70's Vintage Function Gen Question

[Forced Perfect]

Hey all,

I recently scored an IEC F51 Function Generator (10 MHz) on eBay for $7 ($13 of the original $20 was refunded due to it being dead ). It's 1977 vintage (it has some dated ICs).

It's outputting only a flat 24V on the "high voltage" output and nothing on the low. I've tracked down a couple problems, notably some caps which are falling apart, and some dodgy supply rails (-24 @ -2 for example). Hopefully the power problems are simply crapped out regulators, but it's also possible the boards are shorting in some way due to the caps.

My main question is about the 70's capacitors. The wide ones which are EVERYWHERE in this thing are labeled ".1 50V 648", but I'm not 100% sure if they're just plain ceramic caps as the thing also has your more traditional rounded ceramics elsewhere. A few of the .1 50Vs are falling apart, so I would like to replace at least the more damaged ones, if not all. I just need to know if I can just grab a pile of ceramics or if I need some kind of funky cap since I don't have much experience with older packaging styles.

Some of the caps in question. The smaller one behind the leftmost one appears to be a fancy mica cap. The micas all look like they're in mint condition physically, but the ".1 50V 648"'s in front look like they will shatter if looked at too long.



The rest of the board is very nicely done. Very densely packed for the most part. The adjustable caps are a little grungy.





Eat your heart out, oscilloscopes. THIS is a rotary dial.


P.S. If anyone knows of a service manual (or any manual for this thing, really), let me know. I can't find much information at all on this model, let alone manuals.

[SoundTech-LG]

Reminds me of my early 70's vintage Priceton Applied Research model 110 Low Distortion Oscillator/Notch Filter/Selective Amplifier/Distortion Analyzer. Have to get some photos up it has ROTARY SWITCHES too!!

All discrete transistor. No IC's. Had 8 shorted capacitors!!! Measuring 0.0025% Distortion @ 1kHz on my Sound Technology 1700B, but on a better analyzer with a lower noise floor, it could be less  .

[c4757p]

The capacitors you're asking about appear to be film caps, probably relatively common types. Definitely replace them,moisture starts to get in when they've cracked, and then they go leaky.

[Forced Perfect]

The capacitors you're asking about appear to be film caps, probably relatively common types. Definitely replace them,moisture starts to get in when they've cracked, and then they go leaky.

Thanks for the help. These look suitable and not too expensive for the 50 or so I will need.

http://www.digikey.com/product-detail/en/ECQ-V1H104JL/P4525-ND/13790

[c4757p]

Yes, I'd have gone with something like that. Those Panasonic ones aren't bad

[David Hess]

My main question is about the 70's capacitors. The wide ones which are EVERYWHERE in this thing are labeled ".1 50V 648", but I'm not 100% sure if they're just plain ceramic caps as the thing also has your more traditional rounded ceramics elsewhere. A few of the .1 50Vs are falling apart, so I would like to replace at least the more damaged ones, if not all. I just need to know if I can just grab a pile of ceramics or if I need some kind of funky cap since I don't have much experience with older packaging styles.

Some of the caps in question. The smaller one behind the leftmost one appears to be a fancy mica cap. The micas all look like they're in mint condition physically, but the ".1 50V 648"'s in front look like they will shatter if looked at too long.

0.1 microfarad (or whatever value is needed) film or C0G type ceramic capacitors with an appropriate voltage rating will certainly work fine to replace them.  It is difficult to say more without seeing a schematic.

If they are used for decoupling then cheaper X7R type ceramics may be used instead.

The rest of the board is very nicely done. Very densely packed for the most part. The adjustable caps are a little grungy.

The exposed variable capacitors are silver plated and just tarnished.  That is nothing to worry about.

[lowimpedance]

These 'old' style PCB's are a joy to look at with the nice engineering of the day, but can be a real pig to rework if the copper bonding was not brilliant to start with so be careful with the iron. Cut the 'cap' leads and then remove the lead stump's and then suck the solder from the hole last.
One other point to note if you get waveforms that are not quite right, keep in the back of the mind that drifting resistor values will cause problems in DC balance conditions (since these types of generators are full of DC coupled sections). And of course all the other more 'usual' broken transistors stuff etc !.

Looks like nice vintage kit from the internal photos, perhaps some more outside pictures too 

[Forced Perfect]

These 'old' style PCB's are a joy to look at with the nice engineering of the day, but can be a real pig to rework if the copper bonding was not brilliant to start with so be careful with the iron. Cut the 'cap' leads and then remove the lead stump's and then suck the solder from the hole last.
One other point to note if you get waveforms that are not quite right, keep in the back of the mind that drifting resistor values will cause problems in DC balance conditions (since these types of generators are full of DC coupled sections). And of course all the other more 'usual' broken transistors stuff etc !.

Looks like nice vintage kit from the internal photos, perhaps some more outside pictures too 

Thanks for the advice on working on the PCBs, I haven't worked on older stuff too much so I probably would have ripped half the traces off. haha

Outside pictures? No problem-o.

The main frequency dial is SO nice. It's going into a gigantic pot and you can turn it to set the value, then grab the center of the dial and it's geared to do a fine adjustment of the outer ring.


Quite a few different waveforms available.


There is also a large red sticker on the top telling you it is the calibration standard and to never remove it from the lab.


Left all the screws out. One thing to note is it has a voltage selection switch.


I like how you can optionally disconnect the chassis ground by removing the screw from the post in the middle.

[Forced Perfect]

I've ordered replacement caps (picked up 70 to be safe), plus replacements to the 6 TO-220s used in the power system. They all look really crusty. I'm no expert on "compatible replacements", so I tracked down all the damn TIP29A, TIP30A, etc. parts used. All in all about $18 in parts. The learning alone will be worth the money - even if I fail to revive it.

I noticed every single DIP IC is an MC1709L op-amp in a fancy ceramic 14 pin DIP package.

[T3sl4co1l]

Beautiful!

In the 3rd pic, near center, there's a small "tophat" style diode in a yellowed plastic tube -- looks like a tunnel diode like Tek used in those days.  If that's the case, possible uses would be very fast comparison / detection, or flip-flop action.  Once working, this thing probably does a fine job; wouldn't be surprised if some of those features result in lower noise (more stable waveform, etc.) than my Wavetek 193, which isn't too bad of an instrument itself (and is... 80s vintage I think, chock full of 10k ECL logic and various other stuff).

And yeah, all those rounded rectangular looking caps are film.  Odd that they've somewhat.. decomposed, but not hard to replace anyway (assuming the board is in okay shape).

Other things to check for: corrosion (that stuff around Q7 doesn't look too happy, be sure to get in there and clean it up as best you can; re-tin the traces and make sure there are no breaks; find anything else as well), and possible failure around the output circuit (I've had more than one generator die of output-related trauma: one due to backfeeding from an overload, another just from running it into a low impedance load... yeah you'd think they'd have made the thing well enough to handle itself, right?...wrong..).

Tim

[Forced Perfect]

that stuff around Q7 doesn't look too happy, be sure to get in there and clean it up as best you can

I specifically took that picture because I saw how terrible it looked. In some of them, it looks like the center pin isn't even soldered. However, it could just be the angle and a shallow solder blob.

I noticed it pulls ~20-23W despite not working. I didn't see anything melting or whatever from a short, but it does get warm as a whole.

One of the most unique parts I noticed is the 2Nxxx series transistor in the picture showing the Q7 grunge. It's the black and white plastic looking thing. I thought it was a plastic rivet or something, but it's actually Q9. haha There are two of them on the back power supply board.

The board has nicely labeled test points, but not so much on the actual gen board. It has TP1 to TP-42billion but none are labeled. Too bad there doesn't seem to be a service manual around.

[T3sl4co1l]

Yeah, there's some old TO-whatevers of the "plastic blob" style.. they went away pretty well after the molded kind came up.

Likely the 20W is typical operating current and most amp stages or ramp generators or whatever are still doing what they're supposed to be doing, but the voltage somewhere or another is totally screwed up.

Tim

[david77]

Lovely piece of kit. I love this 70's stuff, so nice and colourful to look at  .
Reminds me of my Exakt 7060  generator I fixed a while ago, same vintage and same construction style.
Check all the electrolytics for shorts, tantalums also if there are any in there. I couldn't see any in your pics. I wouldn't worry too much about the 0.1µf caps to be honest. It is rather rare for foil type caps to fail.
Check the PSU section first, obviously. Check if anything gets unusually hot, check for shorts in the supply rails. Without schematic it's a bit harder to guesstimate what the voltages should be but not impossible.
Don't worry too much about the boards, in my experience it's actually much easier to work on this old stuff then on modern multi layer boards. They used proper solder and the holes aren't so small usually. The occasional lifted trace is always repairable and not the end of the world.

I couldn't get a schematic for my function gen either, so I studied schematics of other models and brands with similar features of about the same vintage, after all they were all only cooking with water. These smaller manufacturers surely didn't reinvent the wheel with every new piece of gear. Some Wavetek stuff was quite similar to my Exakt unit, so much so that I could at least about understand how the circuit works.

Edit:
Check the output stage. It's likely it'll be something very similar to an audio power amp, a push-pull stage probably. They can easily be fried, but usually are also easy to fix.

[PA0PBZ]

I wonder who had some fun with the trim caps. I can see one fully closed and 2 fully open, which is strange and I think I can even see where one of the fully open ones used to sit by looking at the color... 

[TerraHertz]

Nice looking bit of gear.
Google found this: http://home.cogeco.ca/~rodemi/Service%20Manuals/Manuals%20TEST%20Equipment.htm
that lists it, but sounds like you'd need to contact the guy.

[Seekonk]

I found a bag of a couple hundred new dipped caps in a bag that dated from probably the early 70's.  They were a shiny tan with the painted stripes.  As I looked at them closely I could see hairline cracks in some.  The leads would easily break off at the weld even on the ones that weren't cracked. 

[Forced Perfect]

The leads would easily break off at the weld even on the ones that weren't cracked.

Sounds like all of the ones on my board. Hopefully I don't go crazy replacing them all.

Likely the 20W is typical operating current and most amp stages or ramp generators or whatever are still doing what they're supposed to be doing, but the voltage somewhere or another is totally screwed up.

Yeah, it didn't seem unreasonable. But all the negative rails seem way off (like -2 instead of -24 and -1 instead of -5). The +24 is at +26, which is reasonable. Even just moving the caps around to read the values caused some to completely fall apart. I think I have to replace all the .1 50V caps before I try anything else.

I wonder who had some fun with the trim caps. I can see one fully closed and 2 fully open, which is strange and I think I can even see where one of the fully open ones used to sit by looking at the color... 

I noticed that! I was like "Why did you put in trimmers if you aren't using them in the slightest on a bit of precision test equipment?", maybe mine is just PERFECT aside from half of the components are disintegrating!

Oddly, there is a giant poly film cap (the big yellow ones which look like layers of paper) rated for 250V way far away from the power supply. I think it is wired into the output in some way, but I haven't traced it out yet. No idea why there would even be high voltage on the board at all aside from the 24V for the HV output and probably op-amps.

[T3sl4co1l]

I noticed that! I was like "Why did you put in trimmers if you aren't using them in the slightest on a bit of precision test equipment?", maybe mine is just PERFECT aside from half of the components are disintegrating!

Although -- at least ideally -- you want to design your equipment so, even if every single trimmer is adjusted to its extremes, the circuit still works.  Maybe not within nameplate spec, but within a looser (design) spec.  And certainly doesn't destroy itself.

An example would be, a voltage sense circuit which is built with 1% resistors -- so it has 2% worst case error -- that's trimmed to 0.1%, say.  Even with the trimmer off in la-la land, it's still usable, and not particularly dangerous.

Whereas, the idiot-mode "I don't know what resistor to use, TRIMMERS FOR EVERYONE!" approach might replace one 1.00k voltage divider resistor with a 10k 10-turn pot: in that case, being off by a factor of 10 (or worse, since the pot can go all the way to zero!) could destroy the circuit!  (Suppose the divider is within a feedback loop, and turning it all the way to zero causes the device to deliver full power -- it never gets to know what the voltage actually is, in a reasonable sense, and has no way to protect itself, if it has any protection circuits at all!)

So, I wouldn't be surprised if your generator was designed with the former, rather than the latter, in mind.  It could very well be a "trimmer for everyone" situation, though... in which case, good luck calibrating it without a procedure (and, perhaps, even with..).

Oddly, there is a giant poly film cap (the big yellow ones which look like layers of paper) rated for 250V way far away from the power supply. I think it is wired into the output in some way, but I haven't traced it out yet. No idea why there would even be high voltage on the board at all aside from the 24V for the HV output and probably op-amps.

That's not uncommon.  I think my Wavetek has 100-250V caps in it for timing -- the 10uF for the lowest range is, well, pretty obvious!

The primary design decision there is, polypropylene (or slightly better yet, teflon or polystyrene) is a low loss and low absorption dielectric, so it makes a good timing capacitor.  Large values in low voltage ratings are uncommon -- and not necessarily desirable anyway (stability, accuracy, electrostriction / piezoelectricity, etc.), hence the oddly huge values.

Unless there is actually a vacuum tube in there and it's filtering a 200V rail. 

Tim

[David Hess]

I noticed that! I was like "Why did you put in trimmers if you aren't using them in the slightest on a bit of precision test equipment?", maybe mine is just PERFECT aside from half of the components are disintegrating!

Although -- at least ideally -- you want to design your equipment so, even if every single trimmer is adjusted to its extremes, the circuit still works.  Maybe not within nameplate spec, but within a looser (design) spec.  And certainly doesn't destroy itself.

...

So, I wouldn't be surprised if your generator was designed with the former, rather than the latter, in mind.  It could very well be a "trimmer for everyone" situation, though... in which case, good luck calibrating it without a procedure (and, perhaps, even with..).

I have a couple of Tektronix oscilloscope mainframes where some of the trimmer capacitors have to be adjusted all the way open or closed for best calibration but in one case recently I found a worn aluminum electrolytic decoupling capacitor (Sprague 30D series) which tested good for value, leakage, and dissipation but had become bad at high frequencies so I suspect a similar problem in the other mainframes which I will eventually track down.

The primary design decision there is, polypropylene (or slightly better yet, teflon or polystyrene) is a low loss and low absorption dielectric, so it makes a good timing capacitor.  Large values in low voltage ratings are uncommon -- and not necessarily desirable anyway (stability, accuracy, electrostriction / piezoelectricity, etc.), hence the oddly huge values.

I have used 50 volt polypropylene capacitors before when low leakage at high temperature was needed but they get physically large quickly.  We qualified the capacitors by batch but never found any which did not meet our specifications.

Tektronix used to make their own hermetically sealed oil filled film capacitors for precision timing and HF coupling applications which confusingly look like wet slug tantalum capacitors.  The timing ones were graded into decade sets with the largest value in the 1 to 10 microfarad range while the HF coupling ones were high voltage.

[Forced Perfect]

I have used 50 volt polypropylene capacitors before when low leakage at high temperature was needed but they get physically large quickly.  We qualified the capacitors by batch but never found any which did not meet our specifications.

Tektronix used to make their own hermetically sealed oil filled film capacitors for precision timing and HF coupling applications which confusingly look like wet slug tantalum capacitors.  The timing ones were graded into decade sets with the largest value in the 1 to 10 microfarad range while the HF coupling ones were high voltage.

Grabbed a picture of the big fella, it's labeled:

American Radionic
10.0 +- 5%
200v 1070
2MFW

The large rotary switch next to it is the multiplier selection of the main frequency dial.



Since we're talking caps, here are the power supply caps which are all rated at 50V - at least the large ones are. I guess there's no input side cap? Quality Mallory branded. Note the "high production" of under 600. I don't know how long they made this model though. They all look like they're in good shape (the crap on the top is just peeling plastic), but it's probably hard to tell with these non-vented ones.

[PA0PBZ]

They all look like they're in good shape (the crap on the top is just peeling plastic), but it's probably hard to tell with these non-vented ones.

Oh they can, but not on the top 

[David Hess]

They all look like they're in good shape (the crap on the top is just peeling plastic), but it's probably hard to tell with these non-vented ones.

Oh they can, but not on the top 

They are not vented so much as not well sealed on the bottom.

I have used 50 volt polypropylene capacitors before when low leakage at high temperature was needed but they get physically large quickly.  We qualified the capacitors by batch but never found any which did not meet our specifications.

Tektronix used to make their own hermetically sealed oil filled film capacitors for precision timing and HF coupling applications which confusingly look like wet slug tantalum capacitors.  The timing ones were graded into decade sets with the largest value in the 1 to 10 microfarad range while the HF coupling ones were high voltage.

Grabbed a picture of the big fella, it's labeled:

American Radionic
10.0 +- 5%
200v 1070
2MFW

The large rotary switch next to it is the multiplier selection of the main frequency dial.

I do not think there is any doubt that the capacitor is part of the integrator used to generate the precision triangle waves and basic timing.

Since we're talking caps, here are the power supply caps which are all rated at 50V - at least the large ones are. I guess there's no input side cap? Quality Mallory branded. Note the "high production" of under 600. I don't know how long they made this model though. They all look like they're in good shape (the crap on the top is just peeling plastic), but it's probably hard to tell with these non-vented ones.

Those big Mallory cans are much less impressive when you take one apart because it failed.  The capacitor element inside is basically just a rolled up wad of paper and aluminum soaked in the electrolyte.  The typical failure mode is for one of the leads to corrode inside and eventually break.  Big can aluminum electrolytic capacitors like these release the "electronics smell" which can be quite strong in an enclosure and is just the electrolyte evaporating.

Here is a photo of a Mallory can capacitor which was replaced in a Tektronix PS503 power supply because one of them failed in the way I described.  Under it are the modern replacements I used which were selected based on diameter so they could use the original mounts.

[SeanB]

The typical failure mode of those big caps is the rubber end bung degrades, and this then allows the electrolyte to dry out and allows oxygen in which corrodes the whole unit The bung stays in one piece even with cracks because it is a thin rubber sheet vulcanised onto a bakelite sheet. The cracks are often very fine, and often you find the vent hole has degraded to powder. Until you unsolder it it looks perfect under the sleeve, as the shrunk sleeve and the back silicone rubber packing keep the metal pristine.

[David Hess]

The typical failure mode of those big caps is the rubber end bung degrades, and this then allows the electrolyte to dry out and allows oxygen in which corrodes the whole unit The bung stays in one piece even with cracks because it is a thin rubber sheet vulcanised onto a bakelite sheet. The cracks are often very fine, and often you find the vent hole has degraded to powder. Until you unsolder it it looks perfect under the sleeve, as the shrunk sleeve and the back silicone rubber packing keep the metal pristine.

What you describe makes sense to me.  The sense of smell may be a useful diagnostic tool in this case because you can smell the outgassing of the electrolyte.

Sometimes I find these can style capacitors with low capacitance and high D/ESR and sometimes I find them with practically no capacitance at all in which case investigation shows that the strap from the pin to the element has corroded through.

The one pictured measured about 5 picofarads but its three brothers were all within their rated tolerance.  I changed them all.