Isolation transformer - Voltage between output and ground

[tony359]

Hi all,

I built my isolation transformer box - with light bulbs for limiting - some time ago. I use it very often even when it's not needed as I can switch on and off the device under test very easily from there.

Some time ago I had a motherboard powered via the isolation transformer (via switching PSU) and I was wearing an ESD wrist-band. At some point I felt a tingling sensation where the wristband was touching my skin. I measured the voltage between what I was touching on the Mobo and mains ground and I could read 30V!

Today I checked and indeed I read 30V between the output of the transformer and ground.

With the box disconnected from mains, I have infinite resistance between the secondary and primary and also between secondary and ground. Ground - of course - is not going to the ground connection of the output which is then completely isolated. So between mains and the output there is the transformer, nothing else.

The case of the transformer is grounded but I read infinite resistance between the metal part and the secondary.

I thought that that was not possible? I thought that because of the isolation transformer in the way, I wouldn't read any voltage between the output and ground?

What am I missing here? Picture of the transformer attached.

Thanks!
Tony

[bdunham7]

The most likely possibility is capacitive coupling somewhere in the system.  This is an example of why you should not break the ground connection by default with an isolation transformer. 

If you want to measure leakage, do it in µA not volts.  You can put a 10k resistor across your multimeter and read it that way, 1V = 100µA.

Just out of curiosity, how is your primary wired on that transformer--the picture is cut off, but it looks like you have a 400V primary? 

[tony359]

Thank you!

It's an optical illusion! It's wired to 0 and 230V (I have 230V coming in).

I can measure current with the multimeter. I thought I didn't want to have ground as that would add a reference to mains ground, making the isolation useless?

[bdunham7]

I can measure current with the multimeter. I thought I didn't want to have ground as that would add a reference to mains ground, making the isolation useless?

If everything is working right and you don't have to change the ground reference point from where it normally is to some other node, then leaving the ground in place is fine.  The isolation is achieved because the secondary output of the isolation transformer is not referenced to that mains ground.  Without the ground connection, the ground plane of the DUT can float away to any random voltage and if there is even a small leakage someplace, you will observe what you did. 

If you want to use a grounded measuring device like an oscilloscope for a measurement where both connections you want to make are not at ground potential, then you can break the ground.  Say you had your MOBO + SMPS sitting on an insulated mat and you wanted to use a small shunt resistor to measure the +5V supply current at the high side.  You would break the regular ground connection, install the resistor, connect your oscilloscope ground clip to the load side of the resistor and the probe tip to the supply side, energize and take your reading.  In this case, even though you have broken the ground, it won't float away as it will always be 5V below the actual ground.

Isolation transformers aren't a magical safety blanket.  You have to think every step through.  Take your small tingling as a warning! 

[EPAIII]

Lets get this straight.

First you completely bypass the safety features of your local power grid by using an isolation transformer.

Then you attach a grounded strap on your wrist. OK, it probably has a resistance built into it to limit the current, but still, IT IS GROUNDED.

Then you are puzzled when you can feel a shock. Granted, a small shock, but still a shock. That resistance in the ESD strap is limiting the current to that small value.

My first thought is YOU DO NOT UNDERSTAND ELECTRICITY! 

I spent over 45 years working professionally with all kinds of electronic and electrical equipment. Voltage levels from 5V or less to 40,000V or even more. I worked on high power transmitters and on flea power logic circuits and virtually everything between. I NEVER WORE ANY KIND OR ESD OR GROUND STRAPS. NEVER!


Why not? Because I valued MY LIFE more than some static sensitive IC which cost mere pennies. If you or I touch a point that is hot, that ground (ESD) strap COMPLETES the circuit and current can then flow.

But I worked with those static sensitive ICs and other components. In 45+ years how many did I destroy with static electricity discharges? ZERO!

That's not an exaggeration. It really was zero devices destroyed with ESD. The need for protection against ESD is tremendously over stated. It is probably only needed in the warehouse where the ICs and other static sensitive devices are handled by people who have no idea of how things work. They are not working with energized circuits so the ESD protection devices are a lot safer to use there.

Likewise, for isolation transformers. Again, in 45+ years of professional electronic work, I NEVER used them. Actually, looking back, I think only ONE SHOP that I ever worked in even had one. And these were professional shops usually with a number of very experienced technicians working in them. If you have an isolation transformer between the mains power and the device you are working on, you have DEFEATED any safety measures that your local power company and local electric code has WISELY incorporated in the grid up to the outlet your transformer is plugged into. And, if you bypass that transformer with a ground wire, then you may as well not have the transformer in the first place.

Somewhere outside of your house or shop, your local power company has, GUESS WHAT. A TRANSFORMER! And it is configured in a manner that was determined by years, decades actually, of experience with electrical safety. You ALREADY have a transformer! Adding a second one is redundant and unnecessary!

If you really want additional protection against electric shocks, I would consider installing a ground fault device either in your breaker box or at the outlet you are using in your shop. But that may drive you nuts with multiple trips if any of your equipment has a leakage path in it.

DO learn how to work safely with electricity! But, in doing that DO understand how it actually works.

PS: I also never liked the ESD mats for the work benches. If I was stuck with one, I DISCONNECTED the ground strap before it allowed me to get shocked.

[EPAIII]

Oh, and I never had any problem with a meter or oscilloscope or other item of test equipment because I made sure that I understood where my grounds were before I connected anything.

If your scope is grounded through the power grid,

And your device under test is grounded through the power grid,

Then they have a COMMON GROUND.

All you need to do is be sure that the ground leads of your scope probes ARE connected to ground points in that device under test.

Or, if you are only doing low frequency work, then you don't even need to use those ground leads at all. The common ground through the power outlets will serve as a common reference. I have done that MANY TIMES.

REALLY, that is all!

[tony359]

@bdhunam7

Thank you, I appreciate your input and yes I am aware the isolation transformer is not a magic thing making everything 100% safe!

I'll study what you said and I'll experiment with ground a bit. Maybe I could add a switch on the box I made to engage/disengage ground as needed. That said, from what I read it feels I might want it connected than not. For measurements where ground is not at the same mains potential I have differential probes.

Thanks for now!

[Zero999]

Yes, it's capacitive coupling. If you measure the current, you'll find it's tiny.

Try measuring the capacitance between the primary and secondary windings, with shorting links between either end of each winding to minimise the effect of the inductance.

Regarding ESD straps: they're not dangerous, when working with low voltages (these are defined as <1kV for the purposes of safety) because they have to have an internal high voltage 1M resistor to limit the current to under 1mA. At the very worst you'll receive a tingle and have a mild muscle spasm in your arm.

Be careful with isolation transformers. They have their place but can make it more dangerous, because they render RCDs useless. It the secondary does become earthed, then there's no protection against electric shock. To reduce the risk, its good to have insulation monitoring on the secondary, which will alert you when the secondary is connected to earth.

[tooki]

I spent over 45 years working professionally with all kinds of electronic and electrical equipment. Voltage levels from 5V or less to 40,000V or even more. I worked on high power transmitters and on flea power logic circuits and virtually everything between. I NEVER WORE ANY KIND OR ESD OR GROUND STRAPS. NEVER!


Why not? Because I valued MY LIFE more than some static sensitive IC which cost mere pennies. If you or I touch a point that is hot, that ground (ESD) strap COMPLETES the circuit and current can then flow.

But I worked with those static sensitive ICs and other components. In 45+ years how many did I destroy with static electricity discharges? ZERO!

That's not an exaggeration. It really was zero devices destroyed with ESD. The need for protection against ESD is tremendously over stated. It is probably only needed in the warehouse where the ICs and other static sensitive devices are handled by people who have no idea of how things work. They are not working with energized circuits so the ESD protection devices are a lot safer to use there.

…

PS: I also never liked the ESD mats for the work benches. If I was stuck with one, I DISCONNECTED the ground strap before it allowed me to get shocked.

I just want to point out that this shows a shocking lack of understanding about ESD risk, and about how ESD protection works.

And of course that the “zero devices destroyed with ESD” claim is not knowable. It certainly isn’t a claim one can take seriously.

[Siwastaja]

Yeah, if you fear the electric shock to the point of randomly getting rid of ESD protection and think you get an improvement in safety that way, something's seriously wrong with the attitude and processes. Seems like false sense of security to me; in reality, removing ESD procedures does not make the environment much safer. You need other ways to guarantee you never become part of the live circuit.

Instead of fear regarding the ratings of the 1Mohm resistor in the mat, why not just verify the resistor is up to specs? You can measure it with high voltage isolation resistance tester for example, or replace it with known good resistor? If you can't trust the rating of such trivially simple safety-critical component, how can you trust transformer isolation ratings, or anything like that?

[Ian.M]

Try reversing the primary connections and measure again.  Its likely that one end of the primary is closer to the secondary, and you want that end connected to Neutral to minimise the capacitively coupled common mode voltage on the secondary.

N.B. Measure AC voltage at each end of the secondary with respect to ground and take the lower reading as the capacitive coupling is almost certainly asymmetric, as it will be between the last layer of the primary and the first layer of the secondary (or visa versa depending on which winding is on the outside)

@Siwastaja
Or build your own ESD grounding point with two 1Meg high voltage resistors, from a 1st world manufacturer,  in series, properly mounted and insulated so they cant (easily) short out.  Then it doesn't really matter if the ESD lead has a crappy resistor!

[Zero999]

1M is low for ESD. I would go for a higher value, perhaps 10M, made with 10, 1M resistors in series. Where I used to work, one of the ESD points just went to a piece of copper foil glued to the wall, via 1M resistor, which measured about 10M to mains earth.

[S. Petrukhin]

But I worked with those static sensitive ICs and other components. In 45+ years how many did I destroy with static electricity discharges? ZERO!

Perhaps your skin has low resistance, or your shoes are soaked in sweat salts, etc. - there is no static charge accumulating on yours. 
However, 45 years of practice is not a reliable refutation of science.

Nevertheless, I agree with you that the precautions for static electricity are elevated to an unreasonable drama. Like ROH, deionized water instead of alcohol, component drying is a field for business and income generation. In the West, they like to create fright and make money from it.

Most often, it is enough to remove the charge once before starting work.

[jonpaul]

Stray capacitances Pri-Sec and Sec -earth  will from a  divicder

EMI filters add capacitances with VDE X and Y capacitors.

The currents are tiny at mains freq and usual capacitances.

Measure all the stry/parasitic, drew an equiv clt and you can see the reaspon

Jon

 

[m k]

In my case personal ESD protection was mandatory, everywhere.
The whole area of the building was also ESD protected.

During winter times static charges were still around.
Personal wrist wrap was clearly needed.

There I also learned that major part of ESD damages are partial.
The machine is still operational, but below its standard.

Later I educated how a computer shop can stop their constant intermittent problems with installed and still warrantied memories.
Unfortunate to end user, after warrantied period the original source of the problem is moot.

[S. Petrukhin]

The whole area of the building was also ESD protected.

The whole building was filled with only iron, like a lathe shop? Did nothing dielectric exist?

[m k]

D as discharge.
Protected against unwanted discharged.

Charges can't be avoided.

[andy2000]

One example of where an isolation transformer is useful is when I need to work on the primary side of a SMPS.  It allows me to connect the ground lead of my scope probe to the hot primary side ground.  Such power supplies often have a large heat sink on the primary side that is tied to the hot ground.  Using an isolation transformer means I don't have to worry about accidentally touching this.  Both of these remain true even if the safety ground is connected.

[S. Petrukhin]

One example of where an isolation transformer is useful is when I need to work on the primary side of a SMPS.  It allows me to connect the ground lead of my scope probe to the hot primary side ground.  Such power supplies often have a large heat sink on the primary side that is tied to the hot ground.  Using an isolation transformer means I don't have to worry about accidentally touching this.  Both of these remain true even if the safety ground is connected.

If you decide to participate in the life of the high-voltage side, then the separating transformer will protect you by galvanic isolation from the phase. A rubber mat under your feet will do the same. Or rubber gloves. 

But this is not the solution!
1. You can touch the two conductors in the high-voltage part.
2. If you detach from the PE circuit, you will receive interference in the air.
3. By connecting the ground terminal of the oscilloscope, you will receive a PE conductor from oscilloscope and the whole earth jumps with interference.

The most reasonable thing is probably to use a differential probe and be careful when performing manipulations only after power is turned off.
Do not rely on an insulating transformer.

[Siwastaja]

One example of where an isolation transformer is useful is when I need to work on the primary side of a SMPS.  It allows me to connect the ground lead of my scope probe to the hot primary side ground.  Such power supplies often have a large heat sink on the primary side that is tied to the hot ground.  Using an isolation transformer means I don't have to worry about accidentally touching this. 

This must be a troll, must be a troll, must be a troll.... I truly HOPE this is a troll.

Note to those who don't understand why this is a troll: isolation transformer here gives you the false sense of security, while the oscilloscope ground lead re-establishes the galvanic connection to building earth, therefore making the primary side, which andy2000 does not worry about touching, lethally dangerous again.

[Ian.M]

OTOH Andy's live heatsink is now grounded.   Its questionable whether this is safer than a heatsink that is live with half-wave rectified mains (and note that many older RCDs don't meet their AC trip current specs in the presence of DC leakage current due to core saturation), as the DC bus  will  have 325V or more on it, and is capable of delivering plenty of current, and has a fair bit of stored energy, so could easily be lethal.   

Anyone who's ex-service trade, particularly CRT TVs, will be familiar with this setup, and back in the day, would have insisted on the rubber floor mat  as S. Petrukhin suggests, + removing or insulating all unnecessary grounds from the bench area (apart from the CRO which had to be grounded), and even then would treat the circuit board with extreme caution.

Nowadays, there's little excuse, as a Cat II rated differential probe is certainly cheaper than the pine box it will keep you out of!

[bdunham7]

Note to those who don't understand why this is a troll: isolation transformer here gives you the false sense of security, while the oscilloscope ground lead re-establishes the galvanic connection to building earth, therefore making the primary side, which andy2000 does not worry about touching, lethally dangerous again.

I don't think andy2000 is a troll nor is he wrong.  There are legitimate methods of using an isolation transformer and accidental touch contact on live chassis was the main reason they were introduced.  For some reason this generates a lot of flipped wigs over the false sense of security issue.  Mitigating signficant touch voltage hazards (chassis, heat sink, etc) doesn't mean you can poke your sweaty fingers into every nook and cranny and it never has.  It doesn't matter if it is an old radio with a 50% chance of a live chassis, a television with a 100% chance of the chassis being live because it is connected to the negative of a bridge rectifier, or a typical SMPS.  Those devices all have significant voltage hazards whether isolated or not, although if you avoid grounding anything you do mitigage single-point contact voltage hazards. 

For an SMPS, if you use isolation and ground both the negative of filter cap (which is likely to also be the big heat sink) and the case, you mitigate both those potential touch hazards and set things up to easily take measurements of the rest of the primary circuit.  Are there now hazardous ground-referenced voltages on the rest of of the primary circuit?  Yes, there are.  So be careful!

[jonpaul]

Rebonjoyr, been a power electronics eng and consultant since 1970s, medical, avaionics, lighting, etc.

Always used 1:1 isolations trsf on 120 and 240 mains, and variacs in the lab from 100 VA to 25 KVA 3 ph.

Staco had metered isolated Variacs for the lab and Signal (now Belfuse) has excellent DU-..DU-5 isolation trasf 1..5 KVA 120/240 dual pro sec.

Ran equip on sec chassis at earth and scope at earth.

With TEK and Pearson CT, current probes and TEK X10 and X1000 voltage probes, never had any accidents or shocks  due to the mains.

Just my experience


Jon

[Zero999]

Note to those who don't understand why this is a troll: isolation transformer here gives you the false sense of security, while the oscilloscope ground lead re-establishes the galvanic connection to building earth, therefore making the primary side, which andy2000 does not worry about touching, lethally dangerous again.

I don't think andy2000 is a troll nor is he wrong.  There are legitimate methods of using an isolation transformer and accidental touch contact on live chassis was the main reason they were introduced.  For some reason this generates a lot of flipped wigs over the false sense of security issue.  Mitigating signficant touch voltage hazards (chassis, heat sink, etc) doesn't mean you can poke your sweaty fingers into every nook and cranny and it never has.  It doesn't matter if it is an old radio with a 50% chance of a live chassis, a television with a 100% chance of the chassis being live because it is connected to the negative of a bridge rectifier, or a typical SMPS.  Those devices all have significant voltage hazards whether isolated or not, although if you avoid grounding anything you do mitigage single-point contact voltage hazards. 

For an SMPS, if you use isolation and ground both the negative of filter cap (which is likely to also be the big heat sink) and the case, you mitigate both those potential touch hazards and set things up to easily take measurements of the rest of the primary circuit.  Are there now hazardous ground-referenced voltages on the rest of of the primary circuit?  Yes, there are.  So be careful!

You could earth it via a Hall effect leakage current sensor, connected to the breaker, to form an RCD on the secondary side, which would trip, if there's any leakage. It would be very handy to integrate such a device into an isolation transformer box.

[andy2000]

I didn't mean to imply that using an isolation transformer somehow makes working with a live mains powered circuit free from danger!  Safe working practices are as important as ever since there are still lethal voltages involved.  It simply turns a line connected chassis into an isolated chassis, so it's the same as if you working on a device with a power transformer. 

There's nothing inherently dangerous about a floating chassis.  Most consumer electronics sold in the US comes with a non-grounded plug, and therefore has a floating chassis.  Isolation is invariably provided by the transformer in the power supply of the device. 

[Oldtestgear]

I read this thread with interest as my isolation transformer was causing the bench RCD to trip intermittently.  I removed the cover & there was the problem. A RIFA capacitor from live to earth was leaky. It was not cracked or obviously damaged but removing it solved the problem.  Just my observation.


Phil

[tom66]

I spent over 45 years working professionally with all kinds of electronic and electrical equipment. Voltage levels from 5V or less to 40,000V or even more. I worked on high power transmitters and on flea power logic circuits and virtually everything between. I NEVER WORE ANY KIND OR ESD OR GROUND STRAPS. NEVER!


Why not? Because I valued MY LIFE more than some static sensitive IC which cost mere pennies. If you or I touch a point that is hot, that ground (ESD) strap COMPLETES the circuit and current can then flow.

ESD straps have internal resistance, usually 1Mohm or so, which limits the fault current in cases of device voltage up to 1kV. You should not use ESD straps on higher voltage equipment without additional precautions, but those situations are very rare.  The idea that you have 'never' damaged ESD sensitive parts is farcical... sorry... I can point to tens of devices I know were only damaged after handling or numerous images of devices with on-die damage from ESD, plus the precautions the electronics industry take.

[Ian.M]

@Phil,
The RIFA was there for a reason, but as you probably know that with age their casings micro-crack and allow ambient humidity to attack their dielectric which  effectively makes them become incendiary timebombs, waiting to burn your equipment or even your whole workshop!  Obviously it works without it, but is likely to be more vulnerable to mains transients and conducted EMI.   You were lucky it was line to ground that tripped the RCD, rather than line to neutral, for which the first warning is a nasty smell if you are lucky, and smoke and flames if not.

I would therefore strongly recommend removing any other RIFA capacitors present, then replacing them with modern non-RIFA capacitors of similar value and voltage rating, class X for Line to Neutral and class Y for either live wire to Ground/PE.