNA. What to check next?
To tear down this device to find the error is not a problem. Yes I have checked this document (Maintenance Manual) also. The document states to find the error anyone should exchange the boards with working one till the error is found. This is not a option for me because I don't have a second working device or reference boards.
Regards
Simon

[vaualbus]

Ya on such instruent the repair is always change boards. Rry yo check the voltages. I thinksome where in the manual there is a list of voltages to check.
Best regards alberto

[smal_edelweiss]

I have checked the voltages which are noted on page 5-7 in the manual. From what I have measured there is no 160Vdc present. I can measure something around 0V. When I disconnect both of the connectors J13, J4 and I can not measure any obvious short circuits on the board regarding this 160V terminals. Where this voltage is going to and what is need for?
There where +4.8Vdc should be I am measuring 5.017V which is higher than specified. By specification should be 4.8V +1% which is 4.848V. Probably something is not drawing as much current as planed. PS_SYNC signal is present with 4Vpp and 109KHz sine wave. Rest of the voltages (+9V, +18V, -18V, +27V, -27V) are present and in specified tolerance field.
I will try to connect 5-10 Ohm resistor on 4.8V power supply to measure all the voltages again on the power supply without connected any of VNA boards.
Best regards
Simon

[vaualbus]

I don't know where the 160v used but seem important.
Could be used in a yig oscillator?
Try to see if is a board that short the rail.

[smal_edelweiss]

Yig is not working. I have disconnected the SMA connector on the yig and there was not a noticeable difference in error messages on the screen. I have already measured ohmic resistance between 160V terminals and it was indefinitely. So no short circuit in that area.
Best regards
Simon

[tube7586]

The 160V supply is needed for the CRT circuit .It seems to be working properly .
I don't know if the YIG osc. has a problem ,you can check its output signal with a good RF counter or a spectrum analyzer if it's possible .
Best greetings ,
Andras

[vaualbus]

it could but right now with have no schematics  so if the problem is not related to the psu it get very hard find the problem.
Now you should remove the boards and see if there is some visible broken component, burn resistor, ecc and check if on the board there some test points with a description and check those.
I wish you would be able to solve the problem.

[vaualbus]

so the yig could be use for generate the frequency so the errors could be related to it.
Try to check the yiG control board (A21 the service manual suggest) and let us know what you find.
Best regards, Albreto.

[smal_edelweiss]

tube7586: CRT has its own 230V power supply. I think 160V is needed somewhere else because the CRT is working flawless.

[vaualbus]

and have to be related to your problem!

[reinhold]

Hello @ all,

Currently, I've the 37247A Model on my desk - reporting nearly the same error messages (5110, 6092, 6094). In difference to pic#3 from smal_edelweiss the red LED dosn't light up. After talking with other guys in our lab we've remembered that the error 5110 has been fixed several years ago by checking the 51 Ohm resistors on this control board. Some of them were broken by overheating. But now these resistors are o.k.
Within the next days I'll check the phase comparision on A8 - they are using an Altera EPM5016-20. I hope that I'm able to check the phase correction output as function of the input signals.
Do you've further results in repairing the 37269A?

[vaualbus]

Have you menaged to solve the problem?

[Staffan321]

Hello,

Knowing this is a really old post, I still thought I'd keep this thread.

I have a 37247A with very similar problems as described above. The 4.8 V is really 5 V and there is no 160 V present. My guess is that this is from an older version and they (Wiltron) never updated the manual.
The YIG runs on -18 and -27 V for main tuning. On my A21, Q7 was burnt and wasn't able to control the current through the YIG main coil. The Q7 is a TIP112 darlington NPN and it dumps a lot of power when running at high frequencies (high current).

I have the YIG running now, but the unit still won't lock more than at a couple of frequencies. These are not very obvious ones and I can't see a pattern.

The power supply is in some way corrupt since ripple is more in the 1 Vpp (110 kHz up to 5 MHz) rather than 100 mVpp and this may be part of the explanation, but I don't see why the unit would lock at any frequency in that case. When it indeed is locking the YIG, there seems to be a ripple in tuning (FM coil?) every 1 ms or every 5 ms - unclear why, but this may be correct.

Both 1st and 2nd LO seem to be very clean (no ripple) and not affected by the ripple from the power supply. The FM driver signal from A8 seems also free from that kind of ripple, but it doesn't seem to be very stable.

Replacing the YIG with a synthesizer locked to the VNA ref signal I tested to insert the expected frequency. Output after mixing before phase comparison is the expected 2.5 MHz and it is apparent that the FM tuning signal is not performing a search, which it does when completely unlocked, but the ripple in the signal is +- 4V, which is full span.

Any suggestions what to look for would be very much appreciated. Will of course try to fix the PSU - may be poor caps only - strange(?) though that all would be bad. Very good to hear that someone has a unit within spec - means that there _is_ something wrong with mine at least.

Would love to hear from anyone with some schematics - I have reversed the main power supply and parts of A21, but it has been a pain. Would so much like to have this one running!

Regards,
  Staffan

[DaJMasta]

Could be tricky, but the challenge can be fun.  Don't know of a source for schematics, even a block diagram could be helpful for seeing what difference there is in areas that are effected or aren't, but I seem to remember a YIG driver repair of an Agilent unit on TheSignalPathBlog on youtube, could be useful for getting an idea of how things are setup if you're not already familiar as the basic architecture is likely similar.


I would definitely look into the power supply - you could try swapping a bench supply for the rail in question or could try disconnecting modules on it to try to see if something else is injecting the noise or if it's from the supply itself.

It's my understanding that changes to the YIG or driver can require realignment, maybe there's a service menu option to sort that out after the initial repair?  It could be that the DAC data table for driving the tuning coils doesn't align with the output of the YIG that's expected now and is giving you your intermittent lock issues.  Could also be worth looking at the output of the YIG with a frequency counter and a power meter - aging can often manifest as degraded signal level, which tends to need to be fairly high for the following circuits, so if you're not seeing like 10dBm+ out of the YIG across its operating band, that could be a problem.  The frequency counter would show you if the output is even in the right ballpark, and if the FM tuning coil glitches you're seeing are having an effect (maybe check coil resistance on the FM coil for the YIG?)

Also could be useful to test everything in a zero span sweep mode, and just monitor to see if voltages or frequencies are changing at all.  If you're seeing errors when sweeping, the FM coil drive noise could be attempts to fix the error, but it should show as clean if at a fixed frequency and not reading an error.

[Staffan321]

Hello,

I've been through most of what you describe and don't think the coarse setting of the YIG is off. The transistor that was changed is only a driver and feedback is through a resistor that seems quite OK. I have also measured the output of the YIG for different CW settings of the VNA and it is in the right neighbourhood.

One thing I just noticed in a previous post was that voltages were supposedly within spec, but was that DC only or has anyone checked the ripple? Don't want to be chasing ghosts right now. Could it be that strong ripple not necessarily is the cause of the problem. If anyone could confirm the ripple of the different output from the power supply that would be great!
I have measured the output capacitance of the 5V supply and it is roughly 12 000 uF as expected when measuring at 1 Hz. Going to 1 MHz, I see more like 100 nF... I guess this is to expect so it would great to know what ripple to really expect.

The strange thing with the locking is that it seems to be working for some frequencies, but not for others - being close by. Difficult to see a pattern. Seems more unlikely to be the phase detector since that one should be independent of frequency. Mixers seem to be working also (haven't checked for the complete range though - should do). Both LOs seem to be doing their job - have seen signal almost over their respective ranges.
Any suggestions welcome!

Regards,
 Staffan 

[Staffan321]

Hello,

Short update since last post.

I have somewhat been able to get the VNA up and running. The unit won't lock around 1 GHz since the internal mixer seems to be deaf to the LO1 signal at low LO frequencies (meaning LO1<400 MHz isn't working).
Nevertheless, sweeping 1.1 - 20 GHz works fine and I have just tried one port calibration with good results with a power supply that certainly is off spec regarding to noise and without the 160 V.
So, anyone thinking about having a broken supply since it is noisy - maybe not to worry too much.
There is also some kind of twitching in LO2 frequency that will give the 2.5 MHz IF a strange behaviour. This too seems to be normal since it hasn't affected the measurements. The same was also discussed in an episode by "The signal path".
By the way, the J4 connector supplies a couple of different voltages in the following way (most is in the manual, but not all):
Pin Colour Voltage/function
1      Orange   +9 V
2      Orange   +9 V
3      Key
4      Grey        9 V Common
5      Grey        9 V Common
6      Yellow  +18 V
7      Yellow  +18 V
8      Brown  +-18 V Common
9      Brown  +-18 V Common
10    Green   -18 V
11    Green   -18 V
12    Pink     +27 V
13    White  +-27 V Common
14    Purple  -27 V   

The common signals are floating with respect to eachother but should be connected on the motherboard.
In the documentation it is stated that there is a 160 V on J13 pin 16. In my opinion this pin is connected (via A16) to the ON/OFF switch on the front and shorting J13 pin 18 and pin 8 will start the unit (load below 3 k Ohms).

Hope it can be of some help.

Regards,
  Staffan