HP-35 power unit: Failing Vgg voltage

[robjordan]

I have an HP-35, it was my Dad's, check my bio for a funny story about how he came to buy it.

It's not functioning and I've determined that there is a problem with Vgg. Powered with 3.75V from a bench supply, I have measured: Vcc = 7.55V, Vss = 5.66V, Vgg = 0.5V. The first two are close to spec and seem clean, Vgg is clearly out of spec and seems a little noisy. The calculator draws a total of 130mA.

I have done some investigation with a scope. The transformer is oscillating at 249kHz. As an example, it is developing an 8V square wave at pin 1, which feeds Vcc.
   

Regarding the faulty Vgg signal, by lifting pins on the caps and diodes in the chain leading to the Vgg output, I have ascertained that the oscillation on pin 4 of the transformer is about 2V peak-to-peak. On the transformer pin it's +/- 1V and after the 2uF series capacitor it's 0 / -2V. (I don't really understand why - I thought capacitors remove DC offset?).
   
   

If I detach Vgg from any load, it develops about -2.8V, but as soon as I attach any load, even a 15K resistor to ground, it collapses to -0.5V.

Testing the diodes and capacitors in the chain that develops Vgg hasn't revealed anything obviously wrong, but I'm guessing one or more of those might be faulty?

Any thoughts?

Thanks!
Rob

[teenix]

Being your dad's it would be something special.

The transistors could be working if the other rails are ok, so I would be checking for failed capacitors and germanium diodes on the Vgg output. Cheap and easy to replace.

Maybe check the 60uF input capacitor as well.

Have a look for corrosion around the transformer.

Laporte's old web site has some interesting HP-35 info.

cheers

Tony

[robjordan]

Thanks for the quick response.

Regarding replacements - do you suggest replacing capacitors with the Kemet T110, which seem to still be available? And the diodes: should they be germanium? I think saw someone recommended Schottky as an alternative with similar forward voltage drop.

Thanks, yes, I've been looking at Laporte's page, also schematics here (Tony Duell's work I think) and Kees Van der Sanden's page.

[teenix]

(11-09-2022 10:58 PM)robjordan Wrote:  Thanks for the quick response.

Regarding replacements - do you suggest replacing capacitors with the Kemet T110, which seem to still be available? And the diodes: should they be germanium? I think saw someone recommended Schottky as an alternative with similar forward voltage drop.

T110 would be good (albeit expensive) if you want to replace the caps with original looking ones, or Kemet 322 a bit cheaper, or Vishay axial even cheaper.

Can't see why the Schottky's wouldn't work as long as the specs are similar. You might just see a slight change in the output voltages. Bat46 maybe?

Transistor equivalents for

1554-0550 - 2N3904
1853-0020 - 2N3702

Depending on the power supply design, I think the later versions changed from PNP/NPN to 2 NPN's like the HP-45.

cheers

Tony

[robjordan]

After some more investigation, I'm now of the opinion that the winding on the toroidal transformer between pins 3 and 4, the one which generates the -12V, is broken. It measures M-ohm range resistance in circuit. Annoyingly it's a version of the toroid which is encased in white gloop, like this picture. Maybe because it's a Singapore manufactured model?



Just wondering on the off chance. Is there anyone who is or has been willing/able to replace these transformers? I guess they are only available from donor calculators, but thought worth checking just in case.

I guess I could try and DIY one but I'm way out of my comfort zone. Or unpick the gloop and see if I can repair this one.

Thanks
Rob

[teenix]

It's amazing after 50 odd years, the board still looks new.

It is quite rare (that I know of) for a transformer like this to go open circuit, unless maybe it was subjected to shock, like being dropped. I would suspect a soldered terminal, as I think it would be unlikely to break somewhere in the middle of the coil.

It may be a long shot but re-soldering the transformer terminals from under the board might help.

If not, and you have a good magnifier, from the top side, you might be able to see a broken wire where it was soldered.

cheers

Tony

[robjordan]

(11-28-2022 10:52 PM)teenix Wrote:  It's amazing after 50 odd years, the board still looks new.

Thanks Tony. The photo isn't my calculator, it was just one I found online that illustrates a similar transformer. But I think all of these classics are beautiful, and such a contrast to modern miniaturised mass-produced circuit boards.

I'll try resoldering the joints and see how we go from there.

Cheers
Rob

[robjordan]

Sure enough there must have been a broken connection between one of the wires and the body of the transformer. After I'd desoldered it from the board and pulled it out, the break became obvious. Fixed that, and Vgg is produced, and the machine came to life! Probably the first time in 40 years that the LEDs had been illuminated. All voltages are a bit low, but that didn't seem to cause any functional problem (7.3V, 5.7V and -10.7V).

Unfortunately it's not the whole story because when I reassembled I find I have problems with the power switch, and some of the keys are erratic, but it's a huge breakthrough.

Incidentally, this calculator has no transparent foil over the keyboard like all the others I've seen pictured. Does the foil serve any purpose other than protective?

Thanks
Rob

[teenix]

Low voltages may be from some of the power supply capacitors starting to fail due to age but can be replaced easily if desired.

The plastic serves to stop dust and a small amount of liquid from getting on the key contacts and is wise to be fitted. Cutting up a zip lock lunch bag should serve ok.

Page 42 of my Classic Notes PDF file describes a cleaning method for the keys.

HP put some switch lubricant on the power switch during manufacture and this has probably hardened over the years and is interfering with proper contact now.

Just clean the contact and circuit board area with isopropyl alcohol. A lot of hobby shops sell silicon switch grease in small tubes, or more expensive ones at Digikey, Mouser etc. If you don't have any proper switch lube, a "tiny" amount of vaseline (petroleum jelly) smeared on the circuit board contacts will suffice.

cheers

Tony