Sizing Mosfet Gate Resistors

[EdV]

From a previous post by Terry Given in "Resistor Current Handling
Capability" circa Thurs, Feb 26 2004 :

"This is a common reliability problem with power electronics -
"designers"
often neglect peak pulse power in MOSFET gate resistors - I saw one
design
with 4.7Ohm Rg, +12V Vg i.e. 31W peak, using 0805 resistors that failed

after a short time, causing catastrophic failure; so much damage was
done
each time that the root cause was totally obscured, and the "designer"
could
not stop it happening. oops. "

I am the lowly test engineer with the 10 fets in parallel test question
from several months ago. I the process of getting my head around how
hard I can pulse my 10 ohm gate resistors when one side itied to
ground. I read Terry's above note.

Anyone care to elaborate or point me to a good app note regarding
sizing Mosfet gate resistors?

The board I am testing (yes "they" decided the mosfets need to be
individually turned on. . .I knew "they" would) looks suspicously like
"oops".

Thanks much,
Ed V.

 

[John Larkin]

Capability" circa Thurs, Feb 26 2004 :

"This is a common reliability problem with power electronics -
"designers"
often neglect peak pulse power in MOSFET gate resistors - I saw one
design
with 4.7Ohm Rg, +12V Vg i.e. 31W peak, using 0805 resistors that failed

after a short time, causing catastrophic failure; so much damage was
done
each time that the root cause was totally obscured, and the "designer"
could
not stop it happening. oops. "

31 watts peak won't hurt an 0805 if the duration and rep-rate are low.

I am the lowly test engineer with the 10 fets in parallel test question
from several months ago. I the process of getting my head around how
hard I can pulse my 10 ohm gate resistors when one side itied to
ground. I read Terry's above note.

One side to ground? Are the resistors connected gate-to-source? That's
sort of strange. "Mosfet gate resistors" are usually in series with
the gate.

John

 

[Mr. J D]

Capability" circa Thurs, Feb 26 2004 :

"This is a common reliability problem with power electronics -
"designers"
often neglect peak pulse power in MOSFET gate resistors - I saw one
design
with 4.7Ohm Rg, +12V Vg i.e. 31W peak, using 0805 resistors that failed

after a short time, causing catastrophic failure; so much damage was
done
each time that the root cause was totally obscured, and the "designer"
could
not stop it happening. oops. "

I am the lowly test engineer with the 10 fets in parallel test question
from several months ago. I the process of getting my head around how
hard I can pulse my 10 ohm gate resistors when one side itied to
ground. I read Terry's above note.

Anyone care to elaborate or point me to a good app note regarding
sizing Mosfet gate resistors?

The board I am testing (yes "they" decided the mosfets need to be
individually turned on. . .I knew "they" would) looks suspicously like
"oops".

Thanks much,
Ed V.

10 ohms works well

 

[EdV]

The gate resistors are in series with the gates but they are all driven
from a single source.

My test stratgey is:

1 - ground the common signal source
2 - apply an appropriate pulse at the test point beween one Fet and its
gate resistor
3 - acquire the Vds data (gppib data from oscioscope) before during and
after the "appropriate gate pulse"
4 - crunch Vds to see that the correct "dip" is there for a Fet sinking
some current.

repeat steps 1-4 for fets through 20

The fets turn on in about 35nsec so I am guessing I can get something
constructive done with a 1 usec pulse. Determining the voltage is
problematic. I am waiting for the designer to tell me which resistor is
being designed in which will have some effect on how hard I can pulse.
I should probably consider a worst case scenario since my purchasing
dept. will gladly "redesign " this circuit for me.

Thanks for being interested in my evolving dilemnas : - >

Ed V.

 

[Joerg]

Hello Ed,

Capability" circa Thurs, Feb 26 2004 :

"This is a common reliability problem with power electronics -
"designers"
often neglect peak pulse power in MOSFET gate resistors - I saw one
design
with 4.7Ohm Rg, +12V Vg i.e. 31W peak, using 0805 resistors that failed

after a short time, causing catastrophic failure; so much damage was
done
each time that the root cause was totally obscured, and the "designer"
could
not stop it happening. oops. "

I am the lowly test engineer with the 10 fets in parallel test question
from several months ago. I the process of getting my head around how
hard I can pulse my 10 ohm gate resistors when one side itied to
ground. I read Terry's above note.

Anyone care to elaborate or point me to a good app note regarding
sizing Mosfet gate resistors?

App notes won't help too much I am afraid. The proper strategy in cases
like this would be:

a. Find out what the highest (worst case) pulse current would be, via
the slope of the drive signal and the gate capacitance.

b. Get pulse load data from the resistor manufacturer, in writing. If
they can't or won't furnish it change to another manufacturer that can.

Should be done by the design engineer but it never hurts to
double-check. We all make mistakes.

BTW, test engineers are not "lowly". I value their work, expertise and
production-related experience a lot.

The board I am testing (yes "they" decided the mosfets need to be
individually turned on. . .I knew "they" would) looks suspicously like
"oops".

Did you see failed resistors?

 

[John Larkin]

The gate resistors are in series with the gates but they are all driven
from a single source.

My test stratgey is:

1 - ground the common signal source
2 - apply an appropriate pulse at the test point beween one Fet and its
gate resistor
3 - acquire the Vds data (gppib data from oscioscope) before during and
after the "appropriate gate pulse"
4 - crunch Vds to see that the correct "dip" is there for a Fet sinking
some current.

repeat steps 1-4 for fets through 20

The fets turn on in about 35nsec so I am guessing I can get something
constructive done with a 1 usec pulse. Determining the voltage is
problematic. I am waiting for the designer to tell me which resistor is
being designed in which will have some effect on how hard I can pulse.
I should probably consider a worst case scenario since my purchasing
dept. will gladly "redesign " this circuit for me.

Thanks for being interested in my evolving dilemnas : - >

Ed V.

Please don't top post.


The average power dissipated in a series gate resistor is about

P = C * F * V^2

where

C is total effective gate capacitance, Miller included

F = drive frequency

V = gate drive voltage.

So calculate that and see if it's within the resistor's power
dissipation spec.

John

 

[Joerg]

Hello John,

The average power dissipated in a series gate resistor is about

P = C * F * V^2

where

C is total effective gate capacitance, Miller included

F = drive frequency

V = gate drive voltage.

So calculate that and see if it's within the resistor's power
dissipation spec.

Don't discount pulse load limits. Metal film types exhibit certain
pathologies related to that. That is a serious concern with most of my
ultrasound pulser designs. I used to prefer Beyschlag because of their
excellent specsmanship. They were bought by Vishay but much of the info
is now online:

http://www.vishay.com/docs/49280/tn0006.pdf

They are also very helpful on the phone but if it's still like it used
to be, ideally you'd have to be able to speak German. Preferably with a
northern accent ;-)

 

[John Larkin]

Hello John,


Don't discount pulse load limits. Metal film types exhibit certain
pathologies related to that. That is a serious concern with most of my
ultrasound pulser designs. I used to prefer Beyschlag because of their
excellent specsmanship. They were bought by Vishay but much of the info
is now online:

http://www.vishay.com/docs/49280/tn0006.pdf

They are also very helpful on the phone but if it's still like it used
to be, ideally you'd have to be able to speak German. Preferably with a
northern accent ;-)

A fet gate will charge up in nanoseconds if you drive it hard, and the
thermal impulse into the resistor will be tiny. If you drive it slow,
the resistor peak power will be low, and that's OK too.

The energy dumped into the resistor on each edge is about equal to the
gate charge energy [1], which will be 10s of mJ for a big fet (several
millicoulombs of gate charge driven to, say, 10 volts.) I've never
seen even an 0805 gate resistor fail from the transient charge
current. At very high rep-rates, the average power could cook the
resistor.

John

[1] assuming the gate driver is infinitely fast, which it usually
isn't. The slower the driver, the less power is dissipated in the
series gate resistor.

 

[Joerg]

Hello John,

A fet gate will charge up in nanoseconds if you drive it hard, and the
thermal impulse into the resistor will be tiny. If you drive it slow,
the resistor peak power will be low, and that's OK too.

That's exactly what we did, drive a gate from a push-pull stage. It's
been a while but IIRC they told me that it'll start eating away right
under the metallization area.

The energy dumped into the resistor on each edge is about equal to the
gate charge energy [1], which will be 10s of mJ for a big fet (several
millicoulombs of gate charge driven to, say, 10 volts.) I've never
seen even an 0805 gate resistor fail from the transient charge
current. At very high rep-rates, the average power could cook the
resistor.

I did (0805). We replaced them with larger ones. Most of the failures
were in the T/R switches though where pulse peaks were higher but still
well under the thermal limit. None of the failed resistors ever got hot
and there were no scorch marks on the boards. Yet fail they did.

BTW the most strange failure I had on a pulser (but this did exceed the
thermal) was when part of a resistor flew off with a loud bang and the
remnants had turned into milky green glass.

 

[Terry Given]

Hello John,

A fet gate will charge up in nanoseconds if you drive it hard, and the
thermal impulse into the resistor will be tiny. If you drive it slow,
the resistor peak power will be low, and that's OK too.

That's exactly what we did, drive a gate from a push-pull stage. It's
been a while but IIRC they told me that it'll start eating away right
under the metallization area.

The energy dumped into the resistor on each edge is about equal to the
gate charge energy [1], which will be 10s of mJ for a big fet (several
millicoulombs of gate charge driven to, say, 10 volts.) I've never
seen even an 0805 gate resistor fail from the transient charge
current. At very high rep-rates, the average power could cook the
resistor.

I did (0805). We replaced them with larger ones. Most of the failures
were in the T/R switches though where pulse peaks were higher but still
well under the thermal limit. None of the failed resistors ever got hot
and there were no scorch marks on the boards. Yet fail they did.

BTW the most strange failure I had on a pulser (but this did exceed the
thermal) was when part of a resistor flew off with a loud bang and the
remnants had turned into milky green glass.

I've seen this with 0603 and 0805 gate resistors. Its worse with IGBTs
than FETs, for 2 reasons:

- IGBTs tend to be bigger (well at least the big ones are), and have
more gate capacitance. 100nF is nothing!

- IGBTs often have bipolar gate drive (to reduce evil effects of Cm),
+/-15V is pretty typical, which quadruples the resistor peak power.


and of course in most smps the rep-rate is not small (although really
big IGBT drivers rarely switch above 10kHz)

The high rep-rate drives up the average power dissipation (100nF 30Vpp
at 1kHz = 90mW into 10R), and therefore both the average and peak
temperatures.


And I should clarify "seen": I havent actually measured a dodgy smt gate
resistor, but rather on several occassions surmised this as the cause of
clusters of field failures - the numbers certainly supported the
hypothesis, and suitable component choices made the problems disappear
completely. because the units had already failed, Rg was always toast.

I have however destroyed a number of resistors in this manner. It all
started when I turned on a drive in the test lab and noticed a flash,
many years ago. I banged the juice off PDQ, and along with another
engineer we tested it carefully, but found no "fault" - it still ran.
But there was definitely a flash, so we dismembered the drive and lo and
behold there it was - a dead PR02 10R resistor, used as damping in an
EMI filter. It wasnt our "design", it was in common usage there. some
calcs showed that at 415Vac a 10R sees 34.5kW peak pulse, which was
clearly a bit much for the PR02. So we went walkabout thru the service
area, and lo and behold, there was many an EMI filter whose caps were
disconnected.

similar issues apply with DC bus soft-charging resistors; a cheap 10W
w/w wont have the mandrel completely surrounded by ceramic material, so
there is an air gap and hence a hot spot; hit it with a sizeable thump
and you get a pretty red flash from that spot, it work hardens and
eventually breaks.

Likewise for Dynamic Braking resistors. Now that is fun; on a cold windy
night a 250kW DB resistor is great!


As an interesting aside, in addition to the substrate solder joint, IGBT
bond wires are vulnerable to thermal cycling.


Cheers
Terry

 

[Joerg]

Hello Terry,

Likewise for Dynamic Braking resistors. Now that is fun; on a cold windy
night a 250kW DB resistor is great!

During a vacation job back in the college days I build many of these.
About the size of a Westinghouse fridge when complete. I guess they are
still doing their job in locomotives (in Brazil).

 

[Eeyore]

Capability" circa Thurs, Feb 26 2004 :

"This is a common reliability problem with power electronics -
"designers"
often neglect peak pulse power in MOSFET gate resistors - I saw one
design
with 4.7Ohm Rg, +12V Vg i.e. 31W peak, using 0805 resistors that failed

after a short time, causing catastrophic failure; so much damage was
done
each time that the root cause was totally obscured, and the "designer"
could
not stop it happening. oops. "

I am the lowly test engineer with the 10 fets in parallel test question
from several months ago. I the process of getting my head around how
hard I can pulse my 10 ohm gate resistors when one side itied to
ground. I read Terry's above note.

Anyone care to elaborate or point me to a good app note regarding
sizing Mosfet gate resistors?

The board I am testing (yes "they" decided the mosfets need to be
individually turned on. . .I knew "they" would) looks suspicously like
"oops".

Q^2 . R . rep rate surely ? Where Q = Vdrive / Ciss

Graham

 

[Eeyore]

10 ohms works well

What's your design experience Mr J.D. ?

Graham

 

[Eeyore]

The gate resistors are in series with the gates but they are all driven
from a single source.

Sounds sensible.

My test stratgey is:

1 - ground the common signal source
2 - apply an appropriate pulse at the test point beween one Fet and its
gate resistor
3 - acquire the Vds data (gppib data from oscioscope) before during and
after the "appropriate gate pulse"
4 - crunch Vds to see that the correct "dip" is there for a Fet sinking
some current.

repeat steps 1-4 for fets through 20

The fets turn on in about 35nsec so I am guessing I can get something
constructive done with a 1 usec pulse. Determining the voltage is
problematic.

Which voltage ?

I am waiting for the designer to tell me which resistor is
being designed in which will have some effect on how hard I can pulse.
I should probably consider a worst case scenario since my purchasing
dept. will gladly "redesign " this circuit for me.

Thanks for being interested in my evolving dilemnas : - >

This all sounds a bit arse about face to me.

Graham

 

[Eeyore]

Eeyore wrote:

Correction !!!! But still not right. This assumes the gate drive current is a
constant.

( Q . rep rate . 2 ) ^2 . R Where Q = Vdrive * Ciss

The 2 comes from the charge and discharge of the gate capacitance..

But the energy in the series R is created mainly in short pulses therefore
increasing the power dissipated in the series R.

If you can solve it for a single pulse then it just needs multiplying by the rep
rate. It's doable but I'm not in the mood for integrals involving e right now.

Graham

 

[Eeyore]

The average power dissipated in a series gate resistor is about

P = C * F * V^2

where

C is total effective gate capacitance, Miller included

F = drive frequency

V = gate drive voltage.

So calculate that and see if it's within the resistor's power
dissipation spec.

Where did R go ?

Graham

 

[Terry Given]

Eeyore wrote:

Correction !!!! But still not right. This assumes the gate drive current is a
constant.




The 2 comes from the charge and discharge of the gate capacitance..

But the energy in the series R is created mainly in short pulses therefore
increasing the power dissipated in the series R.

If you can solve it for a single pulse then it just needs multiplying by the rep
rate. It's doable but I'm not in the mood for integrals involving e right now.

Graham

dont bother, its wrong. The basic maths is:

E = 0.5*Q*V = 0.5*C*V^2

P = E*f

do a dimensional analysis (IOW look at the units) of your equation, they
dont match.

When you put 0.5*C*V^2 J into a cap thru a resistor, the resistor also
dissipates 0.5*C*V^2 J. When you then discharge the cap into the same
resistor, again it dissipates 0.5CV^2 J, leading to a total resistor
energy loss of CV^2 J to charge C to V Volts then discharge C. Repeat at
some frequency F, and voila: P = CV^2F

The resistor doesnt appear in the equation because it is not controlling
the voltage to which the cap gets charged.

The above assumes the cap is both fully charged and discharged. If it is
not, *then* R comes into play. In the case of a FET driver, this is an
extremely good assumption; if the gate isnt charging and discharging
fully, there is most likely a problem.

The same maths usually applies to snubbers, too, but not to RCD voltage
clamps, as the clamp capacitor generally does not discharge between
successive clamping operations.


Cheers
Terry

 

[Terry Given]

John Larkin wrote:




Where did R go ?

Graham

see my post below.

Cheers
Terry

 

[Jim Thompson]

dont bother, its wrong. The basic maths is:

E = 0.5*Q*V = 0.5*C*V^2

P = E*f

do a dimensional analysis (IOW look at the units) of your equation, they
dont match.

When you put 0.5*C*V^2 J into a cap thru a resistor, the resistor also
dissipates 0.5*C*V^2 J. When you then discharge the cap into the same
resistor, again it dissipates 0.5CV^2 J, leading to a total resistor
energy loss of CV^2 J to charge C to V Volts then discharge C. Repeat at
some frequency F, and voila: P = CV^2F

The resistor doesnt appear in the equation because it is not controlling
the voltage to which the cap gets charged.

The above assumes the cap is both fully charged and discharged. If it is
not, *then* R comes into play. In the case of a FET driver, this is an
extremely good assumption; if the gate isnt charging and discharging
fully, there is most likely a problem.

The same maths usually applies to snubbers, too, but not to RCD voltage
clamps, as the clamp capacitor generally does not discharge between
successive clamping operations.


Cheers
Terry

QED: Eeyore has been proven incompetent ;-)

...Jim Thompson

 

[Eeyore]

dont bother, its wrong. The basic maths is:

E = 0.5*Q*V = 0.5*C*V^2

P = E*f

do a dimensional analysis (IOW look at the units) of your equation, they
dont match.

Eh ? Q . t^-1 = I

When you put 0.5*C*V^2 J into a cap thru a resistor, the resistor also
dissipates 0.5*C*V^2 J.

OK. Neat trick. Hadn't figured that one.

When you then discharge the cap into the same
resistor, again it dissipates 0.5CV^2 J, leading to a total resistor
energy loss of CV^2 J to charge C to V Volts then discharge C. Repeat at
some frequency F, and voila: P = CV^2F
Yup.


The resistor doesnt appear in the equation because it is not controlling
the voltage to which the cap gets charged.

Just the rate. Of course it it was very slow the gate would never charge to V.

The above assumes the cap is both fully charged and discharged.
Yes.

If it is
not, *then* R comes into play. In the case of a FET driver, this is an
extremely good assumption; if the gate isnt charging and discharging
fully, there is most likely a problem.
Sure.


The same maths usually applies to snubbers, too, but not to RCD voltage
clamps, as the clamp capacitor generally does not discharge between
successive clamping operations.

Interesting.

Graham