Exceeding Vgs rating

On Friday, April 6, 2018 at 9:21:53 PM UTC-4, mike wrote:
> On 4/6/2018 1:33 PM, Jim Thompson wrote:
> > On Fri, 06 Apr 2018 12:18:32 -0700, mike <ham789@netzero.net> wrote:
> >
> >> On 4/6/2018 7:21 AM, John Larkin wrote:
> >>> On Fri, 6 Apr 2018 16:31:38 +0530, Pimpom <Pimpom@invalid.invalid>
> >>> wrote:
> >>>
> >>>> I'm designing a small simple circuit in which a MOSFET drives a
> >>>> low-power load. The very low frequency gate drive may, on rare
> >>>> occasions, exceed the max Vgs rating of 12V by about 1V, possibly 2V.
> >>>>
> >>>> There are a number of ways to limit the gate voltage but I want
> >>>> to avoid them and keep the circuit as simple as possible, and it
> >>>> won't cause a disaster if the transistor fails. What do you think?
> >>>
> >>> The few mosfets that I've tested failed at Vgs around 70 volts. There
> >>> could be long-term effects.
> >>>
> >>> Gate-protected ones zenered in the low 40's.
> >>>
> >>> Test a couple to destruction. It's fun to blow things up.
> >>>
> >>> I'd estimate that a couple of volts over Vgs_max will have no effect
> >>> on reliability.
> >>>
> >>>
> >> You guys talk like ignoring specifications is acceptable practice.
> >> IT IS NOT!
> >> Testing a few samples under controlled conditions is NOT reason to
> >> exceed specifications.
> >>
> >> Have you never underestimated the consequences of a decision?
> >> Have you never had a vendor make a design change that still meets
> >> the original specification?
> >> Have you never had a part go obsolete and get replaced by an
> >> "equivalent" part?
> >> Have you never had a purchasing manager switch parts on you
> >> without even telling you so he could save a buck and get a bigger bonus?
> >>
> >> I've never attended a seminar on
> >> "it won't cause a disaster if the transistor fails".
> >>
> >
> > You're just exhibiting your inexperience.
> 
> You're correct.  I'm inexperienced at designing systems that exceed
> the specifications of components.  I taught my engineers to derate
> components.
> 
> This statement gave me chills:
> The very low frequency gate drive may, on rare
>  >>>> occasions, exceed the max Vgs rating of 12V by about 1V, possibly 2V.
> 
> So kinda maybe we might, on rare occasions see 1V...maybe 2V beyond the 
> spec.  But don't worry, it ain't no big thing if it fails.
> Why not 3V on rarer occasions?   How about 10V? How rare?
> The OP is asking for encouragement, permission, to build in a failure
> mode and supporting it with kinda/maybe/possibly...
> 
> If an engineer came to me with that, I'd send him back to the drawing
> board.  "We'll discuss it when you come back with words like absolutely,
> definitely, here are the options and costs..."
> 
> If you read the fine print,
> > VGSmax is
> 
> usually
> I split out that word so you can't miss it.
> Curiosity isn't the only thing that can kill a cat.
> 
> 
> specified as "guaranteed by design"... that is, it's
> > NOT measured during manufacture, so there's
> usually?
> a HUGE margin.
> > 		
> >                                          ...Jim Thompson
> >
> I once was forced to build a logic system to operate in
> an environment with limited power availability.
> We literally designed in what would fit the power budget.
> Problem was that it was TTL and there was a huge difference
> between typical and maximum power specs.
> If you used the max numbers, we were over the power budget
> by 20% or so.  Prototypes ran at about 20% under the max.
> 
> The QA dept refused to sign off on the design.  I did a Monte Carlo
> power analysis that included batch effects per part showing
> that the probability of excess was near zero.
> 
> QA signed off on the condition that power was measured at
> final test and that we supply test fixtures to the service
> centers and included the power measurement after every repair.
> I don't think we ever had a non-conforming unit.
> 
> I didn't "kinda/maybe".  I did the rigorous analysis.  Validation
> procedures were in place.  The process worked.

Sure, I think all of that makes sense.  
But there are all sorts here living in different design 
(worlds/ niches/ markets)  
I'm not sure if this is exactly true, but my impression of Pimpom 
is he's mostly a home hobbyist, making 1- 10 of something.  
(That could be wrong.) 

I live in a small volume/ high markup world.  
The last project had a temperature range of 77 to 400K. 
I don't think any of the parts I used as heaters, sensors 
were rated for that low temperature.  So I torture 
tested and used 'em.  Should I have insisted on only using 
'low temp' parts?  They cost ~x100 more. (if they exist)

Knowing which specs are a bit (or very) squishy is part of design.
I've been noodling with a fet product idea, and was worried 
about Vgs (and Vgd) maybe I can find some fet's that I don't 
have to worry about. :^)

George H.   

On Saturday, April 7, 2018 at 1:10:59 AM UTC-4, mike wrote:
> On 4/6/2018 8:34 PM, tabbypurr@gmail.com wrote:
> > On Saturday, 7 April 2018 02:21:53 UTC+1, mike  wrote:
> >> On 4/6/2018 1:33 PM, Jim Thompson wrote:
> >>> On Fri, 06 Apr 2018 12:18:32 -0700, mike <ham789@netzero.net> wrote:
> >>>> On 4/6/2018 7:21 AM, John Larkin wrote:
> >>>>> On Fri, 6 Apr 2018 16:31:38 +0530, Pimpom <Pimpom@invalid.invalid>
> >>>>> wrote:
> >>>>>
> >>>>>> I'm designing a small simple circuit in which a MOSFET drives a
> >>>>>> low-power load. The very low frequency gate drive may, on rare
> >>>>>> occasions, exceed the max Vgs rating of 12V by about 1V, possibly 2V.
> >>>>>>
> >>>>>> There are a number of ways to limit the gate voltage but I want
> >>>>>> to avoid them and keep the circuit as simple as possible, and it
> >>>>>> won't cause a disaster if the transistor fails. What do you think?
> >>>>>
> >>>>> The few mosfets that I've tested failed at Vgs around 70 volts. There
> >>>>> could be long-term effects.
> >>>>>
> >>>>> Gate-protected ones zenered in the low 40's.
> >>>>>
> >>>>> Test a couple to destruction. It's fun to blow things up.
> >>>>>
> >>>>> I'd estimate that a couple of volts over Vgs_max will have no effect
> >>>>> on reliability.
> >>>>>
> >>>>>
> >>>> You guys talk like ignoring specifications is acceptable practice.
> >>>> IT IS NOT!
> >>>> Testing a few samples under controlled conditions is NOT reason to
> >>>> exceed specifications.
> >>>>
> >>>> Have you never underestimated the consequences of a decision?
> >>>> Have you never had a vendor make a design change that still meets
> >>>> the original specification?
> >>>> Have you never had a part go obsolete and get replaced by an
> >>>> "equivalent" part?
> >>>> Have you never had a purchasing manager switch parts on you
> >>>> without even telling you so he could save a buck and get a bigger bonus?
> >>>>
> >>>> I've never attended a seminar on
> >>>> "it won't cause a disaster if the transistor fails".
> >>>>
> >>>
> >>> You're just exhibiting your inexperience.
> >>
> >> You're correct.  I'm inexperienced at designing systems that exceed
> >> the specifications of components.  I taught my engineers to derate
> >> components.
> >>
> >> This statement gave me chills:
> >> The very low frequency gate drive may, on rare
> >>   >>>> occasions, exceed the max Vgs rating of 12V by about 1V, possibly 2V.
> >>
> >> So kinda maybe we might, on rare occasions see 1V...maybe 2V beyond the
> >> spec.  But don't worry, it ain't no big thing if it fails.
> >> Why not 3V on rarer occasions?   How about 10V? How rare?
> >> The OP is asking for encouragement, permission, to build in a failure
> >> mode and supporting it with kinda/maybe/possibly...
> >>
> >> If an engineer came to me with that, I'd send him back to the drawing
> >> board.  "We'll discuss it when you come back with words like absolutely,
> >> definitely, here are the options and costs..."
> >>
> >> If you read the fine print,
> >>> VGSmax is
> >>
> >> usually
> >> I split out that word so you can't miss it.
> >> Curiosity isn't the only thing that can kill a cat.
> >>
> >>
> >> specified as "guaranteed by design"... that is, it's
> >>> NOT measured during manufacture, so there's
> >> usually?
> >> a HUGE margin.
> >>> 		
> >>>                                           ...Jim Thompson
> >>>
> >> I once was forced to build a logic system to operate in
> >> an environment with limited power availability.
> >> We literally designed in what would fit the power budget.
> >> Problem was that it was TTL and there was a huge difference
> >> between typical and maximum power specs.
> >> If you used the max numbers, we were over the power budget
> >> by 20% or so.  Prototypes ran at about 20% under the max.
> >>
> >> The QA dept refused to sign off on the design.  I did a Monte Carlo
> >> power analysis that included batch effects per part showing
> >> that the probability of excess was near zero.
> >>
> >> QA signed off on the condition that power was measured at
> >> final test and that we supply test fixtures to the service
> >> centers and included the power measurement after every repair.
> >> I don't think we ever had a non-conforming unit.
> >>
> >> I didn't "kinda/maybe".  I did the rigorous analysis.  Validation
> >> procedures were in place.  The process worked.
> >
> > That's all a reflection of the type of products you and many folk here design. Cheap consumer electronics is a different market.
> >
> >
> > NT
> >
> I disagree with that mindset.  Start with the mindset that everything
> should work as advertised and be reliable.  Stated another way...put
> on your customer hat.  What do you expect from your product?
> When you're done, go back
> and try to cut cost.  Weigh the savings against the loss of reliability.
> 
> If you start with slipshod design, it's hard to fix little things like
> overvoltaging a FET.
> 
> Thought experiment:
> Do two designs.
> Start one with a quality mindset.
> Start the other with a minimum cost mindset.
I'm always doing both at once... biggest bang
for smallest buck.  That is the value I deliver to 
my customers.  (well, part of it.) And then you've 
got to factor your own time into the equation.

George H.   
> I'd bet that if you massage each design toward the middle,
> you'll end up at price parity with the quality design resulting in a 
> higher quality product.
> And your next design will be better for it.
> 
> If  you came here for validation, you already know  you shouldn't be 
> doing it.

On 4/7/2018 9:38 PM, George Herold wrote:

> But there are all sorts here living in different design
> (worlds/ niches/ markets)

Very true.

> I'm not sure if this is exactly true, but my impression of Pimpom
> is he's mostly a home hobbyist, 

You're both right and wrong. I'm both a pro and a hobbyist and I 
sometimes blur the line between the two. I just rarely talk about 
the pro side of my work here.

"mike" spoke about his engineers. Well, I don't have an army of 
engineers working under me. But I *have* trained several 
engineering graduates and most of them are in key posts in both 
private and government agencies. Elaborating on that might be 
seen by some as bragging, so I'll refrain from doing it.

 > making 1- 10 of something.
 > (That could be wrong.)

Mostly correct. I've seldom produced more than a hundred or so 
units of my designs. The reason is a combination of two factors. 
One is severe restrictions imposed by my location. The other is a 
conscious decision made long ago not to get entangled in the 
management of a sizeable company.

On Friday, April 6, 2018 at 6:01:50 AM UTC-4, Pimpom wrote:
> I'm designing a small simple circuit in which a MOSFET drives a 
> low-power load. The very low frequency gate drive may, on rare 
> occasions, exceed the max Vgs rating of 12V by about 1V, possibly 2V.
> 
> There are a number of ways to limit the gate voltage but I want 
> to avoid them and keep the circuit as simple as possible, and it 
> won't cause a disaster if the transistor fails. What do you think?

Does this source have an impedance? And you do have quiescent current loading limitations? What is the range, both minimum and maximum, of your input signal? What is this low power load and how mow much voltage variation across it can it take?  There's not enough context to offer a fix.

On 04/07/18 12:19, George Herold wrote:
> On Saturday, April 7, 2018 at 1:10:59 AM UTC-4, mike wrote:
>> On 4/6/2018 8:34 PM, tabbypurr@gmail.com wrote:
>>> On Saturday, 7 April 2018 02:21:53 UTC+1, mike  wrote:
>>>> On 4/6/2018 1:33 PM, Jim Thompson wrote:
>>>>> On Fri, 06 Apr 2018 12:18:32 -0700, mike <ham789@netzero.net> wrote:
>>>>>> On 4/6/2018 7:21 AM, John Larkin wrote:
>>>>>>> On Fri, 6 Apr 2018 16:31:38 +0530, Pimpom <Pimpom@invalid.invalid>
>>>>>>> wrote:
>>>>>>>
>>>>>>>> I'm designing a small simple circuit in which a MOSFET drives a
>>>>>>>> low-power load. The very low frequency gate drive may, on rare
>>>>>>>> occasions, exceed the max Vgs rating of 12V by about 1V, possibly 2V.
>>>>>>>>
>>>>>>>> There are a number of ways to limit the gate voltage but I want
>>>>>>>> to avoid them and keep the circuit as simple as possible, and it
>>>>>>>> won't cause a disaster if the transistor fails. What do you think?
>>>>>>>
>>>>>>> The few mosfets that I've tested failed at Vgs around 70 volts. There
>>>>>>> could be long-term effects.
>>>>>>>
>>>>>>> Gate-protected ones zenered in the low 40's.
>>>>>>>
>>>>>>> Test a couple to destruction. It's fun to blow things up.
>>>>>>>
>>>>>>> I'd estimate that a couple of volts over Vgs_max will have no effect
>>>>>>> on reliability.
>>>>>>>
>>>>>>>
>>>>>> You guys talk like ignoring specifications is acceptable practice.
>>>>>> IT IS NOT!
>>>>>> Testing a few samples under controlled conditions is NOT reason to
>>>>>> exceed specifications.
>>>>>>
>>>>>> Have you never underestimated the consequences of a decision?
>>>>>> Have you never had a vendor make a design change that still meets
>>>>>> the original specification?
>>>>>> Have you never had a part go obsolete and get replaced by an
>>>>>> "equivalent" part?
>>>>>> Have you never had a purchasing manager switch parts on you
>>>>>> without even telling you so he could save a buck and get a bigger bonus?
>>>>>>
>>>>>> I've never attended a seminar on
>>>>>> "it won't cause a disaster if the transistor fails".
>>>>>>
>>>>>
>>>>> You're just exhibiting your inexperience.
>>>>
>>>> You're correct.  I'm inexperienced at designing systems that exceed
>>>> the specifications of components.  I taught my engineers to derate
>>>> components.
>>>>
>>>> This statement gave me chills:
>>>> The very low frequency gate drive may, on rare
>>>>    >>>> occasions, exceed the max Vgs rating of 12V by about 1V, possibly 2V.
>>>>
>>>> So kinda maybe we might, on rare occasions see 1V...maybe 2V beyond the
>>>> spec.  But don't worry, it ain't no big thing if it fails.
>>>> Why not 3V on rarer occasions?   How about 10V? How rare?
>>>> The OP is asking for encouragement, permission, to build in a failure
>>>> mode and supporting it with kinda/maybe/possibly...
>>>>
>>>> If an engineer came to me with that, I'd send him back to the drawing
>>>> board.  "We'll discuss it when you come back with words like absolutely,
>>>> definitely, here are the options and costs..."
>>>>
>>>> If you read the fine print,
>>>>> VGSmax is
>>>>
>>>> usually
>>>> I split out that word so you can't miss it.
>>>> Curiosity isn't the only thing that can kill a cat.
>>>>
>>>>
>>>> specified as "guaranteed by design"... that is, it's
>>>>> NOT measured during manufacture, so there's
>>>> usually?
>>>> a HUGE margin.
>>>>> 		
>>>>>                                            ...Jim Thompson
>>>>>
>>>> I once was forced to build a logic system to operate in
>>>> an environment with limited power availability.
>>>> We literally designed in what would fit the power budget.
>>>> Problem was that it was TTL and there was a huge difference
>>>> between typical and maximum power specs.
>>>> If you used the max numbers, we were over the power budget
>>>> by 20% or so.  Prototypes ran at about 20% under the max.
>>>>
>>>> The QA dept refused to sign off on the design.  I did a Monte Carlo
>>>> power analysis that included batch effects per part showing
>>>> that the probability of excess was near zero.
>>>>
>>>> QA signed off on the condition that power was measured at
>>>> final test and that we supply test fixtures to the service
>>>> centers and included the power measurement after every repair.
>>>> I don't think we ever had a non-conforming unit.
>>>>
>>>> I didn't "kinda/maybe".  I did the rigorous analysis.  Validation
>>>> procedures were in place.  The process worked.
>>>
>>> That's all a reflection of the type of products you and many folk here design. Cheap consumer electronics is a different market.
>>>
>>>
>>> NT
>>>
>> I disagree with that mindset.  Start with the mindset that everything
>> should work as advertised and be reliable.  Stated another way...put
>> on your customer hat.  What do you expect from your product?
>> When you're done, go back
>> and try to cut cost.  Weigh the savings against the loss of reliability.
>>
>> If you start with slipshod design, it's hard to fix little things like
>> overvoltaging a FET.
>>
>> Thought experiment:
>> Do two designs.
>> Start one with a quality mindset.
>> Start the other with a minimum cost mindset.
> I'm always doing both at once... biggest bang
> for smallest buck.  That is the value I deliver to
> my customers.  (well, part of it.) And then you've
> got to factor your own time into the equation.
> 

Yup.  My favourite technological sport is doing something amazing(*) 
with nearly no apparatus at all.

Folks that review my designs are sometimes surprised at the percentage 
of LM358s among the op amps--they work fine as bias generators, 
temperature controllers, and stuff like that.

Cheers

Phil Hobbs

(*) for suitably modest values of 'amazing'. ;)

-- 
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

On 4/7/2018 10:17 AM, Pimpom wrote:

> "mike" spoke about his engineers. Well, I don't have an army of
> engineers working under me. But I *have* trained several engineering
> graduates and most of them are in key posts in both private and
> government agencies. Elaborating on that might be seen by some as
> bragging, so I'll refrain from doing it.

I've been out of the business for almost 30 years, but
back in the day, there was nobody to train engineers.
People got moved up in the organization with zero training
for their new post.  They didn't have the skills, so they
couldn't train those who followed them up the organization.
The result was pervasive incompetence at all levels.

I was fortunate to intern under an engineer who understood
the problem and taught me what I didn't learn in school.
I did the same for others.

If you actually trained  your engineers, you deserve some
bragging rights.

On 4/7/2018 11:17 PM, bloggs.fredbloggs.fred@gmail.com wrote:
> On Friday, April 6, 2018 at 6:01:50 AM UTC-4, Pimpom wrote:
>> I'm designing a small simple circuit in which a MOSFET drives a
>> low-power load. The very low frequency gate drive may, on rare
>> occasions, exceed the max Vgs rating of 12V by about 1V, possibly 2V.
>>
>> There are a number of ways to limit the gate voltage but I want
>> to avoid them and keep the circuit as simple as possible, and it
>> won't cause a disaster if the transistor fails. What do you think?
> 
> Does this source have an impedance?

Not enough to protect the gate if it does break down.

> And you do have quiescent current loading limitations?

Not sure what you mean by that.

> What is the range, both minimum and maximum, of your input signal?

Normally not much above the gate threshold. The circuit is about 
proportional control, not hard switching. Exceeding Vgs(max) is a 
remote possibility under abnormal conditions.

> What is this low power load and how mow much voltage variation across it can it take?

It can take the full supply voltage continuously.
  > There's not enough context to offer a fix.

I'm not really looking for a fix. I can think of more than one 
that will cost no more than 1 US cent. The aim is for elegance 
with the least possible parts count.

On Fri, 06 Apr 2018 14:19:49 -0700, John Larkin
<jjlarkin@highland_snip_technology.com> wrote:

>
>I've seen LEDs and optocouplers that have reverse zener voltages in
>the 35v sort of range. The 5V spec is probably to permit multiplexing.

Yes, that's probably why they spec it at 5V, but there are mechanisms
where there is obvious degradation of the LED over a relatively short
time span (1000 minutes) with moderate (40V) reverse bias. 

http://ieeexplore.ieee.org/document/5247076/?reload=true

Something to do with hot carriers. Maybe they could design a better
LED for reverse bias if anyone needed it, but they generally don't. 

--sp 

On Sat, 7 Apr 2018 14:29:30 -0400, Phil Hobbs
<pcdhSpamMeSenseless@electrooptical.net> wrote:

>
>Folks that review my designs are sometimes surprised at the percentage 
>of LM358s among the op amps--they work fine as bias generators, 
>temperature controllers, and stuff like that.

I have never understood why people are are overcome by the vapors at
the use of an LM358, but not by a BJT of similar cost and worse
performance in the application. 

--sp 

On 4/7/2018 11:59 PM, Phil Hobbs wrote:
> 
> Folks that review my designs are sometimes surprised at the percentage
> of LM358s among the op amps--they work fine as bias generators,
> temperature controllers, and stuff like that.
> 

I just love LM358s. I use them all the time - as amplifiers where 
high linearity is not essential, and as comparators when speed is 
not important and an open-collector output is not suitable.