Using SSR to switch transformer

I think - just opposite. With fast SSR you will have gradual increase current with voltage on the transformer primary - do the calculations and simulation. dI=3DE*dt/L The secondary will follow. Zero voltage crossing switching is popular in heating, lightning etc.

Probably correct- the core will tend to saturate most under those conditions.

I guess you could phase control it on.

Worst case 230V/1 ohm = 230A if you don't put a thermistor in series. Probably translates to a 50A or higher 'puck' SSR.

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

With a resistive load, you will not have an inrush current with zero voltage crossing. It will be easier to find a relay that will handle the 3.17A primary current at 230V than to find one that will handle the secondary current of 27A at 27V.

"stp"

I think - just opposite.

** YOU do not think at all.

Learn a thing or two about transformes beforee making as ASS of yourself here again.

......... Phil

With a resistive load, you will not have an inrush current with zero voltage crossing.

** That is ABSOLUTE BOLLOCKS !

YOU need to learn a thing or two about transformers too - pal.

....... Phil

"Arlet Ottens"

** Correct. ** Designing a drive cct for the SSR that will NEVER accidentally switch near zero volts is not that easy. ** Needs to be safe with a 300 amp, one cycle surge.

Better get a 40 amp one.

........ Phil

Phil is spot on.

The magic number you need to look for is usually called Ifsm - the amount of single-cycle surge current it can cope with. Phils recommendation ought to be about right, but check the SSR datasheet to be sure. or post the part you think is right, and we can check for you.

Cheers Terry

I'll bite, why does switching on the voltage zero cross cause inrush current in a transformer?

My thinking is from the old NASA incident with tape wound transformers without a gap. Sometimes when the device was turned off the core was high up on the BH loop. The next time if the startup voltage pushed it even higher into the same direction, the core would saturate for 1 cycle. In their case it blew the fuses randomly from the high inrush current. They solved the problem with a gap so the remnant flux density returned to a low value any time the device was powered off. (from Magnetics Inc literature)

If you switched at the zero cross for a low frequency signal the remnant flux density should be close to zero and therefore have high permeability for the first startup cycle. (assuming you switch on at zero cross as well.)

Where am I off on this? Is it because B laggs voltage due to B being a volt*second product?

"Mook Johnson" "Phil Allison"

** Cos it maximises the low frequency transient included in the initial few cycles of the wave train.

For zero crossing, it take one cycle for the average value to equal zero.

For peak crossing, this is met in one a half cycle.

This low frequency transient drives the core into saturation during the initial half cycle - much more so if it is extended.

If the core is already magnetised in an unfavourable direction at switch on - the surge duration so much the worse.

...... Phil

It is nearly impossible to answer unless you mention what you think is the secundary current.

Main point with semiconductor switching in the primary is the fact that the slightest imbalance will cause a DC component, so DC current to flow. Also may well need snubbers etc. Consider a normal relay :-) If you want proportional control, secondary is what I personally would prefer, a resistive load is simple, could be SCR's or a bridge rectifier followed by some MOSFET, perhaps with PWM. It will need a housing anyways, metal housing makes good heatsink.

** Can't you divide 27 by 1 - f****it ???

** Drivel.

** Yawn.....

** But the OP has his heart set on using a SSR.

** But the OP don't.

** Totally INSANE CRAP !!

At 27 amps rms !!!!!!

....... Phil

On a sunny day (Tue, 18 Mar 2008 08:31:57 -0500) it happened Spehro Pefhany wrote in :

My error, I only grabbed 1 Ohm (that man) primary. Too much speed reading :-)

At 27 A I think secondary SRC would be no problem, neather a bridge + MOSFET.

There once was a man from Erlangen..

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

That 1R is reflected into the primary circuit as 1/(27/230)^2=72R making the maximum possible inrush current due to the secondary loading

1.41*230/72=4.5 Amps peak. This 4.5 Amps then adds in quadrature to the transformer magnetizing current, both currents passing through the series ohmic resistance of the primary winding.

The big unknown is the residual magnetization of the core at turn on, but the worst case is usually zero flux because the initial permeablity, and therefore the inductance of the core, is smallest under that condition.

As you can see from the calculations, the 1R in the secondary contributes little to the inrush compared to the core initialization or possible supersaturation effects, both of which may last for many line cycles before reaching steady state. There are lots of rules of thumb on overrating the SSR, bit if you /have/ to get it right the first time, then you need a soft start module:

After reading the various bits of advice, I decided to order a big SSR (50A continuous) to build a prototype.

I'm using a small MCU to produce the control signal for the SSR, with a simple zero crossing detector for synchronization. The current in the primary coil is measured by a current probe on the scope.

After some experimenting, I now have:

- Initial turn-on done using a soft start mechanism, as suggested by Fred Bloggs, but instead of using a secondary module, the MCU is doing this.

- When the core is fully magnetized, the MCU keeps track of the direction of the magnetic field.

- Further proportional control is done by switching at zero crossings, but only when core is magnetized in favorable direction.

Even without a snubber, switching at voltage peaks only produces small transients (*much* less than switching a 100W light bulb). Inrush current during soft start phase is only slightly higher than normal. Once the core is magnetized, and the MCU switches with alternating polarity, currents are the same as steady-state currents.

Significant DC currents are easily avoided, and small ones that I added for testing don't seem to have much effect on the performance.

The SSR has no heat sink and only gets slightly warm. Much better than switching on secondary side which produces more than 25W of heat. The soft start mechanism is a nice bonus.

On a sunny day (Mon, 24 Mar 2008 17:36:05 +0100) it happened Arlet Ottens wrote in :

Some solid state relais already have a zero crossing chip. Does yours?

Beep

No, I picked one with random turn on (Crouzet 84 134 320), otherwise there's no way to control the inrush currents.

The ones with zero crossing logic seem to be more prevalent, but there are still plenty of choices for random turn on types.

I picked this big one for my prototype because it had a better chance of surviving any timing errors. For production, I'll look for something smaller and cheaper.

Your solution seems to be ideal. For my application, I cannot use a soft start, because I must apply a full current waveform to a circuit breaker to simulate a fault, and measure its response time accurately. The best I can do is time the initial application of the test voltage to coincide approximately to the expected zero crossing of current, based on the impedance of the source and the load, which are both mostly inductive.

Also, in my case, the load itself opens while the test set has voltage applied, and it may open at any point on the waveform. Thus arcing and inductive spikes are inevitable, and the transformer may have some remanant magnetism after the breaker trips, which will then cause a high inrush on the next test.

But I have considered adding a demagnetizing sequence after the trip is detected. I think a series of phase modulated pulses with gradually reduced duty cycle might reduce the magnetization of the transformer to a minimal level for the next test.

Any thoughts on that?

Thanks,

Paul

On a sunny day (Mon, 24 Mar 2008 18:21:21 +0100) it happened Arlet Ottens wrote in :

OK. I wonder if I had to do it myself if I would not simply have used a normal relay. As the load is Ohms, I would expect little problems, you will see a resistor transformed at the primary. My microwave switches a thousand times about a kW with a very small relay... Right now while it is doing the French fries.