HP 3582A Repair help

[bsgd]

Hi guys!

I need some help from the experts in here.

I have spent a few days trying to find many issues Im having with a HP 3582A spectrum analyzer. It powers on but all I get on screen is garbage. ALL front panel LEDS are on, no Data Loading blinking, so I know the processor is not running. I have narrowed the problem down to the A7 Processor board, since I have managed to get another board borrowed from a friend whose 3582A works fine and my unit does power up with this board installed.

I have tried my best to find the source of the problem on my board, but nothing helped so far. Here's what I've done on the processor board:

- Tested and replaced all electolytics (the originals were fine but I replaced them anyway since I had already desoldered them).
- Dessoldered all ICs on the board, except for U19, U20, U22, U6 and U1 (the processor itself). I have tested every IC using the TTL/CMOS test function of a universal programmer and all of them pass the test, even though Im not sure how realiable these tests are. U19, U20, U22, U6 are all HP part numbers so I wouldnt be able to test them with my programmer, and Im not sure how often these fail to test them on a breadboard, and even if such test would give useful results since I wouldnt be able to replicate the speed and conditions these usually work on.
- Tested all transistors (there are only 3 of them on the board)
- Tested -5V regulator U24
- Clock signal seems fine at TP1 and TP2

I dont know what else to try or to look for as I have been banging my head on this for a week now  with absolutely zero progress so far...

[PA0PBZ]

Are the 4 supply voltages correct?
Does the processor get the reset (POP?) pulse?
Is there any activity on the address/data lines?
Did you inspect the pcb thoroughly?

[bsgd]

Are the 4 supply voltages correct?
Does the processor get the reset (POP?) pulse?
Is there any activity on the address/data lines?
Did you inspect the pcb thoroughly?

Yes, +5, -5, +12 and +7V are spot on.

Processor does get the POP voltage. When device is turned on voltage is 0V and after a second or so goes to +5V and stays at +5V. Pressing the reset button makes voltage go to 0V and back to +5V when released.

I did inspect the PCB a thousando times and saw nothing wrong. It was in 100% original condition, no leaky caps and nothing out of ordinary. I will take another look today with a magnifier.

Regarding the activity on data lines, I wouldnt know what to look for so I havent checked. What pins should I look and what should I expect to see?

[PA0PBZ]

Looks like they call it the IDA bus (Instruction Data and Address), there's a nice connector on the top of the PCB with all the signals.
You do have a scope I hope?

[bsgd]

Yes, I do have a scope   

I have made some measurements on the IDA connector, the one on top of the Processor board. There is activity, but my impression is that the waveforms make no sense for a communication bus.

I do not know which pin is 1 on the IDA connector, so I called pin 1 the first pin on the left, staring ate the component side of the board. I only probed the pins on this side, even thoug the connector is double sided.

PIN 1 reads +5V
PIN 3 reads 0V
PIN 7 reads +5V
PIN 10 reads +5V
PIN 11 reads 0V

All pins after pin 11 are interconnected to pin 11.

The other pins do have some waveform present, and I have attached all print screens for each of these pins. Sorry for the very small images, not sure why my scope is doing this. I have also attached a picture of my 3582 screen so you get the idea of whats going on here.

 NewFile1.png (5.47 kB. 320x234 - viewed 288 times.)

[PA0PBZ]

Do you have the service manual? What kind of processor is this anyway?
Pin 5 doesn't look too healthy, and I can't even explain what seems to be a sine wave on pin 9.
According to the service manual pin 0-15 are IDA0- IDA15 which should all have the same levels, by the looks of what you are showing 0-5 volts.

[bsgd]

Yes, have the service manual here.

I dont know whats pin 9, but I decided to put in the working board I borrowed and compare. Here is a screen shot of what pin 9 should look like.

To me it looks like some kind of clock signal that is pulled high from time to time. Not sure what this is, but I do know that on the not working unit this signal never get pulled high.

I also probed the other pins on the working board and signals are square waves from 0 to 5V with varying 'bits' going through the communication bus, as expected. On the defective board, most comunications is just not happening as far as I can tell, although some signals are present. 

[sparkfun]

I just repaired the processor board in my device. The analyzer was not operated for about 8 years and then it was broken.
Screen looked  correct in test mode but  otherwise it shows just weird/ no characters and lines.  (Actually it is the same screen you get when holding the reset button pressed on a working set. There were all lights on  at the front panel and no reaction to any controls.
I found  no  signals at all at the connector above the processor. It was not getting warm despite all voltages and clocks were present, Reset pin was also working.
 I pulled the board and found that the -7V line that goes only into to the processor had no contact (There was no diode to be measured to GND).
So I removed the die cast with the processor below. It is contacted with a golden thick thread. The processor  is bonded on a ceramic substrate with golden pins but the looked grey and corroded.
Hence I cleanded the contacts on the ceramic  with an eraser that they looked shiny again.
I mounted the processor and everthing worked immediately again as expected.