AC leakage in switching power supply 2

[Greenpower]

I would like some opinions on this matter. We currently use Idec PS5R series switching power supplies ranging from 12 to 24 volt dc at 30 to 50 watts with 120-volt ac input. The problem that we have found is that when powered up if you read your dc out from + and – you will read the proper out put 24vdc when you pot you meter on ac to measure ripple you read 1.4 millivac, this is not bad, but when you read from either your + or – dc out to the earth ground or neutral you will se anywhere from 9 to 15vac. I consider this possible leakage and feel that this can cause problems in analog circuits and could cause spikes and transients in the system. I contacted Idec and was told this was not a problem because it was on the secondary side of the transformer. We have some idec supplies that we purchased about 6 months prior and this problem does not exists, only in the ones we have recently purchased, also idec makes a 240 watt supply this one does not have this leakage. Any input would be appreciated.

 

[IRstuff]

Depending on your definition of leakage...

A truly isolated secondary can be at any potential w.r.t. to earth ground; since there is no path to ground, there is no reason for it to be at ground potential. If anything, I'm tempted to conclude that your other supplies are leaking to ground through the AC neutral.

The AC that you are seeing is simply the capacitive coupling to the only potentials that exist in the power supply, which is the AC main.

If your circuitry is isolated from ground, there should be no problem. Your single point grounding should eliminate this altogether. There should not be a ground loop issue since the current flow is from the main and not the supply.

TTFN

 

[busbar]

If the 9-to-15vac reading is measured with a 10MÙ DMM, connect a ~10k resistor to ground from one of the supply-output terminals and if all else works OK, that should cure it. Looking at the voltage with a scope may indicate high-frequency [and/or line-frequency] noise, and have not much effect on the powered system.

 

[peebee]

Can you solidly ground the DC output rather than using the resistor? That will definitely eliminate your common-mode ripple.