@Sensacell,
so, like a mcu -> opamp -> mosfet -> peltier route then ?

 

@crutschow,
i figure the RC filter would actually be creating a linear dc voltage out of the pwm signal. driving with such would get the mosfet pretty hot.

so, mcu -> mosfet driver -> a real Logic-level mosfet -> inductor with a diode -> peltier ?
please could you point to a circuit schematic of this configuration ?

@Sensacell, i meant mosfet driver in place of the opamp: mcu -> driver -> mosfet -> peltier

 

could you suggest a feasible logic level mosfet ?

the mosfet i mentioned earlier, IRLML6344TRPbF, has it's threshold voltage between 0.5 to 1.1V ; Vds = Vgs. It's max on-resistance is 37mohms when Vgs = 2.5V; Id =4A, and 29mohms when Vgs = 4.5V; Id = 5A

^ I thought this would make it "logic level" compatible ??

 

Hmm... or perhaps something with a lower on-resistance, like the IRL2203N ... http://search.digikey.com/us/en/products/IRL2203N/IRL2203N-ND/244638

That has an on-resistance of 10mohms for Vgs = 4.5V, Id =48A ; more than adequate for my needs. what do you feel ??

 

I still feel the IRLML6344TRPbF will be sufficient for what i want to do; to switch it on with 3.3V logic. I will of course use a mosfet-driver between It and the mcu. with an on-resistance of 37mohms at 2.5V ;having a Id of 3A from the peltier would be about 0.3W of dissipation. 0.26W if Vgs was 4.5V. But using the mosfet driver should solve all this. could someone please suggest a feasible driver to use ?

 

gate charge = gate current x required switching time..... so, given that i'm using a pwm of 2KHZ (500uS) and the total charge of this mosfet is 6.8nC.... i need a gate current of 0.0136mA ? is this correct ? the mcu has a maximum pin current supply of 15mA , i think this would be sufficient to drive this mosfet (although a current limiting resistor would be really needed)
giving that it has a Vgs =2.5V of 37mohms, i still think the mcu's 3.3V logic would be suitable. Or should i still go with the mosfet-driver ??

 

Like the attached..

 

Hmm.. PWM frequency and switching time are rather quite different..

 

i think i'll go with the mosfet driver at the end of the day. i hear the microchip TC44xx series are pretty good. so a setup like the one below should be okay ??

I have another question. Say i want to drive two small dc motor pump (supplied external 3.3V and 100mA), with a dual mosfet driver using this same mcu. the schematic attached below, would it be correct ?

 

Here is the simplest solution I can think of.

(the NPN transistor can be anything - 2N2222 is good- the part shown was chosen randomly for the symbol only)

1) Great big 12 volt gate drive- almost any power mosfet will work

2) Happy with 3.3 V pwm input.

3) Very efficient

 

a mosfet such as the tc4422 driver, is it okay supplying it a 12V, 4A regulated supply ?

 

@Sensacell,
the irlml644Pbf's Vgs(max) is 12V though. wouldn't be too good running it at that max.

any schematic with a n-channel mosfet ??

i think perhaps the mosfet driver idea would be better: less components needed.
the tc4422 will accept 3.3v pwm input and drive the mosfet at 10V (which should be more than enough for full turn-on). already took a look at the datasheet, Vin = 4.5 - 18V, so i intend to supply 12V to it.
However, i'm still skeptical about it's current. The datasheet says under "Power Supply" that its 'power supply current' is 0.2mA (1.5mA max.). Is this it's quiescent current or what ??
assuming i supply 12V, 4A to it; would it blow up ? or is 0.2mA just the current that it would take out of the 4A and pass the remaining to ground ?

 

so after modifying, i want to do something like the attached. the pump requires 20KHZ pwm, while the peltier's 2Khz pwm.


however i'm wondering if 4A supply to the tc4422 may be too much.

 

so after modifying, i want to do something like the attached. the pump requires 20KHZ pwm, while the peltier's 2Khz pwm

Where are you getting these numbers for "required PWM"?
The PWM frequency should be chosen based on the system time constant, I don't think your motor has a mechanical time constant of 0.00005 seconds?

A frequency that is too high only increases your switching losses.

BTW- your peltier will be fine IF you use a big enough heatsink, otherwise it's gonna be toast. (overheated) Read the data sheet carefully!

 

the 20KHZ pwm for the pump was the one recommended by the manufacturer

 

i'm running the peltier at a low voltage and current, so i don't it would overheat. though it *will* have a heatsink attached.
I'm abit bothered about the pwm frequency myself. the manufacturer's rep. for the pump said it was recommended a pwm frequency of 20khz to operate it. it's a small 3Vdc pump : http://ph.parker.com/webapp/wcs/sto..._-1_14099_14099_14097_ProductDisplayErrorView

the peltier is a tellurex module. in their appnote, 60Hz or higher was suggested....

 

in tellurex's peltier appnote, http://www.tellurex.com/technology/peltier-faq.php , 120Hz was suggested as a pwm frequency to use.

 

@Sensacell, regarding the earlier 'ripple issue', i read that here: http://electronics.stackexchange.com/questions/28634/how-to-drive-a-peltier-element

I think driving the peltier at either pwm 60Hz or 120Hz should be fine ? As for the pump, i might look for another pump with better well-stated operating specifications. the datasheet of this one is pretty vague regarding pwm-operation

 

Interesting! I read about the response of the peltier junction with regards to PWM, thanks for sharing that, I learned something new!

Here is a linear version of the driver, it is a voltage-controlled-current-sink, this will give you a peak current of 3.3 A, and it's transfer function is linear.

Use a good single supply opamp that can go to ground at the output and inputs, the RC at the input should be 10X the period of your PWM.

This time you need a BIG heat sink for your FET.

 

actually i intend using the peltier to move a small load, about 1W (eh.. 400mW to be exact). Checking the curves (at both 27oC and 50oC) in the datasheet, running 1A,4.0 - 4.2V is sufficient to move 1W . So, intend to supply 4V , 2A and then PWM 50% (so it sees 1A).
about the tc4422, is supplying it 12V, 4A still okay ??