12V control PWM? Rheostat? Pot? Help? Please?

Hi, Im only at chapter 4 of volume one of the book but have been building some simple enough projects while reading through. Dunno the best way to deal with this though.

This is for a TEC / peltier.

I would like to control a 12v supply and make it go either between 0 and 12 or a range like 5 and 12.

The first thing I thought was pot.

Then I saw this LM317 and pot, but specs say 1.5A -is that max?

Then I read something on a PWM and dunno what that is.

At 12v it has to handle around 20A, at 5v its about 5A it uses.

I also plan to buy a bigger one later and be able to vary the voltage and therefore the temp. so its not something I want to preset.

Nothing refused, except blows

PWM means "pulse width modulation." One is turning the power on and off very quickly, thousands of times per second or more. By varying the proportion of on time and off time, we can control the average voltage.

Peltiers don't like PWM. But if you filter the PWM, things should be okay. You'll want an inductor in series and a capacitor or two in parallel with the load. You can get some good ideas by skipping ahead in the text a little: http://www.allaboutcircuits.com/vol_2/chpt_8/1.html You are looking for a "low pass filter."

If you want such quite high currents use a switched mode power supply which uses PWM to regulate the output voltage and it is very efficient compared to linear voltage regulators.

OK, not quite as simple as I imagined.

Looks like I might have to read the whole book before I start this, which isnt a bad thing but I really like some 'practical' to break things up -especially since I learn a lot better that way.

Read the bit on the filter and in another TEC thread the poster says:

"the manufacturer recomends using linear voltage regulation or a high frequency type pwm ,10khz+ min, recomended 50khz+, to reduce stress on the TEC" http://forum.allaboutcircuits.com/showthread.php?t=16201

Frequencies, another thing I havent yet got onto, didnt expect this to be as complicated.

Dunno if it helps but its an ATX PSU I will be using.

ATX form factor PSU's generally are not capable of 20A on the +12v output; most of them are less than 12A.

Note also that you will need to put a load on the +5v supply to get proper regulation on the other outputs.

The 300W I have only has 12A on the 12V rail but the 550 I have has a single 32A rail and the 560 has 20A and 12A.

For the 5v and regulation Ive decided I may as well put some sort of lighting feature on it and also drive a whole bunch of fans and replace a couple of PSU's. They were to be mains run 220vac inline fans with a simple 12v powered relay switch so the fans were running anytime the PC was, however, I found a couple of industrial peltiers I bought a few years ago so now the computer really will be permenately on if I use them (both different sizes for the stacking) and it will need to be 24/7 ... I was thinking whats the point of putting in a huge resistor to burn the power when it could be used for something eating power anyway or have I got that 'theory' wrong? Probably wouldnt stack them though, well not right away

Actually ... could use the smaller peltier for cooling the GPU as well and thats at least 5A at that voltage.

Using an LED controller Ive came to what I have here:

But Im unsure if a) it will work as supposed to ... errr do the job I want it to or b) if the values are anywhere near correct. Most of the things I have lying about so I used them basicaly. The inductor, I havent a clue about that. Im not sure if the 0.17A is max operating or if its the amount needed to change it, I read what I could about the inductors but I didnt see a bit on how to choose which one for things like this or I missed it.

I have to buy the MOSFET, assuming thats what I actually need, the large caps and the inductor, everything else I have, its just what to get to finish it off, assuming the rest will work.

Thanks Alberto.







Also I found 20A and 25A inductors, well found the part need the supplier now.

The IRLZ34 is a logic-level MOSFET. Since you are using a 12v supply, you should be usnig an IRFZ34, which is a standard MOSFET. Also, there needs to be a resistor (39 to 62 Ohms) from pin 3 to the MOSFET's gate to prevent high-frequency oscillations and to limit current.

With VR1 being 1 megOhm and C1 being 1 milliFarad, the RC time constant is going to be very large, and it's going to be days between toggles on pin 3.

The way VR1 is connected (as a rheostat), the ratio between the ON and OFF times will never change. It needs to be connected as a potentiometer; one end to D1, the other to D2, and the wiper to C1. Reduce VR1's value to 50k Ohms. Reduce C1's value to 10nF.

Disconnect pin 7 from D2. Disconnect R1 from +12v, and re-connect it to the output (pin 3) of the 555. Reduce it's value to 120 Ohms.

The recommended value for C2 is 10nF.

L1 appears to only be rated for 0.17A. You need at least 20A.

C4 and C5 should be rated 25v.

Thanks SgtWookie.

I hadn't noticed I put in a variable resistor instead of a pot, everything is new and shiney I did increase it to 1M to give it greater range, I thought anyway. Ratio was too out whack then?

I've upped C3, C4 & C5 caps to 2200μF 25V, is that OK? I read that 'the bigger the better' when it comes to smoothing, is that right? At least in this case with low pass filters?

Has 7 to be left with no connection?

Hope I got these changes right, thanks again, sorry for all the follow up questions, I hope to learn quickly and eliminate the need for them.

Oops!
You still have D1/D2 connected to 12v on top - cut that single wire. You need a 120 Ohm resistor connected from the 555 output (pin 3) to the top of D1/D2. The idea is that the output of the 555 timer charges/discharges C1 via it's own output, instead of using pin 7 (discharge). The way you have it wired now, there is no discharge path for C1.

R1 needs to be somewhere between 39 and 62 Ohms.

1uH is probably way too small a value for L1, but I don't have time at the moment. It should be more in the range of 60uH to 220uH; otherwise it will rapidly saturate.

It really needs to be simulated before you try to build it. Also, there needs to be a "flywheel" diode between the drain of the MOSFET and the +12v side of L1 (cathode to +12v); otherwise when the MOSFET cuts off, the continued current flow through L1 will be a problem, creating reverse EMF with noplace to go. This could even be the body diode of a MOSFET with the gate hard-wired to the source terminal; high-current Schottky diodes are getting to be hard to find (and expensive).

OK, I will try and find something bigger and post back with an updated version ... and look up all that other stuff you just told me about.

Can I ask though, Ive found the link for the circuit I based mine on, its an LED one but I thought 'PWM is PWM as long as you have the output right'. Can you tell me the big difference between it and my modifications please?

http://www.reuk.co.uk/LED-Dimmer-Circuit.htm

Not for this project. You're driving a Peltier device; they don't "like" PWM.

So, to keep the current flow relatively smooth, an inductor is needed.

One problem is that if the inductor reaches saturation while the MOSFET is conducting, the capacitors will charge very rapidly - almost like you connected a straight piece of wire right to +12v. The inductor needs to be a value high enough for the operating frequency to prevent saturation.

Another problem is the back-EMF of the inductor. When the MOSFET turns off, the inductor will try to keep the current flow constant - this leads to the voltage polarity across the inductor to swap, and the voltage level can get to be quite high; enough to damage the MOSFET.

I'm on my way out of town for a day or so; will have to re-visit this later.

OK, back to you.
The circuit you linked to is using pin 7 to discharge the timing capacitor, and R1 is 1k.

Pin 7 in the 555 timer is an open-collector transistor. It takes a good bit of digging in various datasheets to find, but this pin is only good for about 15mA.

Since Vcc=12v and R1 = 1k, when pin 7 turns on, current flow through R1 will be 12v/1000 Ohms = 12mA, leaving only 3mA to drain the charge on C1, which limits the minimum/maximum PWM duty cycle. Using the output (pin 3) to charge/discharge C1 gives a wider PWM range.

It's still necessary to use a 120 Ohm resistor from pin 3 to the upper junction of D1/D2 in case the pot is turned all the way up or down. This will limit peak current to 100mA, even though actual peak current will be less due to the Vf (forward voltage) of D1/D2 (around 0.8v when carrying moderate current) and the cycling of C1's voltage between 1/3 and 2/3 Vcc.

Thanks SgtWookie.

Had to redo the whole thing, not once, not twice but three times because I somehow kept managing to lose it.

I also added terminal blocks, since it would be more realistic and not make it so rigid.

I cant find an inductor bigger than 11A 22UH.



Edit: Looking around I found, heres the noob question of the day, suppression chokes ... is that another name for inductors? I can get 100uH, 77A versions.

Hello !
The suggested circuit has a big mistake! There is no path for the continues current of the inductor. You should add a flywheel diode between the FET drain and the +12V net. Otherwise, you will have very high voltage right after switching off the FET. Cathode should be at +12 side.