Maker Pro
Maker Pro

MOSFETs IN SERIES

arg733

Dec 14, 2010
89
Joined
Dec 14, 2010
Messages
89
http://scholar.lib.vt.edu/theses/available/etd-12092010-221223/unrestricted/Wang_W_T_2010.pdf

Hi guys.
A while ago i run across this document which apparently tells how to wire mosfets in series to get higher voltages (pg 10-14). This is the only schematic i found so far that is actually working. The problem is that i can't make more than 3 mosfets work, when i add a fourth ( with the same circuitry) it remains OFF no matter what. I tried different resistor values. At first i though that if R2 has 2ohms and R3 20ohms that the next should be 200ohms. When it didn't work i tried 20,50 and 1k but with no luck. Can you please tell me what i did wrong?
Thank you.
 

Attachments

  • 1.jpg
    1.jpg
    105.1 KB · Views: 523
Last edited:

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,505
Joined
Jan 21, 2010
Messages
25,505
Interesting. I've seen 2 mosfets connected to do this, but not three or more.

I can't help with your problem though, sorry.
 

arg733

Dec 14, 2010
89
Joined
Dec 14, 2010
Messages
89
It's ok steve.
May i ask if the link to the pdf is visible?
Thank you.
 

pwdixon

Oct 14, 2012
52
Joined
Oct 14, 2012
Messages
52
You need to post your circuit and indicate voltages used, I would suggest looking at the voltage V3 mentioned in the paper relative to V1. I wouldn't really expect there to be a resistor value problem with this circuit.
 

arg733

Dec 14, 2010
89
Joined
Dec 14, 2010
Messages
89
Here is the circuit sorry i didn't attached it in the first post. The MOSFETs have a resistive load of 1A at 350VDC.
 

Attachments

  • 1.jpg
    1.jpg
    105.1 KB · Views: 1,329

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
I don't think your circuit will work. You've omitted the diode from the 12VDC supply to the MOSFET gates. I would use an individual diode from the +12V supply to each gate, and I would replace the gate-source diodes with zeners to protect the gate. Then I think it would work. But why are you doing this? MOSFETs that can switch 350V are readily available.
 

arg733

Dec 14, 2010
89
Joined
Dec 14, 2010
Messages
89
I don't think your circuit will work.

But it IS working with 3 MOSFETs the question is why it works with 3 but not with more??? strange...

and I would replace the gate-source diodes with zeners to protect the gate

Hey i'm no expert but isn't it a "protection" matter? i may do it so not to blow up anything, but the problem is that i can't make Q4 even switch on, thats why i stripped everything out to try it as the paper suggests.

You've omitted the diode from the 12VDC supply to the MOSFET gates

Tried this ... no luck.

I would use an individual diode from the +12V supply to each gate

Nope... that didn't do the trick either.

But why are you doing this? MOSFETs that can switch 350V are readily available.

as you may have figured this is a "low voltage experiment", after this works i will rev engineer it and use it with 4 or more 1000v MOSFETs :) , yeah i know the risks involved and i take the necessary precautions.

Thanks.
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
(zener diodes from gate to source)
Hey i'm no expert but isn't it a "protection" matter? i may do it so not to blow up anything, but the problem is that i can't make Q4 even switch on, thats why i stripped everything out to try it as the paper suggests.
Oh sorry. I foolishly thought that it would be helpful and important to prevent you from blowing up your MOSFETs. Of course, you can just add the zeners later, if you can be bothered.
as you may have figured this is a "low voltage experiment", after this works i will rev engineer it and use it with 4 or more 1000v MOSFETs :)
My apologies again, I hadn't guessed that. My crystal ball has been a little cloudy lately.

If you make the changes I have suggested (diode to each gate, and zeners from gates to sources), and the MOSFETs remain OFF when +12V is applied to the gate of the bottom one, either you've wired something wrong, or you have a damaged MOSFET. If you can't find the problem, post the voltages on all three terminals of each MOSFET.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,505
Joined
Jan 21, 2010
Messages
25,505
Looking more closely at this, I would be careful about exceeding Vdg for these mosfets. Whilst not often specified (but can be assumed to be Vds + Vgs), it is something you need to be careful about. This is especially true when all your slave mosfets gave their gates at essentially the same potential.

I can't see where that paper you've linked to explains how this won't be a problem.

I also note that the lack of a couple of diodes in your circuit is going to make it a lot worse.

If you're simulating this, try graphing Vdg for each device.
 

arg733

Dec 14, 2010
89
Joined
Dec 14, 2010
Messages
89
My apologies again, I hadn't guessed that. My crystal ball has been a little cloudy lately.

:) Sorry i didn't tell you that, but lets try to make it work at 350v for now.

If you make the changes I have suggested (diode to each gate, and zeners from gates to sources), and the MOSFETs remain OFF when +12V is applied to the gate of the bottom one, either you've wired something wrong, or you have a damaged MOSFET. If you can't find the problem, post the voltages on all three terminals of each MOSFET.

Ok i will add those components and measure the voltages at each terminal of the mosfets and post the results.

Looking more closely at this, I would be careful about exceeding Vdg for these mosfets. Whilst not often specified (but can be assumed to be Vds + Vgs), it is something you need to be careful about. This is especially true when all your slave mosfets gave their gates at essentially the same potential.

So what is Vdg for these mosfets? 100+20? i don't get it... and my English is not perfect...


So if i just add a diode at each gate and put zeners (20v?) it should be ok?

Hope to try it tomorrow (if my fever lets me :) )

Thanks
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
Looking more closely at this, I would be careful about exceeding Vdg for these mosfets. Whilst not often specified (but can be assumed to be Vds + Vgs), it is something you need to be careful about. This is especially true when all your slave mosfets gave their gates at essentially the same potential.
Right, that would definitely be a problem. That's why I suggested he uses a diode from his +12V supply to each gate separately, and a zener from gate to source on each MOSFET.

I'm still not very confident that the MOSFETs will voltage-share. I would put a resistor arcross each one just to make sure, and use diodes with very low capacitance (and high voltage specification, of course).

I THINK with those changes, the circuit ought to work...
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
:) Sorry i didn't tell you that, but lets try to make it work at 350v for now.
What else haven't you told us? You'll save time in the long run if you give us a thorough description of what you want to do. Don't try to guess what's important and what isn't. We can make that decision.

I saw your comment that "i discovered that when i talk about voltages exceeding 500v people are less willing to help me". You deleted it from your post, but it was in the email I received that told me of your post. And you're quite right. You don't have the experience to work safely. People make the decision to keep out of your project because we don't want you to electrocute yourself.

So if i just add a diode at each gate and put zeners (20v?) it should be ok?
I would use 15V or 12V zeners (1W is fine) and for the diodes, use high-voltage, relatively high speed diodes like 1N4936, 1N4937 or BA157 with separate series resistors of say 10K.

Please TRY not to kill yourself.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,505
Joined
Jan 21, 2010
Messages
25,505
So what is Vdg for these mosfets? 100+20? i don't get it... and my English is not perfect...

Probably something like that.

So if i just add a diode at each gate and put zeners (20v?) it should be ok?

It's hard to say. This circuit seems to rely on the mosfets breaking down slightly to balance the voltage between them.

Some of the component values (such as the zener across the gate bias source) do not seem to be specified in the paper, or at least I couldn't find it.

Hope to try it tomorrow (if my fever lets me :) )

Good luck.
 

arg733

Dec 14, 2010
89
Joined
Dec 14, 2010
Messages
89
Some of the component values (such as the zener across the gate bias source) do not seem to be specified in the paper

It's on page 15. But everything is different...



People make the decision to keep out of your project because we don't want you to electrocute yourself.

Please TRY not to kill yourself.

Don't worry that's probably isn't going to happen. I work with my left hand in the pocket , and i have a transformer and a fuse so even if i get a shock (don't think so) it won't be lethal. Well... if you know of any other precaution i could take it would be good.

use high-voltage, relatively high speed diodes

I have a bunch of 1000v 45ns diodes (can't remember the p/n) will those do?.

with separate series resistors of say 10K.

Will be done.
Thank you.
 

arg733

Dec 14, 2010
89
Joined
Dec 14, 2010
Messages
89
So that's the complete circuit then?
 

Attachments

  • 2.jpg
    2.jpg
    258.3 KB · Views: 549

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
(Electrocuting yourself)
Don't worry that's probably isn't going to happen. I work with my left hand in the pocket, and i have a transformer and a fuse so even if i get a shock (don't think so) it won't be lethal. Well... if you know of any other precaution i could take it would be good.
Do a Google search. A thorough one. And when you're working with voltages of 500V or more, a transformer (I assume you mean an isolating transformer) may not help you. And a fuse is no use at all - they're there to prevent fires, not electrocution.

I have a bunch of 1000v 45ns diodes (can't remember the p/n) will those do?
Yes.

So that's the complete circuit then?

D15 across the 12V source isn't needed.
R1~3, in series with the gate drive diodes, should all be 10K.
R4~6 and D2, D4 and D6 are now in parallel with the gate-source zeners, and aren't needed.
D1, D3, D5 and D7, the four diodes across the drain-source of the MOSFETs, aren't needed - the MOSFETs include a parasitic diode inside them that does this job.
R7~10 (10k resistors across the MOSFET drain-source) should probably be much higher values. What is the maximum leakage current your load can tolerate when the MOSFETs are all OFF?
 

arg733

Dec 14, 2010
89
Joined
Dec 14, 2010
Messages
89
Thanks.
i will do what you suggested but i still have a few questions

R7~10 (10k resistors across the MOSFET drain-source) should probably be much higher values

1M or so?

What is the maximum leakage current your load can tolerate when the MOSFETs are all OFF

Well... it would be best if it was under 10mA.

R4~6 and D2, D4 and D6 are now in parallel with the gate-source zeners, and aren't needed.

So just the zeners?
 

GonzoEngineer

Dec 2, 2011
321
Joined
Dec 2, 2011
Messages
321
Keep in mind, the gate voltage has to be relative to the source voltage.

I did a project like this once where we ended up switching 200kV at over 10A.

We had to use a seperate, ISOLATED, gate drives for each device.

This is the only way it will work flawlessly!

I can only reveal this, because that company is now out of business.:D
 

arg733

Dec 14, 2010
89
Joined
Dec 14, 2010
Messages
89
Keep in mind, the gate voltage has to be relative to the source voltage.
Oookay... i have no idea what that means.

I did a project like this once where we ended up switching 200kV at over 10A

with 1000v mosfets? really?

We had to use a seperate, ISOLATED, gate drives for each device.

you mean one of those ic's? and what do you mean by "isolated"?
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
(maximum leakage current)
Well... it would be best if it was under 10mA.
OK, so from Ohm's Law, R=V/I. R=350/0.01 = 35000 ohms. Since there are four resistors in series, each one needs to be at least 8750 ohms. How about 22k each. Power dissipation will be P=V^2/R =(350/4)^2 / 22000 = 7656 / 22000 = 0.35W so they need to be 1W resistors.
Total resistance will be 88k so leakage current will be I=V/R =350/88000 =4mA.
So each of the four resistors across the MOSFETs need to be 22K 1W.
(Schematic changes)
So just the zeners?
Yes, just the zeners from gate to source.
I did a project like this once where we ended up switching 200kV at over 10A. We had to use a seperate, ISOLATED, gate drives for each device.
That's the "proper" way to do it. This circuit cheats. I'm still not entirely sure it will work properly, but I THINK it will. Especially with the changes I've suggested.

To the OP, what switching frequency do you plan to use?
 
Top