HP 3325B Repair (Shorted 5V Supply)

[RaymondMack]

I picked up an auction lot of test gear from a university surplus sale a few months ago and never got around to testing anything till recently. (Apparently my ability to procrastinate knows no limits.)

In the mix were two function generators that were marked as faulty: an HP 3325B and a Wavetek 90. On a whim, I decided to try my hand at the HP first. (It was the lowest hanging fruit, seeing as it was at the top of the pile.)

I find performing a visual inspection before powering on to be the best policy when working with used gear that is known to have problems. After pulling the top and bottom panels off, the unit showed no obvious signs of damage or past repairs, which was a nice surprise. At this point I saw no reason I shouldn't try a power on test and noticed the fan turned on but the display remained dark. Thankfully no smoke or explosions were encountered during the test. However, I did notice the case near the pass transistors got pretty damn hot. At this point I decided to check the power supply.

To do this, you connect to power ground at pin 5 on J700 and then check either the test points TP200 +5V, TP300 -15V and TP400 +15V or J700 pins 1 (+15V) 2 (-15V) and 6 (+5V). The +-15V supplies looked okay (if a touch high) at +-15.1V but the +5V supply was pretty low at 3.4V. I figured the +5V supply was shorted somewhere, but to test that I needed to disconnect J700.

If you unplug J700, the function generator will not turn on since the power switch is connected through the front panel to A26 and then to A22 through J700. To fix that, you just connect J700 pin 10 (standby) to pin 5 (gnd). But since each supply has sense connections, you should also connect pin 1 to 2, connect pin 3 to 4, and connect either pin 6 or 7 to pin 8. After doing that. I saw that the +-15V supplies still read the same, but the +5V supply went up about 400mV. Okay. Looks like the fault is somewhere on A22. Better in the PSU than one of the other boards.

After examining A22 more closely I couldn't find anything that looked damaged--no discoloration, delamination or leaking/bulging caps. But the side panel got roasty toasty hot. At this point I decided to look at Q200, a TIP36A PNP power transistor which is used to regulate the +5V supply.

Conveniently, HP remotely mounted the pass transistors and used pins on the PCB with crimped wires to easily connect/disconnect them. After digging through my parts box I managed to find a TIP36C which has the same general specs but a higher breakdown voltage. Using a bunch of Pomona mini grabbers, I hooked it up and fired up the function generator to find the +5V supply reading 5.1V. Success! Looks like the pass transistor died.

Hold on there, cowboy! After pulling it off and comparing it with the replacement, I noticed that it actually looked good! Removing the junk box part and plugging the original transistor back in, with it still disconnected from the chassis, showed that indeed it was okay. The plastic insulator had turned brown from overheating, but the transistor itself hadn't given up the ghost yet. How curious. Looking more closely at the transistor showed that the insulator didn't actually reach the other side of the metal mounting tab. It looked like it was made for a TO-220 tab; unfortunately, a TO-218 tab is about twice as thick. So the mounting screw must have shorted the tab to chassis.

The rear panel around the BNC ports where an adapter is used to feed the option 001 reference back into the unit looks to be bent inward like the unit was dropped at some point. So maybe it got jarred hard enough to cause the mounting screw to nudge close enough to touch the mounting tab of Q200? Shit happens I guess.

Anyways, digging through my parts bin managed to drudge up a longer insulator. I also decided to swap out the original pass transistor since it had gotten pretty damn hot and looked to be a bit leakier than the one I found in the junk bin. Better to replace it and not worry about it dying later down the road--though that might be my paranoia talking.

Sure enough, when I swap in the new Q200 and mount it with the longer insulator the +5V supply is back in action. I then adjusted R352 so the supply voltages were within spec, as the +-15V were a little high previously.

Now was the moment of truth. I disconnected the remaining mini grabbers, plugged J700 back in and powered up the unit. Lo and behold, the display came on! Nice! Even better, I didn't see any error codes. Looks like it was just an improperly sized insulator.

Overall, this was a pretty uneventful, but still satisfying, repair. Another BER scrap destined for the recyclers or landfill was turned back into useable gear.

[RaymondMack]

Here are some more photos.

[RaymondMack]

Here's the PSU without all the jumpers everywhere.

[lowimpedance]

Nice kit , and save !. Its always satisfying to keep nice gear like this out of landfill and on the bench where it belongs.

[bd139]

Nice post and debugging session. Thanks for taking the time to write it up 

[Runco990]

That is a damn fine synthesizer.  I have 3 of them!

Were you lucky enough to have the ovenized time base and high voltage option?

Although it's easy to add later.   

[RaymondMack]

That is a damn fine synthesizer.  I have 3 of them!

Were you lucky enough to have the ovenized time base and high voltage option?

Although it's easy to add later.   

Thankfully it came with option 001 (high stability time base), which is nice. While going through the service manual I saw a note that phase noise is only specified if 001 is installed, so that's a big plus to having it. I do wish it had option 002 (HV output) but maybe a parts unit will pop up on eBay that isn't over priced and I can add it to this one. One can hope anyway. But with the covid-19 economy as it is, things are priced higher than I have ever seen. It's pretty odd how things entirely unrelated go up in value.

As a testiment to the 3325B's quality, I noticed that a lot of calibration manuals reference using the 3325B as a signal source. I just need to find a thermal converter to calibrate it and I can finally tune my Racal Dana 5002 wideband voltmeter. The 5002 also needs some modding to the peak detection circuit since it over shoots and gives wrong readings for fast rise time signals. This is a systemic flaw on all units from the factory. Really throws off the crest factor and form factor readings. To prove this, simply apply a symmetric saw tooth wave and then adjust symmetry until it is a ramp with sharp rise time. Mathematically, both the crest factor and form factor are identical for each waveform, but the 5002 will give wildly different readings from overshoot of the peak detector. Maybe I should open up a new thread to demonstrate this and find a cure... Some extra series resistance may be needed to dampen the output of the opamps or slowing down the inputs. Probably need to experiment and see what works best.

[Runco990]

I have to check my stash....  I MIGHT have a spare HV option board still.

I'll look.  I had to part out a few absolutely hammered units a few years ago.

MIGHT....  so don't get yer hopes up.  Front panels I have plenty... for the buttons and switches.

[RaymondMack]

Please do. If you've still got one and a power cable to go with it, I'd love to buy it off you. But if not, I'll just play the waiting game...

[Runco990]

Well, I know for certain I don't have the power cable.....  The 3325A has it nicely curled up in the machine, the 3325"B" does not.
The HV boards are the same, But I had to make a power cable for my "B" unit.  I have 2 "A" and one "B" unit.

Don't forget to check those troublesome flat ribbon cables....  they never did completely fix those problems...  they usually need to be re-seated from time to time.

[RaymondMack]

Now that you mention it, I wouldn't be opposed to making the power cable myself.

Would you happen to know what type of connector and crimp terminals are used? Or baring that, what would be compatible?

If you can find a spare option 002 board let me know.

[mawyatt]

Nice save!!

Loved those HP3325s, we had a bunch of them way back, they were our go to function generator until the Stanford Research devices showed up. We had a unwritten lab policy not to consider anything that wasn't HP, Tektronix and maybe R&S, for various reasons but the SR rep dropped off a new function generator and left it with us for a few months. After awhile decided to take a look at the SR unit, the performance was outstanding significantly better than the HP3352 I recall, then took it apart (we took everything apart back then). The construction was 1st class, then studied the schematics, same result. The passive multi-pole filters were some of the best designs we had seen, so this displaced the beloved HP3325 as our new go to function generator.

Keep those old instruments alive, we've saved a couple Tek 2465s and a couple HP34401As

Best

[Runco990]

Well, I did check... sadly I do NOT have another HV board left.   I am sorry... I really thought I had one still.   

[RaymondMack]

Thanks for checking. I knew it would be a long shot. I'll just have to wait it out and see if something in my price range pops up.

[trobbins]

TERRA Operative has plans to make the 002 HV board - he has a list going, but likely still delayed by parts.

[TERRA Operative]

Yeah, I can supply gerbers and BOM etc for the Option 002 HV Output, but shipping from Japan for kits is still all Covid.
Here's the link to all the files if anyone wants to organise a local group buy etc. My version is at least as good as the HP design, but modern parts may make it a little better, maaayybeee.

https://drive.google.com/drive/folders/1l8sT5sR7wLVCqZOov9T9yXTmKLvo_4Qv?usp=sharing

[RaymondMack]

Thanks TERRA Operative! You've done an amazing job with the option 002 board. If I can't find a stock board for a good deal, I'll definitely give your project a try.

[bd139]

Oh now I want a 3325 so I have a project to do