20 Jul
2006
20 Jul
'06
5:02 p.m.
O.K. - now on to the power supply.
The +5v channel 1 supply is good, 69 mV peak to peak AC noise (measured o=
n a scope with AC coupling)
the +-12v channels are ~292 mV of noise
the -5v channel has 500 mV of noise, even with an additional 1000 =B5F ca=
pacitor across the output (to hopefully filter this junk out)
(-5V also has periodic fluctuations at a lower frequency)
All of this is high-frequency, excluding the -5v hum.=20
In short, it doesn't look like adding an external filter cap will help mu=
ch, unless I go all the way and throw in a LC filter.
My first through is that capacitors ahve inducatnce too, and that your
added one might well have significant impedance at the noise frequency as
a result
Pulled the P/S apart, and the board with +-12V and -5V
has peculiar capac=
itors mounted on it - 3 lead Mallory Aerovox.
Are these really the multipart electrolytics that went out of style (I th=
ought) with solid state?
My guess is that the 3 leads (2 on te +ve side?) are there to reduce the
self-inductace, e.g. by connecting to both ends of the rolled-up foil
inside. I think I've seen similar caps in HP stuff.
The wiring is 2 leads connected to one side of the
power rail, one lead t=
o the other. Could twin-lead caps be substituted?
Only if the self-inductance is low enough. Capacitors designed for use in
SMPSUs should be OK, maybe parallel them with a 1uF non-electrolytic for
good measure
-tony