heating problem of mosfet

[roykyn]

irf3205 circuit

hi i have a modem with mc34063 switching regulator.i am trying to use an external mosfet for switching purpose.but the mosfet is getting heated a lot.i am using irf3205 the datasheet says the Rdon (on resistance) is 8.5milliohms.but in practice it is dissipating a lot of heat.the dc current is less than 1amps.i am using 150microH toroidal coil at the output.

 

[pauloynski]

driver mosfet irf 3205 with transistor

Many possibilities but first make shure:
1- Core is not saturating
2- Voltage enough to the mosfet gate (something like 12V unless it will not achieve the lowest Rdson)
3- Switching speed ( that may be related to speed that Vgs changes, the amount of current sourced and sinked from the gate, etc...)
4- Poorly defined snubber at the drain (allmost allways a snubber is required and it depends on the topology, current, etc...)
If this is not enough please post your circuit and tell us what is it expected to do.

 

[roykyn]

do it yourself high current mosfet switch supply

here is the circuit i am using..i have tried irf3205 irf640 and 2n3055.input voltage is 9-15volts.switching speed is around 50kHz.

 

[roykyn]

mosfet 34063

this is another circuit i ve tested.

 

[umery2k75]

understanding the fet

I'm thinking if switching loss is not causing the problem.
The MCM34063 regulator can go at upto 100Khz, you're at 50Khz. It's fine
The IRF3205 has given a graph



See this graph, if your not stressing up your MOSFET.In the datasheet it's also seen that the maximum safe temperature is around 62 degrees Centigrade/W.

Power=12*850mA(Less than 1 Amp)
Power=10.2W



You have 62'C per W
so 62*10.2=632'C

That much temperature 632, this can't be true.Some one explains about thermal thermal resistance and about 62'C per Watt.

 

[pauloynski]

34063 high current buck

To achieve the lowest Rdson you need to source a current up to 1A to the gate of the mosfet untill it conducts. Only then the gate current will go to allmost zero. The large resistors you have in your circuit will never bring the mosfet to its optimal state. The same apllies when switching off. Furthermore, I think you need to place a diode from the inductor´s input to ground. Remember: it is impossible to instantaneously switch off an inductor (the same way you can´t instantaneously discharge a capacitor).

Added after 16 minutes:

Also take a look at https://www.richieburnett.co.uk/mosfail.html

 

[roykyn]

34063 high current

1A current ??? mosfet is voltage controlled device.there is a diode at the output sorry i forgot to include it.

 

[pauloynski]

problem of mosfet

A mosfet is voltage controlled device after you have switched it. While switching it is not. See

Added after 30 minutes:

And also this https://www.fairchildsemi.com/an/AN/AN-558.pdf

Added after 40 minutes:

And if you think 1A is too much take a look at this https://www.linear.com/pc/downloadDocument.do?navId=H0,C1,C1003,C1142,C1041,P8290,D5777

 

[roykyn]

problem with mosfet switch

wow.nice.but in my application i want to be super efficient.even the 12v 1.5a transformer is costly and bulky.smps is the next option but because of the cost i cannot use it.so on a whole my supplying power is 12v and 1.5a.

 

[pauloynski]

mosfet heating isolated supply

What I was trying to say is not that you need to buy a driver but that you need to decrease the impedance of your actual driver to meet the mosfet´s requirements. Google for mosfets high side drivers and you´ll find a few using components similar to the ones you have in your circuit like this **broken link removed**.

 

[Siniša]

cd40106 buck

....but in my application i want to be super efficient...

High current supplied to gate lasts very short in time, which is the point of suppying enough current in first place. Idea is to charge and discharge gate capacitance as quickly as you can in order to keep mosfet either fully open or fully closed.
It is with low gate current that you allow mosfet to spend too much time in mode where it dissipates alot and then you have very high loses.

 

[pauloynski]

mosfet shortcut

Another important thing is that you need to keep the voltage at the gate some 10 to 12V above the source voltage when conducting, so you´ll need a bootstraped driver. If you want to be minimalist why don´t you give a try to the external PNP transistor configuration as shown in the mc34063 datasheet from Onsemi?

 

[roykyn]

gate current on fet watts

i tried a new circuit today using irf3205 and is as shown.the voltage across mosfet is 4v at 1.85 amps continuous current.hence the power dissipation of mosfet 7.4watts and the resistance 2.16ohms and not 8 milli ohms as mentioned in the datasheet.
if it was 8 milli ohms then the voltage across the mosfet should be less than 0.5v.
how can this resistance be achieved because in computer smps the current is more than 30amps.if the same mosfet was used at same resistance then it would dissipate 1.8k watts of heat.

 

[Siniša]

1a mosfet data sheet

For starters, I would move this thead to "Electronic Elementary Questions". Second I would suggest you to find some literature on FETs and MOSFETs by using search on this board and read it. You do not have basic understanding of FET needed for implementation.
Although some errors might occur from time to time in data sheets, I would first question method of measurement before questioning data sheet . Don't waste both your and our time by trying to take shortcut without getting yourself acquanited with theory you need so much. Once you fully understand how FET works, you can get on with experimenting and learn more complex circuits with them. Crawl before walk.

Here is some link for applet that can help you visualise what you read in literature I mentioned above. You have both JFET and MOSFET explained.

 

[roykyn]

solve the heating problem on fet

I found out the solution. I have tested this and the resistance of the mosfet drops to less than 0.1ohms.
The mosfet acts as a thyristor. To switch ON a mosfet you need to create a isolated secondary supply like using a diode and a capacitor. Remember that this secondary supply should be connected to gate and the negative terminal of the mosfet. I have attached a figure that will help.
We have to note that the driver circuit has to be super efficient. when the mosfet switches ON the voltage drops to zero and the driver circuit is powered by capacitor alone. The capacitor should be large enough or the power consumed by the driver circuit should be very less. I suggest using CD40106 or low power cmos opamp used as Schmitt trigger.

Please reply back if I've missed some thing.

 

[FvM]

mosfet heating problem

Yes, pauloynski already pointed out, that you have been missing a bootstrap driver. As a another circuit flaw, you have mixed up pin 6 and 7 of 34063, so the current limit is not working, possibly causing unpleasant high transistor and inductor currents.

 

[roykyn]

I didnt use the resistor because. it results in heat dissipation and hence waste of energy. if you have any other method please suggest that.

 

[FvM]

If you need the current limit function, 34063 doesn't provide another option. It's a simple low-cost switcher, not particularly designed for high performance. But perhaps, current limit is not needed in your application.

As another point, I just saw, that your circuit hasn't a free-wheeling diode, which is required for regular buck operation.

 

[Tralfaz]

Without a freewheeling diode between the Q3/L1 junction and ground, L1 is merely a resistor. Suggest adding a diode with cath to L1. Looks like some capacitance is needed on the output too.