peltier device

Hi,

I hear that using relays to control a peltier device is bad for it (long
term). Does anyone have a circuit that they can send to me with an
appropriate way of switching a peltier? Or some advice?

Also, can a 12V peltier be run at 6V to halve the cooling output?

Cheers,
Phil

 

I have designed a PWM peltier system using a relay for directional control
(heating or cooling) and have sold many without significant problems, the
design being about 14 years old. I think the frequency of the PWM is
several kHz from memory. The switching is done with mosfets in a simple
common source topology. I used 9v peltiers in a 12v system and don't run
100% duty cycle by firmware design.

I have had next to no failures this way, even after all this time. The data
sheets did advise, however not to use PWM techniques though they were
unclear as to why. I assumed it was possible that a low frequency switching
was done then the mechanical expansion effects or magnetostriction effects
or whatever, caused mechanical failure.

You can safely run a 12v peltier with 6v, just expect the performance to be
less than relative.

 

Thanks Geoff,

Sounds promising. Yes, I thought it would be the mechanical expansion
effects, either that or inductive spikes in the power line.

I would also guess that swinging polarity in quick succession would cause
stress.

Did your application stay at a constant PWM level or did it alter the PWM to
control temperature? That is to say, did you have an on cycle then an off
cycle, or did you have a gradual increase/decrease of PWM to stabilise
temperature?


Regards, Phil

 

Phil, I agree that fast switching of polarity at maximum voltage should be
avoided. My application used varable PWM at constant frequency in a PID
controller. Naturally the controller decides what the current duty cycle
should be but it cannot change instantaneously due to the integrator. When
the polarity is changed in this system, it is always with a vary small duty
cycle so the shock is minimal.

I am curious about your application, please tell me more about it.

Geoff

 

I am looking at the feasibility of designing a compact 10 litre incubator.
The relative humidity must be monitored and kept at reasonable levels. To
increase RH is easy, but to decrease it, one needs a dehumidifier. I was
thinking of a setup where a peltier device cools the air and moisture
collects away from the chamber. I'm not sure if a peltier will be sufficient
tho.

Cheers, Phil

 

Phil

Oatley Electronics sell a peltier controller kit. The circuit is on their
web site.
no connection with oatley, just pointing.

From memory, it uses PWM and a temperature sensor to control temperature.

pete

 

I've seen an application where a 5x5cm peltier using about
8A at 12V is dehumidifing around 1000l of air in a (semi)sealed
enclosure of electronics.
In your 10l volume you'll have to be carefull about where your
condensate goes and how you'll get rid of the heat from the peltier.

 

I'll be using a 4A 12V peltier unit. One wall of the incubator will be a
heat sink so there will be no probs with heat buildup inside the enclosure.
Initially I thought that using a peltier would not be enough, as I thought
that I would need quite a lot of cooling power, but it looks like (as you
say) there are compact commercial dehumidifiers available using the same
principle. Now I am thinking of increasing the volume to around at least 40L
with no problems.

On a side note, did you know that peltiers *output* power at the terminals
when subjected to a temperature differential? If the efficiency of the
technology is refined in future, I'd say that there may be a future for
peltiers producing "green" electricity from say cold streams of water on one
side, and the sun on the other.


Cheers, Phil

 

Thanks Pete but I'm making a humidity controller (not temperature). From the
diagram at http://www.oatleyelectronics.com/kits/k140a.html I see that a
MTP3055 was used to control the peltier. I'll connect that straight to the
PWM output on my picaxe processor. A humidity sensor will monitor humidity.
Should be as simple as that. I was more worried about all the hype on
turning these peltier things on and off, but if it is just large temp
differentials that can kill them, then that can be controlled through the
PWM.

Does anyone know of a reasonable accuracy humidity sensor with a variable
voltage output? I am about to use the HIH-3610 but its quite pricy - $60
from farnell. Cat. 3937446

Cheers,
Phil

 

Rockby have the Philips 2322 691 900019 (p/n 30425) for $32.45.
It's capacitance varies from about 110pf to 140pf according to relative
humidity between 10% and 90%, and the slope is not exactly linear.
I used one of these about 15 yrs ago with (I think) a 555 timer, and
it worked ok.

Regards
David

 

boys and girls,

I have recently hooked a sht11 up to a pic micro. Easy enough to port from
their example (8051) to a pic c language (ccs). Worked a treat. If further
info req'd, please feel free.

pete

 

Not sure if the one you are referring to is this one, but have a look at the
Sensirion units, farnell had/have them - I think they are a 4 wire
connection (pwr + 2 x data IO + gnd) and cost about $50. Good accuracy and
range on the RH. Nice & simple to hook to your micro, no analogue support
circuitry needed.

http://www.sensirion.com/en/sensors/humidity/default.htm

rob