#### Ömer Faruk CİN

Joined Feb 4, 2017
6
I cool with 4 pieces of 70 W peltier. For this, I work with a 12V output SMPS. Since Smps draws constant 24A, I wanted to control it and found the circuit I shared in the appendix. I fed it through 12V SMPS in the circuit. I set up the circuit with the resistance values I chose randomly and it worked, but I could not use the potentiometer at its full efficiency.
According to the curve in the MOSFET, I have to keep the VG Voltage between Vth and VG = 5V for VDS = 12V so that current can flow between 0 and 24 A. Here I could not analyze what the values of the remaining resistors should be to use the potentiometer at full efficiency. It is the representation of the resistance values and potentiometers in the circuit. Mosfet's threshold voltage is min 2V and max 4 V. I don't have much experience with analog circuits, I have difficulty analyzing. Can you guide and help me analyze?

#### Papabravo

Joined Feb 24, 2006
15,750
I think you don't understand the circuit very well, and I'm not sure I understand what you mean when you say " use the potentiometer at its full efficiency"

I think you mean that you cannot use the potentiometer over the full range. Regardless of anything else you do you cannot achieve a gate voltage in excess of 6 Volts. Maybe this is OK because you input supply can only provide 24 A and the MOSFET can pass up to 100 A in that configuration. It may get hot because the Rds(on) will be higher than it needs to be. Because of the Vce drop of the transistor you maximum gate voltage might be 5.8V. RC and RE will allow about 600μA of current to flow.

With the wiper on the pot a 0%, bypassing the pot, you will have 6.2 mA of base current and giving the transistor forced beta of less than 1. I didn't think it was possible, but this would be severe transistor abuse.

Somewhere in the middle of the pot's range you will be unable to turn the transistor on and the gate voltage will drop to zero.

This is perhaps the worst control circuit I have ever seen

EDIT: The joke is on me. With a fixed load of 70 Watts @ 12 V (2.06 Ω), the pot will vary the load current by 14 mA (5.791 Amperes to 5.805 Amperes) over it's entire range. The transistor never turns off, and the only thing that is changing is the Vds of the MOSFET by a couple of millivolts. This circuit is not only completely useless -- it is worse than useless. It gave you the illusion that something was being controlled when in fact it is the load that controls the current through it.

Last edited:

#### crutschow

Joined Mar 14, 2008
26,992
Can't you adjust the voltage of the SMPS, which would be much more efficient?

#### Ömer Faruk CİN

Joined Feb 4, 2017
6
I think you don't understand the circuit very well, and I'm not sure I understand what you mean when you say " use the potentiometer at its full efficiency"

I think you mean that you cannot use the potentiometer over the full range. Regardless of anything else you do you cannot achieve a gate voltage in excess of 6 Volts. Maybe this is OK because you input supply can only provide 24 A and the MOSFET can pass up to 100 A in that configuration. It may get hot because the Rds(on) will be higher than it needs to be. Because of the Vce drop of the transistor you maximum gate voltage might be 5.8V. RC and RE will allow about 600μA of current to flow.

With the wiper on the pot a 0%, bypassing the pot, you will have 6.2 mA of base current and giving the transistor forced beta of less than 1. I didn't think it was possible, but this would be severe transistor abuse.

Somewhere in the middle of the pot's range you will be unable to turn the transistor on and the gate voltage will drop to zero.

This is perhaps the worst control circuit I have ever seen
My first attempt at analog circuitry. It is not surprising to me that it is the worst circuit you have ever seen. Regarding the control circuit, what would you do if you were. I am open to any advice.

#### Ömer Faruk CİN

Joined Feb 4, 2017
6
Can't you adjust the voltage of the SMPS, which would be much more efficient?
Even if I lowered the voltage, I would have used it in maximum cooling as it would draw current according to the load. Let's say I set the voltage as 6V, this time it will draw 48 A. Of course if SMPS is capable of this. Wouldn't it make more sense to set current?

#### Ömer Faruk CİN

Joined Feb 4, 2017
6
Even if I lowered the voltage, I would have used it in maximum cooling as it would draw current according to the load. Let's say I set the voltage as 6V, this time it will draw 48 A. Of course if SMPS is capable of this. Wouldn't it make more sense to set current?
not lowered, reduce*

#### crutschow

Joined Mar 14, 2008
26,992
I would have used it in maximum cooling as it would draw current according to the load.
That makes no sense to me.
Isn't the TE cooler a fixed resistive load so will draw current according to the applied voltage?

Last edited:

#### Papabravo

Joined Feb 24, 2006
15,750
I guess I'm not clear on where power is being consumed. Is it 70 watts for the four Peltier devices or 70 watts for each device and they are connected in parallel. So if the devices are 70 amps each and they are in parallel, then 5,83 Amps times 4 gets you to 23.32 Amps for all four, which accounts for the current draw from the 12V supply. The impedance of the loads are fixed. You can control the voltage or you can control the current, but you cannot control both of them independently. Many people who are new to analog design overlook this basic principle. It is called Ohm's Law, and violations are not tolerated, either in this world or the next.

To answer your question there is literally nothing we can do for you.

It might help if you could articulate exactly what you are trying to do, because it is my impression that you are trying to do the impossible.

#### Attachments

• 15.3 KB Views: 2
Last edited:

#### BobTPH

Joined Jun 5, 2013
3,101
Please explain what it is you are trying to do. It sounds like you are trying to control the current to the Peltier devices by using the MOSFET in the linear region. If you do get this to work, it will dissipate a huge amount of power in the MOSFET.

You can control the amount of cooling by using PWM instead, with no wasted of power.

Bob

#### Ömer Faruk CİN

Joined Feb 4, 2017
6
Please explain what it is you are trying to do. It sounds like you are trying to control the current to the Peltier devices by using the MOSFET in the linear region. If you do get this to work, it will dissipate a huge amount of power in the MOSFET.

You can control the amount of cooling by using PWM instead, with no wasted of power.

Bob
This is exactly what I was trying to do. As far as I understand, I chose the wrong method. Thank you for your help and suggestions.

#### crutschow

Joined Mar 14, 2008
26,992
If you can reduce the voltage of the power supply, that's the easiest way to go.

#### Ömer Faruk CİN

Joined Feb 4, 2017
6
If you can reduce the voltage of the power supply, that's the easiest way to go.
Thank you. This will be one of the first ways I'll try.

#### Papabravo

Joined Feb 24, 2006
15,750
Please explain what it is you are trying to do. It sounds like you are trying to control the current to the Peltier devices by using the MOSFET in the linear region. If you do get this to work, it will dissipate a huge amount of power in the MOSFET.

You can control the amount of cooling by using PWM instead, with no wasted of power.

Bob
Are you saying that the amount of cooling is or is not related to the power delivered to the devices? If you deliver 4 x 70 = 280 watts to four devices you get some amount of cooling. If instead you use PWM to cut the power delivered to the four devices to 140 watts you get less cooling. Is that what you are trying to do. You could be asking for 280 watts of cooling for an expenditure of 140 watts of electrical power, but that ain't happening. There is no scheme that will get you more cooling for less power.

#### BobTPH

Joined Jun 5, 2013
3,101
As I understand it, he was trying limit the current using a MOSFET. I said using PWM would be more efficient.

Bob

#### Papabravo

Joined Feb 24, 2006
15,750
As I understand it, he was trying limit the current using a MOSFET. I said using PWM would be more efficient.

Bob
His original circuit was spectacularly un successful in doing much in that regard since the MOSFET was always on. I don't think limiting the current was the intention, I think he wanted the same cooling for less power consumed.