Diode-less Power MOSFETS

[Optikon]

Does anyone know if you can get N & P channel _power_ MOSFETS that do not have the substrate connected to source? If so where? Looking for something like 200Vds or more, Id = 2A or more kind of parts..

I've found small signal types but power versions seem to elude me..

Thanks!

 

[zevon8]

As I understand it, all the "Hex-Fets" from International Rectifier have the intrinsic body diode, it is by nature of their design, possibly all Mosfets are like this. Small signal devices do not require the die area, and it is likely left "floating" as it were.

ST Microelectronics has some interesting FET's , some with pretty unique characteristics, maybe they have something suitable? www.st.com

 

[Optikon]

As I understand it, all the "Hex-Fets" from International Rectifier have the intrinsic body diode, it is by nature of their design, possibly all Mosfets are like this. Small signal devices do not require the die area, and it is likely left "floating" as it were.

ST Microelectronics has some interesting FET's , some with pretty unique characteristics, maybe they have something suitable? www.st.com

This seems to be the case for all power fet manufacturers. No power versions at st. Thanks for the suggestion though!

 

[Styx]

is there any particular reason you have to use a FET?
if not use an IGBT (those rating you state are starting to enter IGBT territory)

they dont have any anti-parallel diode, unless you get one like that - rare in distrete form

 

[Agent 009]

What's a Hex-Fet???

 

[upand_at_them]

Does anyone know if you can get N & P channel _power_ MOSFETS that do not have the substrate connected to source?

Isn't that a JFET?

Mike

 

[zevon8]

Styx makes a good point, not sure if it applies to your situation, but it would be a possible solution.

Agent009, a HexFET is a brand name for Internation Rectifier, it comes from the internal structure of the MOSFET's they produce.

upand_at_them: yes, absolutely, a JFET has no body diode, good point. The slightly more complicated biasing may not allow for a simple design swap, but a JFET does have its advantages, not sure of any of the major makers lines for these devices, probably some out there though.

 

[Roff]

I did a search for "power JFET", not expecting to find anything. I found **broken link removed**.

No, a JFET is not a MOSFET sans body diode. The technology is very different.

 

[zevon8]

Same here, that is all I found. Seems to be a fading ( or maybe replaced ) technology. Looks to be found only included in op-amps and similar linear devices. I guess the popularity of MOSFETS, IGBT, and their siblings simplicity has overtaken the market. I seem to remember seeing JFET's in instrumentation many years ago, but nothing recently.

 

[williB]

Look at how they Show two MosFets connected together , to block current in both directions..

If i have to use two then so be it..!!

 

[Optikon]

is there any particular reason you have to use a FET?
if not use an IGBT (those rating you state are starting to enter IGBT territory)

they dont have any anti-parallel diode, unless you get one like that - rare in distrete form

IGBT will have too high a saturation voltage. My application needs bi-directional controlled current flow.

If I could get one off the shelf, I'd use it but since it seems to be exotic (at best), I'll just use 2 mosfets in series with sources tied together.

 

[Optikon]

Look at how they Show two MosFets connected together , to block current in both directions..

If i have to use two then so be it..!!

Yeah, this is a common arrangement. I' always trying to save cost by not using 2 fets.. oh well..

 

[Optikon]

What's a Hex-Fet???

Just a proprietary name used by IR that they give their particular process for creating power FETS.. every vendor seems to have their own secret ingredients if you will, and they all give it a name.
Hex-FET
Cool-FET
Tiny-FET and so on and so on..

 

[Optikon]

I did a search for "power JFET", not expecting to find anything. I found **broken link removed**.

No, a JFET is not a MOSFET sans body diode. The technology is very different.

yeah this is rare..

Interfet makes a 100V+ Jfet.. which is rare also..

 

[williB]

Interfet makes a 100V+ Jfet.. which is rare also..

Optikon , what is the current carrying capibility of that Jfet ..?

 

[Oznog]

The body diode is not always such a bad thing. It depends on your app.

For example. if you want to build a cool and efficient FWB rectifier, you would orient the MOSFETs so the body diodes act just like a normal diode FWB when the gate drivers are turned off. So turning on the gate driver just brings the vds down from 0.7v to millivolts but otherwise it works either way.

 

[Optikon]

Interfet makes a 100V+ Jfet.. which is rare also..

Optikon , what is the current carrying capibility of that Jfet ..?

Its not much.. 10'mA maybe like the usual small signal types.. so it is not really a _power_ device but the blocking voltage is remarkable.

 

[williB]

hmm
i want to use Mosfets to transfer power from my generator as efficiently as possible..
since you are familiar with using two Mosfets to transfer power ( and block it ) .. what do ya think of my Full wave rectifier , using mosfets..
i have allready simulated a mosfet pair and it does in fact work..
the problem is finding a proper ground in this comfiguration..
the inductors labeled L3&L4 are really coils from my generator..
about the ground , the comparators are going to need a reference to know when one leg is more positive than the other..
When the L4 leg is positive with respect to L3 . then A & B would be on and C & D would be off..
the connection between A & C Labeled Vout is positive , the other Vout is negetive..

 

[Styx]

is there any particular reason you have to use a FET?
if not use an IGBT (those rating you state are starting to enter IGBT territory)

they dont have any anti-parallel diode, unless you get one like that - rare in distrete form

IGBT will have too high a saturation voltage. My application needs bi-directional controlled current flow.

If I could get one off the shelf, I'd use it but since it seems to be exotic (at best), I'll just use 2 mosfets in series with sources tied together.

Well the thing is BJT's have the lowest Vsat and they also dont have a intrisnic diode. Only limitation with them are that they are current-driven and are slow

are BJT's a possibility?

 

[Oznog]

hmm
i want to use Mosfets to transfer power from my generator as efficiently as possible..
since you are familiar with using two Mosfets to transfer power ( and block it ) .. what do ya think of my Full wave rectifier , using mosfets..
i have allready simulated a mosfet pair and it does in fact work..
the problem is finding a proper ground in this comfiguration..
the inductors labeled L3&L4 are really coils from my generator..
about the ground , the comparators are going to need a reference to know when one leg is more positive than the other..
When the L4 leg is positive with respect to L3 . then A & B would be on and C & D would be off..
the connection between A & C Labeled Vout is positive , the other Vout is negetive..

OK, that's what I thought it was for.
You do NOT have any need for more than 2 NMOS and 2 PMOS. In this case you would put the NMOS source to the negative rail and the PMOS drain to the positive rail. The body diodes will only be in the forward bias when you're turning the transistor on anyways.

You should be able to power the driver for the MOSFET gates with the DC power rail. It shouldn't have bootstrapping trouble because the MOSFET diodes act as a FWB to drive the rail when there is no charge on the rail. Consider the idle current issue here however. If those comparators drain a few mA at all times then it will slowly drain the caps down.