First Time poster!

I am designing a heater circuit that will source a large amount of current (5A max) to a grid of resistors (each resistor gets 5A) from multiple voltage sources so that I can change the thermal "image". I am toying around with ways to get this done. By far the easiest is to use a high-current op amp like the OPA549, but that will be expensive. My next thought was to use a BJT in a Common Emmiter configuration to boost the current (I'd hope this would be less expensive).

Please let me know if I am overseeing any pitfalls, or how you think I should approach the challenge.

 

So what wattage is the heater going to be, what are the resistor values and what voltage are you using?

 

What is the voltage?
Where are you getting power? (power line, power supply?)
What are you heating?

There are options like having five 1A resistor strings in parallel. To start out you have all 5 on, then as you approach temperature, drop back to 4 or 3 or 2 strings on to hold the heat.

 

Some numbers would be good. Certainly resistors can be used as heating elements - but it would help to know more about why there are "multiple voltage sources" and how much power you are planning to dissipate. It's certainly sensible to control the power by using a PWM signal to turn transistors on and off rapidly so the average current is determined by the mark/space ratio. You could use BJTs or Mosfets. A Darlington Pair with a gain of 1,000 could be driven from a processor output

 

why use analogs to control heater? that will dissipate a lot of heat on the controls (OpAmps/BJTs) and require massive cooling system. since temperature is a slow changing quantity, the simplest thing is to use PWM. then all you need is something with sufficient number of outputs and one MOSFET per heater. you mention "heat image"... to me that suggest plenty of thermal "pixels". large number of outputs can be obtained using any MCU/FPGA with shift registers. PWM could be done all in software... easy to scale for large number of heaters.

 

You do not want a variable current at that level. The resistors should be either on or off. The percentage of on time controls the heating just as effectively. This is known ad pulse width modulation, or PWM.

Controlling them, then comes down to a single MOSFET for each resistor. And the MOSFET wastes very little energy when it is fully on or off. By comparison, an op amp supplying half the max current produces the same amount of heat as the resistor!

And why resistors? Why not use appropriate heating elements? They act just like resistors but are designed for heating.

 

why use analogs to control heater?

You beat me by minutes.

 

i do not know the nature and size of the "heat image" in this case. maybe all of the heater elements are on a PCB. kind of to defog the bathroom mirror with custom message/image if the goal is to spook a family member...

 

First Time poster!

I am designing a heater circuit that will source a large amount of current (5A max) to a grid of resistors (each resistor gets 5A) from multiple voltage sources so that I can change the thermal "image". I am toying around with ways to get this done. By far the easiest is to use a high-current op amp like the OPA549, but that will be expensive. My next thought was to use a BJT in a Common Emmiter configuration to boost the current (I'd hope this would be less expensive).

Please let me know if I am overseeing any pitfalls, or how you think I should approach the challenge.

Max wattage and voltage 10V / 3.5 A / 35W

So what wattage is the heater going to be, what are the resistor values and what voltage are you using?

Max wattage and voltage 10V / 3.5 A / 35W, so the resistance is

P = I^2 * R
35W / 3.5^2 = ~3 ohms

 

When I read the OP I'm not sure what the issue is.

Is it power control of the resistors or the switching of the resistors in the array...or both.

A block diagram would be helpful.

 

why use analogs to control heater? that will dissipate a lot of heat on the controls (OpAmps/BJTs) and require massive cooling system. since temperature is a slow changing quantity, the simplest thing is to use PWM. then all you need is something with sufficient number of outputs and one MOSFET per heater. you mention "heat image"... to me that suggest plenty of thermal "pixels". large number of outputs can be obtained using any MCU/FPGA with shift registers. PWM could be done all in software... easy to scale for large number of heaters.

There is no requirement to be analog, That was ust the first thing I thought of. That is a good point, and you hit the nail on the head with thermal pixels. I actually really like working with FPGAs and want to use them more often so I will definitely take this into consideration.

 

You beat me by minutes.

I was not expecting the thread to get so much traction, but I will also thank you for this idea. I didnt even consider bringing a digital aspect. But it is a great idea!

 

Good Lord you guys replied fast. Thank you for all the help and I will try to reply to everyone.... I am extremely overwhelmed by this though XD

 

i do not know the nature and size of the "heat image" in this case. maybe all of the heater elements are on a PCB. kind of to defog the bathroom mirror with custom message/image if the goal is to spook a family member...

The Grid is 11x11 resistors... My next question is, I will need probably need a power supply that can source enought current if all the resistors are on... which is a enourmous amount of current 121 * 3.5 = 424Amps. Is that right?

 

The Grid is 11x11 resistors... My next question is, I will need probably need a power supply that can source enought current if all the resistors are on... which is a enourmous amount of current 121 * 3.5 = 424Amps. Is that right?

I may have missed it but I'm still not seeing the power you are trying to dissipate in each resistor or heating element? You mention 5 amps - now only 3.5 amps - and the power dissipated in each resistor will be the square of the current times the resistance. So if each resistor is 1R the power dissipated could be 25.00 Watt or 12.25W Watt. (5.0 or 3.5 Amps respectively). Maybe look at using a higher voltage and a lower current to get the power you want? Again, it's important to know the maximum power you want to dissipate in each resistor to decide on the best resistor and voltage values.

FPGAs would work well as suggested to PWM 121 transistors but a processor like an arduino driving an array of eight 16 bit serial to parallel converter ICs might be simpler and cheaper. Your PWM frequency will be lower but I really don't think that will be noticeable.

 

First Time poster!

I am designing a heater circuit that will source a large amount of current (5A max) to a grid of resistors (each resistor gets 5A) from multiple voltage sources so that I can change the thermal "image". I am toying around with ways to get this done. By far the easiest is to use a high-current op amp like the OPA549, but that will be expensive. My next thought was to use a BJT in a Common Emmiter configuration to boost the current (I'd hope this would be less expensive).

Please let me know if I am overseeing any pitfalls, or how you think I should approach the challenge.

What is the purpose of heating resistors? what is the objective here? reduce your "grid" to a smaller size, say 2x2 and then draw a circuit diagram, I have no idea what it is you're trying to do here. Is the heater distinct from the resistors or do you want to use the resistor array as a heater?

 

The Grid is 11x11 resistors... My next question is, I will need probably need a power supply that can source enought current if all the resistors are on... which is an enormous amount of current 121 * 3.5 = 424Amps. Is that right?

that is correct. but... why are you stuck on high amps? heat dissipation is measured in watts...
so 25W resistor can draw 5A if powered from 5V PSU. but it could also dissipate same heat when powered by 24V by drawing only just above an amp. lower current means smaller and easier to manage wires. you don't want to get into copper bars or really heavy PCB tracks.

 

I can imagine a display based on the "Neo-Pixel" driver chip where the chips drive a transistor instead of an LED.

This would give the ability to turn on or off any "pixel" in the display and if you were clever enough even choose a power level for each.

This is just a wild idea that came to me, but I believe there could be something to it.

At first, I wanted to suggest the chip that drives the APA102 LEDs but I am unable to find chips that are not embedded in the LED COB.

Here is a link to the Neo-Pixel driver chips.

WS2811 NeoPixel LED Driver Chip - 10 Pack : ID 1378 : Adafruit Industries, Unique & fun DIY electronics and kits

 

424 Amps! Thats 4240 watts...are you trying to see this image from the Moon?

Consider ditching the resistors and use proper IR emitters like IR LEDs.

 

What makes the most sense is to drive the heaters with mains voltage and a much shorter duty cycle.Of course then there is the concern of mains connection to hardware that a person might contact accidentally. But higher voltage transistors, up to some point, are cheaper than higher current transistors. And the wires are thinner as well. And with variable duty cycle control and lower current in the on state, heat removal gets simple as well. Then if you go to alpha-numeric characters and character multiplexing the number of control devices gets a lot smaller. AND fewer interconnect wires. But much more complex control code.
So could the TS live with that concept???