Summary: I want to build a phase controlled heater controller, that’s huge – 240 Volt 200 Amps. SCR’s that size cost $150 each (would need two, and I’d really hate to blow one). Is there any way to parallel several cheap, say 25-40 Amp, triacs or inverse pairs of SCR’s?
Details: Helping my son-in-law, glass artist, with a glass melting furnace.
There are 6 heating elements all rated at 6KW, but designed to run on approximately 12 to 18 volts. They are darn near 0 ohms when cold but the R goes up as they get white hot. So the problem is starting up the furnace. (These things are lined with several inches of refractory, so the heat slowly builds up inside.)
He picked up a huge old welder for cheap and I think we can use it alright. The transformer has a 240 Volt, 198 Amp primary, and the secondary puts out about 48 volts (at darn near 1000 Amps), so we can parallel 2 sets of 3 elements in series. I want to phase control the primary of the transformer. To fire up the furnace we’d start out with the thing turned way down and then slowly over an hour or two turn it up. (I would tie in the pyrometer/thermostat too which would cycle the whole thing off and on as it reaches the desired temp, but cools down each time he opens the door.)
I’ve looked at the ST_AN308 Fig 11 diagram referenced on a recent post here. I think that’s a good one to use. What I’d like to know is since there is one small triac triggering the larger one, why could it not trigger several paralleled triacs or inverse pairs of SCR’s (say, 5 40 Amp jobs)? Wouldn’t that make all of them turn on at exactly the same moment (phase angle) - thus not causing an imbalance (which is the reason I’ve heard you can’t parallel triacs and SCR’s)?
Another question about this: Even though the measured DC resistance of the bank of cold heater coils is close to zero, do you think the 120 cycle on-off switching of the primary, along with whatever time it takes for the transformer core to saturate, will sort of act as a buffer and therefore not look like a dead short to the transformer or the line?

Thanks all for your help!

 

PS: Just looked at parts catalogs - would be best to use 8 - 25 Amp Triacs -- good part I think is ST's BTA26-600. Would be easy to mount to heat sink, cheap cost, etc. what do you think?

 

Hello - Can anyone help me with this? Is there a way to parallel triacs or SCR's?
THANKS!

 

yes, you can parallel them. however, if any one fails to function then the load on that 1 is distributed across the rest, possibly damaging the rest.

scr's backwards to each other is a triac.

as for the transformer question about zero ohms, thats DC ohms. place a impedance meter across the coil and you will measure Z, not DC ohms. inductance has Z when oscillating current flows through it. Z is the vector sum of inductance, capacitance, and DC ohms.

 

Caveat: I haven't designed and built something like this, so it's just hypothetical reasoning.

I was wondering about using a MOSFET like this device at All Electronics: STW60N10. It's rated to 60 A and 100 V. The on resistance is less than 25 mΩ, so even dropping 48 - 15 = 33 V, the power dissipation isn't a lot. To me, the attractive part is that you could get 10 of them for about $25 delivered. Then you control all of the FETs with a 0-10 volt DC signal from your temperature controller. Put one device per heating element. Screw them all to a 1/4" thick chunk of aluminum plate that acts as a heat sink.

If one device fails, the load on the others isn't affected (but the furnace won't heat as fast as it used to or get as hot).

The heaters aren't a dead short -- they have a low but nonzero DC resistance that increases as they get hot. If you have a DC power supply, hook it across the heater and set it to a constant current (10 A could be a good choice). If you don't have a power supply, then a car battery could work (you'll need something to measure the current though). Then measure the voltage drop across the heater element. A division gives you the resistance. I like to use this method to find bad connections in wiring, as you just go around the circuit measuring voltage drops with your DMM (set the current to 1 A and with a DMM set to a mV scale, you read out directly in mΩ). Using this method, I found a couple of corroded but hidden wire splices in a trailer I bought about 20 years ago -- these were the reasons the lights were dim and caused me to rewire the whole trailer.