royalmp2001 said:
Hi,
I got hold of a 20yr old dead Adcom hifi power amp recently together
with the service manual.
The fuses on each +ve line have blown, the fuses on the -ve lines were
ok.
I have not tried the amp out - I have taken out these supply line fuses
(between the psu and the main electronics), therefore isolating the
power supply to the two amp boards, and checked the psu with a
voltmeter and oscilloscope.
The psu seems fine.... + and - 49.4V out, and all lines giving out
about 10mV ripple. One rail had 20mV ripple.
As this is quite an old amp, and I understand the electrolytics go
first (after pots, swtiches and connectors), what tests can I do on the
caps to see what kind of life I can get out of them. I can replace
them, but if it is not necessary I don't want to do it.
Thanks, guys.
Hi, Royal. The first check should be in the power supply itself. You
seem to have already begun this. Try no load, then add some power
resistors to approximate the load of the amplifier. Look at the loaded
ripple, and see if it's anywhere near what it should be. For a 1 amp
load, you should get 1 volt of ripple for an 8300uF cap. If your cap
is 8300uF and you have 2 amps, you should have 2V of ripple. Double
the load, double the ripple. If your capacitance is doubled, your
ripple should halve. Work it out on paper beforehand, and compare your
calcs with your observations.
While you're at full load, run it for a few minutes, and feel the body
of the cap. It shouldn't be getting more than slightly warm. If any
of them are getting toasty, leakage current is probably well on its way
to cooking the cap.
One good way to check older large electrolytic power supply caps if
you've got time is to disconnect one end, and then charge them up to
rated voltage with a fairly good sized series resistor. For instance,
if your caps are rated for 63WV, try charging them up to 60V through a
10K 1/2W resistor. If the cap can't make it up to anywhere near
voltage, or the voltage starts ramping up, and then either stabilizes
or goes down on its own, you're SOL -- the cap is bad. Once the cap
voltage stabilizes at close to your applied voltage, read the voltage
across your 10K resistor with your DVM. If it reads more than a couple
of volts (a couple of tenths of mA of leakage), that indicates bad.
You're saying that the PSU seems to work fine when the amplifier is
disconnected, and you'd like some information on troubleshooting the
PSU caps. I'm not sure I follow you. You might want to apply yourself
to the amp itself, which seems to be where the problem is.
It sounds like the power supply lines on the amp board are shorted out
(likely a cap on the amp board?). Now you can get a start on tracking
down short circuits if you've got a floating 13.8V regulated bench
power supply, a DVM with a 200mV range, and some power resistors in
your junkbox. If you set up something like this (view in fixed font or
M$ Notepad):
|
| 100 ohm 3 watt
| + ___
| o-----|___|-o---------------o
| | ,
| | ,´
| | ´
| .-. |
| 13.8V 1 ohm | | /+\
| | | (DVM)
| '-' \-/
| | |
| | \.
| - | \
| o-----------o---------------o
(created by AACircuit v1.28.5 beta 02/06/05
www.tech-chat.de)
you'll be able to make up a crude short finder, which shouldn't damage
even semiconductors in the circuit. Use jumpers to clip the 2 lines
parallel to the 1 ohm resistor on the probes of your DMM, as close as
you can get to the tips of the probes. Now place your probes across
the shorted power lines on the amp board, and move them around to try
to get the minimum mV reading on the DMM. This will work down to a
couple of milliohms, which might get you pretty close to where the
short is, or at least narrow it down by quite a bit. The response is
obviously non-linear, but it becomes most sensitive (as well as useful)
at less than ten milliohms. (Note: you're not reading milliohms here,
you're reading the voltage across the probes in parallel with a 1 ohm
resistor. However, a reading of 0.1mV greater than the minimum reading
you can get with the probes touching each other will correspond to
roughly 1.4 milliohms extra resistance.)
If you've got a more expensive DMM with a 20mV range, you can click
down to less than two tenths of a milliohm per count at the low end,
which should walk you right to the short.
The advantage of this little trick is that you'll never apply more than
140mV or so to the board, so there's essentially no chance of smoking
anything, even a JFET input.
If you narrow it down to one area of the board, and you think a cap
might be suspect, use a solder sucker to open one leg of the cap, and
see if the short is still there.
There are easier ways to do this, and some rather expensive instruments
exist which can do some cap testing in-circuit (Sencore), but for
hobbyist repair work this should be sufficient.
Happy hunting
Chris