Repair of Tektronix PG501 Pulse Generator

[xrunner]

Yes a TM500 series pulse generator - or as Rigol would say a "pluses" generator. 

It came with a TM501 chassis. I originally negotiated a price for just the PG501 because I didn't really want the chassis, but he sent both to me anyway. I don't know if the PG501 works yet, all I have done is plug the system in and turn it on, the power light on the PG501 did not come on. I'm about to finish up the FG501A so I'll get right to this in short order. I'm also going to post some scope screen shots from various circuit points at the end for anyone that may need data from a working unit for troubleshooting purposes.

[Rerouter]

I would not be too suprised if you found the LED had just failed, that was all that was wrong with 2 of my plugins, just wish tek had decided on a standard colour, have a mix of green, lime, red and orangy red

[chris_leyson]

Might be a bad cap in the TM501, I recently had to re-cap a TM503 as my SG503 signal generator wasn't working.

[bitseeker]

Looking forward to the repair. The last time I used plugins was in engineering and physics lab classes.

[xrunner]

Looking forward to the repair. The last time I used plugins was in engineering and physics lab classes.

Should be fun - these things always seem to have a few tricks up their sleeves. 

Just about ready to chill out for the evening, but I wanted to check out the TM501 that came with it. I did a quick check of all the voltages.

They all checked out except the +/- 33.5 supplies, well specifically the -33.5 - or so I thought. I measured the following:

+33.5V --> +36.9V

-33.5V --> -23.8V 

I thought the Neg. supply was bad, but according to the TM501 manual p1.1, the tolerance is -23.7 to -40.0. Hmmm, quite a wide range, but I think something is slowly going kaput since it's so close to the lower tolerance. I checked the schematic for the PG501 and sure enough it does use the +/-33.5V supplies.

Will proceed in the morning ... 

[bitseeker]

Yeah, when already in the fix-it mode, might as well bring things back toward nominal.

[Cyberdragon]

Looking forward to the repair. The last time I used plugins was in engineering and physics lab classes.

Should be fun - these things always seem to have a few tricks up their sleeves. 

Just about ready to chill out for the evening, but I wanted to check out the TM501 that came with it. I did a quick check of all the voltages.

They all checked out except the +/- 33.5 supplies, well specifically the -33.5 - or so I thought. I measured the following:

+33.5V --> +36.9V

-33.5V --> -23.8V 

I thought the Neg. supply was bad, but according to the TM501 manual p1.1, the tolerance is -23.7 to -40.0. Hmmm, quite a wide range, but I think something is slowly going kaput since it's so close to the lower tolerance. I checked the schematic for the PG501 and sure enough it does use the +/-33.5V supplies.

Will proceed in the morning ... 

Uh oh...this might be a sign of...

Are there any ceramic trimmer pots with black plastic knobs marked "Bournes" by any chance? (or any open track trimmers for that matter)

[xrunner]

Uh oh...this might be a sign of...

Are there any ceramic trimmer pots with black plastic knobs marked "Bournes" by any chance? (or any open track trimmers for that matter)

LOL well I don't think so ... but we'll see. 

Meaning I've bigger fish to fry at the moment (the PG501). I'm going to put that TM501 aside and just use my working spare TM504 to proceed with the checkout of the PG501.

I think the TM501 has a bad filter capacitor, but I prolly should quit predicting what is wrong with these things as I have a bad track record on predictions so far 

[calmtron]

I love repairing theese little TM500 modules - my PG501 has been healthy since i got it though. Looking forward to seeing more of this.

[xrunner]

I put the TM501 chassis away for now and tried the PG501 in my spare TM504 - the light is still not on but it does output a positive pulse - but not a negative pulse. However, the positive pulse output seems to work very well. I attached a pic of a 15 ns pulse. Although that missing negative output bugs me right now, I decided to go for the low-hanging fruit today.

So remember the phrase "The lights are on but nobody is home"? We have to reverse that "Somebody is home but the light is not on".

OK, well what have we here - a failed LED? NO! We have a burnt out lamp!  Yes, this has a factory original grain of wheat bulb with a green lens - not an LED. The original parts list specifies a 5V lamp rated at 60 mA. So we have that along with a 270 ohm r to limit the current since it's on the -18V bus. That surprises me because all the other units I have now use LEDs. Is anyone aware of any other TM500 series units with bulbs?

The bulb is hung off the -18VDC supply through a 270 ohm current limiting resistor, so I decided to go ahead and convert it to an LED (although I could have kept it original since I have some of these bulbs). I decided to use a clear white LED and keep the original green lens. I just removed it and glued it on the white LED. It has a nice green light with that lens on it.

Now the light is on and somebody is home. More to come ...


For new players reading that need to know how to do an LED retrofit:

The white LED has a ~3.2V drop at 20 mA. To calculate the required current limiting resistor -

18V - 3.2V = 14.8V (voltage drop of the current limiting resistor)
14.8V / .02 = 740 ohms (required current limiting resistor for 20 mA)

We already have a 270 ohm resistor on the circuit board that the lamp used, so in theory we could just add another resistor of about 740 - 270 = 470 ohms. However, I felt that the LED would be too bright @ 20 mA, so I added an extra resistor of 750 ohms. Total resistance is then 270 + 750 = 1020 ohms which allows a current of 14.5 mA. Also, in this case remember to connect the cathode of the LED to -18V and the anode to ground.

[guido]

I fixed mine recently. Same problem, no neg output. Two dead transistors. Easy to find with a scope. Just compare the positive and negative parts at the same points. At some stage the signals on the negative side vanished.

[tautech]

Too bright for my liking or so it looks in the pic.
I'd be increasing the resistor value to get it down to just a few mA's and it will still be bright enough.

[xrunner]

Too bright for my liking or so it looks in the pic.
I'd be increasing the resistor value to get it down to just a few mA's and it will still be bright enough.

The pic is deceiving - it's not nearly that bright. The camera is sensitive to that color I suppose.

I fixed mine recently. Same problem, no neg output. Two dead transistors. Easy to find with a scope. Just compare the positive and negative parts at the same points. At some stage the signals on the negative side vanished.

Cool, I should be so lucky. We'll see very soon.

[xrunner]

Wow! Wait till you see this. 

So when we last met you may recall that the PG501 had a nice positive pulse but no negative pulse. Well I was checking test points and never saw anything wrong along the way. Lo and Behold look at what I found. The three resistors that are supposed to be soldered to the output BNC are just kindof "up there"! Didn't even notice it till now.

After looking at this very closely with a magnifier, I believe this is an original factory oversight - not an attempted repair. You can see years of dust on the wires. The resistors are the exact same type as used on the rest of the board. They have been carefully bent into shape ready to be soldered but the ends have never had solder on them. The ends that are soldered to the PCB are pristine and IMHO have never been re-done. The ground wire has never had solder on the end but the other end on the PCB is pristine and looks factory original. The output BNC is the exact same type as the rest of the BNCs, with the same age patina on the metal and the solder cup has never had solder in it.

I think it was Friday at the Tek factory and somebody wanted to get this boxed up and get to happy hour! What amazes me is that it could have got out of the factory like that. And even if it did, did no one during it's life ever notice it? Perhaps nobody ever tested or used the negative output. I don't see any other explanation.

But ... it has old calibration sticker seals - how did it get calibrated at all - you cannot go through the cal without checking the negative pulse!

[tautech]

 
That's a real doozy.

One of my mottos with unfamiliar equipment is "expect anything and everything", but something like this stretches that to to limit. 

Time spent studying the layout is never wasted as you learn to spot rework, bodges, dry joints, cracked (yes cracked) components, cracked joints and all manner of little things you'd never expect.
Even air bridge joints. 

[bitseeker]

Nice air bridges!

Yeah, a Friday afternoon job was just what I had in mind, as well. Really weird, though, that it had received cal stickers, as you said. Definitely makes for a good fix-it story to be told and retold.

[xrunner]

Well check this out - There is a sortof black scrawl next to the Neg. BNC. It didn't look like more than the typical marks on 30 year old equipment, but now I'm not so sure. Did that mean somebody made a mark there because they noticed a problem? But why wasn't it repaired? Just simply solder the resistors ...   

Quite a little mystery.

[xrunner]

Troubleshooting waveforms and points.

Front dial settings: Period 20 us, pulse 10 us, variable dials X1.

[xrunner]

Troubleshooting waveforms and points - output amplifier.

Front dial settings: Period 20 us, pulse 10 us, variable dials X1.

Note: Channel 4 vertical should be divided by 10, due to incorrect 10X setting on scope.

[xrunner]

Forgot to add one more screen shot. Rigol DG1032Z (top trace) vs. PG501 (bottom) - 20 ns pulse 5V into 50 ohm load.

Not bad. 

[bitseeker]

That's quite a difference.

[xrunner]

That's quite a difference.

Yep, and here's another one - Wavetek is chan. 1. The Wavetek minimum setting is 45 ns. It's not a bad little pulse generator.