Author Topic: Transistor Heater Issues  (Read 2648 times)

0 Members and 1 Guest are viewing this topic.

Offline noname4meTopic starter

  • Regular Contributor
  • *
  • Posts: 93
Transistor Heater Issues
« on: September 23, 2018, 10:54:44 pm »
Hi

I'm trying to build a water heater based on a 2N3055 transistor's power dissipation.



Please note: R1 is a 10k NTC thermistor, R6 and R8 are part of a 10k pot, R9 is a power resistor mounted on a heatsink, and the 2N3055 is also mounted on a heatsink for testing.

While testing this, the 2N3055 keeps dying when the current reaches about 2A.  I have tried increasing the 0.5 ohm power resistor, but then the power is dissipated in the resistor instead of the transistor.

My intention is to dissipate up to 50W in the transistor so that it will heat up whatever it is connected to.  The thermistor will be used to turn down the power so as to keep the heat at a set level dictated by the pot setting.

Can someone please help me understand how to design something like this?  I did some sums and even used LTspice, but it is not working like how I expect.

Thanks
« Last Edit: September 23, 2018, 10:56:23 pm by noname4me »
 

Online Andy Watson

  • Super Contributor
  • ***
  • Posts: 2074
Re: Transistor Heater Issues
« Reply #1 on: September 23, 2018, 11:30:40 pm »
Look up "safe operating area" with respect to transistors. You need to ensure that you are operating within this area. Perhaps more pertinent, you need to ensure that there are no circumstanses in which the transistor will be operated outside that area. For example, what happens if the 2N3055 is turned hard-on? It appears that you have a 20V power supply driving into the 0.5\$\Omega\$ resistor - so ignoring losses, that is a current of 40A. Off-hand I cannot remember what the upper limit of a 2N3055 is, but I am certain it is less than 40A.
 

Offline Cerebus

  • Super Contributor
  • ***
  • Posts: 10576
  • Country: gb
Re: Transistor Heater Issues
« Reply #2 on: September 23, 2018, 11:32:52 pm »
First guess is that the whole thing is turning into an oscillator and you're not spotting it because you're making DC measurements (or AC measurements of too low a bandwidth to see the effect on a handheld meter). A current limiting resistor or a ferrite bead in the 2N3055 base lead may help to suppress this, a small emitter resistor wouldn't hurt either, but your wiring's innate resistance may be sufficient for the latter at high power levels. Stick a scope on and see what it's doing.

Some capacitive feedback around the op amp to limit its bandwidth is probably in order too, but this won't (necessarily) cure the oscillations, if that is the problem, but the whole physical system will have slow response and you'll need the op amp response to be quite slow too or it will just overshoot and force the output into hard saturation which may be enough in and of itself to put the transistor outside its safe operating area. Again, a scope would let you see what is happening.

Adding a current limiter via a base current robbing transistor is another improvement that's just begging to be there for safe, predictable operation.
Anybody got a syringe I can use to squeeze the magic smoke back into this?
 

Offline hamster_nz

  • Super Contributor
  • ***
  • Posts: 2803
  • Country: nz
Re: Transistor Heater Issues
« Reply #3 on: September 23, 2018, 11:40:53 pm »
Hi

I'm trying to build a water heater based on a 2N3055 transistor's power dissipation.



Please note: R1 is a 10k NTC thermistor, R6 and R8 are part of a 10k pot, R9 is a power resistor mounted on a heatsink, and the 2N3055 is also mounted on a heatsink for testing.

While testing this, the 2N3055 keeps dying when the current reaches about 2A.  I have tried increasing the 0.5 ohm power resistor, but then the power is dissipated in the resistor instead of the transistor.

My intention is to dissipate up to 50W in the transistor so that it will heat up whatever it is connected to.  The thermistor will be used to turn down the power so as to keep the heat at a set level dictated by the pot setting.

Can someone please help me understand how to design something like this?  I did some sums and even used LTspice, but it is not working like how I expect.

Thanks

Maybe silly observation - at best you only get 1/2th the power dissipated in the transistor? The rest will be in the 0.5 ohm resistor.

Perhaps you want the 0.5 ohm resistor on the emitter of Q2... that way it should not only be more stable, but the bulk of the power will be dissipated in Q2.
Gaze not into the abyss, lest you become recognized as an abyss domain expert, and they expect you keep gazing into the damn thing.
 

Offline noname4meTopic starter

  • Regular Contributor
  • *
  • Posts: 93
Re: Transistor Heater Issues
« Reply #4 on: September 24, 2018, 12:10:35 am »
Thanks all, will put a scope on it tomorrow - I did hear a squeal from one of the 2N3055s before it short circuited.

From what I can see at the moment, I should be well within the operating area (the power supply current limits at 4 Amps, or 2 if I set it to that) - see attachment.

The reason why I put the 0.5 ohm resistor on the collector was so that there was no chance that Veb could exceed 7 volts (abs max specs).  It's late, so I not be thinking at the moment very clearly - if the transistor turned on with 2V at the base so that the 20V drops across the 0.5 ohms, would I not exceed the 7V limit?  As I write this, it seems wrong, because the transistor would not turn on in that state... :palm:

I've attached the datasheet to this post if it helps.

I bought the transistors from ebay, so this may be the issue, but even if I limit the current to 2 amps so that they don't get fried, even when I put a lighter on the thermistor, it does not seem to change the current through the transistor by much.
 

Offline alsetalokin4017

  • Super Contributor
  • ***
  • Posts: 2055
  • Country: us
Re: Transistor Heater Issues
« Reply #5 on: September 24, 2018, 12:41:27 am »
Forgive me for asking such a silly question... but why in the world are you using a transistor for a heating element?

And yes, fake 2n3055s are a real issue. They are all over the place. The fakes won't handle the power or the high voltages that a real, old-school 3055 will do.
« Last Edit: September 24, 2018, 12:43:27 am by alsetalokin4017 »
The easiest person to fool is yourself. -- Richard Feynman
 

Offline Nerull

  • Frequent Contributor
  • **
  • Posts: 694
Re: Transistor Heater Issues
« Reply #6 on: September 24, 2018, 02:50:04 am »
That safe area assumes you have enough cooling to keep the transistor at 25C. Have you checked the temperature derating?
 

Offline Cerebus

  • Super Contributor
  • ***
  • Posts: 10576
  • Country: gb
Re: Transistor Heater Issues
« Reply #7 on: September 24, 2018, 10:37:47 am »
 
That safe area assumes you have enough cooling to keep the transistor at 25C. Have you checked the temperature derating?

 |O Of course! Most of us have such an ingrained assumption that one will be cooling the transistor to preserve as much of the SOA as possible that we're missed this. If you're deliberately heating it then most of what we'd regard as normal circuit conditions will be out of the window. I'll have a bet that when all the calculations are done it will be found that it's quite normal for the transistor to give up the ghost at 2A under the conditions that prevail at the time it happens.

I didn't immediately treat the idea of using the transistor as a heater as foolish because it's not an unusual thing to do for the small quantities of heat needed for a crystal oven or to stabilise a voltage reference such as the LM399. However, we're talking about much higher energy levels here and under those circumstances, yup, it's probably a dumb thing to attempt. I think a redesign to make a resistor the heating element is in order.
Anybody got a syringe I can use to squeeze the magic smoke back into this?
 

Offline coppercone2

  • Super Contributor
  • ***
  • Posts: 9234
  • Country: us
  • $
Re: Transistor Heater Issues
« Reply #8 on: September 24, 2018, 02:24:41 pm »
its useful to know how to do it right because you can get stuff like semiconductor coated glass to use as a heater for certain applications
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf