TSP #108 - Teardown & Repair of an Agilent E3632A DC Power Supply

In this episode Shahriar & Rosanah investigate an Agilent power supply which does not appear to power on. It can be quickly observed that the fuse has failed on the unit. Using an isolation transformer a small amount of AC voltage is applied to the unit after the fuse replacement. It is clear that a short is present somewhere in the instrument since even at 10V AC the instrument consumes more than 1A.

After separating the internal transformer from the main PCB, the high current consumption is eliminated. The short is traced to a damaged main rectifier in the instrument. A replacement rectifier is used which allows the instrument to power on.

The VFD display is however extremely dim. After a failed attempt at restoring the VFD brightness, a new display is purchased from eBay. The replacement produces a bright vibrant VFD display. The performance of the power supply is verified under load.

The Signal Path

[waew]

Your attempt to restore the VFD reminded me of the old CRT rejuvenator that we used to use in the TV repair shop that I worked at in the late 1970s.

[dummyvariable]

Your new microphone does wonders to your voice.

[TheRadioShop]

Loved that introduction, it was Great job on the repair. I have tried that over current on those types of displays before. Never worked for me.

[michaelhawthorne]

Nice result which was inspected and approved by your cat at the end. Looking forward to your mentioned up coming instruments....

[PlasmaHH]

Generally in my experience: overall dim means oxidization, and difference between characters means aged phosphor.

Note that "burning off" oxidization only works well when there is enough unoxidized getter metal inside the tube.

[kdeat]

Yay Pooch! I always have to rewind and show my wife when Pooch makes a cameo!

[danielhatch]

I tried that same method for the VFD in my 34401A DMM with pretty much the same result (didn't manage to burn up the filament, though!). Your side-by-side comparison with the new VFD makes it quite apparent that the issue is with the phosphor and not the filament. Funny how so many HP/Agilent products from this era have this issue, yet all the VFD clocks in appliances around my house still look reasonably bright despite being on 24/7 for the last 10-20+ years.

[MaxKoschuh]

excellent job. brilliant.

2:45 I was sure it's the rectifier. But I also expected a broken transistor.
21:25 too bad, I wished we could see if this procedure made any difference at all

[Minifig]

Glad it was a nice and easy repair for a change! Not sure if it effects this generation of HP gear as much, but I'd replace those yellow Rifa brand class X and Y mains caps any time I see them. Had two HP instruments get very sticky and smelly on the inside from where they exploded!

[DrFrank-xjbc]

Nice. Especially the trial for regenerating the display.You can also do analogue, obviously!

[hightechstuff]

I've also experimented with rejuvenating VFD's and have found that getting the filaments to just barely glow for about 10 minutes seems to work the best. There is a point where you you can go to far, even with a faint glow. Also, just letting them operate with the most characters illuminated for a couple days sometimes brings them back. I haven't tried it but I think baking them in an oven around 100c for 24 hours would give favorable results. Great video!

[AFLJSue]

Good Video glad to see another piece of gear saved from the landfill. :)

[marcocosentino]

21:51 AFAIK you're NOT supposed to get the filaments white hot but mild red. And given how hard it is to find replacements I think it's a good idea to test first if a lower current did the job.

[WillArtie]

Awesomeness! Love a good dc supply repair, especially when you get to rerplace the rectifier AND the display :)

[Dyaxxis]

22:31 Drawbacks of purchasing old equipment. Even if the part was already completely dead, the obscurity of it really hits the wallet.

[catt]

It's always nice to watch your videos. Thanks! Well done on the repair

[erikderuiter]

In another video is was told that you have to *slowly* turn down the filament voltage after cranking it up, to prevent the filament wires snapping.

[JackZimmermann]

Love your repair videos. More suspense than a thriller! But you really need a brighter red nail polish on your right hand! ;)

[_PovertyLabs_]

Amazing I never knew you preferred that colour nail varnish.

[stevec]

In some cases you can increase the brightness just by increasing the negative voltage going to the display. HP tends to run them on -5 volts so you could increase that quite a bit and not have to bother with the filaments at all.