Reduction of the magnetising current in a transfomers?

How to control the Magnetizing current..??

I have an toroidal transformer with ratio of 1:10 for 40 amp secondary at 15v.

The transformer primary Resistance is in miliohms.

I am trying to switching the transformer with triacs but the triac of 15 amp rating does not handling the Magnetising current at start up.

Can anyone having solution to avoid the inrush current at randoms.
The problem is off when i am doing the voltage dip test (40% of ac cycle)

Your problem is core saturation with specific input voltage waveforms, magnetizing current in general.

It's well known that particularly toroidal transformers have high inrush current due to core saturation if the input voltage is switched on near the zero crossing. The best way to avoid it is to switch the first half cycle at the voltage maximum (90 degree delayed).

Dear FvM

I am switching the transformers by current sensing of input current.. that is not problem

The problem occured when there is voltage sags in 20ms of cycle.
n this is uncertain. & i dont have any sense circuitry to detect the voltage sags.

Kindly Advice if you have any voltage dip Circuit of highly sensitives so i cantake a sense & trigger it to the controller & Off the Triac & i will be free from this project.

Thanks

This is one of those problems in which you DO REQUIRE the proper measurement equipment.
Which in this case means a current probe and a digital scope to capture the waveforms during the transient conditions.

By studying the nature of the current waveform, you may decide on the strategies to mitigate it.
It may be as simple as selecting a higher rated Triac (or perhaps antiparallel SCRs)

I am switching the transformers by current sensing of input current.. that is not problem

The problem occured when there is voltage sags in 20ms of cycle.
n this is uncertain. & i dont have any sense circuitry to detect the voltage sags.

Click to expand...

Your latest post clarifies that you didn't yet describe your application appropriately. There's nothing I can do about it at the present level of vagueness. I basically agree with schmitt trigger that measurement is a prerequisite for control.

Specifically, I don't understand how sensing the input current could prevent core saturation, because a triac can't disconnect the input momentarily.

A general strategy to avoid core saturation in presence of variable and unpredictable input waveforms can be to monitor the flux by integrating the transformer input or output voltage and define limits for the next cycle's phase angle depending on the present flux.

The problem appears to be one of "flux doubling".

That is where the power is switched off, and a high level of remnant flux remains in the core (it remains strongly magnetized in one direction).

If you then switch it back on, and the incoming current may try to drive the flux in the same direction
It very quickly saturates the core, resulting in an extreme and very nasty current spike.

Strip wound grain oriented silicon steel toroidal transformers are notorious for this problem.
They are not really suitable for what you are trying to do.
A normal E/I transformer with soft iron core would be much better.
It has a much softer saturation characteristic.
That plus a very large and very high current rated Triac or SCRs will be workable.

To eliminate saturation laminate steel caused by sag in voltage or unexpected dropouts, there are solutions for large power transformers but not generally tiny ones.

DC and low frequency, magnetic devices are current-controlled. At the SMPS switching frequency, they are voltage driven.

Core limitations -saturation and losses -put restrictions on flux and flux swing, which translate into volt-second limitations on the applied voltage waveforms. Split coils must be perfectly balanced to prevent flux walking . One can neutralize with a separate winding or sense current in each polarity and/or control the phase in each polarity., So for low frequency integrate the squared current for each half cycle , compare and regulate an offset for dual SCR trigger delays to balance the flux. At high frequency use voltage feedback for each polarity applied. This is a bit complicated so in some cases it is possible to use large plastic film caps in series even on power transmission lines.

The problem here appears to be what happens if you have a missing half cycle, or serous sag over half a cycle.

The core is not reset, it is then driven TWICE sequentially in the same direction, which leads to catastrophic saturation.

Unfortunately with devices that latch such as Triacs and SCRs you cannot turn the damned things off, so a conventional electronic current limit is just not possible.

The problem appears to be one of "flux doubling".

That is where the power is switched off, and a high level of remnant flux remains in the core (it remains strongly magnetized in one direction).

Click to expand...

I agree that remanent flux plays a role, but it's not the dominant problem. We are talking about soft magnetic material, remanence isn't so strong here.

"Flux doubling" occurs by just starting the input voltage at the zero crossing, by avoiding this worst case condition you can get rid of saturation without considering remanence. Details depend of course on the saturation margin of your transformer design.

"Missing half cycle" can be even worse than switching on at zero crossing, if the transformer is loaded and the flux can't reset during the missing cycle. That's why I suggested a flux monitoring circuit. I can't foresee if it works for your application, but it's worth a try.

The other point is that I won't use a toroid core transformer with primary phase angle control. It has less flux margin and lower leakage inductance (= higher saturation current) than a conventional type.

Grain oriented silicon steel is far from a "soft" magnetic material.
But soft is a relative term.

The designers of commercial toroidal transformers typically run the design flux density up to 1.2 Teslas, to minimise the steel and copper volumes for economic reasons.

A continuous wound tape wound spiral produces the maximum permeability and maximal remnant flux, as there is minimal distributed air gap.
Its the worst possible combination for asymmetrical excitation.

Warpspeed said:

The problem here appears to be what happens if you have a missing half cycle, or serous sag over half a cycle.

The core is not reset, it is then driven TWICE sequentially in the same direction, which leads to catastrophic saturation.

Unfortunately with devices that latch such as Triacs and SCRs you cannot turn the damned things off, so a conventional electronic current limit is just not possible.

Click to expand...

Very impressive warpspeed, you get the gold medal.

If the transformer current was low enough a series resistor could possibly be found that would prevent blowing the triac without too much dissipation. Let’s try the math.

40A x 15vac = 600W
600W / .9 efficiency = 666W input
666W / 240vac = 2.78A
15A triac, so limit current to 10A.
R = E/I
I am guessing here on the E, 240v is the safe number, but the transformer does not saturate right away. A lower number would work but I am not sure how low.
240v / 10A = 24 ohms
P = I^2 x R, here I am more concerned with how much power dissipation at the normal 2.78A.
= 2.78^2 x 24 ohm = 185W, WOW, way too much. This is unrealistic!

Try backwards approach, how much power do you want to waste on resistor. My guess 5W.
P = i^2 x R, taking a guess at values
2.78A^2 x 1 ohm = 7.73W
2.78A^2 x .75 ohm = 5.8W
2.78A^2 x .5 ohm = 3.86W
These are very low values. The actual voltage drop at 10A would be.
E = I x R
10A x .75 ohm = 7.5V
I do not believe this low value R would help with the saturation problem. I would say a dropping resistor is not a viable solution.

An inductor on the input might work but it’s size and cost would be prohibitive. I think some kind of active current limiting is needed.

Does anyone know how to do the math for the inductor? Here is my guess. L = di/dt = 10A worst case current, 10A/.020 seconds (20mS was given as drop out time) = 500 Henerys? Yikes!

I have designed more than a few successful commercial battery chargers that used phase controlled SCRs in the primary.
The trick is to use soft iron E and I laminations, and be very conservative with the transformer flux density.

If the SCRs also have a sufficiently high half cycle peak surge current (I squared T rating) the whole mess can then hopefully ride through a flux doubling event, with reasonable safety.

I would never consider using a commercial store bought toroidal transformer for this type of primary phase control application, its just tempting fate too much.

I would imagine using <<50% of the Bmax is a safe design goal.
Microwave Oven transformers (MOT's) are very inefficient but can use a higher % of Bmax. Whereas power Distribution transformers (DT) are often designed with <30% margin resulting in fast trip breakers and

Power Transformers (PT) ( >10MVA) have even less margin due to cost of CRGOS laminate. This margin to saturation is a great concern for all magnetic materials as insulation breakdown and massive forces occur from the saturated coil currents often peaking in tons of force on the wires in large PT's. So expensive protection controls are necessary with smart thresholds.

FvM said:

"Missing half cycle" can be even worse than switching on at zero crossing, if the transformer is loaded and the flux can't reset during the missing cycle. That's why I suggested a flux monitoring circuit. I can't foresee if it works for your application, but it's worth a try.

Click to expand...

Are there any circuit which can monitors the flux.?
Can you share how it can be implemented..?

A flux monitoring circuit can only start ringing the alarm bell once its too late.
You cannot turn your SCR or triac off, so its not really a practical solution.

A flux monitoring circuit can only start ringing the alarm bell once its too late.
You cannot turn your SCR or triac off, so its not really a practical solution.

Click to expand...

The idea of a flux monitoring circuit is not to turn the triac off when the maximum flux is exceeded but to limit the phase angle of the next half cycle, as mentioned in post #5.

This means e.g., that the next cycle is only allowed to be triggered at 0 degree, if sufficient opposite flux is present in the transformer. Starting with zero flux, the next half cycle would be limited to 90 degree phase angle.

O/k, lets assume a perfectly normal half cycle, no flux limit occurs.

Next half cycle is completely missing, still no excessive flux, and no flux reset either.

Then you turn on your SCR for the following half cycle at the normal time (whatever that is) still no excessive flux at turn on.

Then during that half cycle the current spikes up incredibly high because the core has saturated.

Your flux sensing/overcurrent sensor suddenly screams in agony, but what can it do to unlatch the SCR ?

Its too late do do anything.

Sorry, you just didn't seriously think about the solution.

Next half cycle is completely missing, still no excessive flux, and no flux reset either.

Click to expand...

After a missing half cycle, you have remaining flux from the previous half cycle, consequently the next half cycle must be completely blocked. Perhaps "flux prediction" would be a more exact description.

I have used the flux monitoring method (many years ago) and it worked well.

FvM said:

After a missing half cycle, you have remaining flux from the previous half cycle, consequently the next half cycle must be completely blocked. Perhaps "flux prediction" would be a more exact description.

Click to expand...

Its an interesting concept for sure.
I have never seen this done though.
Can you link to any circuit examples of predictive flux monitoring and shutdown for phase control you know of ?

My own preferred method is just to design suitably robust old school magnetics with ample soft iron and turns, and adequate oversized SCRs.
Keeping it simple (and bullet proof) has proven very successful and reliable for me in the past.

All these problems seem to only have occurred in recent times since transformer designers greedy for a dollar, have been pushing these tape wound toroids to their absolute limit.

I agree that the "old school magnetics" approach is preferable.

Unfortunately I don't have schematics of my old project, the flux monitoring/prediction was implemented as part of an analog SCR trigger controller. These days, I would most likely put it into a microcontroller based design.