• Please review our updated Terms and Rules here

Leaking Caps in H7821 PSU

DECtechGuy

Experienced Member
Joined
Jun 22, 2017
Messages
168
Location
Florida
So one of the capacitors in my VAXserver 3100 decided to start leaking. I had it on and picked up that fishy smell. Poked around and found some residue around the base of a capacitor, great...

So these caps are on the far side away from mains input and more on the side of the outputs to the mainboard.

I have no idea what the ESR of these caps need to be! I know they're 25V 1800μF caps, but beyond that I'm lost.
 
Using the Mouser website, put in 1800uf capacitor and pick the closest look alike, at least 25V and anything 1800 uf or more. Back in those days, ESR wasn't a rating of importance to power supply filter capacitors I don't believe.
 
Thanks for the explanation.

Got the caps a few days ago and got them put into today, working like a charm now. All of the 1800uf caps were leaking.

I can also mention that the 1989 VAXstation 3100 maintenance manual that floats around does not have the same pinouts for my PSUs. Evidently, at some point, DEC swapped the header connectors, because my 90s era 3100s are using something different. The pinouts for my 3 PSUs I found in the May 1991 maintenance manual for the VAXstation 3100 Model 76.
 
The "impedance" of an 1800 uf capacitor is what the math formula says it is at a given frequency, more than likely 120 Hz. in a power supply: https://www.allaboutcircuits.com/tools/capacitor-impedance-calculator/ or about .74 ohms, essentially a dead short to the power supply AC frequency.

That is true for the capacitive reactance of 1800uF at 120Hz, but it does not account for any series resistance of the capacitor. For a large electrolytic of this type, it can be in the range of 1 to 10ohms, depending on the ripple current rating of the device. Since this is a series resistance, it is frequency independent and must be added to the 0.74ohm frequency dependent impedance. So the real ESR of the device at 120Hz will likely be in the 2 to 5 ohm range.
 
Thanks for the explanation.

Got the caps a few days ago and got them put into today, working like a charm now. All of the 1800uf caps were leaking.

I can also mention that the 1989 VAXstation 3100 maintenance manual that floats around does not have the same pinouts for my PSUs. Evidently, at some point, DEC swapped the header connectors, because my 90s era 3100s are using something different. The pinouts for my 3 PSUs I found in the May 1991 maintenance manual for the VAXstation 3100 Model 76.
I know its been some time, but does any one have the PSU pin outs? My Vaxstation3100/30 PSU has died, but still dead after replacing caps, and I would like to test outside the MV...
 
Generally an 1800uF 25V electrolytic cap in good order would be expected to have an ESR of between 0.05 and 0.1 Ohms. Any higher than 0.1 Ohms would be very suspicious.

There are generally two types of ESR meter out there, one type uses a 100kHz sine wave test signal, the other type (such as the Anatek Blue meter, designed by Bob Parker) uses a pulse method. Which I think is a better method for assessing the ESR of SMPS capacitors, because they are in circuits with current pulses, not 100kHz sine waves.

Interestingly, the original Bob Parker design ESR meter kit, is still available from Altronics, I regard these as the World's best ever ESR meter:


( not everyone might agree, but keep in mind I am a very harsh judge of everything, including test equipment and seldom easily impressed )

Apart from the ESR, the capacitor should also be checked on a capacitance meter and checked for excessive electrical leakage at it max operating voltage for a full health checkup.

One interesting thing about electrolytics, they can just leak out a little electrolyte, or if they have leaked in the past, sometimes it is not obvious. However if they have leaked in the past, the electrolyte soaks into the surface of the rubber bung on the base, and makes the surface conductive and if you check with the Ohm meter, you can detect electrical leakage by pushing the probe tips into the surface of the rubber bung. There should be zero electrical leakage.

One other thing, if a particular brand and age and type (uF and voltage rating) of electrolytic had leaked electrolyte on a vintage pcb, it is highly likely that its same aged bothers and sisters on the pcb, will do the same thing in the next few coming years. This can be a clue that it might be wise to change all of the ones of the same type, if one has chosen to leak electrolyte.
 
I'm working on an ASUS Socket A motherboard and most of the eight 3300@6.3 caps were bulged - ESR ranged from 1.9 to 3.5 ohms. Replacement capacitor ESR's were all about 0.087 ohms. I use one of the MESR-100 100KHz type meters.
 
Back
Top