EB5AGV's Workbench: Tektronix AWG-2020 repair

This page is devoted to the repair of my Tektronix AWG-2020 Arbitrary Waveform Generator.

I got it with a known fault, described by seller as does not power up. Well, once received, fault was not as simple... read on to find why!


Unit as received

Unit was somewhat dirty but in good condition. Here you have some pictures of it:

Unit has two options installed: Option 03 (12bit digital out) and Option 09 (DSP)

So I have tried to power it. Pressing POWER button unit tried to start but immediately stopped... then I have tried to keep pushing the button... and it has started!

But unit powers down as soon as I stop pressing the button...
Unit has finished SELF-TEST but, as I suspected, there is an error:

Image is clear and there are no burns. Here you have a display sample:


Repair work

I have started dismantling the unit, taking out all the plug-in boards:

This is the main board, without plug-ins:

And these are the different cards:

I have continued cleaning the front panel:

Checking the AWG-2021 Service Manual (which is the closest one I have been able to locate at the Tektronix WEB), Waveform Memory CH1: FAIL (710) means a fault located in the A21 CONTROL CH1 assembly... and then it instructs to replace it. Of course, this is not an immediate option so I need to locate where the fault is or, if not, try to find a suitable parts unit. Any help locating it will be welcomed!

I think that the POWER button problem (unit quits as soon as I stop pressing it) could be related with the reported fail.

I have seen that most SMD electrolytic caps have been replaced and it seems original units had leaked in some cases. But there is no visible damage in any card (I have checked specially A21 for leakage damage).


Problem #1: unit powers down automatically - fixed!

As I have commented, unit runs only while POWER button is being kept pressed. I knew it was a momentary pushbutton but, just in case there was something weird, I have looked at it. This means removing front panel frame:

Once there, I have carefully cleaned display filter. As I suspected, there was nothing wrong with the POWER button, so I have assembled the unit back.

Tracing the POWER button signal from the front panel, it arrives at CPU BOARD (A6 card). I have checked which pins needed to be shorted to start the unit and have done it manually. Unit kept powering down by itself, so definitely, POWER button was OK. So I have taken out A6 card to carefully check it... and have found it had two SMD electrolytics which had leaked! (I don't know how I missed it first time!). You can see here the effect on the PCB (both top and bottom sides):

Once capacitors were removed:

And this is bottom side damage:

Checking one by one, I have found three damaged traces and have fixed them with thin wire:

I have found that the MAX695 just in the middle of the faulty caps was no less than a microprocessor watchdog!. I guess it was generating RESET and/or power down signals.

Once I have put back the card in the unit, it has started up fine and has not needed my finger pressing POWER ON to continue working. Nice :-)!!!

But the other problem (Waveform Memory CH1: FAIL (710)) remains. Here you can see what the unit generates, instead of a clean 2.5MHz sinusoid:

It seems there are some bad RAM chips (or auxiliary circuitry) around.

So I need to keep working on the unit!


Problem #2: Waveform Memory CH1: FAIL (710) - work in progress

Having solved first problem, I have tried to fix second one. First step has been taking out again A21 CONTROL CH1 board. Suspecting of the capacitors, I have found that they were of the same type as the ones found in the CPU BOARD and, yes, they had also leaked, if less than in that case. Once removed, this is what they hide below:

After cleaning and checking for shorted or open traces (none has been found this time), I have replaced them with newer units:

I have checked the AWG again but problem remains the same.

I guess there could be some faulty memory chips (or perhaps is the access to them what fails). Anyway, these are the chips it uses:

I will try to find some Hitachi HM62256LFP-10T or equivalents (if there are any). Meanwhile, I will keep checking for faulty things.

Work continues!


Problem #2 (2nd attempt): Waveform Memory CH1: FAIL (710) - fixed!

After trying to get it working yesterday, I have taken today a similar approach but checking with more detail for PCB trace damage... and, yes, I have located an open one in A21 CONTROL CH1 board :-)!. It is in the inner rectangle, bottom part. A very tiny part of the trace was missing due to acid corrossion due to capacitor leakage:

I have fixed it with the usual wire-wrapping wire:

Then I have reassembled the unit and powered up the unit... to find this:

IT WORKS!!!

I have checked the generator with my Tek 2465B. Now signal is clear.

Compare with what I got yesterday!. I have also run diagnostics, just for some extra fun and joy:

This is the configuration of the unit. Note the installed Floating Point Processor (FPP) Board (DSP option, a 1000+ US$ option :-)!):

Rise time seems good enough, checked with my 2465B:

This is the main Signal Generator display but it has lots of possibilities, even a programming language... I need to learn how to use it!

All in all, I am today A VERY HAPPY CAMPER :-)


Go back to WORKBENCH INDEX