Help needed in Using IRFP150n Power Mosfet.

I have two questions about this device which are not mentioned in the data sheet. I have plenty of them so I have to use them.

1. Can I use these devices in parallel (as I have seen people used in inverters). But have a look at my config.

2. The graph of Vds Vs Id curve shows the relation with respect to pulsed voltage. In my case it will be a pure DC current and voltage. 30V of source and load will be of 10V min and 50 Amps current. So voltage across drain-source will be 20V at worst. If the config is possible then how much mosfets do I need I parallel?

Ambient temperature =40 – 42 degree C.

u can connect two mosfet in parallel for each mosfet iin diagram, but you need to take into account that both mosfet share the equal current.
if each mosfet share 25A current then it will work fine.

MOSFETs share current easier than BJT's since they have a PTC characteristic with Resistance vs T. But you still have a huge problem dissipating the waste heat.

Dear Saj,
Why do you want to waste 1kW of heat to drive a 500W load? Is this a battery load tester?
You know each device is rated for 160W @25C derated 1.1W/'C so if you intend to say run @ 125'C then it is derated 110W so only good for 50W but depends on your heatsink. If you use liquid cooling to boil water say with devices, you can make chai in 5~10 minutes.
So design choose your heat sink for 1kW then then thermal resistance of device, then temp rise above ambient then max dissipation per device and best case you need 6 devices in a large tub of 25'C water. Worst case you may need 20 devices with air cool.

I suggest you use std design Class D PWM with suitable inductor cap to simulate linear load. much cooler. then you might get away with 2 devices running 50% PWM @ 100Amp and choke rated for 100A. @ 100KHz.

In short ( pun intended , you need a series loss device that is delta V/I= 20/50= 400mΩ @ 10V,50A 30V V+ rated at 1kW or
a PWM load 200mΩ @100A 50% d.f. Considering each device RdsON min= 36mΩ but then you need to external series Ω to make up the difference (200 - 36/n) [mΩ] for n devices. This could be a lossy choke. Consider https://www.coilws.com/index.php?main_page=index&cPath=71 consider 1 µH @100KHz = 628 mΩ . These chokes are rated for 70A DC. Ask advice from supplier for PWM mode @ 100KHz for example. DC resistance < 1mΩ. >

Consider a resistor ,as a toaster with nichrome wire and adjust to get suitable resistance. dont need to even use a MOSFET for that matter. You may need to rewire the toaster for parallel operation inside with DC 50A @ 20V instead of AC 10A @ 120V so use say 5 or 6 parallel strings inside toaster and dissipate 1kW easily. using AWG 8 or 10 to load.

If you want to make a progammable load cell. You can use the toaster as a current limiter and external heat loss device and then use several parallel MOSFETS running with limited range of duty cycle and choke Cap then plot V vs I on digital scope.

Sajjadkhan said:

...pure DC current and voltage.... 50 Amps current.... 20V at worst.... Ambient temperature =40 – 42 degree C.

Click to expand...

Wow, that's a lot of continuous power dissipation! 20V * 50A = 1000 watts maximum.
Hmm, Let's see....

If you use 20 MOSFETs in parallel, each one will dissipate up to 50W. According to the datasheet, the case temperature must be no more than 110 degrees centigrade for 50W dissipation.

The MOSFET cases will be hotter than the heatsink, though. The temperature difference will depend on how you mount them, but if we assume a difference of 20 degrees, then the heatsink must be at no more than 90 degrees.

With an ambient temperature of about 40 degrees, the heatsink temperature rise must be no more than 50 degrees. 50 degrees / 1000W = 0.05 so you would need a heatsink rated at least 0.05 degrees centigrade per Watt. That's the bare minimum, with no safety margin.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

The next problem is how to get the MOSFETs to share the current equally. There's two issues here:

• There will be a difference in Vgs between the MOSFETs, so some will switch on sooner, and conduct more current than others. To minimize this, it would be a good idea to measure and match them.
• Vgs reduces with increasing temperature, so the hotter ones will draw more current and get even hotter, while the cooler ones will draw less current and get even cooler. If you put a resistor of say 0.5 ohms in series with the source of each MOSFET, that will help to equalise the currents and prevent thermal runaway.

Good luck with the project. Whatever it's for, I can't help thinking switchmode would be easier and cheaper though.

edit: Sorry for the cross-post with SSg. It's important to note that these devices only have a positive tempco when conducting more than about 5A of current each, which won't be the case here.

I like the toaster solution better than 20 MOSFETS. ( 5 or more nichrome wire wound resistors sharing 50 A in //.

OK so these mosfets are going nowhere.....

SunnySkyguy said:

Why do you want to waste 1kW of heat to drive a 500W load? Is this a battery load tester?

Click to expand...

Its not a battery tester (that is in my hobby project list actually ). It would be a charger actually. Whatever you guys said i know it already.

The problem is i have a transformer of voltage rating at primary of 200 , 220 , 240, 260,270. if you apply 220V at 270 tap then your volatage will be 22V RMS min. and at 200v tap it goes up to 28V rms approx. transformer is of 60 Amps. i cant go below 30VDC which i get after rectifier and filter cap. So this is my first restriction.
second i already have 15000uf x 2 caps of 40V. and 20 pieces of these mosfets. i also have MJ11033 x 4. but problem with that is those MJ's have derating factor of 1.71w/c. so i am going no where. i thought if i use those mosfets in parallel then i would be able to do so. They are of no use to me. but changing Vgs vs temp and there miss matching is surely a problem. in this case mosfet will be damaged one by one and soon they all will be fried. i wonder there are some mosfets in whoms datasheet its mentioned "Ease of paralleling". i wonder they will suit me.

About PWM, in my case that is for charging i know one thing for sure that its not good for batteries plates. i spent approx two years in charging batteries with different method so i m sure of that.

---------- Post added at 19:58 ---------- Previous post was at 19:53 ----------

And yea i can surely loosen up the transformer and pull some turns out but then the ripples will be touching my regulated voltage level and surely there would be some voltage drop in wires. i am using 10 guage of wire.

Also i do have 2n3055 x 24. but they will be having the same issue, missmatching of the forward diode voltage and also there hfe that is beta value.

Sajjadkhan said:

....i already have 15000uf x 2 caps of 40V.

Click to expand...

Remember to check the "Maximum ripple current" rating of the capacitors.

Sajjadkhan said:

....but changing Vgs vs temp and there miss matching is surely a problem.

Click to expand...

You can get very good current sharing if you use a separate opamp to drive each MOSFET. See the circuit below. Five quad opamps will not take much space or be very expensive.

Sajjadkhan said:

About PWM, in my case that is for charging i know one thing for sure that its not good for batteries plates. i spent approx two years in charging batteries with different method so i m sure of that.

Click to expand...

The idea with PWM / switchmode is to put a filter on the output, to provide a smooth continuous current to the load. That may be the best option.