Power MOSFET IRFZ44N

Thread Starter

IssacSutt

Joined Jan 6, 2017
37
78B7A925-BAAD-49C6-B8A6-6DDF39585E43.png

Hi, I’m just curious if anyone might be able to explain a possible over-heating issue with a mosfet/circuit, I designed the above circuit myself? I don’t have much hands on experience using mosfets, but I know how to work with them well enough to design my own decent circuits with them.... I have plenty of experience with bjts on the other hand, whether NPN or PNP I use them just as often as the other.

Alright, so the issue I’m having is the Mosfet heats up and feels like it just gets hotter and hotter gradually but over a short time, can probably only keep the power on for for up to maybe 15 seconds before it gets much too hot to touch. I’m not using a heat sink cause I don’t have one. I have tried adding another mosfet in parallel, but it still gets just as hot (however only one of them gets super hot -(the extra one), while the other doesn’t even seem warm). I’m honestly not sure why it’s heating up so easily if the datasheets say it can handle 49 Amps continuous and dissipate up to 94 or 95 Watt... The mosfet in this circuit is only dealing with near 1 Amp, and dissipating about 5.15 watts across the Drain-Source region. So honestly I’m kinda clueless as to why it’s happening or what I can do to fix it without getting into pwm.


Here’s Some more Info about the circuit above:

it’s supposed to function as a current limiter (not to be confused with current source). The circuit on the left side is the actual breadboard prototype, and the one on the right is the equivalent circuit. The Mosfet I’m using is (IRFZ44N) and the bjt I’m using is (2N3904). Theleft circuit some of you make wonder why I used so many resistors or 2 batteries in parallel... I did that just to obtain precise resistance values and to take care not to exceed their power ratings so they don’t over heat. As for the batteries, I put two energizer 9v batteries in parallel because they can’t handle much current by themselves.... but even with two of them the overall supply voltage still drops to about 5.3v while in use (but that’s a lot better than about 2.4v with just one).

My intended calculated current limit for this circuit is about 900mA (0.61/0.6666... =0.9), however I’ve measured an actual current limit of about 850mA while under test, so that’s a +/-50mA... not too awful bad whith a current so high and a very unreliable power source.

Just for the extra info.... The Bjt, (over a very wide range of testing with different shunt resistors and Loads *including a short as a load*) has proved to hold a pretty constant max of a 0.6-0.61 volt Vbe drop even at almost 1Amp, so I know my designed circuit actually functions pretty well including given that I’ve also tested the load current with every various intended current limits, it’s usually had an accuracy of about +/- 0.0005A (500uA) unless the current is really high, which is not bad I would think considering it’s simplicity.

Please let me know what you guys think, I’m not sure where to go from here yet, I’m a little stuck right now.

If I can get this circuit to function without overheating anymore then I eventually plan to use it as a high power driving circuit, pulsed with a 1% duty cycle @ 1 Amp for an IR emitter I have
 

Attachments

ronsimpson

Joined Oct 7, 2019
812
Circuit on the right; I added a 1K resistor to limit the Base current. Other wise there is no current limit to the Base of the little transistor. Your circuit works. Just saying this is common practice.
1595388010017.png
Heat problem: The data sheet might say 50watts or 94W but this is under conditions you don't have. A heat sink the size of a house. With no heat sink the junction temperature goes up 64C/watt. So at one watt and a room air temp of 25C you get 89C inside the transistor. This will burn your finger. With out a heat sink the MOSFET in a TO-220 package is only a 1 to 1.5 watt part.
If you could reduce the voltage across the MOSFET the heat will go down.
 

Thread Starter

IssacSutt

Joined Jan 6, 2017
37
I appreciate the info. But I don’t understand why you say the 1k resistor for the bjt is necessary if that transistor only sees 0.61v max, because at that voltage, the current will be low on its own and the 1k wouldn’t make a difference, especially considering the voltage across the transistor at higher load resistances would be 0v Regardless because there’s a 0.6 Ohm resistor across its Vbe junction... meaning the transistor will always see less than 0.6v unless there’s adequate load current, and even then only up to about 0.6v. What I’m trying to say is The 0.6 Ohm shunt resistor basically already is limiting the base current by limiting the voltage at its junction right? Also The bjt transistor is cold to the touch the entire time while the circuit is powered on, so how would that help? Don’t take what I’m saying as offensive, I’m only questioning things to help me try to see it from your point of view, I’m not doubting your experience or knowledge by any means
 

drc_567

Joined Dec 29, 2008
867
A possible explanation for the mosfet heat generation is that the gate voltage transient is too slow and causing the mosfet to spend too much time in the so-called 'ohmic region', before it is completely turned on and exhibits the milli-ohm resistance property given in the datasheet. The base resistor that is added to the small npn transistor would be responsible for slowing the gate charge drainage, thereby increasing the gate charge rate.
An oscilloscope view of the two different gate voltage transients might confirm this.
 

Alec_t

Joined Sep 17, 2013
11,509
The circuit on the left side is the actual breadboard prototype, and the one on the right is the equivalent circuit.
It isn't equivalent. The load current of the left circuit is about twice that of the right circuit, because the bjt base on the left is fed via a 2:1 potential divider.
I have tried adding another mosfet in parallel, but it still gets just as hot (however only one of them gets super hot -(the extra one), while the other doesn’t even seem warm).
That imbalance is because no two FETs are identical. The one with the lower Vgs(thr) value will hog most of the current and get hotter.
 

KeithWalker

Joined Jul 10, 2017
1,139
What the voltage between the gate and source on the IRFZ44? It is heating up because you biasing it partially on. It is dissipating the power ( Vds x load current ) as heat.
If this is the way you want it to work, you will need to use a heat sink.
Regards,
Keith
 

Thread Starter

IssacSutt

Joined Jan 6, 2017
37
Alec_t, thanks for pointing that out, that’s actually a mistake in the schematic, I did not actually wire the base in that way on the breadboard, the base of the bjt in the schematic should be tied directly to the mosfet’s source just as it is on the right circuit. I’ll re-post an updated schematic.
 

LesJones

Joined Jan 8, 2017
2,687
My understanding is that the left hand circuit is designed as a constant current supply passing about 2 amps through the two 1 ohm resistors in parallel. If I am correct then the mosfet will be dissipating about 13 watts. This would require a heatsink on the mosfet. The right hand circuit is similar but will generate a constant current of about 1 amp (Through the two 1 ohm resistors in parallel) In this case the mosfet would be dissipating about 7.9 watts. This would also require a heatsink. This is based the power supply being 9 volts. As you seem to be powering the circuits with two PP3 batteries in parallel I don't think they would supply 9 volts with such a high current demand so in practice the dissipation would be a bit less than the values I quoted.

Les.
 

Thread Starter

IssacSutt

Joined Jan 6, 2017
37
Update:

I just tried adding a 1k base resistor not long ago, and the mosfet stopped heating up, but every resistor in that circuit got too hot, (other than the 47k)
 

ronsimpson

Joined Oct 7, 2019
812
I just tried adding a 1k base resistor not long ago, and the mosfet stopped heating up, but every resistor in that circuit got too hot, (other than the 47k)
The 1K should not do anything under normal conditions. I build power supplies. People do bad things to power supplies. (short the output, charge batteries, etc)

If the resistors get hotter and the MOSFET gets cooler then most likely the Junction transistor is not working leaving the MOSFET on hard. (almost zero volts D-S) This will put most of the battery voltage across the resistors.

If you change the two 1 ohm resistors like this it will reduce the voltage across the MOSFET and reduce the heat some.
1595456954793.png
 

Thread Starter

IssacSutt

Joined Jan 6, 2017
37
Thanks for the feedback romsimpson, everything you said makes perfect sense, I thought about it the same way. I originally did have the two resistors in series like that, but they couldn’t handle the power, and so I calculated the difference when rearranging them and found that it reduced the dissipated heat a good bit more than when in series, (assuming that the current will stay constant at 850mA, then with the two resistors in SERIES, the voltage across the two in total should be 1.7v for a current of 0.85A which means 0.7225 watts for each resistor...

...but with them in PARALLEL, each resistor should see a voltage of 0.425v for a current of 0.85A split equally between the two which should mean a dissipated power of 0.180625 watts for each resistor). Please check my math and correct me if I’m wrong, but the two 1 Ohm resistors draw less power when in parallel than it does in series right?
Although, I do realize that even if this is true, the mosfet will dissipate much more heat with the resistors in parallel than in series, so I guess this is just a loose-loose situation; either the resistors get screwed or the mosfet gets screwed.

Although, ultimately all I’m really trying to do with those load resistors is stimulate the IR emitter without actually using the IR emitter because I don’t have as many of those to sacrifice as I do the resistors. So knowing this, do you think there might be a better way to stimulate the IR as a load so I don’t have to risk blowing it up??
 

Thread Starter

IssacSutt

Joined Jan 6, 2017
37
Another update to correct something guys... when I re-wired the circuit to incorporate the 1K resistor, I went back and reviewed it cause it didn’t sound right with the mosfet not even getting warm at all, and it turned out I somehow accidentally left out the source pin, because I had tied the npn’s base/upper lead of shunt resistor direct to drain instead, which completely explains why all the resistors got so dang hot and the mosfet not even warm, cause the mosfet wasn’t even being used. SO, long story short, I fixed that, and re-tested the circuit using the 1k base resistor and got the same results as with no base resistor.
 

LesJones

Joined Jan 8, 2017
2,687
You seem to be doing the electrical calculations but not the thermal calculations. Looking at the data sheet for the IRFZ44N (http://www.farnell.com/datasheets/1916237.pdf) the junction to case thermal resistance is 1.5 Dec. C per watt. The junction to ambient thermal resistance is 62 Deg. C per watt. So the case to ambient is 60.5 Deg. C per watt. So even with the dissipation you quoted of 5.15 watts the case of the mosfet will be 5.15 x 60.5 = 311 degrees C above ambient. So the only solution you have is to add a heat sink with a suitable thermal resistance.

Les.
 
Last edited:

Alec_t

Joined Sep 17, 2013
11,509
As an aside, breadboards are not designed to handle heavy currents. 0.85A may cause internal heating such that the contacts lose their springiness and result in poor connections.
 

Thread Starter

IssacSutt

Joined Jan 6, 2017
37
Thanks for all the input, I see what your saying but I don’t have any other means of building the circuit other than with a breadboard or 9v batteries. As for the thermal calculations, I also don’t have a heat sink... so I guess this just means I have to find another project to work on for now if I don’t have all that stuff. Thanks for all the help and advice from everyone, I appreciate it.
 

ronsimpson

Joined Oct 7, 2019
812
I also don’t have a heat sink...
Aluminum or copper makes good heat sinks. (not iron) Do you have a bolt and nut that will go into the transistor hole? A home made heat sink is not as good but worth trying.
Maybe you can find a heat sink in a old dead TV or radio.
1595560492811.png
 

LesJones

Joined Jan 8, 2017
2,687
Old PCs are another possible source for heat sinks. There will be some in the power supply and one on the CPU. The one on the CPU will probably also have a fan on the heat sink that will help to remove the heat. As you say you don't have a power supply that would be a useful project to build but that will also need heat sinks. There are also heat sinks on Ebay but none of them seem to give the most vital specification. The thermal resistance. Strip board and other prototyping boards are also available on Ebay.

Les.
 
Top