Eeyore wrote:

Fred Bartoli wrote:

Franc Zabkar a écrit :
What is the function of the primary-secondary coupling capacitor (eg
http://tinyurl.com/ha435) in a SMPS? I ask this because I suspect it
may be the cause of a video "hum" issue.

I have a new DVD player that gets along just fine with my 34cm TV but
produces severe hum in the picture of my 80cm TV. The same TV works
well with my other DVD player. I am considering removing the subject
cap from my TV (the DVD is under warranty), but I'd like to know if
there is any reason why I shouldn't. BTW, the TV is a gut-buster so
I'd rather not move it unless I have to.

FWIW, the symptom looks like common mode noise.

SMPS, depending on how the transformer is wound tend to generate some
fair amount of common mode current at switching frequency. (due to
parasitics capacitance between primary and secondary and voltage
distribution between layers and turns).

This capacitor, connected between GND_secondary and "GND_primary"
provides a short path to this current, thus preventing this HF common
current flowing other ways and creating EMI issues.

Removing this cap from your TV set, might or might not solve your pb,
(it might even increase it) but it'll almost surely make your TV not FCC
compliant anymore (if it were not needed the manufacturer would have put
it there).

In case your interference pb is due to your TV residual SMPS CM current,
one way to reduce it further is to add some ferrite ring on one of your
devices mains cable, or on the video signal cable. Several turns help more.

I'm afraid the ferrite won't help.


Well, I didn't take the video "hum" as a pure 50/60 Hz issue.
It's highly unlikely: 1nF, at 50Hz/230V, gives only 70uA leakage
current, which translates to, at most 7uV, in case of a very bad 0.1R TV
to DVD GND path impedance.

I think it's more a CM switching residues problem.
A ferrite toroid with a high AL, could easily improve this by reducing
the CM current between the TV set and the DVD player, and also by
improving the coupling between the signal and gnd path.


The problem is that the Y cap in that location results in a small ac leakage
current when the secondary circuitry is connected to other grounded equipment.


Don't think so. See above.
Anyway, the main frequency can get back there, simply by the means of
modulating the SMPS pulse width.


--
Thanks,
Fred.