Cap. leakage current Was:Re: PDP-5 Rescued!

From: Christian Fandt <>
Date: Sun Apr 30 12:25:46 2000

Upon the date 01:10 PM 4/25/00 -0700, Dwight Elvey said something like:
>Chuck McManis <> wrote:
>> This didn't seem to make it out the first time, so here goes a second.
>> I managed to rescue the PDP-5 that I've "owned" for some time now, but it
>> was stuck in a far away city. Anyhoo, I've put up the obligatory pictures
>> at <> for anyone who cares
>> to take a peek at it. This uses "DEC System Modules" the pre-cursor to the
>> DEC Flip Chip technology. It hasn't been powered on in 15 years so I'll be
>> a bit cautious getting the caps warmed up in the PSU. -15V transistor
>> logic, wild stuff!
>> --Chuck
>Hi Chuck
> If I was worried about the capacitors, I'd disconnect them and
>use a bench supply with a nice current limiting resistor. I
>would think a step rate of 5% of working voltage every 1/2
>hour would be safe. Put a volt meter across the resistor.

Use 15k to 30k at minimum 5 watts.

>Any significant large voltage drop would indicate leakage.

Using Ohm's Law, figure the current throught the resisitor. A rule of thumb
I found shows for capacitance value 6000 MFD and higher at 25 working volts
figure a limit of 8 milliamperes leakage current (0.008 amps) before
planning on replacing them. This is even after reforming them using your
step-increase plan depicted above.

>These large caps are known to have some leakage. How much
>is OK, I don't know. That is something you'll have to make
>a judgment call on. I would think that for large caps like
>I see in the pictures, 1/32 watt is on the large size for
>DC power.
> Don't forget the smaller electrolytics in the circuits
>of the supply. Do a value check on these. I usually lift one
>lead and use an ohm meter. Comparable values should take the
>same time to charge. You might want to connect these to

Rough test but sort of okay to get a ballpark idea of condition.

>a setup like used on the large ones. If they respond
>as having low capacitance on the ohm meter, you'll most likely
>have to replace them, forming won't help.

Uh, "as having low capacitance on the ohm meter" ?? Only thing you'll get a
real hint of is whether the cap is rather leaky or not. No close relation
to capacitance value measurement except that for great step increases in
capacitance (say, from 1000 to 10000 MFD for example) you'll get a very
rough idea of value just from time to charge. Then only if the cap is very
leaky would you see a difference between caps of same capacitance value.
And even this idea is not reliable because the variation in capacitance
value for same-marked caps can be up to +100% for the cheap grade and maybe
+20% or more for the "computer grade" caps.

I've got a writeup I quickly tossed together which I posted to a couple of
members of the Boatanchors List (boatanchor == affectionate term used for
any radio which is used for commercial or amateur radio reception or
transmission which is big, heavy and uses vacuum tubes (valves)). This
writeup discusses capacitor leakage testing and electrolytic reforming on
our beloved boatanchors. Similar application to the electrolytics and a few
other capacitors used on the old computers we all love on this list. I can
send a copy to others here who may be interested. Just email me with the
request in the Subject line.

> While you are fiddling with the supply, you might also add
>some new heat sink grease under power transistors. The silicon
>oil in this grease does eventually go away. It is easiest
>if the transistors are socketed.

Indeed, but the silicone grease is only the vehicle for the tin oxide (or
zinc oxide) filler which does most of the heat transfer. If left
undisturbed there should still be virtually all the orginal thermal
transfer capability as when the unit was built. But, I've not studied the
effect of the loss of the silicone grease too much. Tony, have you seen
much variance here in your experience?

> I don't recommend the slow power on for the entire unit.
>Many circuits will draw higher current until full up to
>voltage. If you want to do a slow power up and the supply
>is a linear, disconnect the load and just bring up the supply.
>If it is a switcher, you may damage it with a brown out.
>Most linears can handle it.
> Watch for any signs of smoke or popping sounds. These are
>almost always bad news. At one of the meetings at Stan's,
>one of the fellows, I don't remember who, had three tantalums
>blow. These were all power line input filters on boards.

Power line input filters? Tantalums?? On the AC mains??! No wonder they let
loose as any reverse voltage above a low tolerance level on any tantalum
will short them and they go boom! Not a pretty picture sometimes.

In this case I would say that they were not tantalums but something else
such as a Mylar film dielectric cap or some such. Since they blew up I
would guess they had gotten nailed with lightning or a significant power
line surge which punctured the dielectric and shorted them. Later, when the
unsuspecting fellow at Stan's applied power, perhaps for the first time
after the damaging incident, he got the surprise.

>After we removed them, the unit worked fine. If we'd left
>the power on, these would have burned the boards.

Regards, Chris
-- --
Christian Fandt, Electronic/Electrical Historian
Jamestown, NY USA
        Member of Antique Wireless Association
Received on Sun Apr 30 2000 - 12:25:46 BST

This archive was generated by hypermail 2.3.0 : Fri Oct 10 2014 - 23:32:43 BST