Tabor 8020 Function Generator - smoke escaped!

[ckm]

All,

I've been putting together an electronics bench with a bunch of older electronics test equipment.   I'm not a pro in any way, but I have been designing electronics for many years as a hobby.

Well, I had an interesting adventure today.   I have a really nice Tabor 8020 20Mhz function generator I got for cheap and I was testing it with another cheap find today, a LeCroy LC534AL 1Ghz DSO....  I was just sitting there looking at the waveform when all of it went kinda funny.  Because the leads basically go across my desk and are just a probe connected to an aligator clip, I though it was just a bad connection.

Then I looked up and saw smoke pouring out of the 8020!  I wish I'd had the presence to record it but my first thought was to yank the power out of it....

Anyway, I opened it up quickly (while still hot) hopping I could find the source of the smoke...  Well, I didn't need to rush, here's what I discovered:


Tantalum caps roasted


Another shot


One of his friends that cooked off before....


Underside of board looks OK


Metal shielding took the brunt of the burn.

So basically, a couple of tantalum caps just lost it.   I'm trying to get the schematics from Tabor right now - luckily it's all one sided through hole.   Looks like I am going to have to replace most of the passives in the area.   

Does anyone have a good way of determining the values of passives if you can't read the markings?   Also, any thoughts on using something other than tantalums in this application?  Of course I have no idea what this part of the circuit does, although it was under a metal shield.

I'm also looking at that large yellow resistor & wondering if it was damaged as well...

Thoughts, ideas, action plans are most welcome.

Chris.

[granfaloon]

Those look like power supply bypass caps…..comparing the schematic to your pics looks like C54 and C52 are 10uF 35V tantalums. They are bypassing +15V and -15V supplies on U43.

I attached the manual with schematics as a PDF file, hope it helps!

[dc101]

Definitely worth trying to fix!  With any luck the failure was just the result of old age and not a sign of something else.  I had the same problem with an Agilent 81110A just the other day.  I was really worried about the FPGA, but it seems intact.  Good luck!!

[ckm]

Those look like power supply bypass caps…..comparing the schematic to your pics looks like C54 and C52 are 10uF 35V tantalums. They are bypassing +15V and -15V supplies on U43.

I attached the manual with schematics as a PDF file, hope it helps!

Wow, thanks for that, it's HUGELY helpful.   

Thoughts on replacing them with electrolytics?  OTOH, they lasted this long (probably 15-20 years) so it should be OK.  I hope the rest of the passives didn't get damaged.   

This will give me a reason to try out my Hakko 470 desoldering station I got a couple of days ago...

[ckm]

Those look like power supply bypass caps…..comparing the schematic to your pics looks like C54 and C52 are 10uF 35V tantalums. They are bypassing +15V and -15V supplies on U43.

I attached the manual with schematics as a PDF file, hope it helps!

@granfaloon - I downloaded the attachment, it's not the manual but a Fluke patent from 1994.  It's interesting, but it would be more helpful if you could upload the 8020 manual....

Thx.

Chris.

[granfaloon]

Oops sorry, I uploaded the wrong file! seems like the file is too big for the forum (2.5MB). If you shoot me an email I can send it to you there.

You could try electrolytic caps, they will work but tantalums do a better job of filtering out high frequency content which is usually why they're seen bypassing IC supplies. But electrolytics of similar ratings should do the job as well in a pinch…Also I would probe those pins to make sure nothing is shorting to ground before replacing them, something caused those guys to heat up abnormally.

Good luck!

[ckm]

Oops sorry, I uploaded the wrong file! seems like the file is too big for the forum (2.5MB). If you shoot me an email I can send it to you there.

You could try electrolytic caps, they will work but tantalums do a better job of filtering out high frequency content which is usually why they're seen bypassing IC supplies. But electrolytics of similar ratings should do the job as well in a pinch…Also I would probe those pins to make sure nothing is shorting to ground before replacing them, something caused those guys to heat up abnormally.

Good luck!

Thanks, will probe them.  Actually, I realized that the schematics & parts list are at the back of the user manual... It's so rare these days I didn't even think to look.

Thx.

Chris.

[ckm]

Stupid question:

I ordered new tantalums and a couple of ceramic caps that appear to be damaged, specifically the blue ones in the first picture.   Looked at the parts list and spec'd exactly what was in it on Digikey. 

Got the order today - the replacement ceramic caps are a different color, the typical flat orange.  The ones on the board are shiny blue.   Now, these are pretty vanilla caps as far as the parts list specs:

CAP CER 33P 20% 50V
CAP CER  .1-20+80% 50V

What I am wondering is whether the color could mean something not in the parts list, like low ESR?   Also, I thought that shiny blue was a poly cap, not a ceramic - but the parts list says ceramic....   Or is this just manufacturing differences?  There are both generic ceramic caps and these blue ones on the board so it could be conscious decision...

Should I just stick them in there and see if I can generate more smoke? 

[lowimpedance]

Color means nothing, they are standard multilayer ceramics for power supply bypass duties (0.1uF) .
The 33p probably is a NPO type.
Also I would suspect your blackened resistors are most likely okay value wise, (have you measured them for a sensible value?). Im sure you already gave the area a good clean!.
Tantalums of this vintage are ticking time bombs and usually go short and let go the magic smoke.
 I would replace all in the unit with new ones of the appropriate ratings. Be suspicious of other vintage gear you have that use similar tant's.
A photo of the front of the generator would be nice if you get the time!.

[ckm]

Well, I replaced all the tantalums on the main board and the two ceramics right near the damaged area (one had a melted case).   

The unit ran for about 30 seconds then shut down before I could check it on the scope.   

I double checked the orientation on the tantalums, and it looks ok.   One thing to note is that there were no indications of the + leg on the silkscreen for the two burnt out capacitors (see photo).   I had to refer to the circuit diagram which indicated that C52 had negative to ground and C54 had positive to ground (see attached).   The others were marked on the silkscreen so I followed that.  C48, which I was worried was damaged, tested OK on a capacitor tester.  Which is a good thing because a 1% axial capacitor in that value is unobtainium.

I did check for a short between the rails, but the resistance was high enough that it seemed there wasn't.

I guess my next step is to power it back up & test the voltage in various places.   I'm thinking that U43 may be damaged, but I don't know for sure, although it does have a fairly high input voltage tolerance.  It could be that something in the power supply has gone south, I suppose.

If anyone has any ideas, let me know.  The manual is available here (PDF on Google Drive) - http://tinyurl.com/o7g4co5

Edit: took a further look around, looks like the input AC fuse is blown - there is likely a short to ground somewhere in the power rails....  Now how to find it?  Also uploaded a pic of the front panel as requested.

[ckm]

I'm thinking that, beyond poking around with a multimeter, I could inject 5V, low amps as a trace.

Otherwise, how do people usually tackle finding a shorted component in a board?

Edit: pretty much isolated the short to the +15vdc rail - now I just need to find out where it is....  It would be easiest to power it up & look for a hot component, but I have to build a bulb-attenuator first otherwise it will blow the fuse in 30 seconds...

[lowimpedance]

The bulb tester is a good start, is it possible to isolate the power supply regulators outputs easily? (so you could connect an external current limited supply). Also might be worth checking the electrolytic's in the supplies, they might have dried out.
Good to know the culprit, post any developments.

[ckm]

Thanks for the tips. 

Power supply is integrated to the main board, so not easy to separate.  I would have to desolder the output regulator (an LM317T, seems OK) and there are five rails total: +/- 15vdc +/-24vdc +5vdc - only +15 is shorted.

There's only one electrolytic cap after the regulator, I removed it and it tests OK.   There is still a short in the circuit after the cap I removed.  I should note that this unit is from 2007, so not that old.   The other electrolytic caps are before the regulator & look pretty much fine.   There are also a couple on the 5vdc rail, but that's not causing a problem AFAIK.

I think it may be an op-amp, I'm looking closely at U28 right now.  It's an OP07CP and is one of the only ones operating on the +15vdc rail.  Both it's + & - rails go through tantalum caps, one of which previously failed.  The other possibility is the other IC served by burnt caps, a CA3102 at U43.

Other than that, I don't know.  I don't have a working variable power supply right now as the one I have was DOA (another fix, perhaps), but I suppose I could power it up with bulb trick & try some freeze tests....  Or I could desolder U28 or U43 & see if the short goes away...

[lowimpedance]

The OP07 is easy to get so i would go for that first. then check for shorts after removal.
Interesting that its not that old and the some tant's decided to have a BBQ !.
The CA3102 could be a little harder to get, (ebay?!.).
The final amp stages use the +15V rail, any shorted transistors there could drag the rail down.
Could do a quick 'diode junction' test on the main output transistors before removing the op amp.

[ckm]

Who knows how old the parts are.  The CA3102 went out of production in 2005....   I don't remember how I know it's from 2007, it possible that was just the copyright on some of the literature, so I could be wrong.

As far as the CA3102 - Intersil still has them in qty. 25 - god knows how much they want but it could conceivably be reasonable...   I suppose Tabor might have some as they have a part number for it.... 

I'll have to look at the final amp stages, but IRC they are on a separate output board, not on the mainboard.   At least, that's where the output connector is located.... I have the output board disconnect and there is still a short.   

I think my next course of action will be to desolder the OP07 and see if I still have short.  I should probably get a socket for it. 

Will do the diode test on the output transistors as well, assuming the are on the main board.

Thanks for helping out - I'm a little out of my depth trying to figure this out.

[ckm]

Well, I think I've isolated the problem section to the schematic attached.  The method I used was to start at the +15vdc ps output and, with my DMM negative probe attached to ground, slowly work my way down the rail until the resistance started to rise.

The theory is that the resistance drops as you get closer to the short, then rises after that.   This only works if you have a high-digit DMM, I have a 6.5 digit, so it shows enough precision to theoretically work.  I'm not sure it's entirely accurate, but since I don't have a way to power up the board safely (yet), this is a start.   Perhaps not coincidentally, I wound up in the same general area where one of the tantalums flamed out.  @lowimpedance suggested that I check a few of the transistors around the outputs, but few of them are on the +15v rail.

That said, there is a discrepancy between the labels on the silk screen and the schematic which adds to the confusion.  The resistor which is labeled Q11 on the silkscreen has some (heat?) damage around the through holes and it's right next to a tantalum that blew up, on the same rail, as a matter of fact.   On the schematic, Q11 is using the +24v rail, but on the board, it's clearly using +15v (see pics, last pic shows Q11 connected to +15 rail).

So, is 'Q11' damaged?  The odd thing is that it's not connected to ground other than indirectly.  Worth a shot removing it?  I already removed OP07 and that was a dead end (so to speak).

[lowimpedance]

Looking at the schematic the output stage is marked as being connected to the +/- 15V rails where there is low impedance's in the cct (unless its an error on the schematics).

[ckm]

I think that there is a delta between the silkscreen and the schematic.  The reason I am saying this is that Q11 on the board is directly tied to +15v in much the same way Q12 and is no where near +24v as is suggested by the schematic.  And there is no resistor between Q11 on the board & +15.  So obviously something is off unless I'm reading it wrong (which is entirely possible). 

I'm guessing that Q11 & Q12 are swapped.  In any case, looking at the board, there is a transistor with damaged through holes that is tied directly into +15v next to the tantalum that burnt out.   I didn't damage the holes replacing the tantalum as I used a Hakko desoldering station & a fine-tip temp controlled soldering iron...   So, the heat to damage the traces must have come from somewhere else (?).   Maybe that's whats causing a 250v .3amp fuse to blow?

I'm going to build a bulb attenuator in the morning, so I'll be able to better test it then.

Chris.

[Peter1964]

All,

Well, I had an interesting adventure today.   I have a really nice Tabor 8020 20Mhz function generator I got for cheap and I was testing it with another cheap find today, a LeCroy LC534AL 1Ghz DSO....  I was just sitting there looking at the waveform when all of it went kinda funny. 

Hi

Here in denmark one is for sale now. I think about getting it, but what should I pay? (They are asking 250 USD!)
Do you know how old it is and if it is usefull compared to new models like http://www.tequipment.net/Siglent/SDG805/Function-Generator/?b=y&v=71471. Same price for a new nice model....

He also has a Hameg 1507-3 scope for sale - asking 275 USD.

I think both are way overpriced but I can make him an offer.

Regards Peter

[jpb]

All,

Well, I had an interesting adventure today.   I have a really nice Tabor 8020 20Mhz function generator I got for cheap and I was testing it with another cheap find today, a LeCroy LC534AL 1Ghz DSO....  I was just sitting there looking at the waveform when all of it went kinda funny. 

Hi

Here in denmark one is for sale now. I think about getting it, but what should I pay? (They are asking 250 USD!)
Do you know how old it is and if it is usefull compared to new models like http://www.tequipment.net/Siglent/SDG805/Function-Generator/?b=y&v=71471. Same price for a new nice model....

Tabor have a lot of information still on their site (which they have just revamped in the last few days I think):
http://www.taborelec.com/8020
If you go to the download section and look at the dates you can get an idea as to when the model was first produced - the drivers date from 2004 but the datasheet and manual date from 2006/7.
Tabor is a good make - they have an OEM deal with Agilent as well as producing their own stuff since 1972 I think.

I've been looking at their AWG (ww1072) as I was offered a good deal on a new one. It is still a current model but the manual dates from 2005 which shows when it first came out.

They have a 5 year warranty on new models so I think their stuff is well constructed - a different class from Siglent I think.

[MadTux]

Anyone with a fully working Tabor/Kontron 8021 who can send me EPROM  data?

I grabbed a couple of Tabor/Kontron 8021s on ebay for cheap a while ago and the output voltage offset in all of these doesn't seem work. It's weird, because the voltage offset is broken on each and every unit and I'm quite certain that I'm not too stupid to use it.

Other than that, it's quite a nice function generator. Lots of features, a bit slower, but similar to HP-8116, but build completely from OTS components without the ASIC madness usually found in HP function generators.

I grabbed my logic analyzer and probed U13/CD4094 shift register that controls voltage offset DAC.
All outputs are low and it always gets 0's shifted into before the strobe pulse appears. Clock and data line seem to work properly (data line has data for other shift registers). It now seems to me as if the firmware was modded to disable voltage offset for some large customer. (Bought these instruments from the same guy and they all seem to originate from the same place)
EPROM is 27C512 and is labeled "8021/ Rev 3.00 / c.s.EC12"

[bitseeker]

@ckm: Thanks for posting your investigation. I have an 8551 from that "brown" era. I don't remember the build year, but likely even older than yours. Looks like I've got a tantalum cap replacement project to look forward to avoid an upcoming benchtop barbeque. Good sleuthing thus far. Looking forward to the happy ending.

Tabor is a good make - they have an OEM deal with Agilent as well as producing their own stuff since 1972 I think.

@jpb and Peter1964: Tabor does or did OEM for many well-known brands including Fluke, WaveTek, LeCroy, Agilent and Racal. I think they made some stuff for Keithley, too. When equipment with the well-known brand labels show up on eBay, they go for crazy prices. The Tabor-branded ones usually are a better deal with less competition.