Electrolytic leakage

T

Theo Markettos

Guest
I'm trying to fix a switched mode PSU, which has the symptom that the output
is dropping out under load at about 10Hz. As the PSU usually runs very hot
(and has done for the 5 years I've had it working) this suggests that an
electrolytic might have dried up. I've removed the main mains electrolytic,
which is rated at 100uF 400V. AFAICS none of the caps show signs of damage.

Is there any good test for leakage on such caps? My multimeter (Fluke 87)
only measures capacitance up to 5uF. I tried it in a basic 555 timer
circuit and the value seems to be in the right ballpark. The series
resistance measures about 6.7M - is this too low? My previous meter only
did up to 2M, so I haven't been able to measure this high in the past. I
tried charging it up to 19.00V, and after 40 mins (/without/ anything
connected in the interim) it was down to 18.35V. This sounds like there's
not undue leakage, but are there better ways to test leakage than this - or
other heat-related failure modes?

Thanks,
Theo

--
Theo Markettos theo@markettos.org.uk
Liphook, Hampshire, UK theom@chiark.greenend.org.uk
http://www.markettos.org.uk/
 
Theo Markettos wrote:
I'm trying to fix a switched mode PSU, which has the symptom that the
output is dropping out under load at about 10Hz. As the PSU usually
runs very hot (and has done for the 5 years I've had it working) this
suggests that an electrolytic might have dried up. I've removed the
main mains electrolytic, which is rated at 100uF 400V. AFAICS none
of the caps show signs of damage.

Is there any good test for leakage on such caps? My multimeter
(Fluke 87) only measures capacitance up to 5uF. I tried it in a
basic 555 timer circuit and the value seems to be in the right
ballpark. The series resistance measures about 6.7M - is this too
low? My previous meter only did up to 2M, so I haven't been able to
measure this high in the past. I tried charging it up to 19.00V, and
after 40 mins (/without/ anything connected in the interim) it was
down to 18.35V. This sounds like there's not undue leakage, but are
there better ways to test leakage than this - or other heat-related
failure modes?

Thanks,
Theo
Click to expand...
Here's a good test to see if the cap is drying up. Charge it up to full
rated voltage and leave it that way, still connected to a voltage source for
some time (5 minutes at least, prolly more). Then, disconnect it and
discharge it quickly with a small resistance or a short across the leads.
Remove the short immediately, wait about 15 seconds and then measure the
voltage across the leads with a high impedance meter. If the cap is partly
dried out you should see the voltage slowly rise, as though the cap was
recharging itself.

This is caused by dried out sections of the cap acting as small capacitances
with high resistances in series. If you charge them up, and then dischare
the cap in much less then a time constant, they will still have most of
their charge, and it will bleed away into the rest of the cap slowly. I have
seen old high voltage caps recharge to many 10's of volts on their own.

Adam
 
Theo Markettos wrote:
I'm trying to fix a switched mode PSU, which has the symptom that the output
is dropping out under load at about 10Hz. As the PSU usually runs very hot
(and has done for the 5 years I've had it working) this suggests that an
electrolytic might have dried up. I've removed the main mains electrolytic,
which is rated at 100uF 400V. AFAICS none of the caps show signs of damage.

Is there any good test for leakage on such caps? My multimeter (Fluke 87)
only measures capacitance up to 5uF. I tried it in a basic 555 timer
circuit and the value seems to be in the right ballpark. The series
resistance measures about 6.7M - is this too low? My previous meter only
did up to 2M, so I haven't been able to measure this high in the past. I
tried charging it up to 19.00V, and after 40 mins (/without/ anything
connected in the interim) it was down to 18.35V. This sounds like there's
not undue leakage, but are there better ways to test leakage than this - or
other heat-related failure modes?

Thanks,
Theo

--
Theo Markettos theo@markettos.org.uk
Liphook, Hampshire, UK theom@chiark.greenend.org.uk
http://www.markettos.org.uk/
Click to expand...
You want to test the ESR (Equivalent Series Resistance) of the
capacitor There a re some simple schematics on the web, as well as kits
and commercial test equipment.

One popular meter in kit form is the Dick Smith ESR and Low Ohms
Meter, available world wide.
--


Michael A. Terrell
Central Florida
 
Michael A. Terrell <mike.terrell@earthlink.net> wrote:
You want to test the ESR (Equivalent Series Resistance) of the
capacitor There a re some simple schematics on the web, as well as kits
and commercial test equipment.

One popular meter in kit form is the Dick Smith ESR and Low Ohms
Meter, available world wide.
Click to expand...
Thanks. I made the '99 cent ESR tester' (a square wave driver into the cap)
- but the caps test out fine. Never mind :-|

Theo
 
You must log in or register to reply here.

Welcome to EDABoard.com

Sponsor

Back
Top