HP-81 Capacitors

[Duane Hess]

Hi:

Spin-off from my HP46/81 power supply case posted earlier.

In a previous thread, about HP-46 printer parts/repair, Kattie Wasserman pointed out a Feb, 26 2001 post from Alex Knight:
http://www.hpmuseum.org/cgi-sys/cgiwrap/...i?read=138

Where a disfunctional 3000uF electrolytic capacitor can cause unpredictable and spastic behavior of the printer. The capacitor in my unit is labelled as:
Sprague
0180-2565
3000uF 30VDC
can neg
68D10407 73302

Evidently this is a Sprague brand 68D series capacitor with four solder points. Three share a common connection, apparently ground (Tony Duell's HP-46/81 schematics) and one being the positive circuit. Can't tell without removing the capacitor, but the three ground tabs appear directly attached to the capacitor body. Makes sense with "can neg". Google-ing about Sprague 68D series capacitors has failed so far. Anyone familiar with this cap? Are the three tabs indeed a common ground connection as opposed to a three segment capacitor? My assumption is a single cap with three ground tabs simply for physical strenth/stability, due to size and mounting geometry.

There are other issues with the unit. The printer is always left off as it misbehaves anyway. Repeatedly powering on & off produces correct startup after a small number of tries. Incorrect startup behavior produces:
- lockup (seldom)
- error condition; after a Clear Entry:
= all/most functions work correctly
= registers 0-19 work correctly or are disfunctional in banks of 10
which bank seems unpredictable
When it starts up correctly, the 81 works 100% (minus printer).

Also, the LED/numeric display is from the 46 and not the LED/numeric buffered keyboard display card from the 81 components. Using the 81 buffered card appears to cause either bank of 10 memories to misbehave. This might be incorrect; as the buffered card appeared to cause a problem, I used the 46 display after a couple tries.

The above produces an ambiguity to me. The 3000uF capacitor seems to involve driving the printer. As a common ground connection on the PCB to the negative end of the 2500uF and positive end of the 5000uF capacitors exist, can excessive ripple from the 3000uF capacitor cause errors in those other circuits? I believe the point of all three capacitors is producing a smoothed DC voltage for the respective circuits they serve.

The 81 guts come from a former HP employee who purchased the parts at an internal year-end excess inventory sale. The parts comprise everything but the top/bottom case and power supply. The case/power supply are from a working HP-46 with correct voltage values. The 81 parts are a field service kit minus the substituted parts. All 81 parts were in original field service boxes and in factory sealed component bags. My conclusion is all parts have never been powered on. i.e. the capacitors are well aged with no use; which evidently is a definitive electrolytic capacitor deterioration issue.

Any comments? Can the 3000uF cap affect other circuits? i.e. the intermittent power-up behavior with the printer off. I forget whether the main PCB circuits are powered from the +6VDC or -12VDC circuits or both. But the main question is replace the 3000uF cap or all three of them? And is the 3000uF a single capacitor? A direct replacement is not available to my knowledge.

There are a number of small tantalum(?) capacitors on the PCB, which I believe are electrolytic also. Should one consider replacing those? Buy/lease an oscilloscope (used one about 37 years ago) and replace one at a time? The latter seems most reasonable starting with the 3 large capacitors.

Other than fiddling in previous eras with electronics, which was very minor, my experience is minor. Anyone have suggestions?

[jebem]

Hi, Duane,

Yes, electrolytic capacitors will degrade with passing time, even when not used for long time. They have several failure modes.

A good capacitor meter would be the ideal tool to measure the capacitance and internal impedance (to do this, I would remove the capacitor from the circuit, or at least disconnect the positive lead in order to have the capacitor in a open circuit).
Or else one just assume the capacitor is defective and replace it.

Despite not knowing that HP equipment, I know about these old capacitors and from the description in the Alex Knight post, I have no doubt that this capacitor is a single section with 3000uF / 30VDC, where the Negative side is the can itself (as usual on these type of capacitors).

The reason to have 3 common tags is to mount the capacitor by twisting them in the chassis or in the PCB, providing a good mechanical fixing, and at the same time providing a good ground connection.

As Alex Knight mentioned, any modern electrolytic capacitor will do; the mentioned 3300uF / 35VDC is a standard these days and will fit.
The capacitance value itself is not that important, as electrolytic capacitors usually exhibit a tolerance in the range of -10% to +50%.

That said, if the capacitor is filtering an switching power supply (where the ripple frequency is on the order of several KHz), then I would use a 105 degree (or a 85 degree as a minimum) rated capacitor in this application, as typically these capacitors exhibit a lower internal impedance and therefore a better filtering and current supply.

On the other hand, if the capacitor is just part of a filter for 50/60&100/120Hz ripple filtering, then a regular capacitor will do, but even here I would use the better graded capacitors (85 degrees minimum).

Mechanical speaking, a Radial type format (like the original one) would be adequate, despite having only two leads.
The negative lead will go to one of the 3 negative points and the other two negative points must be interconnect as well.

If you would like to maintain the original look, then if you are fearless, you could carefully remove the capacitor internal components and save just the can.
Then install the new capacitor inside the original can.
This operation is delicate as the remaining electrolytes inside the old capacitor are contaminants and require special handling.

Tantalum capacitors are not the ideal solution here, and usually they are manufactured for lower values.