Dewheater and Peltiers

[MarkS]

Quick question.

I'm thinking about controlling the peltier on my Canon because I rarely need 35C of cooling drawing 2.5amps of battery power.

I've read that PWM (pulse width modulation) is the preferred way of controlling a pelt.

My dew controller uses PWM and can easily handle 3A.

Is it suitable for controlling the peltier?

[Ian]

Is 3A enough to run the pelt at full pelt? (he he...  ) See data sheets

Pelts are 100% resistive loads so will not present unreasonable loads to a PWM power supply, provided the max current available from the PSU will be enough to get the heat transfer you're looking for.

Be aware of electric noise however. Dew heaters are usually away from cameras minimising noise pickup, but your pelt will be much closer to the camera. Give it a go and see what happens.

[Mike]

If that doesn't work you could easily make a PWM controller with an MCU and a high-speed transistor.

[MarkS]

Is 3A enough to run the pelt at full pelt? (he he...  ) See data sheets

The pelt I'm using only draws 2.5A at 12V

Mike, what's an MCU?

[Ian]

well as long as the output of the PSU is at 12V, you should be good to go

[Mike]

Mike, what's an MCU?

Microcontroller

[Ian]

a couple of 555 timers would work as long as your not too fussed about accurate control and stability. Which for a pelt or dew heater, you're not...

[MarkS]

Pelts are 100% resistive loads

Is that true?  During operation, if the power cycles off there must be a back-EMF from the thermocouple effect.

Mark

[Ian]

interesting thought. Yes, there will be a voltage across the pelt, in reverse, but that doesn't necessarily make the load in any way reactive. A reactive load is one where the changes in voltage and current are not co-incident. If the load is reactive, then voltage either leads current or trails behind it depending on if it's an inductive load or a capacitative one. I think it might be the other way round, I'd have to look it up...

[Mike]

Nothing a high rating diode won't fix.

[Ian]

hmm, I'm not sure if that's a good idea or not. Yes if the transistors need protection from the back EMF, no if you want the heat to stay where it is when the power is removed. The thermal behaviour of the pelt changes if current is allowed to flow when there is a temperature difference across the different sides, effectively making the pelt more thermally conductive. A diode would effectively short the pelt out and you'll get a much less efficient cooler.

This could prove to be an interesting read, if anyone's looking at building a TEC PSU. http://www.datasheetcatalog.org/datasheet/maxim/MAX1968-MAX1969.pdf

This is quite interesting too, particularly when discussing what efficiency actually means when using a TEC http://electronicdesign.com/Articles/Index.cfm?AD=1&ArticleID=6325

[MarkS]

Dew controller worked a treat - gave me full control over the peltier.

Reducing consumption from 2.5amps to 1.5amps still gave me a 20C temperature drop (from ambient) - more than enough for most situations.  And it should reduce the likelihood of the CCD misting up.

Now if I will build a similar circuit but with a thyristor or similar I could get proper thermostatic control.

[MarkS]

This could prove to be an interesting read

Thanks - I'll take a look.

Mark

[Ian]

A thermistor rather than a thyristor would be useful.

The control circuit will need to compare the resistance seen at the thermistor against the setpoint reference resistance. You'll have to take a look at the thermistor datasheet to work out what that should be, and provide a way of trimming.

Then you would control the duty cycle of the SMPS to deliver the temperature. I would also have an ambient temperature measurement to ensure that you're not driving the TEC beyond it maximum efficiency... The current drawn by the TEC heats it up, so you get diminishing returns over a certain supply current.

[MarkS]

A thermistor rather than a thyristor would be useful.

That's what I meant 

[MarkS]

A diode would effectively short the pelt out and you'll get a much less efficient cooler.

Surely the diode should go in series so it allows current to go to the TEC but prevents it coming back out.  So I would need a 3amp(ish) diode.

[Mac]

if the transistors need protection from the back EMF

Correct me if im wrong.
but the diode used to protect things from back emf was mainly for relay coils
mainly because the emf that is generated by the collapsing magnetic field is huge,
which destroys the trasistors, ect.

The peltier is a resistive network, not an electromagnetic coil.
Therefore no back EMF produced.

Yes there is a voltage generated because of the Hot / Cold junction. but this should be the same as the voltage needed to run it. 12v.

This low voltage should not be enough to damage the transistor.

As for the diode protecting against back EMF,
the diode is placed across the coil reversed biased, so that the collapsing field has a current path to follow.


Mac.

[Ian]

Mac, yes. The EMF produced by a TEC in that manner is less than impressive to be honest. I tried to find a datasheet for some numbers, but failed miserably. I'm sure I've got one in my library somewhere...

Mark, the diode is placed in parallel across the device you wish to protect. The plan being it provides a low impedence path for the reverse current, so carries much more current than the relatively high impedance part you're trying to protect.

Putting the diode in series with the TEC is less desirable as it would need to carry the full on current to the TEC and dissipate a proportion of the power. A Si junction diode (which is the only type you'd use in this case) drops 0.6V across itself, meaning you'd not get the full 12v to the TEC, and the diode itself would dissipate 7.2W for no good reason.

[MarkS]

The EMF produced by a TEC in that manner is less than impressive to be honest.

Ian,

I've just measured it by getting the camera down to near thermal equilibrium (25C across the TEC) disconnected it from the dew controller and connected it to a meter:  either 1.45V back EMF or 450mA current - but clearly not both at once! 

The final stage of the dew controller circuit should easily cope with a back EMF of 1.5V.
We clearly don't want to short it out with a diode.

Mark

[Ian]

the datasheet would be able to tell you the maximum review collector-emitter voltage, but generally they can tolerate up to about 5V. The TEC is well clear of this.

[JohnP]

Mark - agree with Mac on this one. I doubt Diode would be needed & certainly shouldn't be placed in series - you normally do this to prevent against accidently connecting battery around the wrong way or when you need to drop volts by .6V for some reason or 1/2 wave rectification in AC ckts..... Like Ian says no point in dissipating .6V X 3A - 1.8W for no reason...

[Ian]

only Ian didn't say 1.8W  I put all the wrong numbers into the right equation. Can't trust my own head

[MarkS]

In the meantime,

2 Peltier-Canons wired in series running off one channel of a dew controller.  0.8 amps total consumption (including fans) keeps their temperatures to 14C below ambient - more than sufficient for the freezing temperatures expected this weekend.



Mark

[Ian]

just add tomato sauce and a healthy number of meatballs

Would you like some cable ties?

[Mac]

JUST
ADD
WATER.