Scope Repair : GOULD 1604

[thinkfat]

That RC network was likely a delay circuit for the reset. They put a NAND with Schmitt trigger inputs to get a clean reset release despite the slow rise of the RC voltage.

[Vince]

Not sure what you mean ? What reset line ? And it's not a NAND it's a OR gate, and it's a 74F32 which according to the datasheet I got, does not have Schmitt trigger inputs.

[thinkfat]

Not sure what you mean ? What reset line ? And it's not a NAND it's a OR gate, and it's a 74F32 which according to the datasheet I got, does not have Schmitt trigger inputs.

I was referring to your  Reply #44.

[Vince]

Ah OK, the CPU reset line !    You didn't quote that old post so I had to assume you were replying to my latest posts.

Yes of course that NAND gate is there to generate a clean reset signal. Not of any use to the CPU though, since it has already a Schmitt trigger reset input. However that reset signal feeds a few more chips as well, not just the CPU. So I assume one of those other chips didn't have a Schmitt input, they had not realized it so they had to modify the board. I don't know...

[jdragoset]

It very clearly does appear that, world wide, this Co-Foo-19 does have everyone working with the patience of St. Isaac.
 

[Vince]

Yes, it's not all bad this virus indeed ! Losing lots of money, soon to lose my job most probably, but at least it gives me plenty of time to work inthe house and on this bloody troublesome scope !

OK I spent a few more hours on it. As planned I probed the signal path for analog operation.  I fed the probe calibration signal to all 4 channels simultaneously so I could follow the signal on all 4 channels at the same time and compare everything to channel #1 which is the one that works.

I probed at the output of the preamplifier, ie the output of the front end...then looked at the output of the CH1/CH2 switch/MUX, then the one for CH3/4 on that daughter board, which connects to the main board via coaxial cable. All good. The square wave makes it through all 4 channel paths, no problem with offset, noise or signal amplitude. All 4 channel look similar.

Then I looked at the output of the final switch, that aggregate all 4channel, plus the readout information, and then goes straight to the Y amplifier on the CRT board. Well even there, it's good : see screen shot.  All 4 channels are there, along with the readout information. At this stage, I can play with the front panel control to modify attenuator setting, vertical position, and it all reacts as expected.

So now at this point, we can say for sure that the signal makes on all 4 channels makes it just fine right up to the last final step, right up to the CRT, and that there is no problem with massive vertical offset that would shoot the traces off the screen.

So that's good to know, making progress...

So now what could cause the traces for channel 2 to 4 not to appear on the screen ?

Updated annotated block diagram attached. In green, I highlighted everything I have probed and that appears to work satisfactorily, ie the ' Y ' analog path, the vertical side of things. That's one of the 3 inputs of the CRT board. Other two of course is the  ' X ' horizontal  side of things, ie the sweep signal, and then third signal is the ' Z ' side of thing, the intensity control, ie  blanking signal. So let's investigate those last two paths...

Two hypothesis at this point, highlighted in yellow :

1) sweep : maybe there is no sweep for channel 2/3/4, and the spot is off the screen on the left or right, invisible.

2) intensity : if sweep turns out fine, then maybe the signal is somehow blanked when comes time to display CH 2/3/4 .

Scenario #2 seems most probable so will start there...

In any case, as shown on the block diagram, both the sweep and blanking signals are controlled by... not the CPU, but by the main ACQ ASIC !  So, we are going back to this area of the board !  I guess that's good in some way, as it means it is now becoming possible that a corroded track is causing the problem. But let's not assume anything... let's probe instead and see what's going on !

[Vince]

OK, one step at a time. Just check the X signal, the sweep ramp. Attached is the schematic sheet for the time base. Highlighted is the ramp generator, whose output goes straight to the CRT board. For the life of me I could not locate that chip !  90% of the component designators on the main board are impossible to read.. because they put them underneath the component. How useful... might as well not put them at all ! 

So I resorted to probing the CRT board to board connector directly. Two rows of 16 pins. Here again, no help from the silk screen to tell you which row is which, never mind tell you on what end of the connector is located pin #1 !

So I first probed each and every pin with my DMM to find which pins carried the high-ish voltages from the PSU ( 75V and 185V) so as to avoid getting a nasty shock and/or blowing my scope...

Then I could safely probe all the other pins with the scope. Found only one pin with a ramp signal on it, see below.

We can see clearly a repeating ramp, clean, with a few "glitches" here and there. IF I zoom on it I see it's shorter ramp, of varying length, which appears at "random" places in the normal stream of ramps. I assume these are the ramps for the readout.

So as we can observe, there is a continuous stream of ramps, which is good I think : If the sweep was responsible for the lack of Channel 2/3/4, then I guess we would see just one ramp followed by a blank signal equivalent to the duration of 3 ramps.. but it's not what we have here.

So I think we can say that as far as sweep is concerned, there is no reason for traces 2 to 4 not show up.

So the only thing that can make these traces disappear now, is the blanking signals / intensity control... next on my plate.

[tautech]

So as we can observe, there is a continuous stream of ramps, which is good I think : .................

I'm not so sure.
While I've never worked on a CRO with readouts those glitches in the ramps would be of some concern to me.
Normally you have a single ramp generator and unless its linearity is perfect....Houston, we have a problem !

Study the sweep gen schematic again 'cause normally there's a expected waveform displayed for reference but maybe not for these Gould's. 

[Vince]

Don't be worried Tautech, I am not.. for a change ! 

There HAS to be some ramp for the readout, mixed in with the " regular " ramps, or else how would it be able to display the readout content that's embedded in the vertical / Y signal ?

If you look at the schematic I attached above, in yellow you see U604, the big chip that generates the ramps for the 4 channels. Then if you look to its left we have U609 which precisely generates the ramp for the readout, and mixes it to the main ramp signal from the other chip, to produce the "composite" ramp signal that goes to the CRT board. I wish I could probe these chips individually so you could see the individual ramps BEFORE they get mixed... but I can't because as I said most of the chips on the main board, I failed to locate because the chip designators are not visible. So the only point I could locate easily to check anything really, was directly on the CRT connector, hence only the final, composite signal.

As for linearity, I could not care less.. all that would do is distort a little bit the waveform, but it would not keep the trace from appearing on the screen, nor keep it from stretching fully from the left edge of he screen to the right edge. Same for glitches... basically as long as there is some kind of ramp happening when the chopping control signals switch from one channel to the next, I should get my 4 traces on the screen, no matter how distorted or glitchy or dirty or whatever, but I should still see traces...

As for example wave forms...this is a GOULD scope, not a Tektronix ! 

Tek service manuals are marvelous, but one should not take their quality for granted : as soon as you work on a different brand scope, you should be prepared to be disappointed one way or another... 
I just consider myself lucky that there is ANY kind of service information on this scope at all. Had I not the little that I have got here, the scope would not be repairable at all...

[tautech]

 
Yeah I studied the schematic some and most certainly the area around U604 is interesting, quite unlike the ordinary RC sweep gen. 

Then I spot above it, End of Sweep Detect that of course one suspects is the source of the beam retrace blanking.

In this respect the schematic is great as it properly breaks everything down into each functional block. 

[Vince]

I found half a page worth (spread over two sheets) of explanation in the manual, about the ramp generator, which confirms what I figured by looking at the schematic.  Have a look.

I don't think even old basic scopes used RC combinations to generate ramps, or it would be an exponential not a ramp 
Even old scopes use caps charged at a constant current, which does give ramps.
This GOULD scope is no different, as you can read in the pages attached. On the schematics, you can see the bank of timing caps in the bottom right corner of the sheet. This area is tagged " Calibrate Time Base Capacitors  " . It revolves around chip  U607 which allows to select the desired cap.

[Vince]

TAUTECH !!  I apologize ! 

You had me thinking all night with your RC generated ramps... the constant current charging is so old and wide spread, simple and efficient, that I couldn't possibly believe scopes were using RC networks.  I mean for decent / modern scopes, that do have a calibrated and triggered time base. Not talking about pre- WWII stuff of course...

But to be sure, I checked the schematics of an old and really shitty scope, a Hameg type HM 412 from the '70s, that I just fixed and sold the other day. It DOES use RC charging ! 

Then I looked at my old Tektronix tube scopes from the 1950's... they also use RC not constant current ! 

So I stand corrected, my apologies ! 


I guess if you use only the start /  first portion of the exponential, it does approximate to a straight line sufficiently that you can make use of it for a sweep ramp.  The specs given for the linearity are not spectacular to be begin with so I guess it's " good enough "...

At least the old Teks had the excuse that tubes were large and space at a premium in the cabinet, so if they could get away with RC and save 2 or 3 tubes  by not implementing a current source, it was a reasonable choice I guess...

[Vince]

OK...I spent a few more hours working on the thing, checking the blanking signals that control the intensity / Z signal to the CRT, the third and last information that goes to the CRT, last hope of finding something wrong.

To cut the story short... I did not find anything wrong, all signals in this area looked as I would expect. None of them had blatant "blank" portions that would correspond to the 3 missing channels / traces on the screen.

So that made no sense. All the signals that go to the CRT should make it display my 4 channels, period.

So I must be missing something, just not competent enough for sure, hey I am not a professional after all, no shame...

So decided to give up on it, having run out of ideas as to why it would not work. Bought if for 50 euros not working at all... at least I improved on it quite a bit, now it powers up, runs, and works in analog and storage mode, albeit on channel #1 only... that's a lot better than the brick it was when I got it. So hopefully I can still sell it for at least what I paid for it !

Proceeded to reassemble it, but wanted to give it one last test anyway just to see, so reassembled it only as partially as possible.. did not even put the front panel assembly back onto the chassis, I just laid it on top of the scope... ribbon cable is just long enough to allow for that, and works just fine ! Hell I should have done that from the beginning, saves time ! 

Plugged the PSU connector to the main board.... BIG SPARK, BANG, FLASH !!!!  Oh no... just because I resorted to give up on it doesn't mean I wanted to kill it for good ! Damn it !   Went to turn the power off.... Ah, power strip was ALREADY OFF hmmm.... maybe the spark came from some big cap on the PSU that got shorted... at any rate the chance of the scope surviving are slim. So turned the power back on... scope still alive and kicking, powers up and runs !!! That's one tough guy !!!   

Then what do I see... some garbage on the screen, fuzzy and all.... let me adjust focus and intensity a bit... OK... WHAT ?!

I have FOUR traces in analog mode ?!   They are BACK ?!   Fed the calibrator signal to all 4 inputs at once, I get 4 square waves no problems ! Even better : played with the attenuators : NO MORE random big offsets problem ! Traces don't move around any more as I change range !   

I proceed to switch to Digital / Storage mode, anxious.... IT WORKS !!!   FOUR traces on the screen, four real traces this time, not " place holders " / test patterns like before !!! IT BLOODY JUST WORKS !!!!!!

Even I, can not believe it, so I took some pictures below to prove it.


Oh boy !!!

Can you believe it ?!

I spent countless hours probing around, studying schematics, pulling hair and scratching head.... but did not actually DO anything to it ... yet it WORKS ! 

So I guess all that time and effort put into it was not in vain after all... I was just rewarded for my perseverance, there is some justice out there in the end ! 

So no corrosion problem any more in fact, not a blown tantalum cap, no nothing...

Only explanation is some cold solder joint somewhere, which I must have triggered while working on the board.
That's bad because that means it can stop working randomly at any time. Might not even work after I am done putting it all back together properly, and even if it behaves for so long as to let me take pics of it so I can advertise it online... it might break again during transport to its new owner, and he will be mad at me : "You said it was working perfectly !!! "..... yada yada yada.....   

Anyway, even it it lasts only for 5 minutes we shall see, I am none the less overjoyed to see it working, and my hard work rewarded !
At least if it fails again, I know there is nothing wrong so to speak, so no need to chase red herrings and waste time anymore... just a bad connection god knows where.

So I am super happy, I managed to fix what the previous owner failed at ! 

So to summarize the repair : was bought "No Power". Found a 5V rail that was way high at 5,40V. Fixed it, was a buggy remote sense feedback line. Then it would power up but not come to life/run. Found the CPU was not running. First found its 8MHz oscillator was kaput, replaced that, still not running. Found that the Reset line got broken in the corroded area due the leaky battery. Bridged that track. Scope now back to life !

...and wasted time by trying to be too thorough and fixing two broken traces that turned out to be factory mods hence did not need " fixing " ! 


Thank you Tautech and Shakalnokturn for your support, much appreciated, we won ! 

Now here is hoping the bloody thing will still work in 10 minutes once I have put it all back together again... cross all your fingers and those of your cats and dogs, I need as many fingers as you can come up with !....

[tautech]

 
Vince's new fix method; shock it into operation ! 

Congrats Vince and what nice sharp traces it has. 
Patience and tenacity is generally rewarded.
Hope you make a buck from it as you've certainly earnt it.

[Vince]

Thanks Tautech.

Trace looks sharper in the flesh, but it's so difficult to take a good / realistic picture of a CRT ! 

Bags of reflection on the screen, and no amount of playing with ISO settings can reproduce the reality, not enough light so lots of noise/grain, and a bit blurry because I am human so hands shake a bit no matter what. With good / additional lighting, and a tripod to keep the camera steady, and a delayed shutter so I can get away from the scene so that I don't reflect on the screen... would help a lot I am sure.

Anyway, I have just finished putting the scope 100% together..... and it decided to be nice with me : it still works, PHEW !!! 

Took a couple pics, I can put it on-line for sale now ! 

Quite happy, turned a 50 Euros brick into a fully working scope with only 5 bucks worth of parts, the CPU oscillator and that's it ! 

Now on to the next repair, another scope undoubtedly... what's left that I could fix and sell... which one(s) would I survive without...

[shakalnokturn]

For sure I wouldn't have kept the Gould...
It seems a lot of bother (and even more potential bother depending on who buys it and how it behaves) for the price you're selling it.
A few months ago there was a 100 or 200MS/s Gould going for 70€ with one defective channel, I hesitated and figured it wasn't worth the risk. Gould is not a name that sells that easily in France.

[Vince]

Oh, you saw my ad on LBC for it ?  120 Euros, seems like a fair price given the poor performance of the thing, can't really ask too much for a 20MHz scope can I !        not to mention as you said, that GOULD is not at all as well known and  " admired " as Tektronix !

So I thought I would never manage to get 120 Euros for it, but go figure, only 12 hours after I put the ad on-line, I already have a buyer, and he did not even try to negotiate the price ! 

Seems like a decent guy, if just because he managed to write a 20 line paragraph with zero spelling error, which is unheard of in this country for the past 20 years. Just for that alone I am willing to trust him ! 

[Vince]

Gee that was quick, guy doesn't sleep it seems, 2AM here and he was still on-line ! Already transferred the money, deal done ! 

Couldn't be happier ! I never thought I would manage to sell it for that price, never mind in less than a day ! I guess I am being rewarded for the work I did on this scope... there must be a god of the oscilloscopes out there somewhere ?! 

I mean, with the virus induced economic crisis, so many people are losing their job or at least suffering major income losses, that I expected it would take weeks or months to sell this scope, and that I would at any rate never get the asking price. I am very surprised indeed. I guess there must still be people out there who don't suffer financially from the virus... lucky them !

[Kleinstein]

I got a Gould 1604 for repair.  So far the repair was easy, an exchange of 3 electrolytic capacitors and 1 burnt resistor. I have not put it all together again, but so far it looks good: the supply is back to normal and traces are visible.

There is one question remaining: how important are the batteries for the memory bachup. As expected the batteiries are old and are starting to leak - so they must go. At least no damage to the baords so far. 

Are the batteries really needed and if so, what would be a suitable replacement ?
As far as the calibration instructions go, there are no calibration constants stored in the battery backuped RAM.
There may be NiMh batteries, but they don't like the constant tickle charge very much and may not last very long.
Super-capacitors may be an option (maybe with faster charging - which could be a plus for a more infrequent use), but maybe with a limited old time.
Holding the settings for 1 day or so would already be not so bad.

[Vince]

Wow can't help you much, I sold this scope a year ago ! 

I have long forgotten the details of it. Just read my thread maybe.

From what I remember both the batteries on the motherboard and the one in the external pod looked like standard Varta batteries, should be pretty straightforward to replace them...

[Kleinstein]

Put is back together and it works for 3 out of 4 channels.  Channel2 has a problem (huge offset and no reaction to trimmer) with the input amplifier, but I leave it like this. The actual input uses a small ceramic hybrid with what looks like a chip on boad for the 2 input fets. The part on the main board uses 2 hard to get transistor arrays. So a repair would be very hard.

It is odd they used an extra hybrid for a 20 MHz input BW. 

[Vince]

Yes, they are really weird scopes... very complex for no good reason because very low specs at the end of the day.
A comparable Tek scope of the day was much better designed simpler and better specs and features. The Gould just doesn't make any technical sense...

I honestly was kinda disgusted after repairing mine, so much so I doubt I will ever buy another one for repair.. they are just not worth the trouble, IMHO...

I bought mine purely out of the curiosity factor, because I thought they looked "cool"... but the coolness factor soon vanished as IW as working on it... 

[serban]

Hi !
I managed, including with your help, to bring to life a DSO1604 oscilloscope. I can't understand how they measure the volts, respectively the frequency !!! I don't have interfaces with it ... and neither does that remote control !!!
Any advice is welcome!
Thanks.

[HobGoblyn]

Completely off topic, in my computer operating days in the 80s, our oldest computer was a Gould/SEL, haven’t heard the name in years.

We had 6 of the old vacuum tape drives on it. 

Had a few good simple games on it like the text adventure colossal cave and a chess program.

What was great about playing chess was that all 6 drives would pause while it computed its next move.

[Vince]

Hi !
I managed, including with your help, to bring to life a DSO1604 oscilloscope. I can't understand how they measure the volts, respectively the frequency !!! I don't have interfaces with it ... and neither does that remote control !!!
Any advice is welcome!
Thanks.

Sorry for the late reply... Gmail somehow decided to send all my EEVBlog e-mail notifications... to the SPAM folder ! 

I don't have this scope anymore so can only go by memory but... I don't think you need any interface/pod to get the voltage cursor measurements, should be a basic feature. I don't recall having any difficulty getting the cursors to work, I just fiddled with the various knobs and buttons until it worked.... it was intuitive enough, don't recall struggling at all.

At any rate just go download the user and service manual to know all about this scope, they are out there on the web, had no problem finding them. So just click click click..... 

I would attach the user manual here but the PDF file is a bit over 4MB in size and the maximum file size allowed for attachment is precisely 4MB... so won't work.