Help restoring 1980s era electronics (Wavetek 190 function generator)

[robca]

I managed to get a Wavetek 190 20MHz function generator on eBay, for basically the cost of shipping. The device is banged (plastic enclosure is missing a few corners), but after a good cleanup of the front panel is starting already to look the part.

I quickly tested it, and it works well up to ~4MHz, beyond that the waveforms start being distorted, but it gets to 20MHz with still a recognizable signal. One of the reasons why I decided to go for it, is that not only the schematic is provided, also the full calibration procedure including waveforms on various test points are included http://bee.mif.pg.gda.pl/ciasteczkowypotwor/Wavetek/190.pdf. As long as the few impossible to get ICs keep working, it's easy to fix any potential problem.

The pots and rotary switches seem to be ok, but I'd like to restore/lubricate them as preventive measure. I have some DeoxIT D5, but I was reading that for older stuff DeoxIT DS100 is safer to use, since there is no solvent.

Given how many people here restore even older instruments, what's the suggestion to renew pots and rotary switches? Do it or not? And, if so, what's a good product to use that won't make things worse by making plastic brittle

[bob91343]

If it ain't broke, don't fix it.  Every time you get into a piece of gear you risk breaking something.

I have a similar Wavetek.  I use it so seldom that I don't remember the model number and probably should get rid of it.

[ChunkyPastaSauce]

The pots and rotary switches seem to be ok, but I'd like to restore/lubricate them as preventive measure. I have some DeoxIT D5, but I was reading that for older stuff DeoxIT DS100 is safer to use, since there is no solvent.

Given how many people here restore even older instruments, what's the suggestion to renew pots and rotary switches? Do it or not? And, if so, what's a good product to use that won't make things worse by making plastic brittle

F100 and DFG-213 are lubricants only (propellant in F100), with no solvent, and supposed to be safest for plastics. No oxide remover like D100, but if the switches work electrically fine then skip I'd think.

[Hexley]

You might take a look around for dipped tantalum capacitors. Wavetek used them in that era, and those caps have a tendency to fail short circuit without prior notice.

Also, check the crimp terminals on J5, and the PCB solder joints on the associated connector for intermittent connections.

YMMV, but I had a Wavetek 148 for many years and experienced faults with all of the above. 

[robca]

You might take a look around for dipped tantalum capacitors. Wavetek used them in that era, and those caps have a tendency to fail short circuit without prior notice.

Also, check the crimp terminals on J5, and the PCB solder joints on the associated connector for intermittent connections.

YMMV, but I had a Wavetek 148 for many years and experienced faults with all of the above.

Very helpful, thanks.

I checked the schematic and part list, and there are ceramic, mica, mylar and electrolytic capacitors, no tantalum. Visually I can't find anything that looks like a dipped tantalum capacitor either, so hopefully I'm good there

J5 looks good, too.

Next step is looking at calibration, and deciding if I want to attempt it or use it as it is

[robca]

Pots are all cleaned with DeoxIT D5, and it works much better. Before cleaning, every time I turned one of the pots even a bit, noise would show up on the scope, now everything is clean looking. In the end I decided that since the generator itself cost $30, spending another $30 for DeoxIT 100 didn't seem appropriate . And these days there are fewer and fewer pots to deoxidize. As a matter of fact, the DeoxIT I used was a good 15 years old, and still basically full.

I noticed that the front panel was supposed to have a 5V micro lamp over the main frequency selector, long since burned, so a new orange LED with proper resistor has taken its place. The lamp was help in place with a blob of transparent silicone, so it was easy to fit a LED in the hole and use hot glue to hold it in place. Power for the lamp came directly from a 7905, so safe to replace the lamp with a lower draw LED (15mA vs 50mA)

While working on it, I realized that the online copies of the Wavetek 190 manual were scanned with decent quality, but never had OCR applied to it, so the PDF was not searchable. I applied OCR and proofread all the manual (just didn't too the schematics, too much work going thru everything there and marginal utility). Unfortunately I can't upload it here, as even compressed it's more than 5Mb. If anyone needs it, or wants to replace some of the online copies, ping me

Next step is calibration. Works well up to ~5MHz, then the waves get asymmetrical. Square waves will always have problems above 10MHz, since the rise/fall time, according to the specs, is <15ns, and I measure 8ns. A trapeziodal would work, but right now there is a lot of ringing, which makes the wave really not that square. Also the triangular wave has a tiny defect at the peaks and valleys above a certain frequency. The manual lists calibration and repair steps, with plenty of expected voltages and waveforms at various points. so a fun project

[robca]

Adding inside pictures. I love how clean everything is, and how everything is clearly labeled, including macro functions on the PCB.

The top part is just one big heat sink (the case has no ventilation slots), and the component side faces the bottom of the case

[cliffyk]

Not really relevant, nonetheless here i go:

Years back I had a Wavetek 288 I REALLY liked, it died from failure of an unobtanium proprietary IC.

A bit later I got a Velleman PCGU-1000 (actually a pretty nice piece for it's time) and created a Windows app, Wavetek-like front-end for it:

[T3sl4co1l]

Nice!

Yes, contact cleaner is likely a good idea.  My 193 needed it badly; it's perfectly usable now.  Aside from the hack I did on it:

The output transistors are prone to burning out, into low or shorted loads. Didn't have any of the TO-39 transistors handy, so did the next best thing, grabbed a pair of the fastest audio driver transistors I had.  Which isn't saying much, they're like 100V 1A 50MHz.  So it rolls off pretty badly over the last decade

Think the differences between the 190 family are: 190 looks to be the base board only; my 193 has a top board with aux oscillator, including sweep, trigger and modulation (AM) options; the 191 I think is a pulse gating/sequencing version?  Haven't seen any others (also haven't even looked at what the full lineup was, so I just don't know).

Don't recall there's much unobtanium inside; functions are mostly ECL (MC10k, it's been around forever) and reasonably common op-amps.  Monolithic matched arrays, mixers and whatnot may be hard to find.  And of course, the fast output transistors.  Although come to think of it, I don't recognize whatever the hell that disc-on-leads, left (in side view; rearwards) of C73 is...

Tim

[SilverSolder]

If it ain't broke, don't fix it.  Every time you get into a piece of gear you risk breaking something.
[...]

This goes against the EEVblog spirit - "Don't turn it on, Take it apart!"

[bob91343]

Mine is a model 193.

[robca]

Nice!

Yes, contact cleaner is likely a good idea.  My 193 needed it badly; it's perfectly usable now.  Aside from the hack I did on it:

The output transistors are prone to burning out, into low or shorted loads. Didn't have any of the TO-39 transistors handy, so did the next best thing, grabbed a pair of the fastest audio driver transistors I had.  Which isn't saying much, they're like 100V 1A 50MHz.  So it rolls off pretty badly over the last decade

Think the differences between the 190 family are: 190 looks to be the base board only; my 193 has a top board with aux oscillator, including sweep, trigger and modulation (AM) options; the 191 I think is a pulse gating/sequencing version?  Haven't seen any others (also haven't even looked at what the full lineup was, so I just don't know).

Don't recall there's much unobtanium inside; functions are mostly ECL (MC10k, it's been around forever) and reasonably common op-amps.  Monolithic matched arrays, mixers and whatnot may be hard to find.  And of course, the fast output transistors.  Although come to think of it, I don't recognize whatever the hell that disc-on-leads, left (in side view; rearwards) of C73 is...

Tim

Wow, good eye  . That weird disc component is R161, part of the Sine converter circuit and according to the manual "Thermister R161 adjusts the input voltage to compensate for the diode voltage change with temperature." Good thing I had just completed OCR on the manual, so I could find it without having to read everything  . And according to the component list is THERMISTER. 10. OOO  OHMS .25%

And, yes, from what I can see, the 193 is simply a 190 plus an additional board

Thanks for the heads up on the output transistors (and good to know replacements can be found). Yes, I took into account the risk of having an unobtanium component die. But given all the functional blocks are explained, I thought that should one die I could try to replace that section with modern equivalents. Given how much I paid for it, either way I'm ahead

The 20MHz square wave is not exactly a square, but after reading thru the whole https://www.eevblog.com/forum/testgear/show-us-your-square-wave/ thread, I have to say I feel pretty good about what I get at 20MHz out of an instrument made before the original Top Gun came out. I still need to see if I can get the triangle wave less wonky, but overall I'm really happy: anything below 5MHz is just about textbook, and slowly degrades above that