Continue to Site

# Mosfet based dimmer - Heat Calculation & Suggestions

Status
Not open for further replies.

#### vishweshgm

##### Member level 4
I want to drive a 3kW resistive loadusing Mosfet dimmer circuit. Circuit I plan to use is the following:

Please ignore the 555 circuit part. I will be replacing it with a mosfet driving circuit which in turn governed by using micrcontroller.
I rigged this circuit on a bread board and found that MOSFET STW34NM60N heat too much. (loadd 250W and input230V,50Hz)

I need your help to see What Mosfet can I choose and how do I calculate heatSink parameters forit. Online tutorials for heating dissipation shows formulae for switching MOSFETS, but here driving technique involves just switching on at a particular phase every half cycle based on dimming requirement.
--- Updated ---

Edit : I just realized, I posted this question Microcontroller section, Can anyone move it to Power Electronics Section please?

Hi,

Safety first:
230V on a breadboard? For me a "no-go".

***
There is no STW34NM60N in your schematic.
--> To avoid confusion: Please post consitent informations only (pictures, text, schematics...). At least mark unused / modified parts.

So I have to guess that you replaced the IRFP460 with the STWxx.

To calculate the heatsink you need to do first things first.
1) requirements (current, switching frequency, ambient temperature ...)
2) calculate power dissipation (different dissipation types)
3) calculate heatink (using ambient temperature ...)

to the points:
1) current: One may think this simply is 250W/230V = 1.1A. But is this really the case? we don´t know. Mabe the heater is made of PTC or made of metal wire. Example: An 60W incandescent lamp may have the 15 time inrush current (cold filament) compared to the steady current.

2) there is
* ohmic condution loss of the MOSFET during ON time
* switching loss
* maybe diode loss

3) consider: ambient temperature, air flow, closed case ...

Klaus

I blurred the parts that I didn't use and highlighted those that I did. Also Mosfets were not munted on breadboard It was kept aside. Only uC based driving circuit was on breadboard. Also load that I used is normal incandacent bulbs not the heating coil.

To calculate the heatsink you need to do first things first.
1) requirements (current, switching frequency, ambient temperature ...)
2) calculate power dissipation (different dissipation types)
3) calculate heatink (using ambient temperature ...)
Thanks for this.
So I prepeared an excel and calculated the temperature to be 38degC. But when I did experiment It was well over 50degC

So please guide me what else am I missing here?

 Max Ambient temp, T_a (ºC) 35​ (ºC) Target Power rating 200​ W Voltage (VAC) 230​ V Irms 0.869565217​ A Iavg = 2*root2*Irms/pi 0.782883753​ A Rdson 0.105​ Ohm P=RDSon * Irms ^ 2; 0.064355232​ W R_thj-a 50​ (ºC/W) T_jmax 150​ (ºC) R_th(j-c) (ºC/W) junction to case 0.5​ (ºC/W) R_thc-h case to heatsink 0.3​ (ºC/W) Rthh-a heatsink to air 10​ (ºC/W) Junction Temp, T_j Amb = Ta+P*Rthj-a 38.2177616​ (ºC) Junction Temp, T_j Heatsink = Ta+P*(Rthj-c+Rthc-h+Rthh-a) 35.69503651​ (ºC)
--- Updated ---

The values in above calculations shows both with and withput heatsink. With no heatsink, heating was too much (above80-85degC) but with heatsink, heating was above 50degC

Hi,

Safety: Breadboard still runs on mains voltage. At least I don´t see any isolating parts.

***

* inrush current / cold resistance
* switching power dissipation
* diode dissipation

Frequency, duty cycle range, driving voltage, driving current.....

****
Also: for power dissipation on a resistor:
always use RMS_current.

Klaus

Your gate R of 560 ohms looks high which would keep MOSFET in active region
too long.

These might help -

The input C of the IRFP huge at > 3000 pF, so I question the ability of 7555 to drive it
with enough current to get it out of active region fast enough. I do see your comments
about replacing it, good idea. And the drive has both High and Low side MOSFETS on simultaneously for some period of time, no dead band timing.....as shown by ph1 and
ph2 in this timing diagram

Regards, Dana.

Last edited:

Status
Not open for further replies.