Capacitance between two windings in the isolation transformer, does it matter ?

[BravoV]

The context here is mainly at the isolation transformer used for supplying from the mains power into the DUT that will be probed using an oscilloscope.

For example my 1000 VA transformer has 1250 pF measured between pri vs sec windings.

When or where this capacitance will become a factor that worth a consideration regarding the measurement results, again, in the context of measuring the dut powered from this transformer using a scope.

An explanation + example cases will be really helpful, thanks in advance.

[AlfBaz]

Hi Bravo
I'm not quite sure what you mean by your question
If you are asking wether the isolation is compromised, you are better off measuring the insulation resistance between primary and secondary with an insulation tester.
1250pf equates to 2.5Gohms reactance so I don't think that will be of any concern.

As for it affecting any measurements there are other losses and parasitics that will swamp that capacitance into insignificance, not to mention the inductive coupling of the 2 coils

[SeanB]

It will appear as a 1n capacitor connected to the mains. this current will be trying to flow through any connection you have to ground, so it will generate a current to mains earth via any probe you connect that has grounded connection at the meter end. Just keep this current in mind when using it, it may or may not be significant depending on application. If you are probing the primary of a switching power supply ( only reason that you will be using an isolation transformer in most cases) it just means you can only connect to the negative of the mains capacitor, and will have to measure from there, you will not be able to eg probe the gate drive for a high side switch easily. If you connect the ground to the high side switch something will blow up.

[AlfBaz]

Sean, what am I missing? how do you blow stuff up with a hand full of pico amps? Or is it something to do with creating a ground reference with the oscilloscope's pwr supply and this leakage current?

[SeanB]

Consider the impedance of this 1n capacitor if you apply 20kHz to it, at a voltage of 300V peak to peak. What current will flow in the scope lead. 40mA or more flowing through the probe lead will have a rather big influence on the display, especially as most probe leads are quite resistive.

Connected to the power supply capacitor the current will be at 100/120Hz and will be quite small, though it will be noisy, as it is going to have diode switching edges on it.

[AlfBaz]

Ok thanks
I immediately assumed a mains isolating transformer with a primary voltage of 120/240 at 50/60Hz feeding the device under test, at these frequencies 1n may as well be considered an open cct

[BravoV]

It will appear as a 1n capacitor connected to the mains. this current will be trying to flow through any connection you have to ground, so it will generate a current to mains earth via any probe you connect that has grounded connection at the meter end. Just keep this current in mind when using it, it may or may not be significant depending on application. If you are probing the primary of a switching power supply ( only reason that you will be using an isolation transformer in most cases) it just means you can only connect to the negative of the mains capacitor, and will have to measure from there, you will not be able to eg probe the gate drive for a high side switch easily. If you connect the ground to the high side switch something will blow up.

Thanks SeanB  , this is what I'm expecting.

At normal grounded scope (not battery/isolated one), at normal probing, am I right if this current flows through secondary winding across the dut and to the probe's ground lead ?

Edit :

Anything else that I missed ?

[AlfBaz]

Bravo, I don't mean to hi-jack your thread but I'm curious and want to understand.

What sort of isolation transformer is this? Is it being used to isolate the mains? Is it 240V or 120V 50 or 60Hz or is it as Sean has stated a 300vpk-pk 20Khz xformer rated at 1KVA?

[SeanB]

The best isolation transformers have low capacitance from secondary to ground and to primary. Unfortunately this means a big transformer with separate windings on different limbs of a CI core, with a large air gap between the secondary and the core to reduce capacitance, in a large case so that overall the capacitance is reduced.The primary will also need a shield wound on the outside of the winding to reduce noise coupled to the secondary. Expensive in both a custom core blank and in the larger copper coils needed, along with the special bobbins required in the secondary.

Alfbaz, this is an isolation transformer for mains use, 120/220VAC 50/60Hz transformer rated for 1kVA. The 300VAC 20kHz is the example of a connection to the midpoint of the power devices of a switching power supply fed from it, where you do not want to connect the probe ground to the midpoint where it will try to drive the whole system to +- 300V at the switching rate, with the charging current for the capacitance having to flow through your scope probe lead. Something will not like that.

[AlfBaz]

... The 300VAC 20kHz is the example of a connection to the midpoint of the power devices of a switching power supply fed from it, where you do not want to connect the probe ground to the midpoint where it will try to drive the whole system to +- 300V at the switching rate, with the charging current for the capacitance having to flow through your scope probe lead. Something will not like that.

Thank you.

[The Electrician]

You could get one (or more) of these:

http://tmi.yokogawa.com/products/oscilloscopes/voltage-probes/701921-100-mhz-differential-probe/

Or you could get a scope with isolated inputs, such as this: http://www.tek.com/datasheet/oscilloscope/tps2000-digital-storage-oscilloscopes

The capacitance between inputs and frame for this scope is about 10 pF.

Or you could get a small isolation transformer for the scope; use it in addition to your 1 kVA unit.  I have a small 85 VA isolation transformer for my scope.  It is constructed with a split bobbin, can stand off 5000 volts primary to secondary, and measures about 40 pF primary to secondary.

[EEVblog]

A properly designed isolation transformer will certainly try and minimise the primary/secondary capacitance, and will usually use an electrostatic shield between the two coils.
My isolation transformer has both using a toroid transformer, I was going to do a video on this.

[BravoV]

Thank you guys for excellent replies, really appreciate it. 

Next question, consider this three measurement conditions :

1. DUT powered directly from mains and without iso-trans, and measurement is using a "battery" powered scope.
2. DUT powered by iso-trans, and the scope is mains powered directly.
3. DUT same as point no. 2. but the scope is powered from another separated iso-trans.

Say these 3 scopes are exactly identical in their spec and also the probes as well, and they are measuring the exact identical points at identical 3 DUTs. 


Will that 3 scopes show identical measurement results ?

[SeanB]

Consider that an isolation transformer only does the following: It isolates the output from connecting directly to the MAINS input, it carries the PE conductor through to the output side without any modification. It allows you to have the one side of the output connected to mains earth without causing a problem, but does not provide anything else. Thus the isolating transformer does absolutely nothing when it powers the scope, it will still have the input connected to mains earth via the scope. No1 allows you to probe the high side of a switcher without much problem, as the capacitance to ground of the floating scope will not be very high. No 2 will have issues with that. No3 will go bang and destroy both the DUT and the scope and probe when you connect it.

[lewis]

The interwinding capacitance of isolation transformers can be significant - you mentioned just over 1nF - which at mains frequencies is nothing. However, if you have HF noise and interference on the mains, that will capacitively couple over to the secondary and appear as a common mode voltage on the Live and Neutral with respect to mains earth. If you're probing your circuit directly, with your scope chassis grounded, you may find this upsets your measurements due to this common mode current draining through your scope probe's ground lead.

As Dave mentioned, an electrostatic screen between the primary and secondary almost eliminates interwinding capacitance. This is simply a layer of foil, connected to mains earth, between primary and secondary that is wound in such a way that is does not form a shorted turn, but gives complete coverage of the primary-secondary insulation. While this almost eliminates coupling between primary and secondary, if you have common mode noise on your mains earth this can still couple from the screen into the transformer secondary which will once again drain through your scope's ground. Again, it might not be a problem unless you're making absolutely critical measurements.

I mentioned that 1nF at mains frequencies is 'nothing' but this is not entirely the case. At 50Hz, 1250pF has a resistance of about 2.5M. If you connect your scope probe tip directly to the live or neutral output of the transformer with the ground unconnected, you'll see about 0.8x your mains voltage on the scope trace because this 2.5M becomes the top end of a potential divider to the mains with the bottom end being your scope's 10M input impedance to earth. In other words, it'll look like the isolation transformer isn't doing anything.

In either case, if you're making precision measurements on a mains-connected circuit, it's a good idea to use a heavy mains filter at the input of the isolation transformer to attenuate any HF before it has a chance to couple through the transformer. Better still, combine this with a proper, high quality, high bandwidth differential probe. While the CMRR of even the best differential probe is not infinite, it still breaks any earth loops formed either directly (if your DUT is earthed) or indirectly (through interwinding capacitance to the primary, or through coupling from the electrostatic screen).

[The Electrician]

No 2 will have issues with that. No3 will go bang and destroy both the DUT and the scope and probe when you connect it.

Why would No 3 go bang, but No 2 wouldn't?  In both cases, the DUT is powered through an isolation transformer.

[SeanB]

Oops, my bad, in no 3 you are right, the second isolation transformer serves absolutely no use other than using more power. If you only have the one transformer though and try to isolate the scope only while powering the DUT direct from mains then it will go bang.

[M. András]

that electrostatic shielding on a toroid doesnt reduce the efficiency of the transformer? if anyone knows the answer how much cost it ads over the costum toroid in %


i was looking in the past days of the agilent differential probes, which would be better in these situation a 25mhz one or a 100mhz one? 1.4k differential input range on the second one if i remember correctly

[BravoV]

@lewis , thanks for the explanation ! 

So, whats the result on scenario no.3 ? Still not possible for high side probing ? or other bad side effects ?

[ftransform]

As a side note the only time I ever heard of this capacitance mattering is in Analog Devices Appnote for nanovolt voltage ripple measurements. Jim Williams uses a special isolation transformer which takes into account the capacitance in order to power a bench power supply.

When I looked it up it cost 1600$.

[lewis]

@lewis , thanks for the explanation ! 

So, whats the result on scenario no.3 ? Still not possible for high side probing ? or other bad side effects ?

No problem!

Scenaios 2 and 3 are identical as far as the scope is concerned, as SeanB mentioned the isolation transformer has its input earth connected directly to its output earth, so the scope chassis will still be mains earth referenced.

<bodge mode>

If you're feeling brave, and I'm sure everyone here recoil in horror, but you could use scenario 3 but connect your scope earth to the 'neutral' side of the isolation transformer secondary. In other words, move the earth wire and connect it with the neutral in the scope's plug (leaving the earth pin in the plug unconnected). Now your scope is not mains earth referenced, but will still blow a fuse if a catastrophic fault develops. DO NOT plug the scope directly into a wall socket if wired this way, only use it with an isolation transformer. And you just have to HOPE the isotrans will not go faulty wile you're measuring. Use at your own risk!

Before everyone gets their guns blazing, consider what happens at the transformer on the pole that powers your home.

There may still be some coupling effect, but it will be significantly reduced.

</bodge mode>