Dear all,

I'm trying to build a controller system for a heating system, based on a microcontroller.

I would like to actually implement a PID control instead of an ON-OFF one by using a temperature sensor and controlling the RMS current of the heating element using the PWM output of the controller and a .... And here is my problem!

I need to be able to PWM the current of a 2kW heating element, but I fear (maybe wrongly) that the current is a bit high to use a TRIAC on an inductive load.

I have also thought of using an IGBT or thyristors in anti-parallel. But in reality, I have a tough time deciding which is the safest, easiest way to go.

What would you recommend?

 

What about using a large solid state relay? I have seen them used in electric thermographers to contol the heating elements.

__________________

 

A quick search shows Triacs as large 100 amps. There are many in the 40 amp 400 volt range and 20 amp 400 volt.
Use a triac.

 

A pure heater isn't inductive, but you'll definitely need a filter to control interference (The filter will contain moderately sized inductors). I'm guessing that at 2kW you'll be controlling 220VAC, which is only 10A. A 20A triac, properly heat sinked, is easily accommodated.

 

normaly if the triac is properly cooled (you better can over do it than to less)
it should not be a problem
make sure you get at least a 600V triac (220V AC is peak to peak is 440V)

a 47 nF capacitor and a 100 hm: resistor over the triac terminals A1 and A2 should protect it from spikes on the net

Robert-Jan

 

You don't need filters on heaters, you use burst-fire power control, and only switch at zero-crossing. So no interference, and no need for filters.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

I've quickly checked the controller specs, and I do not think that it can as low as 100 Hz for the HW PWM output, which either leaves me two possibilities (I think):
Soft PWM using a triac (Do I need a snubber circuit of some sort? I've never measured the cos phi of a standard "commerce" 220V heater)
Using an IGBT connected to the HW PWM output, cycling at perhaps 1kHz. However, I fear that I need to install RFI filters on the mains side to avoid pumping the house mains full of harmonics, and probably some filter on the heater side to avoid having the harmonics break the insulation (as happens in "unprepared" motors when using a frequency converter).

Am I missing something here?

 

Yes, why can't you use burst-fire?, what sort of heater is it?, you also can't use PWM really on AC, you would need to rectify and smooth it first.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

The heater is a standard resistive one (the ones you buy at supermarkets - a spiralled cable - I believe it should actually be referred as inductive rather than resistive).

I guess I can use a triac triggered at the "right" part of the cycle (is that what burst-fire means?), but I would need to "track" the cycle or at least synchonize the controller with it, right? Is there a way around this?

I have this heater controller project, but I'm also trying to learn. So, I have one further question: Why can't you "chop" AC with an IGBT suing PWM at a frequency that is a multiple of the mains to get a lower RMS voltage (and lots of harmonics, that is for sure)?

 

You would be wrong, it's resistive - it will have a slight inductive component (so does a piece of straight wire), but's it's MUCH too small to be of any concern.

Yes, you only turn the triac ON at zero-crossing, so there's no interference generated - burst-fire means you turn it ON for a number of cycles, then OFF for a number of cycles. However, for temperature control such a low level isn't normally needed.

Because if it isn't syncronised to the mains the actual power will vary as it drifts across the mains waveform - and even if it is syncronised, the power won't change in a linear fashion. It's a LOT more work, a LOT more expense, for poorer performance, and seems totally pointless?.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

Umm... If it's a plain old heating element, why can't you just use DC? I'm sure the heater rectifies the AC anyway? Mains is 240VAC-RMS so you need 240VDC, which can be achieved by rectifying the mains - although this will result in a voltage slightly higher than 240V it shouldn't do any damage to your heater. You could then just PWM it with a FET?

Correct me Nigel/Someone Else if what I've suggested will result in pyrotechnics.

__________________
What is a joule per second?

 

You might be 'sure', but you are almost certainly completely wrong - heaters run off AC, there's no advantage to DC at all (and there are disadvantages).

Like I said though, PWM'ing isn't the way to go, it gives no advantages, and lot's of disadvantages.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

Thanks a lot for the clarifications!

I want to be able to do a PID control (Ok, I know, it is an overkill, but the reason I want to do the controller myself is also to learn). So will probably define a control time (for example 200 half cycles), and turn it on a number of cycles out of the total that will be PID (probably mostly proportional, but might implement derivative and/or non linear elements later on just to apply what I've studied in control theory classes ) to the difference between the target and the measured temperatures.

How could I implement a circuit that gives me a TTL 0 (or 1) when the mains are at zero crossing (to avoid using an A/D converter)?

 

Zero-crossing is simple to do, and certainly doesn't require an A2D, here's a simple method I posted ages back.

Attached Images

zero_crossing.GIF (4.2 KB, 32 views)

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

Would a zero crossing detector be of any use for extending relay contact life?

__________________
Bill
Smart Kits build Smart People
http://www.blueroomelectronics.com