Discussion of AC mains isolation transformers

[drjimwillie]

 If the center tap of the two coils on the secondary are required to be grounded for the transformer to be wired balanced, then by definition it is impossible to have a secondary floating?    Is that correct? 

 

 I have read this thread from the beginning, and honestly I'm a little bit confused.  Obviously you have strong convictions that the secondary should not be floating.   J.S. Said that a floating secondary is the normal way the transformer is wired. Will you please elaborate on your perspective?

 

J.S. also said that a GFI is not required.  I am going to get a two pole GFI to use with my project.   Is this only required when the transformer is wired balanced?

 

 I know  we operate out on the fringes, to try to get the best out of our systems.   Sometimes even with the best effort it is difficult to weave it all together to have a complete understanding.  

 I am grateful for your help. Thank you 

[Abtr]

55 minutes ago, drjimwillie said:

 If the center tap of the two coils on the secondary are required to be grounded for the transformer to be wired balanced, then by definition it is impossible to have a secondary floating?    Is that correct? 

 

 I have read this thread from the beginning, and honestly I'm a little bit confused.  Obviously you have strong convictions that the secondary should not be floating.   J.S. Said that a floating secondary is the normal way the transformer is wired. Will you please elaborate on your perspective?

 

J.S. also said that a GFI is not required.  I am going to get a two pole GFI to use with my project.   Is this only required when the transformer is wired balanced?

 

 I know  we operate out on the fringes, to try to get the best out of our systems.   Sometimes even with the best effort it is difficult to weave it all together to have a complete understanding.  

 I am grateful for your help. Thank you 

 

I don't think J.S. said that a floating secondary is the normal way an IT is wired. Where did he state that? My Airlink balanced IT came stock with the secondary center tap connected to mains earth. In that case a GFCI is not required at the output for safety (see my reply in 'the true experimental tweak thread').

 

Floating a balanced secondary is not the same as floating an unbalanced secondary because with a floating balanced secondary the power supply ground (and chassis) of your gear is connected to the zero voltage crossing at the center tap with hot and neutral connected to (opposite) phases. This is different from no ground/zero reference at all and it will still reduce/cancel common mode PS noise (albeit maybe not as effective as with a good earth/ground connection to the center tap).

[drjimwillie]

On 11/20/2016 at 2:55 AM, JohnSwenson said:

 

This is a floating secondary, the safety ground does NOT connect to the secondary. The output safety ground is connected to the input safety ground. In addition the metal parts of transformer and the shield are connected to the safety ground.

 

This is the type of connection I was talking about in my post above, the secondary is NOT connected to safety ground. This does not need a GFCI on the output.

 

John S.

This is the post I was referring to. I am sorry I called it normal. 

 

On 11/16/2016 at 5:04 PM, Superdad said:

Thanks for your post John.

 

Indeed a true isolation transformer is quite safe. In fact, before there were GFCIs, hotels used to have "shaver-only" outlets which contained--you guessed it, a small isolation transformer.

 

But it turns out that many (likely most) of the in-a-case with outlets units presently sold as "isolation transformers" in fact are not. Due to NEC (and probably EU) requirements--along with a desire for even greater common-mode noise attenuation--they are most all bonding one leg of the secondary to ground. This of course defeats the safety aspects of a "floating secondary" as John describes above.

 

Again, I am not talking about just running the input ground to the chassis and through to the third pin of the output outlets. I am saying that the TrippLites, the Eatons, and many others are connecting one side of the secondary to ground! That is why there are lots of videos by techs and old radio repair guys showing people how to "re-float" their TrippLite "isolation" trans units--since that is important for safety for those folks.

 

Of course the old stock industrial Topaz units are not doing this (though if I bought one prewired with cord and outlets I would check it).

 

And perhaps the "grounded neutral secondary" pseudo isolation units offered to the public these days would indeed benefit from a GFCI.

 

Here is a page from the current Eaton "Power Suppress" series (you can see the secondary being grounded), and all the TrippLite "Isolator Series" pages state "Secondary Neutral Bonding."

 

In the post two above Alex's post refers to "true isolated transformer " the secondary is not connected to ground. These are why I thought "floating" is "normal"

 

So there are different degrees of  floating?

1.  Input ground is connected to the chassis the shield in the output ground. This is what is  in the two posts above, correct? This is safe?

 

2.  If the output ground is not connected,  then a two pole GFI I is required ?

[Speedskater]

Threads like this have a problem with the lower case word 'ground'.

Because the word has different meanings in different circuit areas, it's unclear just what is connected to which.

 

Another problem is portable plug-in transformers and permanent hard-wired transformers. There are different rules and connection methods for each.

[drjimwillie]

1.   I find this terminal block very hard to work with. I ordered another terminal block, I hope it fits.  If you look at the terminal wire hole, you will notice that the terminal screw pushes down a tab, that stays halfway across the hall and makes it very hard to put wire in. 

 My plan is to put a crimp ring or spade terminals on the existing wires from the transformer.  The wires are so short that I am afraid I will not be able to get my hydraulic crimper on them.  Do you think I can squeeze my lineman  pliers tight enough to make a sound electrical connection?   I made up a power strip with some nice Hubble industrial receptacles and some Canare 4S11  Star quad cable.  Each of the conductors is 14 gauge together they are supposed to be 11 gauge.  This is a hardwire transformer and I don't think I can fit these two wires in that terminal block.   

2.   It was suggested that I run a new ground wire from the input side to the output side.  I have confirmed that there is no wire running through the chassis. The only ground is made with the chassis ground.   If you look in the cavity of the chassis, it does not look like there is much room to sneak a wire past the coil.  Also, isn't the screen blocking the way in the middle  of the chassis, between the two coils?    I was wondering what the two gray wires were on the output side I thought that maybe they ran in to the shields. But as you can see they do not because  they are connected to what looks like a capacitor. 

3.   It was suggested that all of my grounds tied together in a star ground configuration.    Can anyone help me figure out how to engineer that? Or is the whole idea out the window because there are no leads that connect to the shields or the  Faraday screen. 

 Please, I am looking for whatever wisdom and experience you can contribute. 

 

[Abtr]

On ‎15‎-‎5‎-‎2018 at 5:11 PM, drjimwillie said:

This is the post I was referring to. I am sorry I called it normal. 

 

In the post two above Alex's post refers to "true isolated transformer " the secondary is not connected to ground. These are why I thought "floating" is "normal"

 

So there are different degrees of  floating?

1.  Input ground is connected to the chassis the shield in the output ground. This is what is  in the two posts above, correct? This is safe?

 

2.  If the output ground is not connected,  then a two pole GFI I is required ?

 

OK, so I was under the impression that in an unbalanced IT safety ground is as standard connected to both primary and secondary windings (neutral input and output), like in your own MGE. In that configuration you don't need a GFCI on the output because a short to chassis will trip the central GFCI. Apparently (according to @JohnSwenson) floating an unbalanced secondary by disconnecting secondary safety ground, also doesn't require a GFCI at the output, which I admit makes sense. The purpose of connecting the input safety ground to the output safety ground in such a floating unbalanced configuration would just be to improve chassis shielding.

 

Anyway, while a GFCI may not be needed with an unbalanced IT, in a *balanced* configuration (floating or non-floating) the secondary center tap is connected to the chassis of your gear, so a short to chassis will draw a large current from one phase of the secondary which will trip a 2 pole GFCI at the output of the IT before serious damage is done. A balanced IT without a GFCI on the output may trip the central GFCI (non-floating) or high primary current will blow a mains fuse (floating). 

[drjimwillie]

@AbtrI am planning to wire the transformer  balanced, non-floating. 

( The funny thing is, this is how @one and a half suggested I wire the transformer in the first place. But at least now I understand what I'm doing.)

 In the USA we do not typically use a central GFI.  In this instance, if I do not wire a GFI after the IT, is it in issue?  A short would still blow the breaker, correct? 

 Sincerely – thank you for all your help. 

[Abtr]

1 hour ago, drjimwillie said:

@AbtrI am planning to wire the transformer  balanced, non-floating. 

( The funny thing is, this is how @one and a half suggested I wire the transformer in the first place. But at least now I understand what I'm doing.)

 In the USA we do not typically use a central GFI.  In this instance, if I do not wire a GFI after the IT, is it in issue?  A short would still blow the breaker, correct? 

 Sincerely – thank you for all your help. 

 

I recomment a 2 pole RCD/GFCI at the output even with a non-floating configuration. I have seen a non-floating balanced IT blow a 16 Amp fuse instead of tripping the central breaker. A simple 2 pole output GFCI trips at only 15 milli Amp leakage current which is much nicer and safer. You can of course test for yourself what happens, by shorting one of the balanced phases to zero/ground.

  

[drjimwillie]

@AbtrThank you. 

 I don't think I'll try purposely shorting it out. That sounds scary. I would have to take some serious safety precautions. 

 Do you think I will be able to run a ground wire from the input side to the output side so I don't only have to rely on the chassis ground? 

 Also it is recommended that I try to bring all the grounds to the same point,  as a star ground.  I cannot visualize how to engineer that and keep all the wires on the interior of the transformer. 

 Do you think I will be successful and trying to change out the terminal block? The wires to the existing one seem very short.   I don't think I can get my hydraulic crimper onto them.   But I was thinking I could use my lineman's pliers, and squeeze them with a good wood clamp that I have to crimp a ring terminal for the new terminal block. 

 

[Abtr]

9 hours ago, drjimwillie said:

@AbtrThank you. 

I don't think I'll try purposely shorting it out. That sounds scary. I would have to take some serious safety precautions.

Well, ultimately you want your central breaker (or a GFCI on the output) to shut down the circuit in case of a short. Testing if this works should not be particularly dangerous. Just insert a plug with one phase wire and the ground wire stripped at the end and briefly short them..   

 

9 hours ago, drjimwillie said:

Do you think I will be able to run a ground wire from the input side to the output side so I don't only have to rely on the chassis ground? 

Also it is recommended that I try to bring all the grounds to the same point,  as a star ground.  I cannot visualize how to engineer that and keep all the wires on the interior of the transformer.

I wouldn't worry about a star ground inside the transformer. Theoretically it may be best to have the input and output ground wire connect to the same point at the transformer chassis and you can test if that provides any audible advantage by running a ground wire outside the transformer case from input to output, but I would be very surprised if it did.

 

9 hours ago, drjimwillie said:

Do you think I will be successful and trying to change out the terminal block? The wires to the existing one seem very short.   I don't think I can get my hydraulic crimper onto them.   But I was thinking I could use my lineman's pliers, and squeeze them with a good wood clamp that I have to crimp a ring terminal for the new terminal block. 

 

Personally I would leave the connection blocks as they are. Nothing wrong with screw connectors..    

[drjimwillie]

I have two issues with the screw connectors. Although I may have solved one.

1.  If you look at the picture, there is that metal tab that hangs down across the hole for the wire.   It made it very difficult to insert a wire.  This is why you can see  that  the previous owner used these teeny aluminum wires as jumpers.  I have to believe that a teeny wire would impede the electron flow and mess with the dynamics.  I tried to insert a 14 gage jumper, and you can see that I bent the heck out of it trying to get it in the hole.  Now that I have pulled out the terminal block I should be able to  persuade a wire to fit. 

2.   I made up cables which use a star quad with (2) 14 gauge stranded conductors.  The only way that I think I could get that to fit in the hole would be to solder both conductors solid, and then offset them, and solder them together so that only the end of one conductor goes into the connector.    I don't know if that's a good idea ( electrically ) and I'm not good at soldering.   What do you think?

 

 Also, what do you think that thing is that looks like a capacitor on the output?

[Speedskater]

With so many different setups and situations discussed in this thread, we can't make generalizations. Especially generalizations about Safety Grounds (EGC), connections to Planet Earth (GEC), GFCI's, AFCI's, NEC and UR rules.

[Abtr]

4 hours ago, drjimwillie said:

I have two issues with the screw connectors. Although I may have solved one.

1.  If you look at the picture, there is that metal tab that hangs down across the hole for the wire.   It made it very difficult to insert a wire.  This is why you can see  that  the previous owner used these teeny aluminum wires as jumpers.  I have to believe that a teeny wire would impede the electron flow and mess with the dynamics.  I tried to insert a 14 gage jumper, and you can see that I bent the heck out of it trying to get it in the hole.  Now that I have pulled out the terminal block I should be able to  persuade a wire to fit. 

2.   I made up cables which use a star quad with (2) 14 gauge stranded conductors.  The only way that I think I could get that to fit in the hole would be to solder both conductors solid, and then offset them, and solder them together so that only the end of one conductor goes into the connector.    I don't know if that's a good idea ( electrically ) and I'm not good at soldering.   What do you think?

 

 Also, what do you think that thing is that looks like a capacitor on the output?

 

Maybe you can get larger screw terminals (20mm?) and/or use more screw terminals to connect multiple wires at one side and bridge (jump) them on the other side (with stranded copper wire) to a single input/output screw connector.

 

And that thing is indeed a capacitor (2uF 250V). I don't know what it's doing in your isolation transformer. Can you see how it is connected? 

[look&listen]

19 hours ago, drjimwillie said:

Do you think I will be successful and trying to change out the terminal block? The wires to the existing one seem very short.   I don't think I can get my hydraulic crimper onto them.   But I was thinking I could use my lineman's pliers, and squeeze them with a good wood clamp that I have to crimp a ring terminal for the new terminal block.

Coil wires probably have enamel or other insulating covering. Could be big problem to re-terminate. Even if solve that, many things possible go bad, maybe not repairable. I think solve existing terminal block problem best course.

Original t-block designed for size of IT. Larger wires not needed. If must have bigger wires, maybe these help:

https://www.amazon.com/gp/product/B06XP8J311/ref=oh_aui_detailpage_o07_s00?ie=UTF8&psc=1

Never see capacitors in Topaz, etc ITs. must be custom user add. Remove for parts box.

[drjimwillie]

IOk. Now that I have  pulled the existing terminal block out I can probably fit a 12 gauge wire in as a jumper.  I am I correct in thinking that that teeny aluminum wire as a jumper was an issue?  You can see it if you page up to some of my original photos.

 

 Now my issue is how to work with the star quad cable that has (2) 14 gauge conductors.  What do you think of my idea to offset the two conductors and solder them together so that all I have to do is put the tip of one under the screw terminal?

 

 The capacitor is wired like a jumper from X1 to X4.  I will remove it.   The capacitor is tucked in behind the plate that holds the terminal block.  I only found it because I pulled terminal block off so I could see what was keeping the wires from going in the holes.  This is an MGE Ultra isolation transformer 91097-31T  750VA, so probably built by Topaz. 

[Abtr]

1 hour ago, drjimwillie said:

IOk. Now that I have  pulled the existing terminal block out I can probably fit a 12 gauge wire in as a jumper.  I am I correct in thinking that that teeny aluminum wire as a jumper was an issue?  You can see it if you page up to some of my original photos.

I wouldn't know, it might be a problem at continuous max power.. Is it actually aluminum wire; not some other metal?

 

Quote

 Now my issue is how to work with the star quad cable that has (2) 14 gauge conductors.  What do you think of my idea to offset the two conductors and solder them together so that all I have to do is put the tip of one under the screw terminal?

If done well, I don't see a problem with that, although a tinned tip has a tendency to work itself loose from a screw connector. It's best to leave a bare stranded copper tip for the connection.

 

Quote

 The capacitor is wired like a jumper from X1 to X4.  I will remove it.   The capacitor is tucked in behind the plate that holds the terminal block.  I only found it because I pulled terminal block off so I could see what was keeping the wires from going in the holes.  This is an MGE Ultra isolation transformer 91097-31T  750VA, so probably built by Topaz. 

Really? Then the capacitor was connected as a short or impedance or a filter across the AC output. I'm not sure what that would accomplish..

[Speedskater]

Thinking about it further, for those that have a North American type AC power system. If you don't know the difference between a EGC & a GEC circuit and the reasons why and how those differences are important, then you shouldn't be wiring isolation transformers.

[diecaster]

LOL! You don't need to know jack about EGC and GEC circuits to wire an isolation transformer as long as you get the directions from a proper source.

[Abtr]

18 hours ago, Speedskater said:

Thinking about it further, for those that have a North American type AC power system. If you don't know the difference between a EGC & a GEC circuit and the reasons why and how those differences are important, then you shouldn't be wiring isolation transformers.

I think both EGC and GEC can be referred to as safety ground conductors/connectors and wiring them in an isolation transformer is pretty straightforward with balanced and unbalanced ITs and any single phase type mains power system.

[Speedskater]

On 5/18/2018 at 1:49 PM, Abtr said:

I think both EGC and GEC can be referred to as safety ground conductors/connectors and wiring them in an isolation transformer is pretty straightforward with balanced and unbalanced ITs and any single phase type mains power system.

No the EGC  and GEC have different tasks.

Only one of them is the Safety Ground.

The fact that there is continuity between them doesn't make the the same.

[Speedskater]

On 5/17/2018 at 8:26 PM, diecaster said:

LOL! You don't need to know jack about EGC and GEC circuits to wire an isolation transformer as long as you get the directions from a proper source.

Reading the 1500 posts on this thread alone, many won't get directions from a proper source. Or having those directions, understand just what they mean.

[Abtr]

25 minutes ago, Speedskater said:

No the EGC  and GEC have different tasks.

Only one of them is the Safety Ground.

The fact that there is continuity between them doesn't make the the same.

I don't see that. Can you please elaborate? 

[drjimwillie]

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1. For 120V -> 120V balanced non-floating output, jump H2 and H3 (and remove the jumpers between H1A and H3 and between H2 and H4), jump X2 and X3 and remove the jumper at the top between G and X4 at the output, connect both the green-yellow wire and the solid green wire to G at the output and jump G to X2/X3 (which is now the secondary center tap). This is safe because your gear's chassis is connected to mains earth and a short to chassis will trip your central GFCI and/or blow a mains fuse.

2.  Now, for 120V -> 120V balanced *floating* output, connect everything as was described above for a 120V -> 120V balanced non-floating output, and simply disconnect the yellow-green earth wire from G at the output but keep the solid green wire connected to G. This is *unsafe* because your gear's chassis is not connected to mains earth and a short to chassis will therefore not trip your central GFCI or blow a mains fuse! However, you can make this safe by installing a 2 pole RCD/GFCI at the output of the IT. Your gear’s chassis is connected to the zero voltage center tap (X2/X3) of the balanced secondary and a short to chassis will trip the 2 pole GFCI at the output of the IT.

 

Hi @Abtr - the first two images are original, before I touched anything.  The last is with the output opened.  We have established that the two grey wires go to capacitor attached to the back of the panel, that I am going to remove.  At the input a green wire goes into the interior and a green & yellow wire attach to the chassis.  At the output I do not have the green wire coming from the interior, only the green & yellow to the chassis. 

 

How can I wire as suggested above if there is no green wire from the interior??

 

when I wired with the jumper X2 + X3 + G, I measure: (#1 above)

X1 - X2 = 62.7V

X4 - X2 = 62.7V

X1 - X4 = 126V

X1 - G   = 62.7V

X4 - G   = 62.7V

 

With the jumper from X2 + X3, but the jumper removed from X3 + G, I measure:

X1 - X2 = 62.7V

X4 - X2 = 62.7V

X1 - X4 = 126V

X1 - G   = 32.3V

X4 - G   = 86.0V

 

All of my receptacles are Hubbell 5362i industrial grade.  To safely try #2 above, floating the ground, I purchased a  Hubbell 5362i 2 pole GFI ( I do not want any weak links in the power chain), it was expensive.  What do I do if I only have the chassis ground at the output?  I will have to look at the terminal block to see what the chassis ground is connected to.

 

I have all of my parts and I plan to wire it up tonight.  I got some of those Wago connectors.   I will let you know how it goes.

Thank you

 

 

 

[Abtr]

1 hour ago, drjimwillie said:

...

 

Hi @Abtr - the first two images are original, before I touched anything.  The last is with the output opened.  We have established that the two grey wires go to capacitor attached to the back of the panel, that I am going to remove.  At the input a green wire goes into the interior and a green & yellow wire attach to the chassis.  At the output I do not have the green wire coming from the interior, only the green & yellow to the chassis. 

 

How can I wire as suggested above if there is no green wire from the interior??

The green wire with black patches that goes into the interior at the input is the ground connection to the transformer shield and it isn't present a the output. The solid green wire I referred to is the green safety ground wire of the AC output cable, which is already connected to G.

 

Quote

when I wired with the jumper X2 + X3 + G, I measure: (#1 above)

X1 - X2 = 62.7V

X4 - X2 = 62.7V

X1 - X4 = 126V

X1 - G   = 62.7V

X4 - G   = 62.7V

 

With the jumper from X2 + X3, but the jumper removed from X3 + G, I measure:

X1 - X2 = 62.7V

X4 - X2 = 62.7V

X1 - X4 = 126V

X1 - G   = 32.3V

X4 - G   = 86.0V

 

All of my receptacles are Hubbell 5362i industrial grade.  To safely try #2 above, floating the ground, I purchased a  Hubbell 5362i 2 pole GFI ( I do not want any weak links in the power chain), it was expensive.  What do I do if I only have the chassis ground at the output?  I will have to look at the terminal block to see what the chassis ground is connected to.

 

I have all of my parts and I plan to wire it up tonight.  I got some of those Wago connectors.   I will let you know how it goes.

Thank you

To 'float' the secondary just remove the green-yellow wire at the output side. Don't remove the jumper from X2/X3 (center tap) to G (now only chassis ground). And remove the capacitor!

[drjimwillie]

Thank you so much