Heat Source Creator (couldn't think of a better name!!)

[RustyH]

Good day to all,

Im looking to design a rig that allows me text the thermal capabilities of materials (mainly metals). I,d like to accurately control the heat source so I can calculate the power consumption for heat calculations. I would also need to measure the heatsource.

I wounding if anyone can point me in the right direction on how to do this?

I'm thinking i will need a high power resistor controlled by an accurate adjustable power supply, then a PicAxe than could measure the voltage / current somehow to give me a power value. Then using some thermocouples on the heat source connected to a picaxe to gather temp data

Is this along the right lines?

 

[AllyCat]

Hi,

It would probably be easier to use a Power Transistor (Bipolar or FET), because they are better "designed" to couple their heat into a (metal) heat sink. That would also make the heating (power dissipation) more controllable (e.g. by PWM from the PICaxe). Monitoring the current and voltage is easily done with a few resistors and a PICaxe.

Thermocouples generate a rather low voltage so it would probably be much easier (and more accurate) to use "linear thermistors" or the classic DS18B20 sensors. I guess that temperature differences may be quite small across metal materials.

Cheers, Alan.

 

[RustyH]

Thanks Alan,

I should have mentioned the power (heat) I'd like to achieve. From 0 to say 30w

Is this possible with a Power Transistor?

 

[AllyCat]

Hi Rusty,

Yes indeed, I'd try a "classic" metal-cased 2N3055 which you should be able to find quite cheaply. Rated at 115 watts and with a maximum 200 degrees C junction temperature! But not a very high current gain so you'll probably need a driver transistor, something like a BC337 runing from the PICaxe rail. Then you'd need to subtract its base current from its emitter current (easiest to measure) when calculating the power dissipation.

With a thermal resistance (junction to case) of about 1.5 degrees/watt, the chip will only rise about 50 degrees above the case (mounting face) at 30 watts, so you could safely boil water with it.

Cheers, Alan.

 

[BeanieBots]

I'd look up the properties on the interweb thingy..

but if you really want to do it yourself, I would take a very different approach.
I'd use something like an HSA-50 aluminium clad resistor that can be bolted to whatever you are testing.
I'd use a MOSFET driven via PWM from the PICAXE.
Using a 'scope I would calibrate for the power supply voltage etc to determine what PWM value gave what power level.
From there, PWM is linear and so the power can determined (as apposed to measured) for any given PWM value.
Then use DS18B20's to measure the temperature.

 

[SAborn]

Looks to be a simple solution to me, with Allycat's method, as i do like the 2N3055 transistor, as its close to bullet proof.
Its nice to see a simple solution with using a age old transistor, without the need to reinvent the wheel with a mosfet design.

 

[AllyCat]

Hi,

PWM is linear and so the power can determined (as apposed to measured)

Certainly measuring high currents may not be trivial, it depends how you plan to "make" the 30 watts. 1 amp at 30 volts should be easy, 2.5 amps at 12 volts OK, but I'd probably avoid 6 amps at 5 volts or 10 amps at 3 volts.

But an issue with most power resistors is that they are designed to run "hot", so you may have more problems with the heat "leaking" away from where you actually want it to go (i.e. through the heatsink face). For example an HSA-50 is designed to dissipate 20 watts without a heatsink (convection from its wrinkled surface and connecting wires, etc, maybe even some thermal radiation) whilst a TO3 package (2N3055) is only rated to "lose" 3 watts when used without a heatsink (60 degrees/watt "junction to ambient").

Cheers, Alan.

 

[RustyH]

would something like this not be easier, or would it not be accurate

http://uk.farnell.com/ohmite/teh100m1r00fe/resistor-thick-film-1-ohm-100w/dp/1786190

 

[AllyCat]

Hi,

Well, it depends how you define "easier". Potentially quite accurate, but a 1 ohm version will need an adjustable 5.5 volt, 5.5 amp power supply. Also, it's £13.51 + VAT + shipping and appears to be out of stock.

Cheers, Alan.

 

[RustyH]

ohhhh, I see, its its becomes 1 to 1 on the voltage, Current and Power

So 1 volt at 1amp is 1 watt

So for 30 watts it would need 30v and 30amp!! thus big cables and high current.......no the safest and efficient method!!

 

[premelec]

@RustyH - I hope you aren't designing any critical equipment with that reasoning... 30v & 30amp = 30x30 = 900 Watts

 

[AllyCat]

Hi,

Yes, as I indicated before, 5.5 x 5.5 = 30 watts (almost exactly). But still not very "convenient" values.

big cables and high current......

You must avoid "big" cables because they will draw heat away from the heat source (nearly all electrical conductors also conduct heat very well). One of the biggest challenges of this project may be to ensure that (most of) the heat flows through your test sample and doesn't escape elsewhere.

Cheers, Alan.

 

[RustyH]

sorry, bad mistake by me there 1=1=1 doh!!

Got you, so using the transistor will allow small current, therefore small cables, and thus less heatloss. Keep the power cables from the PSU will also help.

Will the transistor loose heat through it case linearly to the power supplied (so this can be taken as a correct factor on the exact total amount flowing through the heatsink. Could I insulate the transistor somehow to ensure all the heat flows only through the heatsink?

 

[AllyCat]

Hi,

Could I insulate the transistor somehow to ensure all the heat flows only through the heatsink?

Yes, you certainly should. Probably using a "foamed" material such as expanded polystyrene (once popular as ceiling tiles and an internal packaging material).

Cheers, Alan.

 

[RustyH]

sound silly, and depend if I can seal it, but would creating a vacuum chamber around the surface not exposed to the heatsink but overkill?
I guess aiming for that kind of accuracy will just be lost else where in other system inaccuracies

 

[erco]

A power transistor is a fine solution as long as your heated volume/area is small and matches the small heat sink area on the transistor. If not, there's always the old standby: a coil of wire. Been working fine forever in irons, toasters, heaters of all sorts. Power consumption and efficiency is all the same with electrical heating, whether coil or semiconductor.

 

[Goeytex]

Another option might be a thick film resistor such as Vishays LTO100 Series. Can dissipate 100 Watts with Temp to 175C. Comes in a T0-247 Package
Not Cheap @ about $11.00 US for a 1 Ohm resistor.

 

[rossko57]

You could calibrate the losses to air and wire of your heater; mount it as you would on your metal sample, but on the best insulator you can find, foam probably in the absence of Space Shuttle tiles. Then determine how much power you need to pass to keep it at your target temperature.
When repeating with your metal sample, the difference in power applied is down to the metal.