Repairing a free wavetek 178 waveform synthesizer.

[lmester]

I bought a Wavetek 178 waveform synthesizer from EBay. I'd call it a function generator with some fancy features. It was listed as functional. Well, I turned it on and got nothing. A big mistake! More on that later. Fan was not even running. I took off the top cover and found the cause of the fan problem. The fan was missing! Some cables were unplugged. I'm lucky it didn't go boom! 120V fan wires were dangling in the breeze!

This was certainly not as described. I sent the seller some pictures. He sent a refund and told me to keep the unit.

Of course I tried to repair it. I now have it working! It's been an interesting repair.

I plugged in the loose cables and put some tape on the fan wires. Powered up and had very low power supply voltages. The power supply has overload protection and was shut down. I started unplugging cables to try and isolate the short. When I got to a ribbon cable that feeds the front panel circuit board I noticed that it was out of it's clips and had a twist in it. This cable was connected when I got the unit. It's not one that I had to plug back in. I checked the cable and Yes, it was not keyed and was plugged in backwards. I unplugged it and powered up. The power supplies came up.

I expected that the front board would be toast because of the reversed connector. I plugged it in correctly and powerd up again.

I got "Wavetek 178 Self test passed" on the display! Lucky me!

If you work on one of these don't run it very long with no fan. After only a few minutes the heat sinks were getting toasty. Power supply heat sinks at 120 Deg. F, power amp heat sink at 140 deg. I aimed a desktop fan on it to keep things cool.

The unit was basically functional. I had to use the GPIB interface to control it. The buttons were really bouncy and some didn't work. It has those metal dome click switches that have clear tape holding the domes in place. The designer of those switches is evil!

Sine triangle etc. waveforms were worrking but all had a high frequency oscillation riding on them. I poked around with the scope and found that the output power amplifier was oscillating. While probing, I noticed a trim pot that had been adjusted. It had a dab of paint on it. You could see that someone had been twiddling it.

I looked it up on the print and it was listed as "stability control". I gave it a little tweak and the amplifier settled down.

It seemed like it was now fixed with no parts needed. Yay!

I decided to let it run for a few hours and then go through the calibration adjustments. Unfortunately, the repair was not done. It died after a few hours. Unit was running but had no waveform output.


Now for troubleshooting round two.


The output amp had failed. Large DC offset and only a spike at the rising edge of what should be a square wave.

It was easy to track down. Not because I have a lot of skill with analog circuits. I've mostly worked with digital. Wavetek has very good detailed service manuals. Boards are marked with component numbers. Each subsection on the board has a dotted line around it with a label. Amplifier, attenuator etc. Also a troubleshooting guide for each section. Waveforms, voltages to check. Not like modern equipment where the troubleshooting section just says "Send it back to us for repair"!

Using the troubleshooting guide I found that DC voltages in the amplifier were way off from normal. Checking in that area of the circuit I found that one transistor in a CA3096AE transistor array had failed open. This transistor was wired with the base and collector shorted. I assume it's just being used as a diode.

I took a transistor from my parts bin and tacked it across the pins on the IC. I could possibly have just used a diode. The output amp is now working.




Now I have a question for the experienced people on here.



Should I replace the chip? The datasheet for the CA3096AE lists it as a matched transistor array. The A version having the better tolerance. Will there be any major problems running it like this?

The problem is that the CA3096AE is out of production and nearly unobtainable. Otherwise, I'd just replace it. I've only found some China sellers offering it. I can buy one from China for about $4. Probably a fake part and possibly not even functional. It might be no better than just soldering a transistor accross the IC pins.

I'm thinking of just going through my parts bin and finding the best parameter match for the failed transistor. It's working with the first PNP that I grabbed. I can't think that It's really critical. Maybe stick it down on top of the chip with a dab of heat sink grease to keep it's temperature about the same as the chip.



Any suggestions? How critical is this transistor?



Finally, I've learned my lesson! No matter how eager you are to plug it in and power it up. Don't do it! Resist the temptation! Even if it was sold as functional. Take the cover off and inspect it carefully!!! This unit had parts missing, 120V wires dangling, cables unplugged and cables plugged in backwards.

I'm really lucky that it was not toast as soon as I powered it up. Dangling 120V wiring. If that fan wire had been touching a board this thing would have been junk. Also lucky with the reversed ribbon cable. Instead of shorting the power supply it could have killed a load of components on the front panel board.

And, for anyone thinking of buying one. If you don't have the ability to control it with GPIB you'll have to deal with trying to clean the really sucky buttons on the front panel. I've had bad luck with these switches. Clean all of the little domes and the PC board and six months later they start to die again.

I can't complain. The price was right. Free! I'm going to let it run for a few days. If nothing else dies I'll decide what to do with the bad transistor and buy a cooling fan for it. I guess not quite free. Probably about $20 for a fan.



Attached are some pictures.
 

[Kleinstein]

The transistor array seems to be there just for offset control. One can check how bad the DC Offset is. The broken Transistor is used in a current mirror with really large emitter resistors (with something like a 7 V drop), so I would not expect a big influence.
Lets do the math:  A 1 K temperature chance would result in 2 mV difference at the transistor and thus something like a 2/7000 relative change is current for the input transistor. This should be something like 2/7000 * 26 mV of change in offset. So one expects an additional TC of some 7 µV/K for the temperature difference of the external transistor.

If really needed one could use a double transistor like BCV62 to replace the two transistors used as a current mirror.

The local circuit is a kind of low drift OP - today I would chose something like an OP07  instead.

[lmester]

The transistor array seems to be there just for offset control. One can check how bad the DC Offset is.

Thank you.

I was able to adjust the DC offset to 0v. I did all of the amplifier related calibration.  It's working good with the substitute transistor.

I ordered a cooling fan for it today. Total cost for this instrument is $12 for a fan. You can't do much better than that!

Also, I noticed that I had no pictures of the front panel. Below are a few pics.

[Vgkid]

Thanks for documenting that repair. I have always liked those Waveteks.