Searching \ for '[EE]: Paralleling MOSFETs' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=paralleling+mosfets
Search entire site for: 'Paralleling MOSFETs'.

Exact match. Not showing close matches.
PICList Thread
'[EE]: Paralleling MOSFETs'
2002\10\10@210930 by Sean Breheny

face picon face
Hi all,

I have always heard people say "You can parallel MOSFETs without source
resistors". Since I never had reason to do this before with power
mosfets, I always just accepted it without going much deeper. I knew that
the reason had to do with the temperature coefficient being such that a
FET at a higher temperature would tend to take less of the current, so
temperature differences would balance out (unlike BJTs where thermal
runaway would happen).

Recently, I did a quick design for a 20 amp current regulator for some
NiMH battery testing. Nothing very fancy, I essentially took two
resistors in series (large, 300W resistors) and put three MOSFETs in
parallel across one of the two resistors. The idea is that the seres
combination is close to the right value to cause 20A drain, and the FETs
shunt current around one of the resistors to allow for some flexibility
in the input voltage (it only has to comply from 13 to 25V).

Here's the problem I ran into, though. It turns out that my FETs
(IRF1104s) have a POSITIVE temperature coefficient on Id at a given Vgs
in the linear region, up to a certain current value (very high, about 20A
per device) where it switches to a negative TC. In practice, they don't
seem to share the current equally.

So, I did a little research. Apparently when people say "you can parallel
FETs" they are usually talking about the "ohmic" region, otherwise known
as "Fully ON". From what I have seen, they have a positive TC on their
RdsON, so that they will tend to share current equally when fully on and
in parallel.

Just because it was easy to do, I had source resistors in my original
design but I don't think they were large enough in value to force the
FETs to share current. I will have to do some testing to verify this
and increase their value.

My question is this: have I analyzed it correctly? Can FETs only safely
be paralled without source resistors in the "fully ON" region of operation?

Thanks,

Sean

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics


2002\10\10@222903 by Dwayne Reid

flavicon
face
At 09:08 PM 10/10/02 -0400, Sean Breheny wrote:
>Hi all,
>
>I have always heard people say "You can parallel MOSFETs without source
>resistors".
>
>Recently, I did a quick design for a 20 amp current regulator for some
>NiMH battery testing. Nothing very fancy, I essentially took two
>resistors in series (large, 300W resistors) and put three MOSFETs in
>parallel across one of the two resistors. The idea is that the seres
>combination is close to the right value to cause 20A drain, and the FETs
>shunt current around one of the resistors to allow for some flexibility
>in the input voltage (it only has to comply from 13 to 25V).
>
>Here's the problem I ran into, though. It turns out that my FETs
>(IRF1104s) have a POSITIVE temperature coefficient on Id at a given Vgs
>in the linear region, up to a certain current value (very high, about 20A
>per device) where it switches to a negative TC. In practice, they don't
>seem to share the current equally.

I think the problem you will have with this arrangement is that the Vgs
threshold varies significantly between FETs.  Your only options are to
match the FETs for Vgs or use some active electronics to adjust the FETs
individually.  Unless your source resistors are large enough that voltage
across them is substantially larger than the variation in Vgs, they aren't
going to help very much.

dwayne

PS - matching the FETs for Vgs is easy.  You can even add a pot to the gate
of each FET to trim the threshold voltage.  They may track fairly closely
once that has been done.  Then again - they might not.  Only way to find
out is to try it . . .

dwayne

--
Dwayne Reid   <spam_OUTdwaynerTakeThisOuTspamplanet.eon.net>
Trinity Electronics Systems Ltd    Edmonton, AB, CANADA
(780) 489-3199 voice          (780) 487-6397 fax

Celebrating 18 years of Engineering Innovation (1984 - 2002)
 .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-.   .-
    `-'   `-'   `-'   `-'   `-'   `-'   `-'   `-'   `-'
Do NOT send unsolicited commercial email to this email address.
This message neither grants consent to receive unsolicited
commercial email nor is intended to solicit commercial email.

--
http://www.piclist.com hint: The list server can filter out subtopics
(like ads or off topics) for you. See http://www.piclist.com/#topics


2002\10\11@115643 by Roman Black

flavicon
face
Sean Breheny wrote:
>
> Hi all,
>
> I have always heard people say "You can parallel MOSFETs without source
> resistors".
> Recently, I did a quick design for a 20 amp current regulator for some
> NiMH battery testing. Nothing very fancy, I essentially took two
> resistors in series (large, 300W resistors) and put three MOSFETs in
> parallel across one of the two resistors. The idea is that the seres
> combination is close to the right value to cause 20A drain,


Hi Sean, I have built a lot of high current dummy
loads etc very similar to what you are describing.

I really don't think FETs are the answer, typically
when you need linear region and high device dissipation
you are better off with some old fashioned power
transistors.

If I read you right you need up to 25v and 20A, this
is max 500w. My preferred method is to use big TO-3
darlingtons, with the large resistors as emitter resistors.
The resistors do the bulk of the dissipation, and also
do the current balancing. Setting base voltage gives
current control within a few percent. The big darlingtons
will give much better reliability in the linear region
with big dissipations, and easier current sharing. :o)
-Roman

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\10\17@214815 by Sean H. Breheny

face picon face
Hi Roman,

Sorry to take so long to reply. I tried out your idea (using two TO-3
darlingtons and 0.15 ohm emitter resistors) and it works really well. The
nice high slope of the Ic vs. Vbe curve (for current sharing) combined with
the larger device package really helps. Much cooler (the FETs were so hot
that you couldn't even momentarily touch them, you can keep your hand on
the darlingtons).

Thanks again!

Sean

At 01:51 AM 10/12/2002 +1000, you wrote:

{Quote hidden}

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email .....listservKILLspamspam@spam@mitvma.mit.edu with SET PICList DIGEST in the body


2002\10\18@044045 by Michael Rigby-Jones

flavicon
face
> -----Original Message-----
> From: Roman Black [SMTP:fastvidspamKILLspamEZY.NET.AU]
> Sent: Friday, October 11, 2002 4:52 PM
> To:   .....PICLISTKILLspamspam.....MITVMA.MIT.EDU
> Subject:      Re: [EE]: Paralleling MOSFETs
>
>
> Hi Sean, I have built a lot of high current dummy
> loads etc very similar to what you are describing.
>
> I really don't think FETs are the answer, typically
> when you need linear region and high device dissipation
> you are better off with some old fashioned power
> transistors.
>
Why is this out of interest?  I was always under the impression that MOSFETs
had far superior Safe Operating Areas (SOA) than bi-polar's?

Regards

Mike

--
http://www.piclist.com hint: To leave the PICList
EraseMEpiclist-unsubscribe-requestspam_OUTspamTakeThisOuTmitvma.mit.edu


2002\10\18@105702 by Peter L. Peres

picon face
On Fri, 18 Oct 2002, Michael Rigby-Jones wrote:

*>> -----Original Message-----
*>> From: Roman Black [SMTP:fastvidspamspam_OUTEZY.NET.AU]
*>> Sent: Friday, October 11, 2002 4:52 PM
*>> To:   @spam@PICLISTKILLspamspamMITVMA.MIT.EDU
*>> Subject:      Re: [EE]: Paralleling MOSFETs
*>>
*>>
*>> Hi Sean, I have built a lot of high current dummy
*>> loads etc very similar to what you are describing.
*>>
*>> I really don't think FETs are the answer, typically
*>> when you need linear region and high device dissipation
*>> you are better off with some old fashioned power
*>> transistors.
*>>
*>Why is this out of interest?  I was always under the impression that MOSFETs
*>had far superior Safe Operating Areas (SOA) than bi-polar's?

Paralelling FETs with too high Rdson specs will be troublesome.
Paralleling FETs with low Rdson will simply lower it further. A recent
discussion mentioned garden light control (Dale Botkin ?) using antiseries
FETs as switches. His got hot and he said he will parallel more of them to
reduce the heat.

Today you can buy 50V 45A FETs with Rdson under 30 miliohms for under $1
in some quantity. I use a lot of Philips BUK45xx ones and I am happy. I
put 10+A through them without a heatsink (switch duty, not pwm). The key
to reach the specified Rdson is to use the specified gate voltage. Eg. the
BUKs want 12V (not 10) for this afair. I have used them up to 30A peak per
device in SMPSUs (with heatsink this time). At that price it pays to look
at them as high efficiency controlled rectifiers too I think.

Peter

--
http://www.piclist.com hint: To leave the PICList
KILLspampiclist-unsubscribe-requestKILLspamspammitvma.mit.edu


2002\10\18@124302 by Roman Black

flavicon
face
Sean H. Breheny wrote:
>
> Hi Roman,
>
> Sorry to take so long to reply. I tried out your idea (using two TO-3
> darlingtons and 0.15 ohm emitter resistors) and it works really well. The
> nice high slope of the Ic vs. Vbe curve (for current sharing) combined with
> the larger device package really helps. Much cooler (the FETs were so hot
> that you couldn't even momentarily touch them, you can keep your hand on
> the darlingtons).


Thanks Sean! Many of those big TO-3 darlingtons
are good for 250W dissipation each! Let's see that
from a TO-220 super-FET. <grin>

Are your resistors too low? If you have a typical
input voltage for your dummy load you could increase
the resistor size so that 2/3 or more of the power
is dissipated on the resistors when the PSU is at
full voltage and full current. :o)
-Roman

--
http://www.piclist.com hint: To leave the PICList
RemoveMEpiclist-unsubscribe-requestTakeThisOuTspammitvma.mit.edu


2002\10\18@130406 by Roman Black

flavicon
face
Michael Rigby-Jones wrote:
>
> > -----Original Message-----
> > From: Roman Black [SMTP:spamBeGonefastvidspamBeGonespamEZY.NET.AU]

> > Hi Sean, I have built a lot of high current dummy
> > loads etc very similar to what you are describing.
> >
> > I really don't think FETs are the answer, typically
> > when you need linear region and high device dissipation
> > you are better off with some old fashioned power
> > transistors.
> >
> Why is this out of interest?  I was always under the impression that MOSFETs
> had far superior Safe Operating Areas (SOA) than bi-polar's?


As a general rule FETs are good for fast switching
on/off and high currents in the ON state.

Transistors are much better with higher voltages and
for operation in the LINEAR region, and come in
cheap common TO-3 devices for very high dissipations.
Even the new IGBT devices seem to have a FET driver
to give fast switching and a transistor power stage
to give good reliable high voltage switching. ;o)

The nature of FETs for fast SMPS switching etc mean
that most FET types are NOT rated for high package
dissipation as they are meant to be operated on/off
in a nice efficient fashion.

Sean asked about a dummy load that needed to dissipate
around 500W, which really doesn't seem like a FET
application to me. :o)
-Roman

--
http://www.piclist.com hint: To leave the PICList
TakeThisOuTpiclist-unsubscribe-requestEraseMEspamspam_OUTmitvma.mit.edu


2002\10\18@154552 by Sean H. Breheny

face picon face
Hi Roman,

Regarding the resistors: I didn't do exactly what you suggested. You'd have
to see my design, essentially it is two R's in series with one of them
bypassed by the transistors. The transistors just sink some current around
that resistor to maintain the total current at 20A. The main Rs are
selected to cause 20A drain at the minimum voltage (13V). This results in
the maximum transistor dissipation being around 90W (total) at 20V, even
though the whole setup may need to dissipate as much as 500W, as you said.
The emitter R's are just for current balancing.

Sean

At 02:37 AM 10/19/2002 +1000, you wrote:

{Quote hidden}

-------------------------------------------
Introducing NetZero Long Distance
Unlimited Long Distance only $29.95/ month!
Sign Up Today! http://www.netzerolongdistance.com

--
http://www.piclist.com hint: To leave the PICList
piclist-unsubscribe-requestEraseMEspam.....mitvma.mit.edu


2002\10\19@053418 by Roman Black

flavicon
face
Sean H. Breheny wrote:
>
> Hi Roman,
>
> Regarding the resistors: I didn't do exactly what you suggested. You'd have
> to see my design, essentially it is two R's in series with one of them
> bypassed by the transistors. The transistors just sink some current around
> that resistor to maintain the total current at 20A.

Sorry Sean, I misunderstood! :o)
-Roman

--
http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads


2002\10\19@110417 by DFansler

picon face
I have not been following this discussion carefully, but would offer
http://www.robot-power.com as a place to see the schematic of a proven, working
design that has parallel mosfets.  Look under the Sections and click on
Downloads.  At the Download page, click on Schematics and board layout for
the OSMC Power Board.  Their board is an H-Bridge good for 160A using
parallel mosfets.

David V. Fansler
EraseMEDFanslerspamMindSpring.com
http://www.DV-Fansler.com

{Original Message removed}

2002\10\20@091828 by Sean H. Breheny

face picon face
Hi David,

Thanks for the reply. I am well aware of the advantages of paralleling
mosfets in switching applications (like H bridges). Unfortunately, in this
case, I needed some active device (either FET or BJT) to operate in the
linear mode, not fully switched on. I naively thought that paralleling
mosfets would work in this case, too, but it turns out to be a much more
touchy situation because of several factors: #1 manufacturing variations
may mean that your fets don't share current equally even at the same
temperature and Vgs, #2 the tempco of current vs Vgs in the linear region
is actually positive up to a certain point, #3 the transconductance of
mosfets is much less than BJTs, #4 most mosfets are designed for hard
switching applications and as such it is difficult to find mosfets in large
device packages (larger than TO-220).

So, what happens is that if you put two similar mostfets in parallel (in
the fully-on region of operation), their currents balance fairly well
because the tempco of Rds_on vs temperature is positive, so hotter ones
tend to take less current. In addition, since they are operating in the
fully on region, they usually don't have to dissipate much heat (on
average) anyway, so a TO-220 case is adequate.

On the other hand, in the linear region, two similar mosfets may draw
currents that differ by a factor of 2 even at the same Vgs and temperature.
Then, as the higher-current one gets hotter, it begins to take MORE of the
current (just like a BJT does) up to a point, where the tempco goes
negative. The result can be that one ends up a lot hotter than the other,
and drawing more current. Not total thermal runaway since the sign of the
tempco does eventually change (as current increases, IIRC).

If you try to fix this problem with the classic solution (usually used with
BJTs) of source resistors, you'll find that you need rather large source
resistors compared to BJTs because it takes a lot more Vgs change per amp
of current imbalance due to the lower transconductance.

Sean

At 11:01 AM 10/19/2002 -0400, you wrote:
{Quote hidden}

>  {Original Message removed}

2002\10\21@030710 by Michael Rigby-Jones

flavicon
face
{Quote hidden}

You can hardly compare a TO3 with a TO220 package, but there are quite
easily available MOSFET's in TO3 packages that are rated for 300 Watt
dissipation(e.h. IXTM67N10).  However, you don't need the hassle of a TO3
style mounting when you can get a plastic TO247 packaged device rated for up
to 400 Watts, and of course the mini-bloc devices that dissipate around 700
watts.

Not knocking bi-polars, I prefer using them to MOSFET's mainly because I
grew up with them and understand their strengths and weaknesses better, but
technology moves on and there are some fantastic devices around now.

Regards

Mike

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\10\21@053806 by Roman Black

flavicon
face
Michael Rigby-Jones wrote:

> You can hardly compare a TO3 with a TO220 package, but there are quite
> easily available MOSFET's in TO3 packages that are rated for 300 Watt
> dissipation(e.h. IXTM67N10).  However, you don't need the hassle of a TO3
> style mounting when you can get a plastic TO247 packaged device rated for up
> to 400 Watts, and of course the mini-bloc devices that dissipate around 700
> watts.

> technology moves on and there are some fantastic devices around now.


Agreed! There are some fantastic devices around. But
at what cost?? Having been into electronics for over
25 years and having bought electronics businesses that
had been going even longer with extensive parts stocks,
I have massive amounts of older parts that are
essentially free for me to use.

For instance on the parts storage thread, I store my
TO-3 parts in 1/2 litre clear plastic tubs, and measure
my stock of TO-3 transistors in LITRES.

As for the "hassle of mounting TO-3" you have a point,
but I would like to point out that as the big TO-3s
are becoming less popular, many of the hobby shops
*cheapest* heatsinks are the pre-drilled ones suiting
TO-3s, which are being sold off. :o)
-Roman

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


2002\10\21@055918 by Nate Duehr

face
flavicon
face
> For instance on the parts storage thread, I store my
> TO-3 parts in 1/2 litre clear plastic tubs, and measure
> my stock of TO-3 transistors in LITRES.

(Heh heh... the newbies like myself are now green with envy, Roman... I
can still count all my parts in numbers less than 100, and most in tens
and twenties... GRIN.  Even simple caps and resistors... have to start
somewhere!)

Nate

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.


More... (looser matching)
- Last day of these posts
- In 2002 , 2003 only
- Today
- New search...