Swapping transformer I/O...

[John Larkin]

On Sun, 2 Jul 2023 18:43:01 -0700 (PDT), Phil Allison
<pallison49@gmail.com> wrote:

John Larkin wrote:
---------------------------


** The maker tells you which is which because they have adjusted the windings ( number of turns and wire gauges) accordingly.

2-winding power transformers are usually designed for equal copper
loss on both windings. So are thermally symmetric.
Sometimes some windings are insulated better than others.

What other catches might there be?

** As I mentioned, which winding carries the I mag plus the turns ratio have been adjusted by makers to compensate for the full load voltage drop. Reverse the makers intended roles and those built in compensations operate in the opposite directions.

That\'s easy to think through. Power transformer models are simple and
easy to reverse.

** But real transformers are not so simple.

Given a Spice model, you can run it in either direction, with
resistive or diode loads or whatever, and Spice will tell you
everything you need to know. The only gotcha it misses is core
saturation.

** That \"gotcha\" is the biggie here.

With pure AC, just don\'t run it much above its rated voltage.
Whichever winding you care to call the secondary will be affected by a
net DC load current.

Basically, if you buy two equal-power-rated transformers, one sold as
120:240 and one sold as 240:120, they are the same transformer.

You may have to seriously de-rate the transformer in order to use it that way,
mains frequency transformers under 100VA are the most affected.

A 100 VA transformer is happy moving 100 VA in either direction.

** JL has such simple faith in overly simple models.

I wonder why ?

If you are concerned about reversing a transformer, just try it.


** LOL - of *course* I have and that is why I know about the pitfalls.

I haven\'t heard any real pitfalls so far. What problems did you have?

 

[Phil Allison]

John Larkin wrote:
----------------------------

Basically, if you buy two equal-power-rated transformers, one sold as
120:240 and one sold as 240:120, they are the same transformer.

** But not exactly - see below.

You may have to seriously de-rate the transformer in order to use it that way,
mains frequency transformers under 100VA are the most affected.

A 100 VA transformer is happy moving 100 VA in either direction.

** JL has such simple faith in overly simple models.

I wonder why ?

If you are concerned about reversing a transformer, just try it.


** LOL - of *course* I have and that is why I know about the pitfalls.

I haven\'t heard any real pitfalls so far. What problems did you have?

** Same as everyone else\'s !!

When roles are reversed, the previous 240VAC supply winding supplies less under load.
The difference is about twice the regulation percentage, so 20 to 30% less in some cases.
Imag current ( previously easily supplied by the mains ) goes up by the turns ratio, up to maybe 40 times in a low voltage winding.
Applying more than rated voltage to the secondary in order to to fix this results in excessive current and overheating the transformer.

This all derives from how transformer makers engineer * real * transformers and rely on specifying which is the primary etc so all specs are met.



....... Phil

 

[John Larkin]

On Sun, 2 Jul 2023 22:47:47 -0700 (PDT), Phil Allison
<pallison49@gmail.com> wrote:

John Larkin wrote:
----------------------------

Basically, if you buy two equal-power-rated transformers, one sold as
120:240 and one sold as 240:120, they are the same transformer.

** But not exactly - see below.

You may have to seriously de-rate the transformer in order to use it that way,
mains frequency transformers under 100VA are the most affected.

A 100 VA transformer is happy moving 100 VA in either direction.

** JL has such simple faith in overly simple models.

I wonder why ?

If you are concerned about reversing a transformer, just try it.


** LOL - of *course* I have and that is why I know about the pitfalls.

I haven\'t heard any real pitfalls so far. What problems did you have?

** Same as everyone else\'s !!

When roles are reversed, the previous 240VAC supply winding supplies less under load.
The difference is about twice the regulation percentage, so 20 to 30% less in some cases.
Imag current ( previously easily supplied by the mains ) goes up by the turns ratio, up to maybe 40 times in a low voltage winding.

Sure. That\'s obvious from the transformer model. A given amount of
core excitation needs N*I, whichever winding makes it.

>Applying more than rated voltage to the secondary in order to to fix this results in excessive current and overheating the transformer.

Don\'t do that!

This all derives from how transformer makers engineer * real * transformers and rely on specifying which is the primary etc so all specs are met.

That\'s the part I don\'t see. \"Primary\" is an application decision, or
a data sheet convenience.

 

[Phil Allison]

John Larkin wrote:
----------------------------

Basically, if you buy two equal-power-rated transformers, one sold as
120:240 and one sold as 240:120, they are the same transformer.

** But not exactly - see below.

You may have to seriously de-rate the transformer in order to use it that way,
mains frequency transformers under 100VA are the most affected.

A 100 VA transformer is happy moving 100 VA in either direction.

** JL has such simple faith in overly simple models.

I wonder why ?

If you are concerned about reversing a transformer, just try it.


** LOL - of *course* I have and that is why I know about the pitfalls.

I haven\'t heard any real pitfalls so far. What problems did you have?

** Same as everyone else\'s !!

When roles are reversed, the previous 240VAC supply winding supplies less under load.
The difference is about twice the regulation percentage, so 20 to 30% less in some cases.
Imag current ( previously easily supplied by the mains ) goes up by the turns ratio, up to maybe 40 times in a low voltage winding.

Applying more than rated voltage to the secondary in order to to fix this results in excessive current and overheating the transformer.

This all derives from how transformer makers engineer * real * transformers and rely on specifying which is the primary etc so all specs are met.


That\'s the part I don\'t see. \"Primary\" is an application decision, or
a data sheet convenience.

** It is far more than that, but you will never admit it.


........ Phil

 

[John Larkin]

On Mon, 3 Jul 2023 03:45:00 -0700 (PDT), Phil Allison
<pallison49@gmail.com> wrote:

John Larkin wrote:
----------------------------

Basically, if you buy two equal-power-rated transformers, one sold as
120:240 and one sold as 240:120, they are the same transformer.

** But not exactly - see below.

You may have to seriously de-rate the transformer in order to use it that way,
mains frequency transformers under 100VA are the most affected.

A 100 VA transformer is happy moving 100 VA in either direction.

** JL has such simple faith in overly simple models.

I wonder why ?

If you are concerned about reversing a transformer, just try it.


** LOL - of *course* I have and that is why I know about the pitfalls.

I haven\'t heard any real pitfalls so far. What problems did you have?

** Same as everyone else\'s !!

When roles are reversed, the previous 240VAC supply winding supplies less under load.
The difference is about twice the regulation percentage, so 20 to 30% less in some cases.
Imag current ( previously easily supplied by the mains ) goes up by the turns ratio, up to maybe 40 times in a low voltage winding.

Applying more than rated voltage to the secondary in order to to fix this results in excessive current and overheating the transformer.

This all derives from how transformer makers engineer * real * transformers and rely on specifying which is the primary etc so all specs are met.


That\'s the part I don\'t see. \"Primary\" is an application decision, or
a data sheet convenience.

** It is far more than that, but you will never admit it.


....... Phil

The word \"primary\" on a data sheet doesn\'t stop me from using a
transformer in either direction. Like the Coilcraft PL300-100L right
now. Try and stop me!

 

[Anthony William Sloman]

On Sunday, July 2, 2023 at 9:39:31 PM UTC+10, John Larkin wrote:

On Sun, 2 Jul 2023 04:19:47 -0700 (PDT), Phil Allison
palli...@gmail.com> wrote:

John Larkin wrote:
------------------------------

Question is whether they are symmetrical from either side (primary/secondary) in efficiency and load capacity.
I will be driving it from both sides, but not at the same time..

Sure, a transformer works either way.


** But makers nominate the primary and secondary for good reasons.
The primary supplies the magnetising current, which can be significant, plus is wound to allow for voltage drop at the secondary under full load. If the roles are swapped, there are several changes that can catch the unwary.


Any winding can magnetize the core. The primary is the one you decide
to put power into and the secondary is the one you elect to take power
out of.

** The maker tells you which is which because they have adjusted the windings ( number of turns and wire gauges) accordingly.

2-winding power transformers are usually designed for equal copper
loss on both windings. So are thermally symmetric.
Sometimes some windings are insulated better than others.

What other catches might there be?

** As I mentioned, which winding carries the I mag plus the turns ratio have been adjusted by makers to compensate for the full load voltage drop. Reverse the makers intended roles and those built in compensations operate in the opposite directions.

That\'s easy to think through. Power transformer models are simple and
easy to reverse.

** But real transformers are not so simple.

The Spice model of a transformer is simple: each winding has ESR, there\'s magnetizing inductance, and there\'s leakage inductance. Ignore capacitances at line frequencies.

The Spice model of a transformer doesn\'t have leakage inductance. It has primary inductance L1, secondary inductance L2 and mutual inductance M, which happens to be be the product of the coupling coefficient k and the root of the product of L1 and L2. The coupling coefficient k is never greater than one and it\'s rarely less than 0.99 for transformers wound on cores.

Picking off the flux going through L1 that doesn\'t go through L2 (or the flux going through L2 that doesn\'t go through L1) and calling it leakage inductance isn\'t a helpful way of looking at what\'s going on, though some antiquated textbooks used to use it.

Ignoring stray capacitance isn\'t a good idea either. There may not be much current flowing along that path, but if you need to worry about grounding shielding - and you often find that you have to - ignorance isn\'t a good place to start.

> Given a Spice model, you can run it in either direction, with resistive or diode loads or whatever, and Spice will tell you everything you need to know. The only gotcha it misses is core saturation.

And the the John Chan model will let you figure that in, if only for a single inductor.

I had to take TWO semisters of Electrical Machinery at Tulane, which was a huge nuisance, but I did learn about the circuit equivalent of a
transformer, which was almost worth the hassle.

He means two semesters. He was presumably taught about leakage inductance then, which will be where he got his false idea about how the LTSpice transformer model worked. Not every university course is as up to date as it should have been, and Tulane wasn\'t on the bleeding edge.

You may have to seriously de-rate the transformer in order to use it that way, mains frequency transformers under 100VA are the most affected.

A 100 VA transformer is happy moving 100 VA in either direction.

** JL has such simple faith in overly simple models.

I wonder why ?

I design things that work, and verify by experiment if there\'s any doubt.. And the transformer model IS simple.

But misleading. Designing things that work most of the time doesn\'t hack it and constructing experiments that pick out the corner cases where they stop working does call for a good grasp of what might be going on.

One thing that\'s often worth verifying is temperature rise, with actual mounting and air flow. Transformer vendors are vague and the
thermals are hard to math out. [1]

If you are concerned about reversing a transformer, just try it.

[1] the affect of air flow and duty cycle on parts is rarely mentioned.

It\'s pretty obvious. To get anywhere with duty cycle you need to know thermal mass. which isn\'t too hard to work out or measure.

--
Bill Sloman, Sydney

 

[Anthony William Sloman]

On Monday, July 3, 2023 at 9:15:10 PM UTC+10, John Larkin wrote:

On Mon, 3 Jul 2023 03:45:00 -0700 (PDT), Phil Allison <palli...@gmail.com> wrote:
John Larkin wrote:

<snip>

That\'s the part I don\'t see. \"Primary\" is an application decision, or
a data sheet convenience.

** It is far more than that, but you will never admit it.

The word \"primary\" on a data sheet doesn\'t stop me from using a transformer in either direction. Like the Coilcraft PL300-100L right now. Try and stop me!

Not our job. The way that happens is that you build it into something and sell it, and it catches fire on a customer\'s premises, and you get sued.

I\'ll leave out the bit about you trying to explain to the judge that it shouldn\'t caught fire because you\'d tested it to your own satisfaction. Your lawyer would probably be able to persuade you not to do that,

--
Bill Sloman, Sydney

 

[Fred Bloggs]

On Monday, July 3, 2023 at 7:15:10 AM UTC-4, John Larkin wrote:

On Mon, 3 Jul 2023 03:45:00 -0700 (PDT), Phil Allison
palli...@gmail.com> wrote:

John Larkin wrote:
----------------------------

Basically, if you buy two equal-power-rated transformers, one sold as
120:240 and one sold as 240:120, they are the same transformer.

** But not exactly - see below.

You may have to seriously de-rate the transformer in order to use it that way,
mains frequency transformers under 100VA are the most affected.

A 100 VA transformer is happy moving 100 VA in either direction.

** JL has such simple faith in overly simple models.

I wonder why ?

If you are concerned about reversing a transformer, just try it.


** LOL - of *course* I have and that is why I know about the pitfalls.

I haven\'t heard any real pitfalls so far. What problems did you have?

** Same as everyone else\'s !!

When roles are reversed, the previous 240VAC supply winding supplies less under load.
The difference is about twice the regulation percentage, so 20 to 30% less in some cases.
Imag current ( previously easily supplied by the mains ) goes up by the turns ratio, up to maybe 40 times in a low voltage winding.

Applying more than rated voltage to the secondary in order to to fix this results in excessive current and overheating the transformer.

This all derives from how transformer makers engineer * real * transformers and rely on specifying which is the primary etc so all specs are met..


That\'s the part I don\'t see. \"Primary\" is an application decision, or
a data sheet convenience.

** It is far more than that, but you will never admit it.


....... Phil


The word \"primary\" on a data sheet doesn\'t stop me from using a
transformer in either direction. Like the Coilcraft PL300-100L right
now. Try and stop me!

An exception is for industrial grade \"protected\" transformers which have either a fuse or thermal cutout embedded in their designated \"primary.\" That really needs to be on the line side of the transformer drive.

https://www.electronicproducts.com/protecting-transformers-with-pptc-devices/

 

[John Larkin]

On Mon, 3 Jul 2023 06:23:17 -0700 (PDT), Fred Bloggs
<bloggs.fredbloggs.fred@gmail.com> wrote:

On Monday, July 3, 2023 at 7:15:10?AM UTC-4, John Larkin wrote:
On Mon, 3 Jul 2023 03:45:00 -0700 (PDT), Phil Allison
palli...@gmail.com> wrote:

John Larkin wrote:
----------------------------

Basically, if you buy two equal-power-rated transformers, one sold as
120:240 and one sold as 240:120, they are the same transformer.

** But not exactly - see below.

You may have to seriously de-rate the transformer in order to use it that way,
mains frequency transformers under 100VA are the most affected.

A 100 VA transformer is happy moving 100 VA in either direction.

** JL has such simple faith in overly simple models.

I wonder why ?

If you are concerned about reversing a transformer, just try it.


** LOL - of *course* I have and that is why I know about the pitfalls.

I haven\'t heard any real pitfalls so far. What problems did you have?

** Same as everyone else\'s !!

When roles are reversed, the previous 240VAC supply winding supplies less under load.
The difference is about twice the regulation percentage, so 20 to 30% less in some cases.
Imag current ( previously easily supplied by the mains ) goes up by the turns ratio, up to maybe 40 times in a low voltage winding.

Applying more than rated voltage to the secondary in order to to fix this results in excessive current and overheating the transformer.

This all derives from how transformer makers engineer * real * transformers and rely on specifying which is the primary etc so all specs are met.


That\'s the part I don\'t see. \"Primary\" is an application decision, or
a data sheet convenience.

** It is far more than that, but you will never admit it.


....... Phil


The word \"primary\" on a data sheet doesn\'t stop me from using a
transformer in either direction. Like the Coilcraft PL300-100L right
now. Try and stop me!

An exception is for industrial grade \"protected\" transformers which have either a fuse or thermal cutout embedded in their designated \"primary.\" That really needs to be on the line side of the transformer drive.

https://www.electronicproducts.com/protecting-transformers-with-pptc-devices/

Sure, if something is powered off the AC line, there needs to be
protections and the design should be able to be UL and CE certified.
One would always apply 120 volts, or whatever, to a transformer\'s 120
volt primary.

Fortunately, we electronics guys don\'t connect to the AC line a lot
ourselves any more; we can buy listed AC switch/filter/connector
things and listed switching power supplies. Or, whenever possible, use
a wart. There are some monster 24 and 48 volt warts and laptop
supplies around these days.

60 Hz magnetics and giant electrolytic caps and primary voltage
switches are big and expensive anyhow. Ancient tech.

I don\'t want nasty line voltage in my beautiful little boxes!

 

[John Larkin]

On Sun, 02 Jul 2023 11:44:20 GMT, Jan Panteltje <alien@comet.invalid>
wrote:

On a sunny day (Sun, 2 Jul 2023 04:19:47 -0700 (PDT)) it happened Phil Allison
pallison49@gmail.com> wrote in
5f09e2f5-5376-44bd-8d95-f299090aad04n@googlegroups.com>:

John Larkin wrote:
------------------------------

Question is whether they are symmetrical from either side (primary/secondary) in efficiency and load capacity.
I will be driving it from both sides, but not at the same time.

Sure, a transformer works either way.


** But makers nominate the primary and secondary for good reasons.
The primary supplies the magnetising current, which can be significant, plus is wound to allow for voltage drop at the
secondary under full load. If the roles are swapped, there are several changes that can catch the unwary.


Any winding can magnetize the core. The primary is the one you decide
to put power into and the secondary is the one you elect to take power
out of.

** The maker tells you which is which because they have adjusted the windings ( number of turns and wire gauges) accordingly.


2-winding power transformers are usually designed for equal copper
loss on both windings. So are thermally symmetric.
Sometimes some windings are insulated better than others.

What other catches might there be?

** As I mentioned, which winding carries the I mag plus the turns ratio have been adjusted by makers to compensate for the
full load voltage drop. Reverse the makers intended roles and those built in compensations operate in the opposite directions.

That\'s easy to think through. Power transformer models are simple and
easy to reverse.

** But real transformers are not so simple.

You may have to seriously de-rate the transformer in order to use it that way,
mains frequency transformers under 100VA are the most affected.

A 100 VA transformer is happy moving 100 VA in either direction.

** JL has such simple faith in overly simple models.

I wonder why ?

Because it works, did something for the county here like that to replaces a very complex design..
Used existing transformer models in reverse.

My/our job is to build things that work, and we can use any design
tools, or none for that matter, that work.

Spice has removed the necessity for a heap of classic and tedious
tools and equations and papers. But any non-trivial design, no matter
how derived, will probably be simulated and often lab tested.

It\'s good to get away from a screen and solder up stuff now and then,
stress some parts and see what happens.

Last week I toasted a bunch of wirewound resistors and blew up a few
solid-state relays, and learned things not on data sheets.

The Coilcraft PL300 planar power transformers are dynamite. They have
6, optionally 7 windings and they call one set \"primary\", which is
silly; they\'re just windings. Leakage inductance is so low it can be
ignored.

 

[Anthony William Sloman]

On Tuesday, July 4, 2023 at 1:04:08 AM UTC+10, John Larkin wrote:

On Sun, 02 Jul 2023 11:44:20 GMT, Jan Panteltje <al...@comet.invalid> wrote:
On a sunny day (Sun, 2 Jul 2023 04:19:47 -0700 (PDT)) it happened Phil Allison
palli...@gmail.com> wrote in <5f09e2f5-5376-44bd...@googlegroups.com>:
John Larkin wrote:

That\'s easy to think through. Power transformer models are simple and
easy to reverse.

** But real transformers are not so simple.

You may have to seriously de-rate the transformer in order to use it that way,
mains frequency transformers under 100VA are the most affected.

A 100 VA transformer is happy moving 100 VA in either direction.

** JL has such simple faith in overly simple models.

I wonder why ?

Because it works, did something for the county here like that to replaces a very complex design..
Used existing transformer models in reverse.

My/our job is to build things that work, and we can use any design
tools, or none for that matter, that work.

Spice has removed the necessity for a heap of classic and tedious tools and equations and papers. But any non-trivial design, no matter how derived, will probably be simulated and often lab tested.

It\'s good to get away from a screen and solder up stuff now and then, stress some parts and see what happens.

Last week I toasted a bunch of wirewound resistors and blew up a few solid-state relays, and learned things not on data sheets.

The Coilcraft PL300 planar power transformers are dynamite.

They blow up violently?

> They have 6, optionally 7 windings and they call one set \"primary\", which is silly; they\'re just windings. Leakage inductance is so low it can be ignored.

It\'s never that low, and the way they measure leakage inductance isn\'t all that informative. What you actually need are measures of mutual inductance.

\"Leakage inductance measured at 200 kHz, 0.5 Vrms, 0 Adc between pins 2 and 5, with pins 3 and 4 connected,and with all secondary pins shorted\"

This includes the resistance of the secondary windings in the measurement, which isn\'t exactly clever, if cheap and easy to do.

Using bad models gets you bad results.

https://www.coilcraft.com/getmedia/636aef6f-8168-4ece-bb01-e16574e130d8/pl300.pdf

--
Bill Sloman, Sydney