Isolation transformer

[zvonex66x]

Hi all!!! Like every lab need to have isolation transformer, i'm finally get one old... Now.. everything is ok with this transformer, but if I try to measure voltage between ac main ground and secondary 1 or 2 of my transformer, DMM says about 100V.
Is this only induction in transformer, becouse if i hook up 5w resistor on this point of measureing, i got 0V.
I think evrything is ok but still, better to ask ppl who knows more....
Thanks to all...

[Jeroen3]

When unsure, you draw up a simple equivalent circuit. 

You are measuring leakage current.

[TheUnnamedNewbie]

Simple way to be sure is to just test for continuity (obviously when the transformer is unplugged).

[Frankentronics]

I picked up this thread as it is relevant to a project I happen to be doing.

When unsure, you draw up a simple equivalent circuit. 

I drew up a circuit the way I understand it.



So, assuming my drawing is correct, it is normal to measure 100V between one of the transformer's leads and earth ground?

[zvonex66x]

Yes, it's the same in my case, but I think that is only some induction in the transformer, becouse if you put the resistor parallel with your voltmeter, you will get 0V.

[Jeroen3]

@Adinol, your drawing has two live wires. This is very uncommon in europe.
In europe we have only 1 live, and 1 neutral bonded to earth at the distribution point.

[jeroen79]

Jeroen3's circuit with a high and low impedance voltmeters added:
http://tinyurl.com/yb999wna
Note the difference in measured voltage?

[Richard Crowley]

So, assuming my drawing is correct,

Your drawing is NOT correct, assuming you are talking about a common 115V domestic branch circuit.
If you were talking about a North American 220V branch circuit, then your drawing would be correct.
Please specify voltages so that we will know what you are talking about?

it is normal to measure 100V between one of the transformer's leads and earth ground?

No matter WHAT is on the primary (input, left) side of the iso transformer, it is NOT NORMAL to measure ANY voltage between one of the output leads and earth ground.  There may be a TINY amount of capacitively coupled voltage from the primary to the secondary windings, but with any kind of load, it should read essentially ZERO volts.  Else it wouldn't be worthy of the name "isolation".

[Brumby]

it is normal to measure 100V between one of the transformer's leads and earth ground?

No matter WHAT is on the primary (input, left) side of the iso transformer, it is NOT NORMAL to measure ANY voltage between one of the output leads and earth ground.  There may be a TINY amount of capacitively coupled voltage from the primary to the secondary windings, but with any kind of load, it should read essentially ZERO volts.  Else it wouldn't be worthy of the name "isolation".

That response is a bit confusing.

"it is NOT NORMAL to measure ANY voltage between one of the output leads and earth ground"
Actually is IS quite normal ... and you have given the reason why:
"capacitively coupled voltage from the primary to the secondary windings"

The magnitude of this voltage will depend on the load impedance of the measuring device - but with a 10M input impedance of a typical DMM, 100V doesn't seem unreasonable, especially if the voltage on the transformer is circa 220V.

"There may be a TINY amount of ..."  Tiny is not really relevant to the voltage - but it IS relevant to the current capability, which you correctly infer with this statement: "but with any kind of load, it should read essentially ZERO volts"

The OP has shown this to be the case:

Is this only induction in transformer, becouse if i hook up 5w resistor on this point of measureing, i got 0V.

(Not sure of the actual resistance used - but most of the 5W resistors I have are sub 1K and that would be quite successful in pulling the voltage down.)

[Richard Crowley]

@Brumby, thanks for clarifying what I was trying to say.  The diagram that @Adinol posted remains confusing.

[Brumby]

I looked at that diagram and immediately interpreted the centre earthed utility transformer as the common (as I understand it) split phase pole transformer seen around the US.

[Richard Crowley]

Yes, that is what a 220V branch circuit looks like, but 220V branch circuits are not provided anywhere in typical domestic buildings except for large-power loads like electric water heaters, space heaters, etc. And those are typically hard-wired.  Clothes dryers, electric stoves (cookers/ovens) and some window-mount air conditioner units use special plugs for rare 220V circuits/receptacles.  But finding (and using) a 220V circuit in any kind of electronic bench setting seems pretty remote, at least here in North America.

[Frankentronics]

I looked back at my drawing and also realized there was a mistake. I made a new drawing.



I didn't change the voltage on the DMM but I guess in N America that would not read 100V.

So, I am still not clear. What would be a normal voltage at that spot?

Also, should devices that are plugged into the isolation transformer also be grounded through the mains earth ground?

[helius]

So, I am still not clear. What would be a normal voltage at that spot?

It is really a question of ghost voltage. There is no stiff voltage source at the points tested that could supply any kind of shock current. But with only a high impedance DMM in the circuit, the mutual capacitance of the transformer windings can pass a small current that appears as a fairly high voltage. Some DMMs have "Lo-Z" modes to prevent being fooled by this phenomenon.

Also, should devices that are plugged into the isolation transformer also be grounded through the mains earth ground?

In general, equipment that is sealed in a metal case is grounded in case of a wiring fault. If there is a wiring fault inside the case (line voltage wires break off, fray, or burn up and make the chassis live), the fault current from live to ground through the case will blow a fuse or trip a circuit breaker. This protects people from being shocked when they touch a live chassis in one hand and a grounded object in another.

For diagnosis and repair on an electronics bench with an isolation transformer, it is NOT recommended to ground the device under test, because the reason for isolation it is to protect the technician who is working inside the equipment with line (or higher) voltages. The technician is actively poking into the DUT: they can easily see wiring faults, and have been trained to use a one-hand technique when working with dangerous voltages. Grounding the equipment doesn't do anything for wiring faults with an isolation transformer in the way, as there is no fault path to trigger a circuit breaker or GFCI. It can cause unexpected shocks and burn up equipment when grounded instruments like oscilloscopes are connected to a grounded DUT (see "How not to blow up your scope" EEVBlog video).

[Richard Crowley]

Unless your transformer is defective (which you can measure with your ohm-meter), you are most likely seeing the very loosely capacitively coupled voltage from the primary side to the secondary side.  If you were to connect any kind of load there (even a 10K resistor across your volt meter) you would likely see zero (or close to zero) volts. 

Your reading is an artifact of a meter with a very high input impedance (which in most cases for electronics is a GOOD thing). I believe there are some meters (and especially those made for electricians) which have a lower input impedance specifically to prevent this kind of rather misleading measurement.

You can measure the resistance between all four leads of the transformer to see if there is actually any kind of fault resistance path.  The resistance from end-to-end of the windings should be pretty low (10s, 100s, perhaps thousands of ohms).  While the resistance between either of the primary leads and either of the secondary leads should read in the megohms (i.e. "open circuit" for all practical purposes).  Of course, you would not attempt to make this measurement on a transformer which had ANY connection to the power utility mains.

[zvonex66x]

Hi... Thx all for this answers...
All is measured, all it normal like resistance..sorry for mistake when postinh and not give a resistance of 5w resistor. I don't remember it but sure is under 1ohm, 0.22 or 0.33 i think..
I'm in Europe , and you can sketch circuit of main in many ways but the fact is: you have:
1: live wire(220v)
2: neutral wire
3: ground wire
... and in the distribution point(trafostation) the neutral wire and the ground wire is connected together...
Isolation transformer.... hmmm it's notning then just some induction of transformer or something like this... like I wrote, if I put the resistor parallel with DMM, readings are 0V... now, i have go a bit deeper in this so I connect regular 5mm led at this measured point (i think it's about 3V, 25mA)... and ofcourse I got.... nothing... so this measure of 100V is just DMM readings, and I'm sure that is mine isolation transformer good.
Once again, thanks everyone for help..

[Brumby]

Your reading is an artifact of a meter with a very high input impedance (which in most cases for electronics is a GOOD thing). I believe there are some meters (and especially those made for electricians) which have a lower input impedance specifically to prevent this kind of rather misleading measurement.

Dave's BM235 has exactly this - a LoZ voltage range - for exactly this sort of issue.

The fact that there are meters currently on the market with it should tell you that this is a "not uncommon" issue.  Even if you don't have a meter with this feature - understanding what is happening will serve you well in your future endeavours.

Just remember though ... you need to treat these situations as if there is a real voltage with real power behind it, until you can eliminate the possibility - because one day you are going to meet up with such a situation and a wrong assumption could turn your day into a really bad one.

[oldway]

From experience, I have already found many times that insulation faults that trigger the differential switch (so leakage current to ground is > 30 mA) can't be measured with an ohmmeter, voltage is too low... An isolation meter must be used to detect it.

[retrolefty]

To decrease this capacitance coupling the transformers used to minimize this use an internal metal shield placed between primary and secondary windings and bring out a single lead wire out of this shield to be grounded. Hi-end test gear often uses such and show in the schematic drawing of the transformer terminals.

[Frankentronics]

To decrease this capacitance coupling the transformers used to minimize this use an internal metal shield placed between primary and secondary windings and bring out a single lead wire out of this shield to be grounded. Hi-end test gear often uses such and show in the schematic drawing of the transformer terminals.

Good to know. Thanks.

Is there a technical term for this, that I should include in my search terms, to find it?

I found this 1:1 isolation transformer on Jameco's site, for $50.

https://www.jameco.com/z/ITR300-Power-Transformer-Isolation-120VAC-120VAC-300Va_181315.html

The datasheet https://www.jameco.com/Jameco/Products/ProdDS/181315.pdf says "Electrostatic shield internally" but does not show a schematic. Could this internal shield possibly be what you are talking about?

Thanks...

[Richard Crowley]

Yes, it says "Electrostatic shield internally" but it doesn't provide any evidence of it.
The "data sheet" shows only a 2-prong (ungrounded) input cord.
Hard to imagine how an electrostatic shield works with no reference to ground?

[retrolefty]

To decrease this capacitance coupling the transformers used to minimize this use an internal metal shield placed between primary and secondary windings and bring out a single lead wire out of this shield to be grounded. Hi-end test gear often uses such and show in the schematic drawing of the transformer terminals.

Good to know. Thanks.

Is there a technical term for this, that I should include in my search terms, to find it?

I found this 1:1 isolation transformer on Jameco's site, for $50.

https://www.jameco.com/z/ITR300-Power-Transformer-Isolation-120VAC-120VAC-300Va_181315.html

The datasheet https://www.jameco.com/Jameco/Products/ProdDS/181315.pdf says "Electrostatic shield internally" but does not show a schematic. Could this internal shield possibly be what you are talking about?

Thanks...

 Could be if the internal shield is attached electrically to the transformer's metal frame and then you ground the transformer via it's mounting hardware.

[Frankentronics]

Or perhaps they just connected that internal shield to the input's neutral wire. Obviously that would violate the electrical code in North America, but the question is, if it is in fact done that way, could a simple mod fix it?

[retrolefty]

Removing the end covers of the transformer should expose a interwinding shield or not and if so it's attachment method to the frame or the end covers.

[Frankentronics]

I called Jameco. They will contact the manufacturer and request a new datasheet that shows how the internal shield is connected. They hope to have the new datasheet within 2 days.

If this can be modified so that the internal shield is properly grounded I guess it's a decent transformer for 50 bucks, as I see that other isolation transformers are considerably more expensive.

But let's see if that fresh datasheet ever arrives.

[Richard Crowley]

If it really has a shield, it would be trivial to replace that 2-wire power cord with a proper 3-wire, grounded power cord with the green wire safety ground ("protective earth") connected to the transformer case.

[Frankentronics]

I never got that reply from Jameco, asking them to provide a schematic from the manufacturer that would show how the internal static shield is connected. But I did end up taking a $50 risk and purchased one unit. Here's what I found.

The power cord is not polarized.

The inside view does not offer any clue as to how a static shield might be wired. There's no third wire coming out of the transformer, so I think there is no static shield. And all the connections are only "glued together" with solder - no firm mechanical connections.





Further, without a load I measure 119V on the output, while the input is getting 121V.

My understanding is that all transformers have losses and for that reason the 1:1 transformers should not actually have actual 1:1 ratios. As I understand it I should be getting more than 121V on the output, without a load, so that I would get a 1:1 ratio once I plug in a load. Is that correct?

Since this transformer seems to be stepping down, slightly, should I just rewire it backwards?

Or should I just return this item and search for a better one?

Thanks...

[schmitt trigger]

A little hard to tell from the photo, but it appears that there is a red-color external winding, and a green-color internal winding. Am I seeing it correctly?
If so, and based only on the provided photos, it doesn't appear to have any shielding.

The shielding would consist of a single turn ** copper foil with kapton or other insulating tape between the two windings, with a grounding wire attached to it.

** Not a full turn, and definitely with its ends not touching. At all. Otherwise it would create a single-turn short.

[Richard Crowley]

I never got that reply from Jameco, asking them to provide a schematic from the manufacturer that would show how the internal static shield is connected.

Yeah, that was a long-shot.  I wasn't expecting that you would get anything.  How would they know?  Why would they care?
IF it does have a shield, and it doesn't appear to have a separate wire for it, then they probably connected it to the transformer core. (Which is a good place, IMHO)

Further, without a load I measure 119V on the output, while the input is getting 121V.

My understanding is that all transformers have losses and for that reason the 1:1 transformers should not actually have actual 1:1 ratios. As I understand it I should be getting more than 121V on the output, without a load, so that I would get a 1:1 ratio once I plug in a load. Is that correct?

Since this transformer seems to be stepping down, slightly, should I just rewire it backwards?

If the transformer is really wound 1:1 then "rewiring it backwards" would yield exactly the same results.
What exactly is the issue with a 2V drop?  What kind of drop are you seeing with a load?  1A?  5A?  10A?

Or should I just return this item and search for a better one?

What does "better" mean in this context?  Remind us what is the problem here that you are trying to resolve?

If it were me, I would put a proper 3-wire supply mains cord on it.  I would connect the green-wire PE (protective-earth) to the transformer chassis.
If the transformer has a shield, then it will be connected to the core/frame and the input source PE.

Your internal photos show a transformer that appears to be reasonably good for a hobby-level bench isolation application.
My only beef with it would be the lack of PE on the input/source wiring.  And that is trivial to remedy.

[Frankentronics]

I took a closer look. Here's what I can see without taking it apart beyond repair.

Both coils use red wire. What I think might show up as sort of green on the photos is what I think the manufacturer calls a "shield".

I see that the secondary coil is wound close to the core, then there is a "shield" made of the same material as the outside wrapping, which looks like some electrical tape. Then, after this "shield" there is the primary coil.

I think it doesn't have a real static shield, as advertised. My understanding is that a static shield will send the capacitance between the two coils to ground. But I don't know what that means in practical terms. What difference in performance will I see with or without a static shield?

While reading through numerous internet articles about isolation transformers and DIY isolation transformers I came upon some information that DIY isolation transformers are not quite the same as factory made ones. One reason being that there is loss in the coils and that manufacturers wind a few extra turns on the secondary coils to compensate for these internal losses. Thus, the output without any load should be slightly greater than the input voltage.

I quickly tested it under load, to answer the question.  I plugged in an old tube radio through it. The line voltage is 121.7V and under this load the transformer outputs 118.2V. I don't have a quick way to measure how much current my tube radio draws, but would 2A be a good estimate?

I only need this for hobby purposes and I guess it's OK for $50 as I've seen other isolation transformers tend to be considerably more expensive. My main use will be to have an isolation transformer for my variac. The variac will be used to test transformers, power supplies and audio amps.

[Richard Crowley]

Insulation material (like tape, cardboard, etc.) cannot form any kind of electrostatic shield so the preponderance of the evidence suggests that this transformer does not have an electrostatic shield.

The only significant benefit from having an electrostatic shield in a hobby-grade bench isolation transformer is to reduce the capacitive coupling between primary and secondary which caused your original misunderstanding.  However that was only a measurement/perception error and not any kind of practical problem in common and ordinary use of a power isolation transformer. 

Now, if you were designing a permanent power supply for some kind of sensitive gear and/or for use in an unusually noisy power region, then electrostatic shielding might be an important consideration.  But my perception is that it doesn't seem very important for your application.  The difference in "performance" may be that your device under test would be subject to slightly higher power-mains induced noise.  Although going through both the iso transformer AND another chunk of heavy iron (the variac) this doesn't seem like any kind of real-world significant issue.

The voltage variations don't seem any greater in magnitude than normal utility voltage variations that we all see from hour to hour, or daily or weekly, etc.  And if exact mains voltage is important to you, then you have a readily available solution by simply tweaking the knob of the variac.

[Frankentronics]

Thank you, Richard, for taking the time to explain.

I'll keep this transformer and when I have time I'll just make a new housing with an ON/OFF switch and a panel meter. Possibly put it in the same housing with the variac and have a separate output option bypassing the variac. Might as well fit it with a grounded power cord. I have very limited bench space so a tower might be my best option.

Thanks...