Advantest R3361 Display issue (repair)

[ph4as]

Hi all,

Hoping that someone here has some experience that can help to diagnose this.

Have had my Advantest R3361A working fine for over a year after I purchased it second hand. Then my power supply blew out and I fixed the elco's.
It worked fine after that.

Then later an issue occurred that looks like reported in this thread:https://www.eevblog.com/forum/chat/advantest-spectrum-analyzer/msg1010132/

After some warm-up time,the display would sync at an offset. Preset would not clear it; only turning it off and back on.
Initially this was annoying, but not a huge issue. But it degraded over time and after the offset sync, it would also start to garble (see screenshots).
I also noticed that after the display goes blank completely, the sweep LED stopped blinking.

While researching the issue I also found a post on the Yahoo Advantest forum describing the similar issue:
https://groups.yahoo.com/neo/groups/advantestinstrument/conversations/topics/1064
The solution proposed is to tune the 3rd conv 200MHz VCO (-3.5V at TP3). I checked and indeed it was a bit off (-4.7V) but still locked. I fixed this with the trimmer.
Unfortunately the issue did not go a way but steadily got worse.

I suspect the graphic controller or related circuitry, but it is hard to read when the CPU board is in the device. So what I tried instead is to reset the 68000 CPU when the device is powered.
I noticed that sometimes this caused the display to reset back to correct sync, but there were also cases where the screen did not change (either frozen or garbled) and did not recover either.
I also measured the CPU clock but it was steady all the time.

Now I am at the point that when I turn on the device I do hear the relay clicks, but the screen does not turn on at all and the sweep LED is permanently lit from boot onward.

Any pointers on what to try next are welcome, thanks,

Job

[TheBay]

Could this be a RAM Issue, is there a memory test on this device?

[nctnico]

The service manual can be found here:
http://www.ko4bb.com/getsimple/index.php?id=manuals&dir=Advantest/Advantest_R3261_R3361/R3261_Service_Manual

Are you sure you used low ESR electrolytics in the power supply? If you used regular electrolytics they might have gone bad. If you have an oscilloscope you should check the 5V power rail to make sure. I'd say this is a problem where the digital electronics are affected and the power supply is suspect number 1. Did you also clean the boards of the PSU thouroughly before mounting the new capacitors? I used to own a similar spectrum analyser and I recall the capacitors in the PSU had leaked quite a bit.

[ph4as]

Thanks for the suggestions. I indeed have the Service Manual in multiple copies, but there is not much detail on the CPU side of things, such as how the sw initializes etc. Except for the schematics of course.

I did replace with the power supply elco's with almost identical ones and the levels are all within spec. The 12V is a bit low (11.85) but within the 5% spec. 5.2V is almost exact.

[nctnico]

Did you use an oscilloscope to check the ripple on the power supply lines? A DMM will just show the average.

Edit: I'm finding this a strange phenomenon. The video system is a 'simple' bit mapped system which uses special DRAM which has a seperate serial output port. There are 4 banks of memory to have 16 intensity levels per pixel. Given the screen gets garbled it means that somehow the memory contents of the DRAMs is getting corrupted. The AM95C60 is an intelligent video controller chip which can also draw lines and other graphics by itself. It also generates the hsync & vsync signals. The screen going bad can be an indication that it gets the wrong commands or that it doesn't do what it is supposed to do. Given the fact that the spectrum analyser itself (the primary process) also stops working makes me think that the main processing system could be the culprit and the video subsystem is OK.

Did you try to reseat the EPROMs? Pull them up a little and push them back into their sockets? If you have an EPROM programmer you could check the contents against the images on ko4bb's website.

[ph4as]

Good point about the ripple, I had not checked that. I see ~100mV switching spikes, but the rest is well below 50mV. They are equal across all power levels (5v, 15, -15, 12).

The display luckily came back after taking out and reseating the CPU board, however the syncing and garbling remains.

As I now saw 12V line was below 11.7V I decided to feed it directly from my lab supply. However, the symptoms are the same.
Will try to also feed the 5V from a lab supply to take that out of the equation.

I agree with your analysis. It was my feeling it was the video or cpu. However, several people indicated that these symptoms also happen when the clocking gets bad.
I can imagine that the sync offset can occur when the video chip gets a clock glitch. Unfortunately I have not been able to identify such an issue with my analyzer.

Main question is how to go about narrowing down the issue. The CPU board is very had to get to. I can reach one side of the CPU with probes, but that's about it.

Edit: I did reseat the CPU and EEPROMS, also added sockets for the parameter EEPROMs and backed them up. I doubt the contents is bad as there is a clear degrading effect which seems unlikely from faults in the EEPROM.

[nctnico]

The power supply voltages look OK to me and 100mV ripple shouldn't be a problem. Does the problem still progress when the spectrum analyser heats up? You could try to use freeze spray to cool certain parts down.

[Justa]

 Same issue in same model analyzer. Refer to the attached picture. The coil pointed is placed in the rf module feeding 20 MHz clock to the cpu(bottom of instrument see service manual). Re spacing the coil seem to solve the problem for now. The problem seems to be here. 

[Tantratron]

Edit: I'm finding this a strange phenomenon. The video system is a 'simple' bit mapped system which uses special DRAM which has a seperate serial output port. There are 4 banks of memory to have 16 intensity levels per pixel. Given the screen gets garbled it means that somehow the memory contents of the DRAMs is getting corrupted. The AM95C60 is an intelligent video controller chip which can also draw lines and other graphics by itself. It also generates the hsync & vsync signals. The screen going bad can be an indication that it gets the wrong commands or that it doesn't do what it is supposed to do. Given the fact that the spectrum analyser itself (the primary process) also stops working makes me think that the main processing system could be the culprit and the video subsystem is OK.

Did you try to reseat the EPROMs? Pull them up a little and push them back into their sockets? If you have an EPROM programmer you could check the contents against the images on ko4bb's website.

Do you know if it is possible hack or intercept the digital video information then route towards an LCD display in order to replace an old or faulty CRT on these R3x61y spectrum analyzer serie ?

Thanks, Albert