If anyone has an idea of what is going on without that gate resistor to ground - im intrigued to find out.

The Rgs was used to turn off the mosfet when the input is low, sometimes it was used to turn off the mosfet quickly.

 

If you like then you can still using the resistor.


I just drawn the normal driver circuit, If you want then you could take the bjt away, and using the port of uC to drive the LED.

I appreciate the responses!
I am going to place an order for a bunch of random bits and bobs so i will be looking into the zener and getting some more BJTs.

im not sure if i have been going in circles today haha. But i have managed to narrow down a few issues.

My mosfet still seems to get too hot and smoke when it shouldn't be (when i am plopping 10V to the gate with no PWM) so i need to figure that one out.

Also, my variables i am using to control my PWM duty cycle on my Micro seems to be straying and resulting in improper switching.
The output of the opto (when i feed the gate voltage to short it to ground) varies from 50 to 100% instead of 0 to 100% - so i need to figure if this is simply the unclean PWM or how im utilizing the opto.

I appreciate all the responses ! sometimes i feel like a lost child when my brain starts to melt haha.

Cheers again !

 

One issue sorted - the optocoupler does not vary from zero to full.
When the PWM output is at 100% (shorting the gate to ground, therefore should be zero V) i can see that it is only managing to drop to half its max before rising again.

This has to do with the PWM frequency and the optocoupler maybe.

The circuit I drawn was the normal application, if input is low then the mosfet will be turn off, if the input is high then the mosfet will be turn on.

 

If you want to take the Q1 away then you can change the Zd1 to 5.1V or as you wish change to resistor, but you should modify the voltage to 5V.

 

i just went with the method used in the datasheet as it worked for a high voltage configuration doing the same job but with a 240V bulb and with an IGBT not a mosfet.

Also, why 5V? the mosfet requires 10 to be on fully.

 

i just went with the method used in the datasheet as it worked for a high voltage configuration doing the same job but with a 240V bulb and with an IGBT not a mosfet.

Also, why 5V? the mosfet requires 10 to be on fully.

You can also change the mosfet to using the logical level 4.5V for Vgs of mosfet.
Or you have the port of open drain to drive the LED of optocouple.

 

You can also change the mosfet to using the logical level 4.5V for Vgs of mosfet.

haha that makes way more sense!

 

The circuit shown on #26, that was designed for the uC, it is easy to control by 5V logic.
If you choosing the Rds too big then the mosfet will easy to get hot when the current is large.

Vds, Rds, Ids and logical level, sometimes you should make a choice to suit the application.

Edit : The circuit shown on #16.

 

Here are a few possibilities:
Your FET has a pretty high Rds on of .134 ohms. So at 7 amps it will dissipate 6.5 watts. So even if things are working well it will need a pretty good heat sink.
The 4N24 has a minimum transfer ratio of 20%. Meaning if you need to sink the 7 ma thru the 4.7k you might need to drive the LED in the opto with 35 ma. I doubt your micro can do that.
You really should use a current limiting resistor in series with the opto diode. It will make both the micro and perhaps the opto happier.

 

Here are a few possibilities:
Your FET has a pretty high Rds on of .134 ohms. So at 7 amps it will dissipate 6.5 watts. So even if things are working well it will need a pretty good heat sink.
The 4N24 has a minimum transfer ratio of 20%. Meaning if you need to sink the 7 ma thru the 4.7k you might need to drive the LED in the opto with 35 ma. I doubt your micro can do that.
You really should use a current limiting resistor in series with the opto diode. It will make both the micro and perhaps the opto happier.

This post just made me realise my stupidity .... i am well aware my pic can only sink/source 25mA but for the duration of this project - my brain forgot to pass that on haha! thanks
i guess ill need to look into BJT hooked up to the PWM and VCC for the input of the opto. However - i did use this exact set up with no issue using a 1k resistor between the opto input and pwm output though that was sinking 3mA.
As for the on resistance - i was hoping it wouldn't cause too much of an issue running short term. this problem will inevitably be made worse by the massive inrush from the cold filament.

i have been using a current limiting resistor on the input to the opto - 1K previously and now 100 ohms. - though if the MCU can only put out 25mA and the opto can suck in 60 - is there really a need for a current limiting resistor if i hook it up directly? (ignoring the insufficient input driving current)

Edit: im using a 4n25

 

The escape answer is it depends.... If the spec. is 25 ma I would limit it to 25 ma. then your safe.
http://www.ruggedcircuits.com/10-ways-to-destroy-an-arduino/The 4N25 typically has a transfer ratio of 50% but worst is 20. So if you add a 150 ohm in series with the diode you should be ok.
This would let it sink 12.5 ma. or maybe as low as 5ma. You could make your dividers a little bigger and this would also worst case. But my bet is you just need a FET with better Rds on. Usually the lower voltage ones have lower Rds on values.

http://www.mouser.com/ds/2/389/DM00095160-490569.pdf

 

The escape answer is it depends.... If the spec. is 25 ma I would limit it to 25 ma. then your safe.
http://www.ruggedcircuits.com/10-ways-to-destroy-an-arduino/The 4N25 typically has a transfer ratio of 50% but worst is 20. So if you add a 150 ohm in series with the diode you should be ok.
This would let it sink 12.5 ma. or maybe as low as 5ma. You could make your dividers a little bigger and this would also worst case. But my bet is you just need a FET with better Rds on. Usually the lower voltage ones have lower Rds on values.

http://www.mouser.com/ds/2/389/DM00095160-490569.pdf

Great help with the info! thanks alot!

Managed to get it all working (to a level i am happy with to initiate sleep), I removed the resistor in series with the opto (ill stick a low value on tomorrow).
The resistor i added from the gate to ground stopped latching with no gate voltage.
The resistor in series with the gate seems to have improved switching by eliminating some ringing(?).

Now i just need to iron out the mountainous creases.
Thanks for all the input people!

(also, i have done nearly half of the stuff in that link too my Ready for Pic board from shiteroelectronika and it is still going strong haha)

 

................................
EDIT: fixed an issue - with no gate voltage - the mosfet smoked and the bulb was very dim.
I added a 700K resistor from the gate to ground and now it works as expected when applying the gate voltage and removing it!
..............................................
If anyone has an idea of what is going on without that gate resistor to ground - im intrigued to find out.

No gate voltage (unconnected) is not zero gate voltage.
A MOSFET gate looks like a capacitor and has a very high impedance thus, if left open, the gate can be at any voltage (floating) which can turn on the transistor partially and cause the high power dissipation you noted.
Never leave a MOSFET gate floating.

 

No gate voltage (unconnected) is not zero gate voltage.
A MOSFET gate looks like a capacitor and has a very high impedance thus, if left open, the gate can be at any voltage (floating) which can turn on the transistor partially and cause the high power dissipation you noted.
Never leave a MOSFET gate floating.

agreed!
Though it happened when i tied it to ground? however i have thought i tied a few tings to ground and turned to see them floating :S
Thanks!

 

UPDATE:
Optocoupler was not sinking enough current with its old config, this was remedied by altering the voltage divider to lower the gate current the opto was require to sink. Opto seems to be working fine !
Here is my updated schematic which works.

I swapped out the MOSFETS I had for the IGBT i used for my 240V circuit and it seems to enjoy the work a-lot more than the smoking FETS!

If anyone has any tips for making the MOSFET operate at sub hazardous temperatures: im all ears!
I understand it will dissipate relative to current and its RDs ON value however something seems to be taking it beyond that.

Another issue i have is that both the MOSFETS and the IGBT seem to latch when i quickly ground the gate at 95% duty or in other words, when i pulse the gate to 5% duty. Though this is only with a cold filament.
When I have the circuit powered on with the gate voltage at zero (PWM is 100% - so opto is shorting to ground and the bulb is off and cool), if I quickly adjust the PWM duty to 95% then back to 100% the mosfet/igbt latches on. (NOTE that the PWM duty cycle is inversely related to the bulb brightness)
It will turn off again if I lower the duty then return it to 100% (i.e if i increase the bulb brightness then decrease it to zero again).
So my guess is the inrush current from the cold filament is causing this latching - given that this issue is not present when the bulb is allowed to run for a moment.

If anyone can shed some light on the latching observation i would be intrigued to look into it.

Also any insight to what i'm missing with the MOSFET power dissipation would be greatly appreciated.

Thanks AAC!

 

Just hooked up the MOSFET to the scope checking the trace between the drain and source. I can see the pulsed DC however it seems i am seeing +1V across the drain and source, this rises steadily in the next 40 seconds or so to around 4V+ .
I unhooked the power after that. The mosfet is rated for 600V and 20A - i am running no more than 8A steady but probably around 70 Pulsed - though i cant confirm this with the equipment i have at hand.

 

Do you have it on a heat sink?

 

a very very ..... very bad one. Its like a simple metal clip with a rubber grip to open and close it.
I used an igbt that worked alot better. It was showing only 0.8 across the collector and emitter of the igbt and it did not raise rapidly like the mosfets.
It also remained a-lot cooler, i think i may need to reassess my choices of switching components but in the meantime, i want to try figure out what is causing this.

 

i may try a thermistor in series with the bulb to limit the inrush and see how that affects the voltage rise across the drain and source of the mosfet.

EDIT: limited the current for about 3 seconds then "POP" full brightness and the temps started to rise!

Thermistor seems fine though

 

I'm not sure why you want to using the reverse way to control the mosfet or IGBT, but here you haven't given the enough Vge for IGBT, the Vge should be >=15V.

The normal control method was shown on #16, input high then turn on the mosfet, that's a quite normal method and save the energy, could you tell me do you have any special reason to used that way to control?

What's difference between that using zener diode and the voltage divider to supply the power, when the input voltage varies then the output voltage will be follow the changing, that's your circuit did, but when you using zener diode then the output voltage only just has a very small changing.