Mains half cycle skipping to regulate heater element

[ocset]

Hello,
We have an electric heating element connected to the mains. We wish to regulate the temperature by skipping half mains cycles, as in the attached waveform (LTspice sim also attached).
The power factor is great because mains current and voltage are always in phase. But is there any problem caused in the mains electricity system by this ON/OFF type waveform?

[NiHaoMike]

Zero cross switching is the standard way to do it. Only downside compared to phase angle control (generates lots of EMI) is the slower response, but that's rarely important for heating applications.

[Brutte]

What is the time constant of that thing that you heat?
For fast control you have to use triac which requires some additional protection + drive. Things get simpler with dumb relay but then this is for big time constants.
There are triacs which already have zero-cross detection but the cheapest version just uses raw triac + uC that does zero-cross detection.

[ocset]

Thanks, the  attached two waveforms show  the skipping of 10ms half cycles…in one, every fifth half cycle is skipped, in the other, every sixth  half cycle is skipped.
As you can see, when every sixth half cycle is skipped, there is an overall DC current in the mains supply system…do you think this matters?

[Kleinstein]

The overall DC contribution can be a problem, if the distribution transformer is relatively small. A DC component of the mains can cause problems with transformers (hitting saturation to one side).  AFIAK there are limits how much such DC background is permitted. I would not worry about a 50 W soldering iron, but a 10 kW warm water heater.

[IanB]

Thanks, the  attached two waveforms show  the skipping of 10ms half cycles…in one, every fifth half cycle is skipped, in the other, every sixth  half cycle is skipped.
As you can see, when every sixth half cycle is skipped, there is an overall DC current in the mains supply system…do you think this matters?

Why don't you skip whole cycles? Avoids the whole problem of power factor and DC bias.

[Cerebus]

The overall DC contribution can be a problem, if the distribution transformer is relatively small. A DC component of the mains can cause problems with transformers (hitting saturation to one side).  AFIAK there are limits how much such DC background is permitted. I would not worry about a 50 W soldering iron, but a 10 kW warm water heater.

Quite. The easy way to avoid a DC component is to zero-crossing switch whole cycles. For pretty much all applications where half-cycle skipping is acceptable in time constant terms, whole cycle skipping will also be acceptable.

The only thing to watch out for in either scheme is the EMI generated by the 'sharp edges' where you turn the signal on from zero or off to zero, they can have surprisingly high frequency content if you're using a good, fast switch - easy to fix with a little bit of capacitive bypassing though. It pays to have your load well characterised (some heaters can have non-trivial inductance) and, if necessary, to take it into account in devising your EMI suppression and switch protection.

[Gyro]

Thanks, the  attached two waveforms show  the skipping of 10ms half cycles…in one, every fifth half cycle is skipped, in the other, every sixth  half cycle is skipped.
As you can see, when every sixth half cycle is skipped, there is an overall DC current in the mains supply system…do you think this matters?

Why don't you skip whole cycles? Avoids the whole problem of power factor and DC bias.

Agreed, that is the optimal (normal?) way of doing burst-fire control.

[floobydust]

Normally, heater control is done using many cycles on, many cycles off PWM because a heating element has a large time constant (thermal mass). You would not skip a cycle.
Net DC offset is not permitted in the high power heater standards to protect the distribution transformer. It depends on how many kW this is, hot water or what.

[IanB]

Normally, heater control is done using many cycles on, many cycles off PWM because a heating element has a large time constant (thermal mass). You would not skip a cycle.
Net DC offset is not permitted in the high power heater standards to protect the distribution transformer. It depends on how many kW this is, hot water or what.

It might be the instant heat electric shower from the previous thread. If so the action should be reasonably fast or the water will run hot and cold.

[floobydust]

What is the time constant of that thing that you heat?

We don't know the application, wattage, or the thermal time constant.
Your standard treez thread that wastes everyone's time guessing details. 10msec off time is nothing for a heater.

[ocset]

Thanks, yes, its for an domestic electric shower.
Thanks, do you think the attached mains input current waveform would pass EN61000-3-11?
Its 32ARMS.
If this were the mains input current waveform to a device, would it fail regulations?

One whole mains cycle has been "sliced out" at the zero crossing, every 5 mains cycles.

[ahbushnell]

Thanks, yes, its for an domestic electric shower.
Thanks, do you think the attached mains input current waveform would pass EN61000-3-11?
Its 32ARMS.
If this were the mains input current waveform to a device, would it fail regulations?

One whole mains cycle has been "sliced out" at the zero crossing, every 5 mains cycles.

That should work.

[Phoenix]

Its 32ARMS.
One whole mains cycle has been "sliced out" at the zero crossing, every 5 mains cycles.

This looks like how I would go about making flicker if I intentionally wanted to.

As for 61000-3-11, you have a simulation, put in the reference impedance and measure the result. Compare to the limits.

[Syntax Error]

Are there any inductive results from pulsating a heating element with 8Kw of raw power? If so, would this set the current and voltage out of phase enough to break the triac in the blanking gap?

[temperance]

Part from anything else: if this is for a commercial product and you have to pass some regulations, then pulse skipping can be ruled out as a solution because it will probably generate to much flicker. You might want to investigate this before you continue.

As you might find out, passing CE regulations is not a matter of luck.

[NiHaoMike]

The voltage sag is going to be the same whether switched with a relay or a solid state switch. If the supply is up to spec, the sag should remain within limits. If not pulling a substandard supply out of tolerance is important, the only reasonable way is to actually reduce the load such as by dividing the heater into multiple parts and switching them separately.

If you're worried about many units synchronizing and causing a huge peak load, you can randomize the switching based on a PRNG seeded by reading the LSB of the microcontroller's ADC many times on startup.

[David Hess]

At high power levels, half wave rectification can cause excessive saturation current in the transformer.  If the power line cycling is random, then I would not worry about it.

[Cerebus]

At high power levels, half wave rectification can cause excessive saturation current in the transformer.  If the power line cycling is random, then I would not worry about it.

Not to pick on you David, but I'm constantly surprised how quickly on this (possibly any) forum people start rehashing points already made or answers already given. The whole DC saturation/half cycle switching/full cycle switching thing was raised and dealt with in message #4 et seq. It does seem to happen sooner/faster on here than on many of the other fora that I've participated in.

If there were pages and pages of previous discussion I can understand how it might happen, but with all of 16 messages (less than a page) in total when David commented, it seems a bit surprising for it to happen - surprising to me at least. Do people just not read what's already been said before commenting?

[splin]

Part from anything else: if this is for a commercial product and you have to pass some regulations, then pulse skipping can be ruled out as a solution because it will probably generate to much flicker. You might want to investigate this before you continue.

As you might find out, passing CE regulations is not a matter of luck.

For 'it will probably generate to much flicker' I'd suggest a better phrase would be 'it will almost certainly generate too much flicker'. You can get a free simulator here: https://www.solcept.ch/en/tools/flickersim/

If it's a one off then I wouldn't worry too much as very few people use incandescent bulbs any more and if they do the more flicker the better to encourage them to replace them! 

I guess some LED lights could also be susceptible to flicker though depending on the type of driver.

[David Hess]

Not to pick on you David, but I'm constantly surprised how quickly on this (possibly any) forum people start rehashing points already made or answers already given. The whole DC saturation/half cycle switching/full cycle switching thing was raised and dealt with in message #4 et seq. It does seem to happen sooner/faster on here than on many of the other fora that I've participated in.

If there were pages and pages of previous discussion I can understand how it might happen, but with all of 16 messages (less than a page) in total when David commented, it seems a bit surprising for it to happen - surprising to me at least. Do people just not read what's already been said before commenting?

I should have paid closer attention but if the forum software tracked and displayed which replies have been read, like email and news forums have for *decades*, then it would be easier for the user to know which messages are new and which they have already read before.  Discussion forums on the internet have taken steps backwards in usability since Eternal September.

[fourtytwo42]

Not to pick on you David, but I'm constantly surprised how quickly on this (possibly any) forum people start rehashing points already made or answers already given. The whole DC saturation/half cycle switching/full cycle switching thing was raised and dealt with in message #4 et seq. It does seem to happen sooner/faster on here than on many of the other fora that I've participated in.

If there were pages and pages of previous discussion I can understand how it might happen, but with all of 16 messages (less than a page) in total when David commented, it seems a bit surprising for it to happen - surprising to me at least. Do people just not read what's already been said before commenting?

I should have paid closer attention but if the forum software tracked and displayed which replies have been read, like email and news forums have for *decades*, then it would be easier for the user to know which messages are new and which they have already read before.  Discussion forums on the internet have taken steps backwards in usability since Eternal September.

I wouldn't get excited up about it, after all its a treez post and nobody reads those expecting anything serious 

[David Hess]

I wouldn't get excited up about it, after all its a treez post and nobody reads those expecting anything serious 

I think it was Kleinstein's post he was referring to but I am not worried about it when the forum structure encourages it.

[ocset]

Interestingly, The following electric shower..
https://www.mirashowers.co.uk/showers/electric-showers/mira-decor-dual-108kw/
has an extremely fine resolution in its temperature setting, so it must be doing mains burst fire control…….there is pretty well no other economically feasible way to do fine adjustment temperature control in an electric shower.

[Gyro]

More likely, a decent quality thermal wax capsule flow regulator, as used in most modern mixer showers.

Manufacturers will go simpler and cheaper, rather than excessive complexity.