Nice image of front of unit

Full manual on archive.org

The Unit

I acquired this unit non-functioning since someone was going to throw it out. I don’t consider myself a particularly good repair guy but you gotta learn somehow right?

The output of the unit had a considerable DC bias coupled with large 50Hz spikes. Sounded like a power supply thing. Somewhat intermittent initially, but after the first round of messing around, never worked again (until I fixed it of course).

The Repair

Thankfully the HP manuals from this era are particularly helpful in identifying which bit of the circuit is at fault. The output had a heavy DC offset and was jammed outputting some nice 100Hz spikes irrelevant of the front controls. The supply rails measured all weird like. More specifically, the +10, -10, +25 & -25 V rails measured +10.04, -15.74, +36.68 & -25.52 V respectively. Yowzers!

Since the schematic shows that the +25V rail is derived somewhat from the -10V rail, lifting one leg of R75 allows supply of that -10V reference for the +25V rail externally. Powering up the unit with the external supply confirms correct operation of the +25V rail and thereby isolates the -10V rail as the issue.

Once confirmed the rail was not current limiting all the electrolytics were checked. This was not the logical next step, but it’s what I did nonetheless.

One year later I return to this repair…

There weren’t any particular components that would be prone to failure in the power supply so the most obvious component to check was the pass transistor (Thanks Mr Carlson). It was removed and checked absolutely fine.. Drat…

Next up the board was generally checked for connections between boards, no amount of bending and wobbling boards could get it working again. A considerable number of joints were reflowed at this point since I had at least once experienced the unit working shortly after being powered externally. Crazy times.

Next thing checked were the three pnp transistors in the -10V rail that could the culprits. Their internal diodes were able to be tested in-situe. Thankfully, Q26 was found to be open circuit between all its pins, and of course everyone has a drawer of 2N3906 diodes on hand for exactly this scenario so upon replacing the transistor, it worked!

The only other necessary adjustment I made to the unit was to trim the +10V rail to be as close to +10V as I could get. Tidy.

Re-Capping

I can safely ignore board A5 since that was fitted only to the 3310B models. My board A3 which handles the power supply and (something else) is marked Rev B. Thankfully there

Titanium Oxide Capacitors?

Not something i’ve heard of before. Thought it sounded a little similar to Aluminium oxide electrolytics so I had originally planned to replace them. But given the very tiny values used (low nF, pF range) I had a dig.

Nothing really came up bar this site which appears to say they’re used in MLCCs so it’s fair to assume they’ve got alot more life left in them, and don’t need replacing. Good to know.