PWM Controller

[rcheshire]

This is the new version - a new thread as there are a number of differences. I have deleted a potential duplicate post.

The new version uses an opamp or voltage comparator IC to control a 555 timer. In some respects it functions a bit like PID.

The controller can be used to vary fan speed or maintain a TEC setpoint, not both simultaneously. I guess that's possible but not intended.

For fan control, the PWM frequency remains within the specifications required of computer fan control (21kHz - 28kHz and 5.25V max) throughout the full PWM range with a suitable C1 (frequency) value.

Primarily intended for cooling, frequency over the entire PWM range is again determined by the application of a suitable C1 (frequency) value. The MOSFET gate is driven at 10V for low drain resistance.

In auto the output of the comparator is fed to the 555 control pin to vary PWM pulse width at whatever duty cycle is selected. At setpoint, the system defaults to a narrower pulse width and goes to zero when below set point.

Auto would normally be operated at 100% PWM, selected by the other potentiometer. Either way, setpoint control can be fine tuned by the pulse width pot, if required, in auto mode.

Temperature control can be turned off and the unit operated as a basic PWM controller. This is also the default fan control mode - until I get time to test the fan in auto. Although for cooling running the fan at high speed is the usual practice - the fan can be plugged into the board without connecting the PWM lead.

Preset temperatures will be marked on the reference setting pot, calibrated using the linear characteristics of the TMP36 sensor - no need for a digital readout of any sort - one LED indicates the operating mode and the other cooling and modulating about set point.

EDIT: Preset cooling differentials could be provided in manual mode - this would require a specific cooling system model, as each system is different.

I used CircuitLab - EveryCircuit (Android) and KiCad to design and produce the PCB. The 555 PWM circuit is adapted from one of the many on-line examples, as it works well.

Gerber files are available on request... other than the through hole components the SMD components are 1206 - fairly easy to manage with a regular soldering iron. There is no need to build the entire board for a manual system. The MOSFET can be omitted as can the opamp/voltage comparator if wanting fan control only. Similarly, add the MOSFET for TEC control.

[rcheshire]

The screen shots show the behaviour of the system as it approaches set point temperature 0C (500mV) in auto. In this case at ~1Hz.

From the left - the pulse width diminishes getting close to set point, within set point +0.1C. At set point the pulse width further decreases to a lower duty cycle and at -0.1C duty cycle = 0.

Typically, the system will modulate either side of set point.

The second image shows the same behaviour with 50% duty cycle set by the PWM pot in auto.

Fan control at 25kHz is best operated in manual mode.

[rcheshire]

A couple of screen shots of the board in its final draft.

[Garbz]

Looking good. Are you expecting any power dissipation in Q1 during switching? If so it may be wise to create a place to solder it to the PCB to allow the PCB to act as a heatsink.

[rcheshire]

Thanks Chris. No power dissipation to speak of, providing drain resistance is <~1.3mOhm. Although the gate voltage will be around 10v so rds on could be a little higher. It would be safer to provide a solder pad all the same. Using KiCad these days so very easy to create and modify footprints.