Heater Power Supply - SCR Phase-Angle Control

Thread Starter


Joined Aug 7, 2020
Have this schematic for a Heater Control. Please direct me to an explanation for (R1 + C1/C2) and (R2 + C3). I cannot find any online papers about how to implement phase-angle control. I can only find lots of explanations as to what it is.

Specifically, I want to know how would I modify them for a different frequency or voltage input? Would they need to be changed for a different output transformer? Would they need to be changed based on the expected phase angle usage (say 50-80% power; not up to 100%)? Are they a good replacement for a snubber that is only across the SCRs?

I have looked online for controllers to purchase, but they are all 50-60Hz at >8A output and their OEMs declare they will not work on 400Hz.

And a follow on question: this particular "external transformer" is long obsolete. Do you know of a current (haha) manufacturer of similar transformers?

I much appreciate your help!


schmitt trigger

Joined Jul 12, 2010
(R1 + C1/C2) is simply a snubber.
R2 is a dummy load such that the SCRs have a minimum holding current.
C3 is simply a filter.

The problem with SCRs is that they are slooooooow switching devices. Many years ago, close to 40 years, we used inverter grade SCRs when switching at higher frequencies.
I ignore whether they are still being produced, as IGBTs have superseded them for the low to medium power range. They are still being produced for massive voltage and current ratings.

So I would suggest that you consider IGBTs also.


Joined Aug 21, 2017
Any real heater is veeeeery slow device, therefore to apply the phase controlled SCR is the last thing one ought to do. If not want to have at home visit of radio-police inspectors trying to understand who demolishes the mains line signal form use anything other with zero crossing the keyword. So, better instead to evaluate the pulse count method with zero-crossing thing. For small power it will let the one half period goes through and then long and long silence. For near the 100% power it will let the many halfperiods go through and then one of it stopped to. Very nice regulator, simple at build, accurate, and even 100 kW are not demolishing the mains EMI.


Joined Jul 1, 2009
@slewelln @Janis59

Jans59 is correct. The correct way to do this (if you are wishing to effectively use PWM) is by using a zero-cross detector so that you can trip the SCR when everything is at ground, and then disconnect it some time after. This is how many light-dimmer-touch circuits work. However, you have to use an SCR that will respond at the frequency you need. Others have suggested IGBTs... I would examine those as well.

Determine your constraints; Align them with your goals; choose parts accordingly. Always put in safety margin in component currnet/voltage values. Always consider how it might fail and how to make it do so to a safe condition.


Joined Aug 21, 2008
You can use a Triac. Better than an SCR because it can conductor over a full cycle without the need of other components to help it do so. About 10 years ago I had an on-demand water heater for a shower that made the water too hot. The tightwads that managed the construction of our house used an under-powered water pump and very small diameter PVC pipe for plumbing. That, and the only guidance I received when buying a water heater (no this new house did not come with water heaters!) was the maximum current it could pull. The result was too much power going into too little water per unit time, meaning that the shower was scalding hot unless the hot side was turned down, in which case the over-temperature sensor on the heater would trip.

Over-thinking the problem brought about this solution:


Note: The triac is an ST Semiconductor "snubberless triac" which did not require a snubber circuit.

The circuit worked well for years until one day my wife opened up the holes in the showerhead, thus allowing more water per unit time to pass through. So much for the high-tech solution! The power reducer module is now collecting dust in the junkbox.

This has two settings, one for summer which phases back the duty cycle of the conduction time and one for winter (yes, both weeks of it here in the tropics!) in which the dutycycle is close to 100%. To find the values for the RC timing network I measured the water temperature reduction I needed by measuring the temperature of the painful full-power hot water and the temperature of the full-power water mixed with unheated water that felt fine. That gave me a temperature difference in °C.

I knew that the heating element was resistive and the nonlinearity because of the thermal coefficient of resistance could be ignored because the temperature difference was very small.

Knowing that temperature rise over seasonally varying input unheated water temperature is nearly linearly proportional to power, using a spreadsheet I made a table of % power vs conduction angle (related to duty cycle). That gave me the conduction angle.

I don't remember how I came up with the trigger delay but it probably involved the firing voltage of the diac on the triac gate.

I hope this gives you some good ideas though admittedly it is not a complete solution.