# 42v to 12v

Discussion in 'General Electronics Chat' started by SimonTHK, Aug 28, 2016.

1. ### SimonTHKMember

Joined:
Feb 10, 2010
Messages:
345
Likes:
1
Location:
Denmark
I calculated the absolute minimum of the battery pack to be 27,6v. So I want to put two 10k resistors as a voltage divider to add a minimum of 13,8v to gate. If the battery pack reach max potential, it reaches 45v which will give 22,5v to gate. Can anyone tell me if the mosfet can handle this voltage on gate?

In other words.
Gate to source voltage says maximum +-20v so I guess I cant add 22,5v to gate?

Last edited: Sep 6, 2016
2. ### SimonTHKMember

Joined:
Feb 10, 2010
Messages:
345
Likes:
1
Location:
Denmark
Hello again

Sorry for all my confusing posts.

I have drawn a schematic again.

I will be using IPP045N10N: http://www.infineon.com/dgdl/Infine...n.pdf?fileId=db3a30431ce5fb52011d1e8b0cc31586

Max current drawn on my controller= 25 amp.

Voltage from battery pack ranges from 27.5v to 45v. This is also the voltage that has to give a signal to the mosfet, that it should close as a switch (open of for current running through it). It has to be fully closed, so there is a minimum resistance.
Someone said in here, that I should range gate voltage between 12v to 18v, so I tried.

-Can I do it like in this schematic? Using a zener diode?
-Or should I instead be using two resistors? 10k and 5.7k should give me an output voltage on the voltage divider, with the given voltage input (changing on battery pack), between 10v-19v. http://www.raltron.com/cust/tools/voltage_divider.asp

File size:
119.8 KB
Views:
66
3. ### alec_tWell-Known MemberMost Helpful Member

Joined:
Jul 10, 2011
Messages:
9,325
Likes:
1,231
Location:
Cardiff, Wales
Either way will work.
Don't forget that, unless you use an insulation pack to mount the FET on the heatsink, the heatsink is at the same potential as the FET drain.

Joined:
Jan 12, 1997
Messages:
-
Likes:
0

5. ### kinarfiWell-Known Member

Joined:
Jul 7, 2009
Messages:
1,514
Likes:
30
Location:
P. G., Utah, United States of America

I gave zener diodes a thought also, but decided it is a bad idea because when you turn the power off, the fet needs to have the capacitance of the gate to source discharged, a zener would not discharge, use a resistor. Do you know how to check a FET? Use the diode check of your volt meter, on the N Fet, hook the black(-) lead to the source, red(+) to the drain, should be open (OL) on my meter, momentarily, just touch the red lead to the gate, then back to the drain, meter should read zero volts, now, with your finger, touch the source and the gate at the same time and the meter reading will go to 0pen. the do it again, but reverse the leads to source and drain and it will still be zero, now touch gate and source and the voltage will read around .7 volt. You need to be careful not to discharge the gate accidentally while doing this test. If you clip the leads to the fet, just leave it charged and watch how long it takes to discharge by itself.

I would suggest the gate voltage be kept to 20 max, the FET will still be turned fully with a gate of 8 volts.

in response to Alec_T, the FET does need to insulated or the heat sink need to be isolated and I must disagree about the zener, it will not turn the FET off.
Jeff

Last edited: Sep 6, 2016
6. ### SimonTHKMember

Joined:
Feb 10, 2010
Messages:
345
Likes:
1
Location:
Denmark
I dont quite understand what you mean? I will be using a small cooling profile on its back, for to220 with cooling paste. What exactly do you mean b "the same potential as the fet drain?

7. ### kinarfiWell-Known Member

Joined:
Jul 7, 2009
Messages:
1,514
Likes:
30
Location:
P. G., Utah, United States of America
Here's the answer to " the same potential as the fet drain?" see the transistor in the drawing
This schematic would work with a zener

8. ### SimonTHKMember

Joined:
Feb 10, 2010
Messages:
345
Likes:
1
Location:
Denmark
Ah ofcourse, pull down through the latch relay, good idea
Do you believe it turn off the FET quick enough? Or should I rather go for the resistor solution?
Ok I did know that it had the same potential, I just had to figure out what he meant. Thank you. There will be no connections to anything, everything will be dropped into hot resin and dissapear with wires sticking up. Hopefully the circuit should be ebough oversized, so it doesnt need air around the passive cooling profile. Or maybe I should let it stick up above the resin.

9. ### SimonTHKMember

Joined:
Feb 10, 2010
Messages:
345
Likes:
1
Location:
Denmark
Ah ofcourse, pull down through the latch relay, good idea
Do you believe it turn off the FET quick enough? Or should I rather go for the resistor solution?
Ok I did know that it had the same potential, I just had to figure out what he meant. Thank you. There will be no connections to anything, everything will be dropped into hot resin and dissapear with wires sticking up. Hopefully the circuit should be ebough oversized, so it doesnt need air around the passive cooling profile. Or maybe I should let it stick up above the resin.

10. ### kinarfiWell-Known Member

Joined:
Jul 7, 2009
Messages:
1,514
Likes:
30
Location:
P. G., Utah, United States of America
Yes, 10 K will be fine, that is what I use most of the time, and since at max current, the power dissipation will be 105 mW, you should be fine burying every thing in resin. I suggest you build it, try it, then bury it in resin after your sure every thing works, run it and see if the FET gets warm!

11. ### specWell-Known MemberMost Helpful Member

Joined:
Oct 27, 2015
Messages:
5,828
Likes:
710
Location:
Somerset UK GMT
Hy kinaffi,

I would advise using two 10K resistors and an 18V Zener diode. That way you will protect the MOSFET gate from over voltage and also provide a path to 0V when the relay goes open circuit. That will ensure that the MOSFET gate is not floating and the MOSFET is not in an undefined state.

spec

12. ### kinarfiWell-Known Member

Joined:
Jul 7, 2009
Messages:
1,514
Likes:
30
Location:
P. G., Utah, United States of America
SimonTHK, is there a switch in here some where that turns every thing off for when your done with your ride ?
spec do you mean like this?

#### Attached Files:

• ###### Untitled.png
File size:
237.7 KB
Views:
44
Last edited: Sep 6, 2016
13. ### kinarfiWell-Known Member

Joined:
Jul 7, 2009
Messages:
1,514
Likes:
30
Location:
P. G., Utah, United States of America
spec
Do you always design for the highest gate voltage? I usually design for 10 +-, with this FET, it looks like 6 would actually work.

14. ### specWell-Known MemberMost Helpful Member

Joined:
Oct 27, 2015
Messages:
5,828
Likes:
710
Location:
Somerset UK GMT
Spot on: you have the best of all worlds there: voltage protection for the MOSFET gate and a good gate drive.

You have also protected the MOSFET gate from any spikes both positive and negative that may get into the system.

spec

Last edited: Sep 26, 2016
15. ### specWell-Known MemberMost Helpful Member

Joined:
Oct 27, 2015
Messages:
5,828
Likes:
710
Location:
Somerset UK GMT
No, you rarely design for maximum gate voltage, but I do aim protect the gate from voltages of Max VGS - 10%.

The gate drive voltage to design for depends on the characteristics of the MOSFET. All you need to do is ensure that you have enough gate drive voltage to conduct the required current (ID) and have a low enough on resistance (RDS) under all conditions of gate threshold voltage variations from one particular MOSFET to another and variations of RDS with temperature.

With the IPB042 I would suggest a minimum gate drive voltage of 12V for this application. You have provided a minimum gate drive voltage of 13.5V (battery stack at 27V) so the circuit design is optimum I would say.

Also, if you are switching the MOSFET on and of at speed you need to take into account the effective capacitance between gate and source. For small MOSFETs the effective gate capacitance may be as low as 500pF but on some modern, high current, low threshold MOSFETs the effective gate capacitance can be huge, like 20nF. So if you had a 10K drive impedance, as is the case of your drive circuit at turn off, the gate time constant would be 20 nF * 10K = 200u sec.

One last bit of cracker barrel advice: MOSFETs are very prone to oscillate, typically between 2MHz and 6Mhz, especially if the layout is not optimum, so it is always wise to fit a gate stopper of 22R for slow circuits, 10R for medium speed circuits, and 1R for fast circuits. The gate stopper should be mounted directly on the gate terminal. A gate stopper also helps shape the waveform at the gate and also gives the driving circuit an easier time. Note that a MOSFET may not oscillate all the time but it will certainly oscillate under these conditions:

(1) The first time management come to see the equipment working
(2) The first time the customer comes to view the equipment working
(3) When you are on trials miles away from base and halfway up a cliff face
(4) When the unit goes into production
(5) When the first unit is delivered to the customer
(6) At any other time as long as there is an R in the month.

spec

Last edited: Sep 6, 2016
16. ### alec_tWell-Known MemberMost Helpful Member

Joined:
Jul 10, 2011
Messages:
9,325
Likes:
1,231
Location:
Cardiff, Wales
Point taken. I was only considering voltage reduction.

17. ### ChrisP58Well-Known Member

Joined:
Apr 27, 2012
Messages:
1,081
Likes:
131
Location:
Provo, Utah, USA
The power dissipation number of most mosfets is only valid as long as you keep the case temperature at or below 25°C, while it's pumping out that rated power. And the only way that I know of to do that is with refrigeration. Not very practical in most applications.

The RDSon of the IRF3205 is only at 8 milliOhms when the junction temperature is 25°C. That resistance goes up at higher temperature. See figure 4 of the datasheet.
-
The IPP045N10N is a better choice due to it's lower resistance and higher Drain-Source voltage.
But remember, it's resistance will also increase as the junction temperature rises. Things that will contribute to the rise in junction temperature include:
- Ambient temperature
- Power dissipation
- How well you can get rid of the dissipated power.

But I still have a concern with pushing 25 amps through the leads of ANY single TO-220 component.

For the gate drive, I agree on a resistor AND a reverse biased 12-18 volt zener across the Gate-Source pins.

18. ### specWell-Known MemberMost Helpful Member

Joined:
Oct 27, 2015
Messages:
5,828
Likes:
710
Location:
Somerset UK GMT
That is so true. In fact, the power dissipation figure for a transistor is meaningless in practical terms and is really a marketing ploy. Apart from the obvious silicon and mechanical (bond wires, case leads etc) limits, the two controlling factors are, safe operating area (SOA) and junction temperature. It is often assumed that the ambient temperature is 25 Deg C because that is the standard on data sheets, but the effective ambient inside an equipment can be anything. A good general ambient figure to work with is 50 Deg C.

There is another factor that is rarely considered and that is maximum rate of change of voltage (dv/dt) and current (di/dt).

spec

• Like x 1
19. ### SimonTHKMember

Joined:
Feb 10, 2010
Messages:
345
Likes:
1
Location:
Denmark
lol

I will add a gate stopper thank you.

I come to think. The circuit that Kinarfi has made, with the pull down to 0v through latch relay and an eventual switch that also connect to 0v. Wouldnt the "newly added" 10k resistor in parallel with the zener diode pull down any float current before the gate? So that I dont actually need the 0v through latch relay and switch.
I ask this only because the wiring is simpler.

Thank you

20. ### Nigel GoodwinSuper ModeratorMost Helpful Member

Joined:
Nov 17, 2003
Messages:
39,331
Likes:
653
Location:
Derbyshire, UK
21. ### kinarfiWell-Known Member

Joined:
Jul 7, 2009
Messages:
1,514
Likes:
30
Location:
P. G., Utah, United States of America

That would work well also, just forgot to dump it after Specs suggestion, still not sure I (PO)(personal opinion) agree with 18 v zener, I'd stay with 12V, your choice, of course, and only use this in May, June, July, and August! NO Rs

Nigel, that for AC