Estimating heat desipation of MOSFET

Thread Starter

anishkgt

Joined Mar 21, 2017
458
hello All,
Is there a equation/formula to estimate the current flowing through a mosfet when Vgs = 8.1v can't seem to figure it out from the datasheet to estimate the heat desipated. My average current requirement would be about 70A for the below CV schematic.

Annotation 2020-05-23 162058.png
Thanks in advance.
 

OBW0549

Joined Mar 2, 2015
3,409
No, there's no handy formula you can use, other than Ohm's Law. Your best course of action is probably to estimate the power dissipation from some of the graphs in the datasheet showing Rds(on) vs. gate drive:

Rds(on) 1.png
Rds(on) 2.png
The information on these graphs, plus your estimated drain current, will allow you to calculate the power dissipation-- by my mental arithmetic, about 20 watts at a junction temperature of 25 °C, or twice that at 150 °C.
 

ronsimpson

Joined Oct 7, 2019
684
estimate the current flowing through a mosfet …………… to estimate the heat desipated
What is the input voltage to your circuit? Vcc=?
Power loss is Volts X Current.
Current = 70A Stated in post #1.
Voltage on the MOSFET = Vcc-8.1V
Power lost in the MOSFET is 70 x (Vcc-8.1)
 

Thread Starter

anishkgt

Joined Mar 21, 2017
458
That appears to be a high-side N-mosfet. Are you sure you know Vgs? Very odd circuit with a voltage regulator on the gate.

Here's just one of several reference in my files about mosfet heating: http://www.fairchildsemi.com/an/AN/AN-558.pdf In short,it is Ohmic.
Can't seem to come up with a better way to get a CV and higher current. My initial thought was this
Annotation 2020-05-23 162058.png

Then that limited the charge current. Charge time 3s2p bank = 210F*8.1/70A = 24seconds
 

OBW0549

Joined Mar 2, 2015
3,409
None of the schematics you've shown make even a little bit of sense. Have you looked at the data sheet for the PMBT2222A? Did you notice its maximum rating for collector current is only 600 mA? How in the world do you expect to get 70 amps through it??

You need to throw out what you've done so far and start over from scratch using some common sense.
 

ronsimpson

Joined Oct 7, 2019
684
Lets start over. It looks like you want to charge up some very big capacitors.
Your supply is 12V 70A. Is it current limited?
Do you know if the capacitors will handle 35A each?
What are you doing with the caps when they are charged?
 

Thread Starter

anishkgt

Joined Mar 21, 2017
458
Lets start over. It looks like you want to charge up some very big capacitors.
Your supply is 12V 70A. Is it current limited?
Do you know if the capacitors will handle 35A each?
What are you doing with the caps when they are charged?
The caps are these. The power supply is an HP server power supply that has all the bells and whistles in it. I know these are current limited because i work in the IT industry and i use them to charge my LiPo batteries for my RC car.

The caps are gonna be used for my spot welder project that i plan on to build.
 

ronsimpson

Joined Oct 7, 2019
684
power supply
The link is to a power supply. I think it will buck the 12 volts down to a maximum of 8.1 and current limit to 20A. Both V and I are adjustable.
I just saw you capacitors. They can handle 20A to 38A all the time. (will get hot)

Because you power supply current limits, I think you could use 2, 3 or 4 caps in parallel to get high current. Then you don't need to current limit with a power supply like in the link.

Is there a way to turn off the supply? Some have a external "enable" When the voltage gets to 8.1V we could kill the supply!
 

Thread Starter

anishkgt

Joined Mar 21, 2017
458
The initial charge from 0-8.1v would need a high current but once the voltage on the caps have reached 8.1v then on each weld it consume 1 or 2v max, topping-up would not require much current. That i what i presume.

I was planning to get a constant 8.1v and have a bypass mosfet for higher current but doing it on the high side would be bad idea with that much current and N-MOSFET would be even worse because of the high Rds(on).

Damn i wish knew a little more about electronics.
 

Thread Starter

anishkgt

Joined Mar 21, 2017
458
moreover i was planning on powering the other electronics with individual LM317 buck converters for 5v 600ma, 10v 2A, 12v 600ma and the 8.1v would be used to charge the caps and shut them down when done. Initially i thought the AP3211 would take care of the CV 8.1v part and eliminating the inductor and adding a mosfet at the SW pin would do the job control the mosfet based on the feedback from the FB pin.
 

jpanhalt

Joined Jan 18, 2008
9,435
Here is a more conventional layout for what you seem to want:

1590400705501.png

The gate resistors (RG) are optional. N-mosfet needs to be compatible with your enable voltage. That is, if the enable voltage is 5 V, then you need a logic level device. R1 is often something like 10 k. Once the mosfet is off, there is virtually no current flow.

You will need a common point ("ground") for the power supply, load, and enable voltages. Finally, if you want several amps for charging (e.g., 70A) a linear regulator will produce a lot of heat. I would control the power supply to the proper voltage.
 

Thread Starter

anishkgt

Joined Mar 21, 2017
458
I came across one of these from RICHTEK. Is this worth a try, can source 10A too. At that current a 3s2p cap would charge up from 0 to 8.1 in about 2.8minutes, consider 3minutes.
2.png
 
Last edited:

jpanhalt

Joined Jan 18, 2008
9,435
The specifications says 10A max with output up to 15 V (set by R1/R2) for the RT7259. I suspect either might work. As they are switching regulators, their heat dissipation would be a lot lower than the linear regulator you showed earlier. Both are no-lead packages, so hand soldering might be tricky, but reflow will be easier.

EDIT: I should add that those large capacitors may/will act like a short on the output of either supply. That could cause an over current condition or shut down. It was mentioned earlier that the HP supply you have also has current limiting.

1590420563792.png
Some devices (many of the lithium chargers) are made to operate with thermal shutdown being normal at the start of charging. It appears that chip is not so easy to restart nor is shutdown expected.
 
Last edited:
Top