Thermal control schematic

controlled

Use one of these (or similar) to control a relay that switches a heating element..

Use a motorised paddle to stir the water to ensure no hot spots exist and reduce any temperature under/overshoot.

Take extreme care if mixing water and mains electricity!

Well, I don't know if it's enough. The basic idea was a temp sensor, such as an LM35 or similar, and a temperature control circuit (don't know if analog or digital yet) with a precision of a tenth of degree. The precision is important, because it is needed for photographic film development. If it's better to make it digitally, i would have some problems, because i don't have experience with PICs and similars.

I thought it could be a good idea using PWM driven peltier cells as actuators and a PI regulator. But I don't know how to build a PI regulator... :-/

I'll surely do!

--

il Clod! ICQ UIN 97056271

"ilClod" wrote

Sure. Put a load of ice cubes in it and stir. Keep adding ice cubes as necessary.

Oh? You wanted to be able to _change_ the temperature?

That'd be quite good. :-))) Would you suggest me to have a pot of warm water ready to pour in the tank? ;)

Seriously, need to build this thermally controlled tank for my dark room, and I need an extremely precise control (1/1O °C), vith water temp adjustable between 15°C (59°F) and 42°C (108°F) in steps of 1°C.

Is that so difficult?

--

il Clod! on VFR RC36 I "il Daily" ICQ UIN 97056271

Son giunto alla conclusione che, almeno una volta nella vita, bisogna usare Windows a fondo, perche' e' un'esperienza che rende uomini. Almeno, io, dopo, ho sempre due palle cosi'. -- Leonardo Serni

tank?

What is the volume of water that you need to control? What kind of container will the water be in? Is the container insulated? How well is it insulated? Will the water be circulating?

To get temperature control of 0.1 degrees, those things become very important. Trying to control temperature on any useful volume of liquid in an uninsulated container will be nearly impossible. And because you want to cool it below room temperature, that means that a large peltier array is probably your best first choice as the active element, or if the volume is too great for a peltier array, a refrigerator/heater will be needed. Then, you're in a whole new arean, because of the combination of two disparate systems trying to control the same environment.

Are you sure you need 0.1 degree precision?

Problems with accurate temperature and heat flow control (poor "calorimetry") led to the downfall of several scientists who thought they had (and may have actually) discovered Cold Fusion :-)

My understanding is that professional photographic systems use a thermostatically controlled valve to control the mixing of cold and hot water that flows continiously. The resulting mixed water at a controlled temperature goes into the insulated tank and the overflow goes down a drain. The temperature of the incoming cold water obviously needs to be below that of the min temperature you want to achieve (15C) with this type of system.

I think this web log appears to describe how one person thinks it could be done at home..

Not being a photographer I looked up what other people use.... This site seems to cover the popular options but I doubt any of those shown achieve temperature control to +/- 0.1C - althought they may measure the temperature with a display that has a 1/10 C digit.

The Jobo CPA processor..

..doesn't even appear to have an insulated cover over the top of the bath.

Sorry if this isn't what you are talking about - like I say I'm no photographer.

Colin

That's a good starting point. Thanks a lot :)

--

il Clod! on VFR RC36 I "il Daily" ICQ UIN 97056271

Hawkeye Pierce writing home: "Korea's pretty much the same story: the fighting goes on, the hatred, the violence, the senseless brutality, men behaving like animals - and then there's the war..."

..but I do have experience with electronics, control systems and thermal management.

That's good. My idea is building a (more or less) insulated container (ie. a glass fishtank covered with expanded polystyrene) with an external water pump that moves water from the top of the tank to the bottom of it, by two pipes, set at different depth. Just after the pump, i thought to apply a heat exchanger built with an alu heatsink and a pair of peltier cells. This cells are driver by a dual supply PWM circuit, so that I can achieve heating and cooling easily. The temp sensor should be placed in the middle of the tank, where presumabily I would put the film container filled with developers etc. I thought something like LM35...

The problem is building the control circuit, because I don't know nearly nothing about MCU. I could be able to do it analog, but I'd prefer to avoid thermal drift phenomena.

--

il Clod! ICQ UIN 97056271

"You know you're beautiful when he's angry" -- Hawkeye Pierce

Beware of the difference between 'accuracy' and 'resolution' when you're calibrating your system.

avoid

What did you have in mind for the sensor?

Something simple to build and calibrate, and precise enough for my purpose. I thought LM35 could be fine. Maybe it could be better if applied on a aluminium heatsink, to increase the thermal exchange surface. Or is it a bad idea?...

--

il Clod! ICQ UIN 97056271

Hawkeye Pierce on the war: "In case you haven't noticed, this is ugly. This is not exciting."

purpose.

If you calibrate it then you should be ok (otherwise the datasheet suggest the accuracy will be around 0.75 degrees).

the

Personally I doubt you can get +/- 0.1C control overall but circulating the water as you describe is the right way to go. Adding a heat sink reduces the time constant but hopefully you aren't looking for rapid temperature changes with no overshoot - just steady state accuracy.

Try this for a starting point of a design...

As shown it would only give control to around 1C but you could improve this by changing the gain of the input amp and rescaling the ADC. Currently the input amp has a gain of 2 so that 0-150C at the sensor maps to 8 bits of the ADC range. If you change the gain to 6 that would map 0-50C to the 8 bit range. This would give a resolution of 50/256 = 0.2C. That's not enough to meet your 0.1C spec but you could further improve this by using more bits of the ADC (eg 9 bits or all 10 bits). Notes: I've not checked that all 10 bits of the ADC are "usable" on the micro they specify. Nor have I looked to see what language they wrote the code in - if it were C rather than assembler you would have to find yourself a compiler. The assembler is free I believe.

You might also look at using the LM35C or LM35CA part - if I've read the data sheet correctly these have better linerarity.

The PIC 16F8xx series are easily programmed using a home made in circuit programmer - essentially this is just a modified serial cable connected to the serial port on a PC.

There is a huge community of engineers using the Microchip PIC series so it's not hard to find guides on how to get started programming these devices. The PIC web site is:

Colin

You could reduce the size of the devices you need if you are prepared to prime the system with water close to the target temperature - otherwise you need quite powerful devices. I guess it depends on how much water is in the tank and how long you want to wait for it to reach the right temperature.

Say you want to heat the water from 20C to 30C and are prepared to wait 10 mins for it to get there. If I've done my sums right I calculate you need

Thank you so much! :-), really helpful.

In case of problems (hope not :) I'll ask for help again

Thanks by now.

--

il Clod! ICQ UIN 97056271