Need professionally made circuits to control heaters using PWM and MOSFET

[Eskay]

Hi, this is the first time for me to post on this forum. Also, I consider myself a newbie in the electric circuit in general. I am a software engineer but also need to take a part in electronic design. This is a bit general question rather than specific one.

My project requires to use 1kHz PWM signal to control ceramic heater power. After some research, I was able to create a simple circuit board using resistors, MOSFET, diode, and perfboard. It is a basic circuit that can handle 1kHz pwm. The circuit is designed to support up to ~300W and it is working well. Our target is now changed to support heaters up to 1000W. I think I can achieve that by using a higher power MOSFET and heatsink. The project will be eventually integrated into a product which will be used by a customer.

There are many concerns though, to make further progress in this project. The reason is because no one is expert in electric design in the group including myself. There is no way to ship a product that includes my own soldered perfboard I believe there are many many companies out there who can:

1. Create circuit board for me based on my design
2. Create circuit boards for me based on my requirements (without me having to design circuit)
3. Sell existing circuit boards that already do the same thing (because this particular circuit should be very basic).

However, I do not know where to start. I need this design to be 4 channels, so I can individually control 4 different heaters using individually controlled PWM signals. It means I need to 4 MOSFET circuits to be placed in a single board.

Please advise how I should make progress. Thank you in advance.

 

[AnalogKid]

Many people on this forum have the tools and experience needed to do this, but you won't get any useful feedback without posting the schematic. Also, what is the environment for the resulting product? Indoors / outdoors / extreme cold / no ventilation / power source / etc?

ak

 

[Arouse1973]

A data sheet of the heaters would also be good.
Adam

 

[Eskay]

Our controller uses NI 9474, which is 24V Sourcing Digital Output Module. This is used as a 1kHz-PWM generator. Our current heater is 24V180W DC heater. I don't have the spec sheet for the heater. I use the diode to protect from the reverse current (I learned this through Google). 1KOhm resistor was originally 1MOhm, but it did not work well because the switching time was too slow for 1kHz PWM. I replaced it with 1kOhm to make the discharging quicker.

Here are some concerns:

- This worked well for the specific heater, but will it work for 1000W heaters?
- How can I prototype such circuit for 1000W heater? Perfboards can handle that much current of 40A?
- Is this a good design first of all?
- Is it good idea to drive a heater like this using 1kHz PWM?


Thank you in advance!
Eskay

 

[duke37]

Give us some more information.
1000W at 1kV = 1A
1000W at 10V = 100A
The catch diode needs to be able to pass the load current for a short time.
If this drives a heater, then 1kHz seems to be plenty fast enough, what is the thermal time constant?
You could parallel several MOSFETs to lower the heat loss and to spread it. What are the data for the FETs you intend to use?

No printed circuit board will be able to take the higher currents you want.

 

[Colin Mitchell]

"No printed circuit board will be able to take the higher currents you want."

What do you think all the 40 amp printed circuit boards are doing???????

 

[Eskay]

Thank you everyone for posting to this thread. As you may have found out by now, I am still on the serious learning curve.

ak, it will be used indoors
Adam, we do not know what the heater will be yet. We are looking at this
duke, thermal time constant affects the circuit design?

Up until now, I have been only considering DC circuit as it is much easier for me to understand. But using a DC circuit means I have to live with low voltage like 24 Volts. Does a use of 120VAC make life much easier because then I can lower the current?

I am looking at heaters like this one:
http://www.omega.com/pptst/CSH1_SERIES.html?pn=CSH00020

The heater will be connected through a wire (~ 1m long) to PCB.

Thank you
Eskay

 

[Colin Mitchell]

Look . . . I think this whole requirement is way over your head and we are not going to be able to help someone who is transferring 40 amps in a design.
You have to look at electrical emissions and get the whole project certified before you can sell it to the public.
Where are you going to get 40 amps from??????
There are just too many complications for you.

 

[AnalogKid]

Sticking with a 24 Vdc system:

1. Increase the 10 ohm resistor to something greater than 200 ohms to limit the peak gate voltage (Vgs) to fewer than 20 V. 10 V is the typical value for specifying the minimum Rdson, but check the FET datasheet.

2. 42 A is a lot for a typical pc board construction, but the necessary parameters are well within most board house capabilities. The simplest thing is to increase the copper weight to something like 4 oz. and pay attention to the trace widths. If your favorite shop can't do the extra weight, add extra layers. There are simple layout techniques to make sure the current gets to the interior layers. 42 A isn't cheap, but it also isn't tremendously difficult.

3. Yes, a single 75 A MOSFET can handle 42 A, but you have a significant and unnecessary dingle point of failure. Better to spread the device power dissipation among 2 or 3 50 A or 75 A parts. You can direct-parallel the drains (and of course the sources), but not the gates. Use independent resistor networks for each transistor.

3a. Ballast resistors - something for another post. Critical for bipolar power transistors, not nearly so much for MOSFETs. Others will weigh in.

4. Anderson and others make connectors specifically for low voltage high current DC. Fir long term reliability, think 75 A minimum rating per contact.

5. AC powered heaters will use proportionally less current, but now you should vary the conduction phase angle of a TRIAC, just like a standard light dimmer. The 9474 is fast enough, but it will take different programming.

ak

 

[duke37]

Since the heaters have a time constant of several seconds then the switching does not have to be rapid.
I think that the best solution would be to run the heaters on 110V AC and switch with a solid state relay. There will be no massive currents on your PCB and no large power transformer or rectifier.

DC does not have to be low voltage. Most switch mode power supplies use rectified mains which in the UK is 340V. If you want to go to DC, then it would be possible to use it unsmoothed so not needing a lot of large expensive capacitors.

Phase angle control would work well but may give interference problems If you switch once every few seconds, the problem will be less and if switched at zero voltage, there should be no problem.