properly increasing current for a temp controller / TEC circuit.

Discussion in 'The Projects Forum' started by savlaka, Nov 20, 2008.

  1. savlaka

    Thread Starter New Member

    Nov 20, 2008
    9
    0
    Ok so I'm looking to provide temp. regulation to a TEC, I have a digital fan controller that will vary the voltage to a fan (max 1A) basied on temp. I'd like to use the same controller to run my TEC but I need about 14A for it at max power. as I under stand things I should be able to use a Transistor ,or even a few, to do this but I don't really know how to set it up? the controller I have varies the voltage from 3.7v to 12v (or power supply rail voltage).

    altertively I could use a PWM if there is one that will act basied on the controllers output voltage if there is something like that availabe out there.

    I can read and convert scematics to pcb layout and physicly build this, but my electronics experiance is mostly in repair not in design...

    Thanx ya'll
     
  2. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    What's a TEC? :confused: Thermoelectric cooler, like a Peltier device?
    You mention up to 14A, but not the voltage.

    You can use a 555 timer as a PWM controller; the output controlling a MOSFET's gate. Pin 5 is the control voltage input.

    Go to National Semiconductor's site, and download the datasheet for an LM555.

    There are many application circuits shown.
     
    Last edited: Nov 20, 2008
  3. hgmjr

    Moderator

    Jan 28, 2005
    9,030
    214
    Is that TEC as in Thermo-Electric Converter?

    hgmjr
     
  4. savlaka

    Thread Starter New Member

    Nov 20, 2008
    9
    0
    sorry ya'll when you order them (thermoelectric cooler's - peilter's) most of the manufactures refer to them as tec's, I had to talk to more than a few to find the right one for this job... this is a 168w @12v nominal thermoelectric cooler, max input 16.2v <-lessens life of device hence the 12v rating. the manufacturer recomends using linear voltage regulation or a high frequency type pwm ,10khz+ min, recomended 50khz+, to reduce stress on the TEC

    so a 555 timer wil adjust it's frequency baised on a input voltage? (i'll be looking up the data sheet right after this reply)....

    again sorry for the confusion.
     
  5. savlaka

    Thread Starter New Member

    Nov 20, 2008
    9
    0
    ok, correct me if i'm wrong but from what i've read i'd need 2 555's one to generate the operating frequency and one to act as the pwm controller?

    so i'd build a 10khz+ frequency generator and feed it's output into pin 2 of the second one and then feed pin 3 to the gate of a mosfet or three. and to controll the output i'd feed the voltage from my fan control into pin 5 of the second one....?
     
  6. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
  7. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    It would be a lot easier on your TEC if you used a synchronous buck circuit.

    Synchronous buck circuits have two N-ch MOSFETS configured in a half-H bridge with an inductor in the loop between the MOFETs and the load; in your case a Peltier device. The inductor keeps the current flowing through the load in a relatively constant manner. If the current through the load (and inductor) builds up beyond the limit, the upper MOSFET gets turned off, and the lower MOSFET gets turned on. The current then "flywheels" around through the load and the lower MOSFET back to the inductor. When the current through the load drops below the limit, the controller turns the lower MOSFET off and the upper MOSFET back on to "charge" the inductor back up. This can happen at a very high frequency. It's very efficient.

    Wikipedia has an entry on buck converters, with a section on synchronous buck converters:
    http://en.wikipedia.org/wiki/Buck_converter

    Microchip's MCP14628 is a synchronous buck MOSFET controller. Have a look at a datasheet to see if it makes any sense to you.
     
  8. savlaka

    Thread Starter New Member

    Nov 20, 2008
    9
    0
    Allright sorry for my limited understanding of this...
    i've attached a jpg that is of the basic app schematic and some qusetions about it's use...
    it looks fairly strait forward to build and set up if i'm understanding it right, then again I'm probably not.
    Thanx again for yall's help and Information... didn't really expect it to be this mutch to increase the current from the temp controller...no problem i'd really like this working right, so i'd rather fight thru tring to design the proper system than accept a half*** solution...:)
     
    Last edited: Nov 21, 2008
  9. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    FCCM Control: Read the datasheet for the actual operation of this pin but if you connect it to 5V will be fine.

    Current sensor: This may be a current transformer

    Oscillator: Yes, it can be an input from a 555 if the 555 can provide the frequency the PWM chip needs to work.

    Reference voltage: How you will feed the voltage to the chip depends on the temp sensor output. Do you have a datasheet for it?

    MOS: They can be used for your current needs, 14 Amp max right?
     
  10. savlaka

    Thread Starter New Member

    Nov 20, 2008
    9
    0
    I didn't really understand the fccm in the data sheet thats why i asked...

    current transformer?... i've worked with a few devices that used what looked like a small tiroid that slid down the output wire for current detection, is this what you mean? <- don't worry I will look current transformers up to try and answer this myself :) - yep that what what it is... any particular size / type i should look for? -- the mcp1630 says it wants 1.7v? for this input i'm a bit lost here as to how to supply this current sense to it...

    I'll verify the frequency it needs and chk that in the 555 sheet. - looks like the mcp1630 operates at a max frequency of 1mhz and the 555 timer will put out the required 10-50khz+ (max somewhere near 100khz) that I was recomended by the TEC manufacturer, so a 555 should work ok...

    data sheet no sorry.... it's a standered computer (pc) digital (in the controls anyway) fan control that provides analog voltage regulation (3.7v to 12v @1Amp max) to each of its outputs (4x) for running case, cpu, vga # 1 & 2 fans even if the fans in question do not allow pwm control (some of the ones with integrated controllers)... the one I have is a bit outdated but is very similar to the kaze master 5.25" model and works about the same. you set a temp on the control panel and it adjusts the voltage to the fan to maintain the set temp, in my case +/- 3deg C. I'm using this Tec on my overclocked cpu and I really don't like the way that most of the OC'ers solve their condensation problems.... basicly they run the tec at max and permantly seal the cpu / tec / bottom of heatsink with spray foam, RTV or similar items to provide a airtight seal... Instead of just regulating the tec's output to keep the cpu running just above the dewpoint no matter the load on it... <---AMD told me that maintaining a constant temp would be better for the cpu anyway.

    aye the irfz44's are rated at 50A max or 15000mw disipation (and yes i have some BIG heatsinks for them out of old audio amps) and i could allways double them up, actualy I probably will just in case I ever need to get a larger TEC...I asked about these B/C I keep them on hand for repairing car amplfiers this is the part in their dc-dc switiching supply (usally a pwm) that fail when folks short the + input to the case ,ground / zero ref., 2 are usally used for a 400w amp (usally fused for 35-40A).- and also i've found that the irfz48's are rated at 19000mw

    any outher info I should provide ya'll?
     
    Last edited: Nov 21, 2008
  11. scubasteve_911

    Senior Member

    Dec 27, 2007
    1,202
    1
    Just use PWM to control the TEC. I designed one a few months ago that was meant for hot and cold operation, so I used an h-bridge configuration.

    You can simply use a low-side mosfet and connect the TEC as the load. It was good up to 30A at 12V, use SMT, and used the PCB has the heatsink. Just be sure to find low RDSon mosfets and drive the gates with a high current driver.

    Steve
     
  12. savlaka

    Thread Starter New Member

    Nov 20, 2008
    9
    0
    Ok this is where i'm at right now: (see pic)

    so opinions?... should this work?...anyone know how to find the rest of the passive componet(s) values?...

    it's a bit more than i'd envisioned when i'd started this thread, i was figgureing there was a way to use a transistor (and a few passive componets) to supply more current following the voltage changes of the controller...:) (don't get me wrong if this pwm setup is the right way to do this then no problem, I was mostly just seeing how far we'd come from my original idea)

    - scubasteve_911, I think thats what we are working on... but i need more than a power supply for it I'm after automaticly controlling the temp of the cpu so it remains constant under any load and never drops below the dewpoint , so no condensation problems to deal with. And not permantly sealing my cpu, tec, and heatsink to my motherboard (i like to upgrade stuff when i have the spare cash)...

    thanx again yall
     
  13. savlaka

    Thread Starter New Member

    Nov 20, 2008
    9
    0
    hmmm I ran across some info for using a opamp and or transistor as a voltage follower to supply a higher current... would something like that work here?... since that basicly sounds like what i'm trying to do.?. oh well i'll check back with ya'll later :)
     
  14. scubasteve_911

    Senior Member

    Dec 27, 2007
    1,202
    1
    What I was saying is that you don't need an inductor or the high-side FET. That gives a smooth DC voltage, something that you do not need. The thermal lag of the system is much greater than electrical time constant, which means PWM is the way to go. You can get away with a single mosfet controlling the TEC. Trust me, I have spent about a month doing research with these devices..

    By using a switching supply, you are only introducing losses and lag into the system. Although, I admit, the lag isn't a problem because of the thermal lag of the PEC.

    For the controller, you can get away with a very simple proportional control.

    Steve
     
  15. savlaka

    Thread Starter New Member

    Nov 20, 2008
    9
    0
    ok... sorry for the misunderstanding. So how do I go about implmenting this basic pwm your sugesting? any chance you have a schematic (even a basic one) to modify / use? like I stated in the begining I'm no designer, I do some repairs and I've built some audio and test equipment over the years. mostly thru adapting schematics to suit my needs...
     
  16. scubasteve_911

    Senior Member

    Dec 27, 2007
    1,202
    1
    savlaka,

    After some more research, please ignore what I had said regarding the 'lack of need for smooth DC'. I was basing my information on experience instead of hard theory. From experience, I did make a full temperature control loop that went from -10 degrees to 100 degrees for a small test fixture. I didn't use any sort of switching converter, just simple PWM to a mosfet h-bridge. It worked extremely well, but we were getting a maximum 45 degree negative swing, instead of an advertised theoretical 67. We thought that was pretty on par with reality.

    Apparently, PECs do need smooth DC voltage. Supposedly, ripple currents cause some sort of internal loss, therefore a lowering of delta T that the device can do. This can be accomplished with what you are already doing.

    Sorry for the misinformation!

    Steve
     
  17. savlaka

    Thread Starter New Member

    Nov 20, 2008
    9
    0
    I'm haveing trouble finding the chips and current sensor for the circuit above... allong with the values for most of the passive componets:confused:... ya'll think something like one of these circuit alterntives would work ok?... I think I'll probably need a filtering circuit going to the tec, maby a 2 stage LC (two inductors four caps) to reduce any ripple like a standered swithing power supply uses...

    I have everything to try this out on hand... But before I started puting one together I figgured I should get the schematic looked at by someone more experianced than me...

    In fanpwm-555-741mod pic it's setup to use a thermistor to control the output and a pot to tune the controlled temp...

    in the fanpwm-555-741mod2 pic thats to run it off the controller...

    so opinions? sugestions?
     
  18. jofoeus

    New Member

    Sep 30, 2010
    1
    0
    Hello Steve,

    I'm doing a similar project with the TEC device. Could you please post a schematic of how the h-bridge is connected to the TEC and microcontroller. And also code to reverse the polarity of the TEC. Thanks.

    James.
     
Loading...