HP 721A - Lab Power Supply
Good evening and welcome to Tinkering with Atkelar! Since I'm working in high gear on getting my workshop and storage rooms up to speed, I can't tackle any larger projects for now. But I think the smaller ones should also be quite interesting if only for the issues during the restoration process. What I have for this episode is a vintage HP lab power supply. It is
almost the granddaddy of all DIY power supplies. The classic 0-30V range with a bit of current limiting. So absolutely nothing to write home about by today's standard. But I just couldn't pass up this cute little device, barely bigger than an average potato. Let's dig in and
see how the build quality is! The case was designed for both horizontal and vertical use. The feet on the bottom and on the rear are a dead giveaway, even though several of them are broken off and missing. I think the style is a holdover from the tube era. The front panel and chassis are connected and put into an outer shell. Unscrewing the rear holding screws will release that. Taking everything apart so I can start cleaning the case and order replacement capacitors.
The ones on the PCB have developed some sort of a sticky rash and their rubber bungs appear to be bulging. The device arrived from the US, as is evident by the US style plug. And since the cable is turning into sticky goo, I'll just replace the whole combination.
I am going for a complete cleanup here, so the parts need to come off the chassis and panel. Getting rid of the sticky cable. There are two strain reliefs on it on the inside.
Note that the output voltage is completely isolated from the case. The plus minus sockets can be raised by about 400 volts relative to ground for creating relative supplies. The chassis ground is connected to mains earth and available on the third binding post. Straightening out the case. It clearly had a tumble at some point in its life. Getting rid of the old - ahem - "rubber" feet was a mixed bag. Pretty easy on the rear ones.
Just drill out the rivets. But the bottom ones... I had to get a bit creative. And finally, I found a reasonably close can of paint for the shell. I tried cleaning, but with that structured paint I ended up without the structure or without paint before the dirt disappeared. So covering it up. The front panel also had a few dings in it. Time for some percussive adjustment! Please note, anvils are usually used on a steady surface, I know. But
holding this one in hand while whacking on stuff keeps the noise down at night and the neighbors happy. I am trying to desolder the minimum number of connections. The PCB is clearly a very early one. They
added rivets to the holes for soldering. This is still quite common for prototyping multi-layer PCBs, but I've never seen it done on a single layer one. All the wires are connected to the PCB with a strain relief. There's a hole, without a rivet, next to the wire's target and the wire gets threaded through and looped around. Quite odd. Did HP expect lots of tugging on these? And I decided to put the LCR meter to good use and check just how far off that tin can cap is.
And much to my surprise, it was closer to its nominal value than my modern replacement! Whoops! I guess this one can go back in after all. The PCB mounted ones were slightly out of spec and had a higher ESR than the replacement ones. New ones for those it is then! The meter is also a bit on the rough side, but after removing the front cover, the inside is actually very clean. So I leave that assembled and just clean and polish the cover.
And back together again, making sure to align the zero adjustment peg. The knobs and potentiometer are gummed up and feel not nice at all; adding fresh lubrication and cleaning the contacts. To keep the discs aligned, I usually flip the switch all the way to the leftmost position before taking it apart. Removing the old grease.
Adding new grease. I did make one slight change when putting the switches together. The little board that carries the shunts originally was loose on the end of the set of screws. I decided to tighten it down as part of the assembly. I might revert that change, since I thought somebody was careless and didn't add a set of nuts, but I later discovered the manual also shows it as floating, so it is by design. I think this might
have been done for thermal expansion reasons, but I doubt that the phenolic paper will expand so much that it would bend. The grub screws in the potentiometer knob left a dent in the axle. I need to smooth that out so I don't mar up the bearing surface. The power switch? Hmm... riveted. I was hoping to leave it,
but it did show several ohms and contact spray did not change much. Yet another style of one of these switches. I expected one of these roller style contacts, but this one has a sliding sheet metal frame. A bit fiddly, but no real problem here. I got fitting rivets. Starting the assembly with the meter movement. The nuts
are a bit hard to access if the other parts are already in place. I am tightening up the nuts for the meter by hand only. The case shows a few cracks already, and going overboard with a wrench might turn that into dust. Neon power indicator and power switch. Double checking if the switch goes on where it says it should. The binding posts had some... "stuff"... in the
threads. Yuck! Maybe old grease. But since these are captive type and don't unscrew all the way, I had to flush them out with IPA again and again until eventually I got it clean enough. I have seen devices with all black binding posts and I have seen colorful options. This is the first one that I see that has multiple colors and still uses the same for the key outputs. Both
plus and minus are red. The chassis ground is black. I wonder... was this common back in the day? The output sockets have the usual capacitor to give some last chance filtering and somehow a diode snuck in in parallel on my unit.
This is not mentioned in the schematic at all! Fitting the potentiometer and the selection switches. Connecting the meter. The eyelets were also in need of a good clean. The transformer is a dual voltage model.
It currently is wired for 110 volts. I'll remove the bridges and put the windings in series for 220 volts. The manual shows how to connect them but it seems that the unit I have uses a slightly different transformer setup. Nothing adds up. The transformer pin numbers are wrong and the mentioned colors of the wires are nowhere to be found. I had to measure the primary side and try out for the correct polarity.
Now since I had the power transistor out, I thought it might be a good chance to check it. If it's broken, it's easier to replace now. Whoops! A case of identity crisis? Transistor identified as resistor? That's not good. It also shows a 0.07 volt diode drop. That doesn't appear healthy!
But a quick test on the transistor tester shows it still does something. Disregarding those huge loops, it at least functions like a transistor. So let's go with it and see! Fitting the power transistor. With a dab of thermal paste. That black pad is anodized aluminum as an insulator.
The chassis is completely floating after all! And here comes the new power cord! Soldering the front panel components. I want to try out the primary of the transformer first before connecting anything else to it. But the front panel has to be in place for the power switch and indicator. Those wires are tight. And it's easier to solder in the in-between things when the front panel is still off.
New fuse. Half the rating of the 110 volt one again. Powering up with my isolation transformer. Neon works! The secondary voltages also check out.
Now here I ran into a little snag. I checked the isolation between the mains connectors and the case and also between the output and the case. Just to make sure that no solder bridge is causing trouble. The primary side was
fine, but the output shows 4k in one direction and a capacitor charging in the other. Whoops! That is not supposed to happen. I triple checked visually if any of the wire contacts got bent, but no. To narrow down the issue, I removed one component from the front panel at a time and checked again. To find out that the potentiometer is the culprit! Somehow one of the pins is shorted to the case. Let's have a look! Oddly enough, everything seems in order.
Until I took a closer look at the wiper! As a cat, I'm partial to whiskers of course, but tin whiskers? On the end stop? Yikes! I did clean up the axle and the wiper, but even with all that handling, those little buggers didn't register. They must have bridged out the wiper to the axle, and thus to the case. The fiberglass eraser makes short work of them, followed by a wipe with IPA.
And closing the pot for the second time. Now there's no detectable resistance between the binding posts and the case anymore. Good! A quick PSA. If you have such a unit, be sure to check that potentiometer. It is likely that this is a common ailment of these, and since the pot is directly connected to the positive output, any external offset input up to 400 volts are in spec according to the manual also carries over to the case. That might give you a buzz when you least expect it.
Adding the PCB. HP was kind enough to label the wire colors on the silk screen. Some wires need a bit of a trim to fit, but there is enough slack.
And after wiring up the two transformer outputs, the main regulated circuit and the control circuit, let's give it a try! Since components were swapped, it is time to calibrate the unit again. The service manual is very helpful here. We calibrate the meter reading with the meter trimmer on the frame, set the maximum no-load voltage output to 31 volts, and adjust the current limiter on the highest range. A new grommet for the new and slightly smaller cable. New feet! I can't set rivets in that case, but putting countersunk screws in from the inside makes them flush enough. And closing the case again. Hmm, that
grommet is a bit too big. I need to nip a bit off from it. And a new European style plug. And that's it! Another random piece of old-school lab equipment brushed up. For daily use, I'll keep with my Philips one, but if I need an additional voltage again, this will do nicely. I hope you enjoyed this little episode! See you next time! The no-ups and potentiometer are gummed up and feel not nice at all.
Adding fresh lubli- lubli- lublication, heh. Adding flesh lubli- cation, mmm, it gets worse!
2025-05-04 21:27
