low-cost 1800-amp heating source

[Spehro Pefhany]

This is assuming there is not mains ripple getting into the triggering
circuit.

Yeah, and that the diodes are pretty much the same.

It only ensures that it is as symmetrical as the triggering
is. A triac running from that same trigger signal would be about as
good.

Triacs, especially those of the sensitive-gate persuasion, don't like
to commutate when there is much dv/dt. Alternistors are better, SCRs
are the best thyristor.

We could put a large capacitor in series to ensure no DC component.
It leads to some fun waveforms and interesting shopping trips at these
current levels, however.

I don't think it's an issue with a reasonable design. although
toroidal xfmrs are supposed to be worse in that regard than E-I lam
types.

Heck, if you want to get fancy, why not detect any miniscule DC
component and actively null it?


Best regards,
Spehro Pefhany

 

[MooseFET]

Yeah, and that the diodes are pretty much the same.

At the voltages involved, we would have to use some strange diodes to
make enough mismatch to matter.

Triacs, especially those of the sensitive-gate persuasion, don't like
to commutate when there is much dv/dt.

Even an SCR would likely need a snubber in this circuit.

Alternistors are better, SCRs
are the best thyristor.


I don't think it's an issue with a reasonable design. although
toroidal xfmrs are supposed to be worse in that regard than E-I lam
types.

Neither really do I. At worst we will have a small fraction of the
peak current as an average.

Heck, if you want to get fancy, why not detect any miniscule DC
component and actively null it?

I think that is a somewhat silly idea so lets do it

We know that the current stops when the SCR is off. We can use a
current transformer and restore the zero when the current stops. I
assume we are tripping the SCR with a tranformer so we can have the
whole circuit isolated from the mains.

The current sense transformer would drive a circuit like this:

-----/\/\---------
( !
( !
( !
! --/\/\---+
-GND ! !
! !
---/\/\-+-!-\ ! C1
( ! >----+---!!--+---!+\ To DC servo
( GND-!+/ ! ! >--+---
( / -!-/ !
---GND SW1 O ! !
! ------
GND

The first opamp lets us use a cheap design for the current transformer
and get a big enough signal to drive C1.

C1 is snapped to ground by SW1 when the SCR is off to reset the zero.

 

[Tony Williams]

If the scr/triac conduction isn't perfectly symmetric, you can
wind up with net DC into the primary, which can get weird. And
sometimes an scr won't stay triggered when driving an inductive
load, ie the transformer leakage inductance. There are hazards,
that's all.

Perhaps no need to use a semiconductor. This could be
an opportunity to investigate the design of the old GE
theatre light dimmer, based on the saturable reactor.
There should be no net DC component from that, and
relatively soft waveform modulation.

La
ACin-----------/////////---------+
========= |
DC Control---////////------------- DC Control
========= |
ACout----------/////////---------+
Lb

The saturable reactor can be wound on E-I laminations or
on two toroids stacked together, with the control winding
wound on the stacked assembly.

Effective inductance control is just a matter of a variable
DC constant current source driving the control winding.

 

[MooseFET]

Perhaps no need to use a semiconductor. This could be
an opportunity to investigate the design of the old GE
theatre light dimmer, based on the saturable reactor.
There should be no net DC component from that, and
relatively soft waveform modulation.

La
ACin-----------/////////---------+
========= |
DC Control---////////------------- DC Control
========= |
ACout----------/////////---------+
Lb

The saturable reactor can be wound on E-I laminations or
on two toroids stacked together, with the control winding
wound on the stacked assembly.

Effective inductance control is just a matter of a variable
DC constant current source driving the control winding.

Just take two transformers and wire the primaries in series and the
secondaries in series bucking as the control. It worked fine some 40
years ago to control a light bulb so a bigger one should control a
heater quite nicely. DC in the secondaries makes the cores
saturate.

If you want to be a purist about it, you need to place a large
capacitor across the control winding. When you feed DC into the
control some AC currents flow in the control windings. The capacitor
keeps this out of the control circuit.

 

[Frithiof Andreas Jensen]

The question then becomes, what's the best way
to control the total output current, to obtain a
precise oven temperature.

Phase Control - I.M.O.

 

[Frithiof Andreas Jensen]

Perhaps no need to use a semiconductor. This could be
an opportunity to investigate the design of the old GE
theatre light dimmer, based on the saturable reactor.
There should be no net DC component from that, and
relatively soft waveform modulation.

La
ACin-----------/////////---------+
========= |
DC Control---////////------------- DC Control
========= |
ACout----------/////////---------+
Lb

The saturable reactor can be wound on E-I laminations or
on two toroids stacked together, with the control winding
wound on the stacked assembly.

Effective inductance control is just a matter of a variable
DC constant current source driving the control winding.

Boat Anchor ;-)
But .... In a switcher I can testify that "the old way" is really Neat!

 

[Frithiof Andreas Jensen]

"Spurious Response" <[email protected]> skrev i en meddelelse

The real question would be whether you can even pass a "wire" that can
handle 1800 Amps through the toroid.

Well, since we are into plumbing supplies anyway, one might as well run
cooling water through the "wire" - i.e. make it from household copper tubing
;-)

 

[Tony Williams]

Frithiof Andreas Jensen

"Tony Williams" <[email protected]> skrev i en meddelelse


Boat Anchor ;-)

Brick Outhouse more probably.

 

[Winfield Hill]

Just take two transformers and wire the primaries in series and the
secondaries in series bucking as the control. It worked fine some
40 years ago to control a light bulb so a bigger one should control
a heater quite nicely. DC in the secondaries makes the cores
saturate.

If you want to be a purist about it, you need to place a large
capacitor across the control winding. When you feed DC into
the control some AC currents flow in the control windings.
The capacitor keeps this out of the control circuit.

One transformer with two secondaries, wired
in series bucking, wouldn't work, I suppose,
since the dc fields would cancel.

 

[MooseFET]

One transformer with two secondaries, wired
in series bucking, wouldn't work, I suppose,
since the dc fields would cancel.

Yes, you need to have a net field in the core(s) from the control
winding.

 

[Winfield]

One transformer with two secondaries, wired
in series bucking, wouldn't work, I suppose,
since the dc fields would cancel.

I'd appreciate a response on this. The saturable reactor is
appealing to me because it's a linear way of dealing with an
ac-transformer power-control problem. One thing that slows
me down is considering the two transformers needed to create
a saturable reactor - how big do they need to be? Thinking
about this, with no DC current, they create an open circuit,
no current flows and no power is dissipated. Alternately,
with maximum DC current the two transformers are saturated,
and simplified, look together like a length of copper wire.
So the intermediate condition provides the greatest stress.
That should be the region to evaluate.

Another issue, how much power will be involved generating
the DC-current to control say 1.0kVA? Will it be a similar
amount, say 500W dc? Ahem.

 

[Tony Williams]

I'd appreciate a response on this. The saturable reactor is
appealing to me because it's a linear way of dealing with an
ac-transformer power-control problem. One thing that slows
me down is considering the two transformers needed to create
a saturable reactor - how big do they need to be? Thinking
about this, with no DC current, they create an open circuit,
no current flows and no power is dissipated. Alternately,
with maximum DC current the two transformers are saturated,
and simplified, look together like a length of copper wire.
So the intermediate condition provides the greatest stress.
That should be the region to evaluate.

I was interested in Ken's idea, so sketched a
possible circuit, as below.

+ Pri - Load current
230V high +------////////---------->-----+
T1 ======== |
DC-------////////-----+ |(
+ Sec - | |( L(leak)
| |(
.----------------' | Win's Load.
| \
| - Sec + /Rload
+---////////---------DC \
T2 ======== |
230V low +-------////////----------------+
- Pri +

Each transformer's primary winding has to be able to
take the final max load current. So if you had a
230V/2500VA load then you would possibly use a pair of
115V/1250VA transformers.

I suspect that if the transformers were designed for
this application then they would be smaller because
more window area could be allotted to the AC side.

As a first pass, the load current, Iload = Vin/Z.

Where Z^2 = (w.(2Lp + Lleak))^2 + (2Rpri + Rload)^2.

The effective '2Lp' is the variable that is controlled
by the DC current, by varying the incremental permeability.
DC control will probably be non-linear, more or less
following the shape of the B-H loop.

Another issue, how much power will be involved generating
the DC-current to control say 1.0kVA? Will it be a similar
amount, say 500W dc? Ahem.

I suspect that the control power will be far less than
the normal rated output power of the secondaries.

The B-H loop will show what DC field strength is required
to push the flux density into saturation. An old graph
that I have shows that a sample of 0.014" Silicon Steel
required a polarising field strength of about 5 Oersteds
to reduce the incremental permeability from over 2000 to
about 100.

Polarising field strength, H = 4.pi.N.I/10.le Oersteds.

So the required N.I can be estimated. It would seem
useful to buy unpotted toroidal transformers so that
a few turns can be wound on, to get some idea of the
turns/volt the secondaries were wound at.

 

[Winfield]

Just take two transformers and wire the primaries in series and the
secondaries in series bucking as the control. It worked fine some
40 years ago to control a light bulb so a bigger one should control
a heater quite nicely. DC in the secondaries makes the cores
saturate.

If you want to be a purist about it, you need to place a large
capacitor across the control winding. When you feed DC into
the control some AC currents flow in the control windings.
The capacitor keeps this out of the control circuit.

Actually, why can't the two transformer primary windings be
in parallel? That'd reduce the copper resistance. And
L(leak) for that matter, right? Editing Tony's drawing:

AC high + Pri - Load current
--+-----------////////-------+--->---,
| T1 ======== | |
| DC --+---////////---, | |(
| | + Sec - | | |( L(leak)
| +_|_ | | |(
| --- | | |
| | + Sec - | | / Win's Load.
| DC --+---////////---' | \ Rload
| T2 ======== | /
'-----------////////-------' |
+ Pri - AC low

OK, now I see that Tony took advantage of a series primary
connection to use transformers with half the primary voltage
rating. Hah, it appears there's no free lunch.

 

[Winfield]

I was interested in Ken's idea, so sketched a
possible circuit, as below.

+ Pri - Load current
230V high +------////////---------->-----+
T1 ======== |
DC-------////////-----+ |(
+ Sec - | |( L(leak)
| |(
.----------------' | Win's Load.
| \
| - Sec + /Rload
+---////////---------DC \
T2 ======== |
230V low +-------////////----------------+
- Pri +

Each transformer's primary winding has to be able to
take the final max load current. So if you had a
230V/2500VA load then you would possibly use a pair of
115V/1250VA transformers.

I suspect that if the transformers were designed for
this application then they would be smaller because
more window area could be allotted to the AC side.

As a first pass, the load current, Iload = Vin/Z.

Where Z^2 = (w.(2Lp + Lleak))^2 + (2Rpri + Rload)^2.

The effective '2Lp' is the variable that is controlled
by the DC current, by varying the incremental permeability.
DC control will probably be non-linear, more or less
following the shape of the B-H loop.


I suspect that the control power will be far less than
the normal rated output power of the secondaries.

The B-H loop will show what DC field strength is required
to push the flux density into saturation. An old graph
that I have shows that a sample of 0.014" Silicon Steel
required a polarising field strength of about 5 Oersteds
to reduce the incremental permeability from over 2000 to
about 100.

Polarising field strength, H = 4.pi.N.I/10.le Oersteds.

So the required N.I can be estimated. It would seem
useful to buy unpotted toroidal transformers so that
a few turns can be wound on, to get some idea of the
turns/volt the secondaries were wound at.

A nice bit of work, there, Tony!

For quick estimates, could we guess that the DC current
would be 3 to 5 times higher than the peak AC current at
the VA rating of transformer? Given that the DC winding
resistance of the transformers is quite low, the DC-
control requirement doesn't look too bad. A fair amount
of DC control current, sure, but not too much voltage...

 

[Tony Williams]

For quick estimates, could we guess that the DC current
would be 3 to 5 times higher than the peak AC current at
the VA rating of transformer?

I don't think it is even as high as that Win.

If you were to leave the Pri open circuit, connect a
voltage source to the Sec, and turn it up until the
current waveform had nice wide flat tops then the
measured Ipeak could be a reasonable estimate of the
polarising DC amps needed to saturate the core.

3 to 5 times the secondary magnetising current peak
is probably much lower than 3 to 5 times the peak
secondary load current.

The control power would then be the DC amps times
the winding resistances.... but see below on what
could buggerate that low control power.

Given that the DC winding resistance of the transformers is quite
low, the DC- control requirement doesn't look too bad. A fair
amount of DC control current, sure, but not too much voltage...

There could be a little ambush in the two-transformer
saturable reactor. Those two series'd secondaries
may not cancel completely, so there could be a
residual AC voltage at their terminals. I suspect
that the DC current source must have a high enough
compliance/supply voltage such that no AC current
flows in the secondaries.

The same (or worse) AC current flow could occur
with the circulating currents that would arise
from the paralleled windings in your previous post.

This is getting interesting.......

 

[Tony Williams]

Actually, why can't the two transformer primary windings be
in parallel? That'd reduce the copper resistance. And
L(leak) for that matter, right? Editing Tony's drawing:

AC high + Pri - Load current
--+-----------////////-------+--->---,
| T1 ======== | |
| DC --+---////////---, | |(
| | + Sec - | | |( L(leak)
| +_|_ | | |(
| --- | | |
| | + Sec - | | / Win's Load.
| DC --+---////////---' | \ Rload
| T2 ======== | /
'-----------////////-------' |
+ Pri - AC low

The Lleak in my sketch is to allow for the possible
leakage inductance in your transformer Win.
It could be significant.

 

[Tony Williams]

If you were to leave the Pri open circuit, connect a
voltage source to the Sec, and turn it up until the
current waveform had nice wide flat tops then the
measured Ipeak could be a reasonable estimate of the
polarising DC amps needed to saturate the core.

Oh bum! Wrong way round.

Drive the Sec from a resistive source until the
voltage is seen to have nice wide flat tops.
Then measure the peak of the current waveform.

 

[MooseFET]

I'd appreciate a response on this. The saturable reactor is
appealing to me because it's a linear way of dealing with an
ac-transformer power-control problem.

This depends on how you define "linear". The cores of the two
transformers don't start to saturate at low currents.

As you increase the control current the saturation starts to happen
near the end of each alternation. The transformer in which the AC and
the control adds, saturates first. An AC current flows in the control
winding coupling the low impedance of the saturated one over onto the
unsaturated one. This is why I suggested a capacitor across the
windings.

As you increase the control current, the saturation point moves
towards the peak of th ewave form. This means that the control gain
is still increasing in the early part of the range.

Once you pass the peak, the gain of the control signal starts to
decrease. When you get to the point where the core is saturated with
no AC, the circuit is nearly fully one and the gain is very low. As
you go above this point, you start to saturate the core that the AC is
bucking in. This takes the leakage inductance out of the picture.

The transformers will have the I^2 * R losses due to the current. For
this reason, I don't think you could go with a much smaller
transformer.

 

[MooseFET]

I was interested in Ken's idea, so sketched a
possible circuit, as below.

+ Pri - Load current
230V high +------////////---------->-----+
T1 ======== |
DC-------////////-----+ |(
+ Sec - | |( L(leak)
| |(
.----------------' | Win's Load.
| \
| - Sec + /Rload
+---////////---------DC \
T2 ======== |
230V low +-------////////----------------+
- Pri +

Both transformers can be in the same leg. This would mean you only
need to break into one wire.

Each transformer's primary winding has to be able to
take the final max load current. So if you had a
230V/2500VA load then you would possibly use a pair of
115V/1250VA transformers.

Yes I agree.

I suspect that the control power will be far less than
the normal rated output power of the secondaries.

The B-H loop will show what DC field strength is required
to push the flux density into saturation. An old graph
that I have shows that a sample of 0.014" Silicon Steel
required a polarising field strength of about 5 Oersteds
to reduce the incremental permeability from over 2000 to
about 100.

It would be only several times the no load current of the xformer.
The core is sized so that it doesn't saturate in the no load case.
The current for the full load case doesn't represent an increase in
the field in the core.

 

[MooseFET]

I don't think it is even as high as that Win.

If you were to leave the Pri open circuit, connect a
voltage source to the Sec, and turn it up until the
current waveform had nice wide flat tops then the
measured Ipeak could be a reasonable estimate of the
polarising DC amps needed to saturate the core.

3 to 5 times the secondary magnetising current peak
is probably much lower than 3 to 5 times the peak
secondary load current.

The control power would then be the DC amps times
the winding resistances.... but see below on what
could buggerate that low control power.


There could be a little ambush in the two-transformer
saturable reactor. Those two series'd secondaries
may not cancel completely, so there could be a
residual AC voltage at their terminals. I suspect
that the DC current source must have a high enough
compliance/supply voltage such that no AC current
flows in the secondaries.

When they are not saturated, and if they came from a good maker, the
match should be more than good enough to keep this current quite
small. When you start to saturate, you want a capacitor between the
windings to pass the AC that develops because one core saturates when
the other doesn't.