Anritsu MS2651A repair (need service manual)

[jimon]

Hi, got a broken unit from ebay, hoped it will just work, but not so fast ... so my first repair let's see what we can do.

Symptoms so far :
- no/very small signal (was unable to track anything reliably from rigol dg4102, don't have any other reliable signal sources atm, anything I can buy cheaply for it?)
- huge (like 50db or more) amplitude level difference between 0-3GHz full span and 0-100MHz span, when spectrum analyzer in full spane mode picture looks like normal (with noise level on -70db or so), in 0-100MHz mode noise is somewhere around 0db, no idea whats going on there, is there even a different path for different frequencies?
- no other visible issues
- cpu board shows self check is ok.

Looks like there is no internal test signal to check if this thing works, 'cause I've tried to disconnect whole RF stage and there were no complains on a screen.
Checked input DC block and attenuator - they seem to be fine. Checked almost all test voltage points - they look fine.

Does anyone have an idea where to look for service manual ? Tried to google it, but was not so lucky.
And any tips what to check next ? and maybe how IF should look like on oscilloscope ?

[nidlaX]

Hi, got a broken unit from ebay, hoped it will just work, but not so fast ... so my first repair let's see what we can do.

Symptoms so far :
- no/very small signal (was unable to track anything reliably from rigol dg4102, don't have any other reliable signal sources atm, anything I can buy cheaply for it?)
- huge (like 50db or more) amplitude level difference between 0-3GHz full span and 0-100MHz span, when spectrum analyzer in full spane mode picture looks like normal (with noise level on -70db or so), in 0-100MHz mode noise is somewhere around 0db, no idea whats going on there, is there even a different path for different frequencies?
- no other visible issues
- cpu board shows self check is ok.

Looks like there is no internal test signal to check if this thing works, 'cause I've tried to disconnect whole RF stage and there were no complains on a screen.
Checked input DC block and attenuator - they seem to be fine. Checked almost all test voltage points - they look fine.

Does anyone have an idea where to look for service manual ? Tried to google it, but was not so lucky.
And any tips what to check next ? and maybe how IF should look like on oscilloscope ?

If you haven't already, I recommend watching some of Shahriar's spectrum analyzer repair videos to get an idea of how the input stage typically works and where you might have a problem. If you have a scope, have you tried following the signal from your Rigol through the input stage?

[jimon]

Not really sure what's wrong with that unit, I've tried to bypass input attenuator and put 1Mhz -40 dBm into input of RF front end ... and I found nothing at with oscilloscope all around first mixer. Also nothing shows-up on output of LO ... it might be the same as here http://www.antiqueradios.com/forums/viewtopic.php?f=8&t=274605

Will I even be able to see LO output on 100 MHz oscilloscope ? Or should I look for whatever driving LO ? (there is probably some sort of PLL for it)

[sherlock]

I don't have that analyzer, however I had bought an Advantest SA a few years back that didn't display any signals on the screen.  In my case the mixer diodes were shot, along with a couple of burned resistors in the attenuator.  I too had questions about the LO circuit working.  You're correct in that a 100 MHz scope won't show anything.  Since I was mostly interested to see if the LO was working or not, and not on frequency at this time, I used an old radar detector for the car that I had lying around.  I powered it from a small 12 volt supply and used the detector as a probe.  In my case I was able to pick up the microwave LO signal so I knew it at least worked.  You might have to remove some shielding to pick up anything.  You can also check if you have the large "spike" at the left side of the screen (0 MHz).  Or set the center frequency for zero Mhx, full frequency span, and see if you get a zero indication in the center of the screen.  If you get that, then the LO and mixer is probably working.  Then you can troubleshoot where the input signal is getting lost.  Good luck with your purchase.

[jkf1000]

Hi Jimon, did you manage to get any further with your analyzer? I have just purchased the "B" model of the same analyzer as a spares or repair unit. The fault with my unit is a display issue, the LCD displays nothing but vertical lines, I have scoped the ACRTC chip and have found signals are missing for the Vsync and Hsync and have ordered a new PLCC chip for it. It appears as if it is functional as far as the RF circuits are concerned as I can get the unit to run its own self cal. routine by pressing the shift-cal-F1 while monitoring the IF out on the scope.

Did you manage to get hold of any service data for your unit as I have found nothing online and Anritsu UK could only supply the user manuals!!

Given that we are looking at 2 completely different faults we may be able to help each other out given the lack of supporting manuals. Karl

[kanuk]

Does anyone have an idea where to look for service manual ? Tried to google it, but was not so lucky.
And any tips what to check next ? and maybe how IF should look like on oscilloscope ?
[/quote]

Look here: https://www.anritsu.com/en-US/test-measurement/support/downloads    Good Luck.

[jimon]

jkf1000, awesome that you only had LCD issue! nice score!

I only found service manual for MS2601A version, I believe they should be similar in RF part. I played with it for few days, was unable to detect if LO1 is present or not (meh I just don't have any tools for that  ), maybe I should revisit it.

jkf1000, if you have any tools for that, would you please measure what kind of signal is going on from RF block (from port X3) to LOCAL-B board. I suspect it's just 2nd IF after 2nd mixer, but my oscilloscope doens't pickup anything. If it's indeed 2nd IF I wonder if I can just inject something to LOCAL-B board and see if it works.

Also would you mind to check if you noise level changes if you change RBW? My noise level jumps from ~-80db to ~-30db when I change from 3 MHz to 1 KHz RBW :/ I wonder if integrator is fine there.

[eeadata]

Check all RF mixer diodes in mixer sections.

[jkf1000]

Mine is still a work in progress unfortunately, I still have no display and without any documentation I am still a little lost as to which direction to go. I am trying to draw out a block diagram of the CPU board by following the traces on the board but it is at least 4 layers thick which makes it quite time consuming. what I have established so far is that the signals from the Toshiba processor pass through a pair of 3 state octal transceivers before the ACRTC video controller chip. To further complicate matters the output signal line from the ACRTC then go into the Altera Max PLD device which I am guessing is handling some of the video processing before feeding the LCD display. Unfortunately video signals are new to me as I am more into radio diagnostics.

Jimion, I will see what signals I can find at various points the next time I pull the unit apart but at the moment it is waiting for parts to be delivered so is all boxed back together.

It would be so much easier with some real documentation to follow. Karl

[jimon]

If you want me to trace some signals - just ask, in my case screen subsystem is perfectly fine.

[jimon]

So finally took time to make photos of RF module, unfortunately the very first part between input (X1) and 1st LO is hidden behind a board and really hard (desoldering) to get to.

1) Photo of top of RF module
2) Close in on 1st LO-ish part
3) This is the closest thing between YiG and input ... but is it really 1st the mixer? It sure doesn't look like one :/

[jimon]

4) Bottom of the RF module, looks like 2st LO is there!
5) What this thing might be It was even pressed firmly by a cover, and looks like after I removed the cover (or poked into something else) I started to get 0dbm+ peak on 45 MHz instead of 0 Hz (stupid me  )
6) PLL, wonder where control frequency comes from, maybe I should poke that
7) Output board, 100 MOhm resistor O_o, but looks like it outputs 10-20 MHz signal (based on manual for MS2601A)

PS. So where the infamous mixer diodes are?

[jkf1000]

Hello again Jimon, have you looked at the block diagrams in the version 1 operators manual? Have a look from around page 201 to 211 and it gives you quite a good idea of the signals you should be looking for. As a side note all the high frequency appears to be only in the 3Ghz unit, all the IF and mixer frequencies look to be a lot lower which you should be able to sniff out with a scope if you can go upto around the 100Mhz range.

If you want a simple and cheap higher frequency sniffer have you tried one of the cheap SDR dongles as a wide frequency receiver? The one I use can cover from 24Mhz to over 1.5Ghz and can make a simple spectrum analyzer with software such as SDR# but the resolution is only around 3.5Mhz bandwidth but still makes a useful tool. Karl

[jkf1000]

Jimon, (or any one else that may have the same spectrum analyzer) Considering the fact that I am not able to see any information on my LCD screen, can you list in sequence the button press order to enable the composite PAL output at the rear of the analzer in the menu settings? I have read through the manual but it assumes that you can follow the onscreen prompts.  Many thanks, Karl

[jimon]

jfk1000, please find a video here (with instruction in the video description)
Not sure if this what you need, because looks like my analyzer doesn't have TV-out option (this is what it prints when I try to enable it), though it does have "composite out" on back panel. If I plug oscilloscope to "composite out" channel I'll get some signal by default without doing/pressing anything. On a board itself it marked as "Y output", and X and Z are not populated.

[jkf1000]

Thanks for uploading the video but unfortunately it does not help much. I don't have the tv option installed only the rear composite out which I am unable to activate without a screen. Looking at my user manual the setting is in the system menu 3, composite mode-composite-PAL-return.

I finally had chance to do some measuring for you (assuming my unit is working ok on the RF side of things) on the coax to connector X3 on the 3Ghz unit there is no output, just an input from the Local board of around 96Mhz which matches the info on my block diagram. It list the local signal on X3 to be from 94 to 106Mhz I measured the signal on my 100Mhz DSO at around 3.56v Peak to Peak.

On your scan board with the "Y" output board I have 3 flashing LEDs, on yours are they flashing or constantly lit as I suspect mine are showing a fault condition relating to my failed display condition?

[jimon]

So X3 is an INPUT not output woah thanks for the info! that explains a lot ... especially why I was not getting anything out of X3

On my board all three leds are flashing on a same time with frequency around 2Hz. If I connect a scope to Y output I get something like on the screenshot.

[jkf1000]

I don't know what technical differences there are between your A model when compared to my B model. Do you have the block diagram available? In the absence of any sort of in-depth technical documentation or schematic it will at least give you an idea of what signals to scope for and the direction of the LO and signal paths. As I mentioned previously, apart from inside the 3Ghz convertor, all other signals should be low enough in frequency to trace with your Rigpl DSO.

Just as a long shot, but if yours is just a level error have you tried running the full auto calibration routine within the menu. Also check the backup battery for 3v just in case it is not able to retain its stored values.

Not made any progress with mine as yet but cleared most of the other jobs off the bench so should get back to it later. So far I have confirmed digital communication between the processor and the video chip but no sync output between the video chip and the CPLD. Without the Vsync and Hsync signals there will be no timing signals for the LCD module. I will try and source another video controller chip first and take it from there, A process of elimination without the schematic and I am in unknown territory trying to to trace and understand digital signals.

Let me know if you make any progress, Karl.

[jkf1000]

Just done a quick google search for you and to see what differences there are in the models and I found a manual specific to the "A" models. Unfortunately it is too large to attach here but it was easy to find with google, I just searched MS2651a and it was one of the first results. Download it and check around page 181 for the block diagrams which gives you the signal paths and frequencies. Karl.

[jimon]

Found it, awesome, thanks!

I'm still a bit busy with other stuff so had little time to do anything with it. Looks like their internal connectors represent PATA-ish pitch. Need to make a small cable extension so I can run RF module and be able to access back side of it (where 2nd mixer is).

[jkf1000]

Post back when you have more time spare or anything more you need me to measure. I have questions about the signals around the ACRTC chip which you could scope for me when time permits, Karl.

[jimon]

Sure, I'll do it tomorrow evening, anything in particular you want to know? I've tried to look for ACRTC-ish chip/marking, but can't find it ... board next to LCD mostly contains 74-ish chips and CPU board doesn't look like have this marking. So where I should look?

[jkf1000]

The ACRTC is the HD63484 on the main processor board. It is the main process chip for the display. What I need to know is what if anything is happening on your Hsync and Vsync lines which are pins 13 and 14. I was expecting to see cmos level timing signals to control the display timing but not getting any thing like it. I am still trying to diagnose the root of my display issue but not making much progress so far. Karl

[jimon]

Thanks for the waiting, here hsync and vsync signals. Please let me know if you want me to probe more!

[jimon]

jkf1000, do you have IF 110.695 MHz anywhere ? I was poking all around RF block ... and meh .. sure I found 100 MHz from getting from X3, and than it's was all over RF traces (was poking around 2nd mixer and VCO) with level around -30 dBm. Looks like it was leaking to VCO output trace with level around -35 dBm.

Looks like 110.695 MHz should go through backplane, but I cannot find anything with oscilloscope going through RF block connector to backplane :/ I would really appreciate if you poke there and check if IF is present.