Use of mains isolation transformer for offline SMPS testing?

[cupoftea]

Hi,
When testing offline SMPS’s primary side…when using an isolation transformer, I deliberately ensured that the Earth wire did not connect to the secondary whatsoever…in other words, secondary was totally floating, and this enabled me to put the scope probe ground clip anywhere on the circuit and not cause a problem….however, it does dawn on me that this is unadviseable?...as such, earth could be connected to the secondary “neutral” via a 1MEG resistor. Would you say this was reasonable?

 

[betwixt]

If it is transformer isolated the whole secondary side should 'float' and you could earth any point you wish safely. Voltages are measured relative to another node, it doesn't matter what the absolute voltage common to both is. However, because of leakage through the isolating transformer, it wouldn't be sensible to earth a sensitive signal node. Earthing one side of the supply lines should be fine but be careful if you attach a load that it too is isolated.

Brian.

 

[cupoftea]

However, because of leakage through the isolating transformer, it wouldn't be sensible to earth a sensitive signal node.

Thanks, so for example, if we wanted to scope the primary side current sense pin of a PWM controller, (that was fed by an isolation transformer) then we should clip the scope ground clip to the dc bus rail, and put the scope tip to the current sense pin, rather than vice versa?

Earthing one side of the supply lines should be fine

....thanks, i do notice that a lot of mains isolation transformers have one of the (isolated) AC outputs internally connected to earth....and so then the first thing you need to do is rip the earth wire out of the (primary side) mains plug that they come with.

Isnt it actually a little dangerous to have the secondary of the isolation transformer totally floating?...and that you need to earth it through say a 1MEG resistor?

 

[betwixt]

How you probe it is up to you but remember your probe is essentially measuring the differential between clip and probe so it makes sense to connect across the sense components. If they are at high voltage, use two probes and the "add and invert" function on the scope so you measure the difference between them, that way the scope can be grounded to the 'ground' side of the supply.

Whether or not a resistor to bled transformer leakage away is needed depends on how good the isolating transformer is. A good one with isolated windings and an earthed screen and core should be OK without a resistor but adding one wouldn't hurt. In most cases, the test equipment itself would be earthed so it would hold any connected equipment at earth potential anyway.

Brian.

 

[cupoftea]

the test equipment itself would be earthed so it would hold any connected equipment at earth potential anyway.

Thanks yes, thats a good point, and im referring to the times when the test equipment isnt attached......totally floating secondaries scare me a bit........as in to what voltage will they float up to.

 

[Easy peasy]

A 2W 2Meg2 ( 600V ) resistor to EARTH on the secondaries is commonly used to set the 0v output of a psu at near earth potential - if a mistake is made and the earth becomes phase the limiting current is 105uA rms - which should be safe.