Seems to me that you need to consider a couple of things for temperature control:
1) The maximum practical temperature differential; TEC vs ambient. After you've reached that point, you're just wasting power to try to cool it down more. Right now, the maximum practical temperature differential is an unknown.

2) The lowest temperature that you would want the beverage cooled to.
If it's a carbonated beverage you're cooling down, you probably wouldn't want to go below around 36°F. If it freezes, you'd have a mess on your hands.

Both of those things could likely be handled by temp sensors (such as LM35's) and comparators driving MOSFETs, but that would be a somewhat ambitious project for a n00b.

 

The weight of the full can in a 45' tilted alloy tube should be plenty of thermal conduction. They will get wet in there from condensation.

It's a great idea! I would look at ways of sinking the hot side of the peltier to the engine block, then cool the whole block with the fans underneath in the crank tunnel.

Also, since you have a insulated gap between the cooler tube and the piston bore, put some red/blue leds in there. Then run individual temp control for each can, and make the lights go from red to blue when that can is cold. That would be a great party effect, maybe with a "ding" sound or a sampled "VROOOM" sound to say a can is done and you can all dive on it and drink it.

Some microcontrollers have 8 ADC inputs, you just need 8 cheap thermistors for sensors.

 

Hi,

this looks like an article here.

http://forums.bit-tech.net/showthread.php?p=1073304#post1073304

I can't see the pic's, hope they have not been removed.

Cheers

PS: The article mentions large amounts of condensation that need to be dealt with so keep that in mind

 

Some microcontrollers have 8 ADC inputs, you just need 8 cheap thermistors for sensors.

I like that LED idea! Shame my electronics skills don't really extend past soldering! Is there a simple/cheap way to regulate temperature (and control the LEDs)?

http://forums.bit-tech.net/showthread.php?p=1073304#post1073304

I really wish the pictures still worked.

 

Both of those things could likely be handled by temp sensors (such as LM35's) and comparators driving MOSFETs, but that would be a somewhat ambitious project for a n00b.

I found this simple circuit (plus another similar one) using a thermistor, a potentiometer, a comparator and a relay. Would it be better to use a relay so the current for the TEC doesn't have to go through the transistor/MOSFET?

Instead of using 2 LEDs, I would like to use a bi-colour LED to change colour when the temperature has been reached. I am worried that the LED might 'flick' between the two colours as the temperature fluctuates around the desired temperature. Would using a separate comparator to control the LEDs (setting a slightly higher temperature as acceptable) be the easiest way to solve this?

Any help would be appreciated! Do you think a simple circuit like this would be suitable for my project?

Edit: just found some information on MOSFETs being used for fan control. It did say on another page that this circuit is not really suitable for Peltiers, as the current required is too great. Are higher-rated MOSFETs available? What is the difference between a MOSFET and normal transistor?

 

MOSFETs are generally better than BJTs for high current switching. They can do analog, but that isn't what they are really good at.

 

...
Instead of using 2 LEDs, I would like to use a bi-colour LED to change colour when the temperature has been reached. I am worried that the LED might 'flick' between the two colours as the temperature fluctuates around the desired temperature. Would using a separate comparator to control the LEDs (setting a slightly higher temperature as acceptable) be the easiest way to solve this?

Any help would be appreciated! Do you think a simple circuit like this would be suitable for my project?
...

Yes it should be ok, but you will need 8 circuits for the 8 coolers, unless you are ALWAYS going to insert 8 cans at the same time and pull 8 cans out at the same time.

Once you get rolling it's not much harder to make 8 circuits than it is to make 1.

I would consider using relays, you might find that easier than FETs and easier to connect bi-colour LEDs to.

Most comparator temp-control circuits have hysteresis built in, that means there is a gap between the cut-in and cut-out temperatures. So it won't flick on/off all the time, well actually it will but it will be slow, like once a minute or so after it gets to temperature.

 

I've attached a simple circuit that I drew out based on some other circuits that I found (for a single can). The LM339 chip has 4 comparators, so I'd only need 2 in total for my project (or 4 if I have separate circuits for the LEDs).

I understand most of the circuit but I have a few questions:

1. Would I be able to use a single potentiometer to set the reference voltage for every circuit?
2. What is the 3.9k resistor for?
3. I think the 5.6M resistor provides "snap action" feedback to the comparator. Is this necessary?
4. I believe that the diode around the MOSFET is for protection. Is it necessary?
5. Is the capacitor required when using a converted ATX power supply? I read that it filters the power supply.

Thanks for the help so far.

 

Hi,

- change the 50k pot for one 25k and one 5k (coarse and fine settings)
- the other side of the pots needs a 22k resistor
- the thermistor needs to be a 4k7 ntc (at 2.5°C will measure ~14.5k)
- the 5M7 hysteresis resistor will result in lots of switching, fit a 1Meg pot so you can tweak it.
- wont you need a transistor to drive the Mosfet?
- keep the diodes

where will you be placing your thermistor?

Andrew

 

where will you be placing your thermistor?

I'm not sure yet. Either on the copper block or part way up the aluminium cylinder.

Thanks for the help with the circuit.

 

Hi,

this is the controller I will be building soon. It basically was designed by someone very helpfull for my Gas / 12V / 220V fridge.

You may be able to adapt this to your requirement. The differences are that I was going to supply the circuit from an automotive battery and the power for the element comes from a 220V source. I will be making one for the 12V element for when I drive.

The PCB is 98% complete and could make some minor mod's if required for you circuit and send you the flles.

A few changes and we should be OK. It is a simple comparitor using a 4k7 thermistor and intended to control temperatures in the range of -8 to 0°C. Experience on my fridge has taught me that it is best to have the thermistor in the "evaprator" which in your case probably amounts to the copper intermediate block.

The simulation shows R 8, 9 and 10 as fixed values but these will be pots. R8 being the pot to control hysteresis.

I needed to drive 220V ac so there is an optocoupler there with a triac, This could be removed and an SCR substituted.

Cheers
Andrew

 

Thanks for sharing your circuit. It really is a big help for a newbie like me! The left side is very similar to mine.

I am still unsure of the best way to add the TEC into my circuit. Could I just add the TEC next to the LED on your circuit and remove everything to the right? Or would I need to use some kind of 'switch' (like an SCR, relay or MOSFET)?

The circuit and the TEC both require 12V DC so I'm unsure why a switch would be required; is it just to stop the transistor from overheating?

 

Thanks for sharing your circuit. It really is a big help for a newbie like me! The left side is very similar to mine.

I am still unsure of the best way to add the TEC into my circuit. Could I just add the TEC next to the LED on your circuit and remove everything to the right? Or would I need to use some kind of 'switch' (like an SCR, relay or MOSFET)?

The circuit and the TEC both require 12V DC so I'm unsure why a switch would be required; is it just to stop the transistor from overheating?

I am also a newb so hopefully the gurus will help us out.

The MOC3081 is designed to switch 1.2A if my memory serves me correctly. The application note for the MOC3081 suggests that the device should not be used to drive loads directly but via an SCR or Triac.

As my circuit uses an a.c. powered heater this circuit works. What I have found out that switching d.c is another matter entirely. I spent last night trying to get some device to switch d.c. as co-incidently I also need to make a second unit that will switch a 12V element. I am discussing this on another forum.

The switch. The circuit works like this, the way I understand it that is. The LM339 is used to compare the two voltages, one at the thermistor and the other at the pots. Depending on which voltage is greater the LM339 will cycle ON and OFF sending a signal to the low power transistor to turn ON or OFF (not enough power to drive your load). This signal then turns the MOC ON and OFF. The reason for the MOC is that is allows the swithing of a.c. as the sine wave crosses the zero point. Finally a Triac able to cope with the current of my heater is used to do the switching. In essence there is a control stage which uses low power and a power stage.

When I have the d.c. switch done will let you know.

Cheers
Andrew

 

When I have the d.c. switch done will let you know.

Thanks. Will be good to have a look at.

Although, thinking about my cooler a bit more, I might not bother with the temperature controller. I've heard that it's not good for TECs to be switched on and off too much and drinks would probably cool a bit quicker if the TEC was always at 100% when switched on. It'll use a bit more power, but the cooler won't be used all that much anyway, not like a normal fridge.

Just need to thinking of a way to deal with condensation. I might just make sure that each cooler 'tube' is watertight so it doesn't leak through to the electronics, then let evaporation deal with it when the cooler's off. Hopefully ice won't be a problem.

 

I thought about it and I'll have 12VDC ~45A running inside this table if every cooler/fan is on. There won't be more than 12VDC 5.5A in each wire, though. I read that as little as 0.01A can be fatal. With the table being metal, and the possibility of condensation, how safe will this be?

I'll electrically insulate all the connections (actually I'm going to use connectors for most of them, so I don't need to modify the PSU). I'll also make the inside of the cooling tubes watertight and add thermal insulation to the outside of the tubes which should eliminate condensation there.

Just another couple of questions regarding safety:

1. Should I connect the metal table to earth?
2. Would using a 13A circuit breaker where the PSU plugs into mains make it any safer? I know that the PSU already has over-current protection and should draw no more than 10A from the mains.

 

Hi,

I would not have thought that 12V at 0.01A would be any problem. Perhaps you would be in trouble at 120/240V under certain circumstances.

Table to earth, I would have PSU to earth, if it is mechanically attached to the table then the table will automatically be to earth. This is to protect you against an earth fault as ther will be mains voltage floating around.

I would individually fuse each board (to protect the electronics) and each peltier device to protect the device as they are not that cheap. Your PSU is protected and the exixting circuit breaker will take care of the rest as will your earth leakage.

You will only be drawing ~ 2.5A @ 220V or 4.4A @ 120V.

Cheers
Andrew

As a footnote. I know this is a novelty but it is remarkably innefficient, 540W to cool 8 cans when a 180l fridge and freezer only take 175W?

I have been doing more work on my circuit and it can now switch d.c. through Mosfets. If you want it, shout. One is going to need a heatsink as these dissipate a fair amount of power.

It must be said that with all my experimenting I have proved to myself that direct switching is not the ideal circuit to control a cooling device. The thermal lag is just too much to guarantee accuracy. I will, with the help of a friend, be producing a proportional control model when he gets back from vacation in January. This current unit is a great improvement on what I had but is merely a temporary measure.

 

Hi,

have a read of this article, it is a simple solution that you may find you could adapt to your needs.

http://www.electrokits.com/downloads/pdf/82.pdf

Cheers
Andrew

 

I actually took apart a mini cooler and it was put together almost exactly like that -- and that is the design I'm copying. Like that design, I will probably just connect the fan and TEC directly to the power supply (without a controller), although adding a fuse is probably sensible.

I know that this design is horribly inefficient. I could also just buy a £25 mini cooler to do the same job, but as you said, it's the novelty of a V8 drinks cooler that I'm after!

I assume that I'd need to mount the PSU to the table with a conductive bracket to ensure that it's safely earthed?

 

Hi,

would do no harm to attach the PSU wire an earth.

You know what, you have got me so interested in this peltier device I am going to build myself a small car refrigerator for fun, straight after finishing the current projects.

Cheers
Andrew