Hi All

I'm not great with electronics but have a project that's going to force me to improve!

I've got a water bath that's heated/cooled by a 12V thermo-electric element (peltier). What I'd like to happen is:

1. run the peltier (12V DC from the power supply).
2. Once a thermostat/thermometer detects that the bath temperature has reached 45°C, stop the peltier.
3. Wait an hour.
4. Run the peltier with the polarity reversed until the temperature reaches 25°C.
5. wait an hour.

Would that be complicated and or expensive to do? The thermostat doesn't need to be hugely accurate - +/-2°C or so would be fine. I've been reading up on it and believe I'll need a time-delay relay but I haven't got much further than that so far!


Thanks,
Danny

 

About US $ 10.00 in parts not counting postage. Thermister, window comparator, some 555's, 2 relays, some resistors & caps.

 

I've got a water bath that's heated/cooled by a 12V thermo-electric element (peltier).

And this is working well already? Is this now a manual process, or do you have some control system already? That would likely simplify what needs to be added.

 

Hi, no - it's not currently working other than having tested that the PSU is happy running the Peltier. The PSU is an old PC ATX power supply and the peltier is 15V max, but running it at 12V is good enough for what I need and plenty of amps available on the 12V rail.

I'm just looking at window comparators - not heard of them before so am investigating on the RS website (I'm in the UK).

 

Window comparator uses 2 reference limits to switch on/off

http://www.google.co.uk/imgres?q=wi...&ndsp=15&ved=1t:429,r:0,s:0,i:68&tx=131&ty=92

 

...it's not currently working other than having tested that the PSU is happy running the Peltier....

Cooling a bath from 45°C down to 25°C is going to require the TEC to move quite a lot of the stored heat capacity, and it will add at least that much heat (and more likely several times as much) as electrical power that needs to be dissipated from the surface of the TEC. It'll need a big heat sink and cooling fan to move that heat.

My point is, I'm not sure cooing by the TEC will do much more than the other paths of heat loss; conduction, convection and radiation to ambient. Using the fan to blow over the surface of the water would likely give a LOT more cooling by evaporation of a small amount of the warm water.

 

Cooling a bath from 45°C down to 25°C is going to require the TEC to move quite a lot of the stored heat capacity, and it will add at least that much heat (and more likely several times as much) as electrical power that needs to be dissipated from the surface of the TEC. It'll need a big heat sink and cooling fan to move that heat.

My point is, I'm not sure cooing by the TEC will do much more than the other paths of heat loss; conduction, convection and radiation to ambient. Using the fan to blow over the surface of the water would likely give a LOT more cooling by evaporation of a small amount of the warm water.

Yeah - they're not very efficient! We've actually got a buffer tank to pump the excess heat into when cooling, and then there's another tank with a peltier running in reverse of the first one, so when one is heating, the other is cooling and vice versa. As you say though, there'll be a net gain of heat so we've got a big fat PC heat-sink ready to attached to the buffer tank. That'll need a thermostat to fire it up when necessary but I think I can get that working easily enough as a separate loop.

 

As far as 'ease of use' goes could not the water be cooled much faster by the addition of a set amount of ice, and then using a direct immersion resistance heater(120VAC) to add the heat?

More information on what you need the water for would get you the best advice.

 

As far as 'ease of use' goes could not the water be cooled much faster by the addition of a set amount of ice, and then using a direct immersion resistance heater(120VAC) to add the heat?

More information on what you need the water for would get you the best advice.

It would be faster, but it's for an experiment and needs to cycle continually for four hours or so each way, for say a week at a time.

The TEC should work fine from our calcs though - there'll be some excess heat in the buffer tank but it won't be a huge amount (the heating and cooling tanks are much smaller than the buffer tank) and the excess can easily be dumped into the room with the heat-sink.

 

Do you buy window comparators pre-made with variable resistors already connected, or is that usually part of the build?:

If I got this comparator:
http://uk.rs-online.com/web/p/comparators/0649475P/
how do I determine what impedance potentiometers to get?

For the 555's would this be ok?
http://uk.rs-online.com/web/p/timer-circuits/0638942/

I have no idea what Maximum High Level Output Current, Maximum Low Level Output Current, or Maximum Input Frequency I should be going for! I guess it depends on the other components?

I gotta get to bed, but thanks for the help so far!
Danny

 

If you need more than 2 comparators, look at the LM339 quad. Extra comparators are handy to have on board, since they can be used as a way to indicate connections, voltage levels, all sorts of handy things. For the timers, consider the 556 (two 555s together).

 

Here's my outline: Power is always applied via RY 1 unless interupted by one hour timer U1. Hesting- Cooling controlled by RY2, DP DT, driven by R-S flip-flop U2. U3 out is normally hi, goes low when 45 C is reached, setting U2 & triggering U1; U2 going high turns off RY2. When pot cools down to 25 C , U4 goes hi, resetting U2, turnin RY2 on for another heat cycle. U3, U4 not really a window comparator, just two singles. U5 is pulse inverter. Can't do more untill my eyes un-cross.

 

Thanks Bernard - you're a legend! I'm going to be looking at this over the weekend