Re: [EE] Triac phase control power controller designs

My thanks to everyone who replied.  I am glad to know this list is still 
active.

My project is to produce a triac power controller for a large mains 
solenoid which is for a lab demonstration of electromagnetic induction. 
The solenoid's existing controller currently uses a Variac which is not 
ideal for various reasons.

The first attempt used an off-the-shelf controller; a CSR2-15E from United 
Automation. It is a two-wire device, like a light dimmer but for up to 
15Amps.  The problem was the solenoid often randomly received an initial 
pulse when powered on, even when the power control pot was wound down to 
minimum.  No similar pulses were seen when the load was just a 100W GLS 
mains lamp, used for testing purposes. I believe this was due to the 
inductive nature of the solenoid and the point in the mains cycle at the 
instant the controller was powered on. If/how its phase/timing capacitor 
was already charged at the instant of power-on might also have been a 
factor.

After measuring the electrical parameters of the solenoid I found it was 
highly inductive. Its j*w*L component was 6 times its DC resistance. It 
seemed reasonable to conclude the solenoid was just too inductive for the 
CSR2-15E. The CSR2-15E's datasheet says it is 'suitable for some inductive 
loads such as transformers, fans and motors.'

After that I read ST's AN308 application note "CONTROL BY A TRIAC FOR AN 
INDUCTIVE LOAD HOW TO SELECT A SUITABLE CIRCUIT". It discusses 3 basic 
designs; Firstly the classic two-wire light dimmer type where the triac is 
triggered by the voltage across the load. Secondly a three-wire design 
where the triac is directly triggered by the mains voltage.  It is 
recommended for inductive loads because the triac triggering waveform is 
not determined by the I/V phase lag etc. across the load.

Using the 3-wire circuit from AN308 does work. I found a couple of 
problems.

The first problem was hysteresis in the power as the control pot was 
turned up and down.  This is not a result of the inductive load as the 
same results are seen when the load is just the 100W lamp.  There is a 
neat circuit using 4 diodes in Littelfuse's "Teccor brand Thyristors" 
AN1003 application note (Figure AN1003.12b) to deal with this problem. It 
turned out to be quite effective.

The second problem is a discontinuity / sudden jump seen in the 
voltage/power applied to the load as the control pot is turned up and 
down.  This one I have been unable to get rid off.  The effect is very 
noticeable with the 100W bulb and would doubtless affect the solenoid too 
in an unacceptable way.

I have tried using both a 'standard' BTA26-600BRG triac and a 
'snubberless' type BTA24-600BWRG in case 4th quadrant triggering etc. was 
a factor but it made no difference.

I also note in passing that rewiring the circuit (with & without the 
additional 4 diodes anti-hysteresis circuitry left in place) to move the 
100W lamp load from the triac MT2 circuit to the mains input (i.e. 
changing from a 3 to a 2 wire design) gets rid of the discontinuity / jump 
in power (JIP) -- with the lamp brightness/voltage varying smoothly 
throughout.

Observing what the circuit is doing when the discontinuity / JIP occurs 
with a scope is difficult without either a mains isolation transformer to 
power the circuit or a battery powered (and therefore not grounded to 
mains earth) scope.

I did get a very basic single channel battery powered DSO and monitored 
the triac gate voltage with it. What it showed as the power was adjusted 
from just below the discontinuity / JIP to just above it was interesting. 
With the control down at minimum and the triac untriggered the gate 
voltage is just a 50Hz sinewave.  Once the triac starts to trigger and the 
lamp comes on dimly a single dip appears e.g. in the +ve & -ve going mains 
1/2 cycles of the gate voltage sinewave. The control pot is advanced until 
the JIP (lamp suddenly become brighter) occurs, then the triac gate 
voltage acquires a 2nd dip in voltage during the +ve/-ve going 1/2 cycles.

This observation made me wonder if the problem was related to the triac 
holding current.  However the datasheet gives I_L as 70mA to 80mA whereas 
the JIP occurs at significantly higher currents, e.g. 250mA.

The other thing I wondered about was whether the triac gate triggering 
pulses could be too short to work properly. ST's AN303.pdf "Thyristors and 
TRIACs: latching current" application note discusses this.  It recommends 
a low value, e.g. 33R resistor in series with the gate to stretch the gate 
pulse, e.g. by approx. T=R*C where R is the gate series resistor and C the 
capacitor in the phase control circuit which discharges into the triac 
gate via a diac.  I tried a range of series resistors between 39R and 1k. 
Subjectively the JIP appeared less marked for some of the higher value 
resistors but it certainly was not eliminated.

I also tried simulating the circuit in Spice, using ST's models for the 
triac and diac.  The simulation of the power delivered to the 100W lamp as 
a function of control pot resistance showed a smooth curve with no 
discontinuities/jumps.

I have not used Spice before starting this project. My question is should 
I reasonably expect the problems in my circuit to be reproduced in a 
simulation, or are the manufacturer's models effectively an idealised 
version of these semiconductors, especially if they are being used in an 
unusual/unintended part of their characteristic?

Apologies for this rather long initial post. I have been working on this 
project on and off for months now with limited success.

Am I missing something obvious? Any thoughts on how I might best proceed?

Thanks
Tom Crane


On Tue, 13 May 2025, Harold Hallikainen wrote:

> As others, I check the list every day.
>
> On phase control dimmers, I look forward to your questions. I designed all
> the dimmers (and controllers) at https://www.dovesystems.com/ .
>
> Harold
> https://w6iwi.org
>
>
>
>
> -- 
> Not sent from an iPhone.
> -- 
> http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> View/change your membership options at
> https://mailman.mit.edu/mailman/listinfo/piclist
>