What happened to toroid power transformers?

[Spehro Pefhany]

On Mon, 12 Mar 2007 20:21:03 -0800, the renowned John Larkin
<jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

On Mon, 12 Mar 2007 23:09:26 -0500, Spehro Pefhany
speffSNIP@interlogDOTyou.knowwhat> wrote:

On Tue, 13 Mar 2007 01:19:48 -0000, the renowned dplatt@radagast.org
(Dave Platt) wrote:

In article <gtsbv2923v8c5p7oevdl4r3f081c66nt3t@4ax.com>,
John Larkin <jjlarkin@highNOTlandTHIStechnologyPART.com> wrote:

For those of us not familiar, 'splain, please?

A conventional laminated core has corners and stuff. Some parts run at
lower flux density than others, so are sort of buffers against hard
saturation. Toroids have nice uniform cores, so can be designed to
have all of the core material run near saturation. That's one reason
they are so small and light. The geometry favors low copper
resistance, too.

So switch off a piece of gear that uses a toroidal line transformer.
If you're unlucky, the switchoff will happen at maximum flux density
in one direction, and leave some residual magnetization. Now, more bad
luck, turn it on at the ac zero crossing in the same direction. All
the core saturates and a huge primary current flows. This cheerfully
takes out mdl or even slo-blow fuses, and sometimes power switches.
We've measured 1000 amp peaks on modest-sized transformers, and you
could hear the wiring jump inside the wall.

Seems like a good application for an NTC-thermistor inrush current
limiter, with a few ohms of "cold" resistance?

What happens if the power blips with the NTC hot? Short blips in AC
power are pretty common, and there would be negligible time for the
NTC to cool.


We were concerned about that, and did some tests, on a 1000 watt CAMAC
crate power supply. It ate power switches before we installed NTCs,
and after that was fine. We tried teasing the power switch all sorts
of ways, and it still worked. Ditto on an NMR gradient driver. Don't
quite understand why.

John

You didn't happen to measure the current peaks before and after, did
you? Probably not or you'd say so. That would be an interesting bit of
info.


Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

 

[Eeyore]

John Larkin wrote:

John E. <incognito@yahoo.com> wrote:

We use toroidal power transformers in some of our products. They're
small, don't leak much field, and don't cost much more than regular
ones. But they are sure hard on line fuses.

For those of us not familiar, 'splain, please?

A conventional laminated core has corners and stuff.

It also has a small distributed air gap.

Some parts run at
lower flux density than others, so are sort of buffers against hard
saturation. Toroids have nice uniform cores, so can be designed to
have all of the core material run near saturation. That's one reason
they are so small and light. The geometry favors low copper
resistance, too.

So switch off a piece of gear that uses a toroidal line transformer.
If you're unlucky, the switchoff will happen at maximum flux density
in one direction, and leave some residual magnetization. Now, more bad
luck, turn it on at the ac zero crossing in the same direction. All
the core saturates and a huge primary current flows. This cheerfully
takes out mdl or even slo-blow fuses, and sometimes power switches.
We've measured 1000 amp peaks on modest-sized transformers, and you
could hear the wiring jump inside the wall.

CE requirements don't allow over-rating fuses a lot, so that can be
really nasty. The super-slow TT fuses help, but are sometimes hard to
get.

John, you can fix this by running toroids at a *lower* flux or you can fit an
inrush current limiting device / circuit.

The absence of an air gap in toroids is a contributory factor to the problem
btw.

Graham

 

[Eeyore]

Spehro Pefhany wrote:

What happens if the power blips with the NTC hot?

You hope the PSU filter caps hold up !

Graham

 

[Eeyore]

Winfield Hill wrote:

John Larkin wrote:

We use toroidal power transformers in some of our products. They're
small, don't leak much field, and don't cost much more than regular
ones. But they are sure hard on line fuses.

They have poor line-to-output AC isolation. They usually have low
leakage inductance, that's bad for direct bridge-rectifier storage-
capacitor setups. Plus, it's not so easy to add a grounded primary-
secondary inter-winding shield. But hey, what the hell, I like 'em.
Low ac magnetic fields spreading out into my sensitive electronics.

In which case you'd probably like R-cores and O-cores even more.

Graham

 

[Eeyore]

Paul Mathews wrote:

John Larkin wrote:

So switch off a piece of gear that uses a toroidal line transformer.
If you're unlucky, the switchoff will happen at maximum flux density
in one direction, and leave some residual magnetization. Now, more bad
luck, turn it on at the ac zero crossing in the same direction. All
the core saturates and a huge primary current flows. This cheerfully
takes out mdl or even slo-blow fuses, and sometimes power switches.
We've measured 1000 amp peaks on modest-sized transformers, and you
could hear the wiring jump inside the wall.

CE requirements don't allow over-rating fuses a lot, so that can be
really nasty. The super-slow TT fuses help, but are sometimes hard to
get.

Other approaches commonly taken include a series power resistor
shunted by relay contacts or a triac turned on after a delay.

The omly reason I've never done that is the issue of dissipation in the triac.

Another approach is to use SCRs in 2 legs of the secondary bridge rectifier,
using phase control to ramp up the secondary current.

A certain designer uses that technique to modulate the output voltage.

Graham

 

[Eeyore]

Spehro Pefhany wrote:

John Larkin wrote:

We were concerned about that, and did some tests, on a 1000 watt CAMAC
crate power supply. It ate power switches before we installed NTCs,
and after that was fine. We tried teasing the power switch all sorts
of ways, and it still worked. Ditto on an NMR gradient driver. Don't
quite understand why.


You didn't happen to measure the current peaks before and after, did
you? Probably not or you'd say so. That would be an interesting bit of
info.

John probably ought to speak to Plitron.

Graham

 

[Winfield Hill]

John Larkin wrote:

We use toroidal power transformers in some of our products.
They're small, don't leak much field, and don't cost much
more than regular ones. But they are sure hard on line fuses.

They have poorer line-to-output AC isolation the conventional
types with separated split windings. They usually have lower
leakage inductance, which is bad for direct bridge-rectifier
storage-capacitor setups as it leads to higher peak currents.
Plus, they don't welcome adding a grounded primary-secondary
inter-winding shield. But hey, what the hell, I like their
low ac magnetic fields spreading into sensitive electronics.

 

[Eeyore]

Winfield Hill wrote:

John Larkin wrote:

We use toroidal power transformers in some of our products.
They're small, don't leak much field, and don't cost much
more than regular ones. But they are sure hard on line fuses.

They have poorer line-to-output AC isolation the conventional
types with separated split windings. They usually have lower
leakage inductance, which is bad for direct bridge-rectifier
storage-capacitor setups as it leads to higher peak currents.
Plus, they don't welcome adding a grounded primary-secondary
inter-winding shield. But hey, what the hell, I like their
low ac magnetic fields spreading into sensitive electronics.

R-cores are vastly better.

Graham

 

[John Larkin]

On 12 Mar 2007 22:06:44 -0700, "Paul Mathews" <opto@whidbey.com>
wrote:

On Mar 12, 5:58 pm, John Larkin
jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
On Tue, 13 Mar 2007 00:24:45 GMT, John E. <incogn...@yahoo.com> wrote:
We use toroidal power transformers in some of our products. They're
small, don't leak much field, and don't cost much more than regular
ones. But they are sure hard on line fuses.

John

For those of us not familiar, 'splain, please?

A conventional laminated core has corners and stuff. Some parts run at
lower flux density than others, so are sort of buffers against hard
saturation. Toroids have nice uniform cores, so can be designed to
have all of the core material run near saturation. That's one reason
they are so small and light. The geometry favors low copper
resistance, too.

So switch off a piece of gear that uses a toroidal line transformer.
If you're unlucky, the switchoff will happen at maximum flux density
in one direction, and leave some residual magnetization. Now, more bad
luck, turn it on at the ac zero crossing in the same direction. All
the core saturates and a huge primary current flows. This cheerfully
takes out mdl or even slo-blow fuses, and sometimes power switches.
We've measured 1000 amp peaks on modest-sized transformers, and you
could hear the wiring jump inside the wall.

CE requirements don't allow over-rating fuses a lot, so that can be
really nasty. The super-slow TT fuses help, but are sometimes hard to
get.

John

Other approaches commonly taken include a series power resistor
shunted by relay contacts or a triac turned on after a delay. Another
approach is to use SCRs in 2 legs of the secondary bridge rectifier,
using phase control to ramp up the secondary current. This often
works, since part of the turn-on surge....sometimes a big share of
it....is actually the charging current for the secondary side
capacitors reflected back to the primary side, with very little
leakage inductance in series. ST makes a part designed to switch the
line at zero crossings.

Paul Mathews

In one of our products, we use a primary-side resistor-triac as both a
surge limiter and a crude bang-bang regulator, to reduce the stress on
downstream stuff as line voltage changes.

John

 

[John Larkin]

On 12 Mar 2007 21:59:12 -0700, "Winfield Hill" <hill@rowland.org>
wrote:

John Larkin wrote:

We use toroidal power transformers in some of our products. They're
small, don't leak much field, and don't cost much more than regular
ones. But they are sure hard on line fuses.

They have poor line-to-output AC isolation. They usually have low
leakage inductance, that's bad for direct bridge-rectifier storage-
capacitor setups. Plus, it's not so easy to add a grounded primary-
secondary inter-winding shield. But hey, what the hell, I like 'em.
Low ac magnetic fields spreading out into my sensitive electronics.

This was for an nmr gradient amp, in a rack with lots of stuff that
doesn't like 60 Hz fields. We finally talked them into letting us use
switchers - it took about 10 years - and they work great.

John

 

[Terry Given]

Eeyore wrote:

John Larkin wrote:


John E. <incognito@yahoo.com> wrote:


We use toroidal power transformers in some of our products. They're
small, don't leak much field, and don't cost much more than regular
ones. But they are sure hard on line fuses.

For those of us not familiar, 'splain, please?

A conventional laminated core has corners and stuff.


It also has a small distributed air gap.



Some parts run at
lower flux density than others, so are sort of buffers against hard
saturation. Toroids have nice uniform cores, so can be designed to
have all of the core material run near saturation. That's one reason
they are so small and light. The geometry favors low copper
resistance, too.

So switch off a piece of gear that uses a toroidal line transformer.
If you're unlucky, the switchoff will happen at maximum flux density
in one direction, and leave some residual magnetization. Now, more bad
luck, turn it on at the ac zero crossing in the same direction. All
the core saturates and a huge primary current flows. This cheerfully
takes out mdl or even slo-blow fuses, and sometimes power switches.
We've measured 1000 amp peaks on modest-sized transformers, and you
could hear the wiring jump inside the wall.

CE requirements don't allow over-rating fuses a lot, so that can be
really nasty. The super-slow TT fuses help, but are sometimes hard to
get.


John, you can fix this by running toroids at a *lower* flux or you can fit an
inrush current limiting device / circuit.

yeah, but it has to be 50% of Bsat, or the problem still occurs.

So Np doubles.

And given a full winding window, Rp quadruples.

The absence of an air gap in toroids is a contributory factor to the problem
btw.

Graham

I've since sold it, but I used to have a little 100kVA transformer I
bought for $200, brand new. It ran at 250mT peak flux density. was
designed for a motor test application, where it was switched on & off
about once per minute, hence the tiny Bpeak. but the customer went broke
and never picked it up, so it sat in the factory for several years,
until I came along.

I ended up selling it for $1000

damn shame though, I could use it now

Cheers
Terry

 

[John E.]

In one of our products, we use a primary-side resistor-triac as both a
surge limiter and a crude bang-bang regulator, to reduce the stress on
downstream stuff as line voltage changes.

Diagram, please (c:
--
John English

 

[Joerg]

Winfield Hill wrote:

John Larkin wrote:

We use toroidal power transformers in some of our products.
They're small, don't leak much field, and don't cost much
more than regular ones. But they are sure hard on line fuses.


They have poorer line-to-output AC isolation the conventional
types with separated split windings. They usually have lower
leakage inductance, which is bad for direct bridge-rectifier
storage-capacitor setups as it leads to higher peak currents.
Plus, they don't welcome adding a grounded primary-secondary
inter-winding shield. But hey, what the hell, I like their
low ac magnetic fields spreading into sensitive electronics.

We use toroids almost exclusively for medical. A shield layer is no
problem at all. I also use them here in the office and in the lab for
120V/230V conversion because they are almost completely silent.

--
Regards, Joerg

http://www.analogconsultants.com

 

[Lostgallifreyan]

Joerg <notthisjoergsch@removethispacbell.net> wrote in
news:uryKh.8079$Um6.6155@newssvr12.news.prodigy.net:

Winfield Hill wrote:

John Larkin wrote:

We use toroidal power transformers in some of our products.
They're small, don't leak much field, and don't cost much
more than regular ones. But they are sure hard on line fuses.


They have poorer line-to-output AC isolation the conventional
types with separated split windings. They usually have lower
leakage inductance, which is bad for direct bridge-rectifier
storage-capacitor setups as it leads to higher peak currents.
Plus, they don't welcome adding a grounded primary-secondary
inter-winding shield. But hey, what the hell, I like their
low ac magnetic fields spreading into sensitive electronics.


We use toroids almost exclusively for medical. A shield layer is no
problem at all. I also use them here in the office and in the lab for
120V/230V conversion because they are almost completely silent.

While following this thread, I saw a couple of people mention R-cores.
Would they not be better?

I'd once seen one and thought it was some eccentric variant on a toroid
that someone made so they could mount it where their design once called for
a chassis mounted E/I type, or had some other odd space restriction. I was
so wrong.

From what I saw via Google once I'd seen the name 'R-core', I see that easy
fitting of split bobbins directly round the straight long sections allows
either a commercial firm OR a hobbyist to not only build their windings
quickly and easily, but to modify them, as an assembled bobbin can rotate
freely if wanted. Electrical isolation between windings can be better than
in a toroidal type, which could be important for use in a medical device.
The efficiency is good, and the flux well-contained, and they'd probably
run as quietly as toroidals. Waste heat can escape from them more
efficiently that either E/I types or toroidals. I'm surprised they aren't
much more widespread than toroidals.

 

[Joerg]

Lostgallifreyan wrote:

Joerg <notthisjoergsch@removethispacbell.net> wrote in
news:uryKh.8079$Um6.6155@newssvr12.news.prodigy.net:


Winfield Hill wrote:


John Larkin wrote:


We use toroidal power transformers in some of our products.
They're small, don't leak much field, and don't cost much
more than regular ones. But they are sure hard on line fuses.


They have poorer line-to-output AC isolation the conventional
types with separated split windings. They usually have lower
leakage inductance, which is bad for direct bridge-rectifier
storage-capacitor setups as it leads to higher peak currents.
Plus, they don't welcome adding a grounded primary-secondary
inter-winding shield. But hey, what the hell, I like their
low ac magnetic fields spreading into sensitive electronics.


We use toroids almost exclusively for medical. A shield layer is no
problem at all. I also use them here in the office and in the lab for
120V/230V conversion because they are almost completely silent.



While following this thread, I saw a couple of people mention R-cores.
Would they not be better?

I'd once seen one and thought it was some eccentric variant on a toroid
that someone made so they could mount it where their design once called for
a chassis mounted E/I type, or had some other odd space restriction. I was
so wrong.

From what I saw via Google once I'd seen the name 'R-core', I see that easy
fitting of split bobbins directly round the straight long sections allows
either a commercial firm OR a hobbyist to not only build their windings
quickly and easily, but to modify them, as an assembled bobbin can rotate
freely if wanted. Electrical isolation between windings can be better than
in a toroidal type, which could be important for use in a medical device.
The efficiency is good, and the flux well-contained, and they'd probably
run as quietly as toroidals. Waste heat can escape from them more
efficiently that either E/I types or toroidals. I'm surprised they aren't
much more widespread than toroidals.

They used to be quite popular in TV sets. I believe I still have a few
cores. Nowadays often just called U-U cores. For those who haven't seen
them yet:
http://www.electroassemblies.com/r-core.htm

One challenge with these is proper clamping. You can't inspect how snug
the core halves are joining because it is inside the packets.

BTW your follow-up settings aren't right, was missing three NGs. That
would have broken the thread for those folks.

--
Regards, Joerg

http://www.analogconsultants.com

 

[Eeyore]

Joerg wrote:

Lostgallifreyan wrote:

From what I saw via Google once I'd seen the name 'R-core', I see that easy
fitting of split bobbins directly round the straight long sections allows
either a commercial firm OR a hobbyist to not only build their windings
quickly and easily, but to modify them, as an assembled bobbin can rotate
freely if wanted. Electrical isolation between windings can be better than
in a toroidal type, which could be important for use in a medical device.
The efficiency is good, and the flux well-contained, and they'd probably
run as quietly as toroidals. Waste heat can escape from them more
efficiently that either E/I types or toroidals. I'm surprised they aren't
much more widespread than toroidals.

Expensive ! You also can't bump up the copper as you can with a toroid.

They used to be quite popular in TV sets.

??????

I believe I still have a few cores. Nowadays often just called U-U cores.

Not the same thing at all.

For those who haven't seen
them yet:
http://www.electroassemblies.com/r-core.htm

One challenge with these is proper clamping. You can't inspect how snug
the core halves are joining because it is inside the packets.

An R-core is made from continuous strip like a toroid. No clamping is invoved.

Graham

 

[Joerg]

Eeyore wrote:

Joerg wrote:


Lostgallifreyan wrote:

From what I saw via Google once I'd seen the name 'R-core', I see that easy
fitting of split bobbins directly round the straight long sections allows
either a commercial firm OR a hobbyist to not only build their windings
quickly and easily, but to modify them, as an assembled bobbin can rotate
freely if wanted. Electrical isolation between windings can be better than
in a toroidal type, which could be important for use in a medical device.
The efficiency is good, and the flux well-contained, and they'd probably
run as quietly as toroidals. Waste heat can escape from them more
efficiently that either E/I types or toroidals. I'm surprised they aren't
much more widespread than toroidals.


Expensive ! You also can't bump up the copper as you can with a toroid.



They used to be quite popular in TV sets.


??????



I believe I still have a few cores. Nowadays often just called U-U cores.


Not the same thing at all.



For those who haven't seen
them yet:
http://www.electroassemblies.com/r-core.htm

One challenge with these is proper clamping. You can't inspect how snug
the core halves are joining because it is inside the packets.


An R-core is made from continuous strip like a toroid. No clamping is invoved.

Hmm, so how does that make winding easier then?

--
Regards, Joerg

http://www.analogconsultants.com

 

[Eeyore]

Joerg wrote:

Eeyore wrote:
Joerg wrote:
Lostgallifreyan wrote:

From what I saw via Google once I'd seen the name 'R-core', I see that easy
fitting of split bobbins directly round the straight long sections allows
either a commercial firm OR a hobbyist to not only build their windings
quickly and easily, but to modify them, as an assembled bobbin can rotate
freely if wanted. Electrical isolation between windings can be better than
in a toroidal type, which could be important for use in a medical device.
The efficiency is good, and the flux well-contained, and they'd probably
run as quietly as toroidals. Waste heat can escape from them more
efficiently that either E/I types or toroidals. I'm surprised they aren't
much more widespread than toroidals.

Expensive ! You also can't bump up the copper as you can with a toroid.

They used to be quite popular in TV sets.

??????

I believe I still have a few cores. Nowadays often just called U-U cores.

Not the same thing at all.

For those who haven't seen them yet:
http://www.electroassemblies.com/r-core.htm

One challenge with these is proper clamping. You can't inspect how snug
the core halves are joining because it is inside the packets.


An R-core is made from continuous strip like a toroid. No clamping is invoved.

Hmm, so how does that make winding easier then?

Only easier with the correct machinery.

The formers are made in 2 pieces that clip together and the winding machine spins
the bobbins on the limbs of the core. The bobbins have 'gear teeth' to engage with
the winding machine.

Graham

 

[Lostgallifreyan]

Joerg <notthisjoergsch@removethispacbell.net> wrote in
news:%jCKh.10629$jx3.8853@newssvr25.news.prodigy.net:

Hmm, so how does that make winding easier then?

It does if you can wind the bobbin round the former it's clamped round.

Eeyore wrote:

I'm surprised they aren't much more widespread than toroidals.


Expensive ! You also can't bump up the copper as you can with a toroid.

What does 'bump up' mean? Re expense, if the bobbin can be rotated round
the straight part of the former it's built onto, it would be a lot less
awkward than winding a toroid, it would not be much more awkward that
winding onto any spool. So why would it be more expensive than a toroid,
given that the former is made the same way, and the windings are easier to
wind?

 

[Joerg]

Eeyore wrote:

Joerg wrote:


Eeyore wrote:

Joerg wrote:

Lostgallifreyan wrote:

From what I saw via Google once I'd seen the name 'R-core', I see that easy

fitting of split bobbins directly round the straight long sections allows
either a commercial firm OR a hobbyist to not only build their windings
quickly and easily, but to modify them, as an assembled bobbin can rotate
freely if wanted. Electrical isolation between windings can be better than
in a toroidal type, which could be important for use in a medical device.
The efficiency is good, and the flux well-contained, and they'd probably
run as quietly as toroidals. Waste heat can escape from them more
efficiently that either E/I types or toroidals. I'm surprised they aren't
much more widespread than toroidals.

Expensive ! You also can't bump up the copper as you can with a toroid.


They used to be quite popular in TV sets.

??????


I believe I still have a few cores. Nowadays often just called U-U cores.

Not the same thing at all.


For those who haven't seen them yet:
http://www.electroassemblies.com/r-core.htm

One challenge with these is proper clamping. You can't inspect how snug
the core halves are joining because it is inside the packets.


An R-core is made from continuous strip like a toroid. No clamping is invoved.

Hmm, so how does that make winding easier then?


Only easier with the correct machinery.

The formers are made in 2 pieces that clip together and the winding machine spins
the bobbins on the limbs of the core. The bobbins have 'gear teeth' to engage with
the winding machine.

Ah, thanks. To be honest I have never seen one. But it sure looks like a
cool trick.

--
Regards, Joerg

http://www.analogconsultants.com