Exceeding Vgs rating

And that is primarily what consumer product engineering is about, bang for buck. Consumer and industrial versions of the same product have a lot of differences, and they're mainly for that reason.

NT

I can think of more than one

an elegant design would remove all doubt of a long term reliability issue for 1 cent

mark

Oh dear. My latest design has microprocessors, FPGAs and lots of analog stuff, almost all of which will be used outside the manufacturer's specifications, in some cases way outside. Luckily, the customers are aware of this.

The connectors and the PCB are within spec though. Phew.

Cheers

That is NOT ACCEPTABLE and you are a BAD PERSON.

Lol. I won't say a thing about unknown & thus effectively unspecified transistors then. Someone's in-house part numbers maybe.

If there are transistors that can't do a few milliamps at a handful of volts I've not met them.

Hi-rel electronics did teach me one thing: how inefficient it was & how to do it all cheaply.

NT

Can you say more about inefficient? Do you mean there was testing and lots of rejects?

George H.

I recall a study that found JAN-TX transistors to be less reliable than regular ones.

If you're exceeding the breakdown voltage specs of your FPGAs and microprocessors by a large margin (way outside), you must be a lucky designer. Even the worst of the clueless designers I've known knew better.

There once was a design for a lithium ion battery charger inside an expensive instrument. When everything was working right, there was no problem with voltages. But, if the battery voltage was allowed to drop below a certain point, the FET that managed the charge disconnect broke down and failed shorted. The fire department was not amused. They didn't learn about that problem until there were many in the field. Big recall. SAD.

It's very hard to teach engineers to do bad things to their designs (at least in thought experiments) and see what happens. I used to sit down with engineers and say things like, "What happens if that cap shorts?" "What happens if those two un-keyed ribbon connectors get swapped in production?" "Put on backwards?" "How much smoke comes out if this device if I short that power supply to ground...or if the negative supply goes open and the voltage gets dragged positive by the loads?"

There's no excuse for an un-keyed ribbon connector to reverse Vcc and Gnd, if you reverse it or put it on "off-by-one". It's amazing how many ribbon connector users put Vcc on one end and ground on the other. It doesn't cost anything to eliminate that possibility if you're used to thinking about how the real world works. Shit happens. Design for it if it costs nothing.

They thought I was crazy, but after the first few averted disasters, they decided that it wasn't such a dumb idea after all.

There are plenty of disasters waiting to happen. There's little reason to purposely introduce others. Saying things like, "nothing serious happens if the FET is destroyed," just adds insult to injury.

alog

re

ransistors then. Someone's in-house part numbers maybe.

volts I've not met them.

to do it all cheaply.

In what way is hi-rel not inefficient? Eliminating some components ups the price Eliminating most suppliers ups the price Eliminating some circuit configs because you can't prove them in every conc eiveable situation costs, eg asynchronous sm volt reg. Training staff to follow numerous extra rules & restrictions consumes money & time & makes it harder to replace them Then there are the many mechanical requirements. Eg what I worked on all ha d to have laced wire looms, laced in a relatively slow way of course, with every single wrap individually knotted. Production needs far more controls, tests & checks Some product designs get dropped because they don't meet some detail or oth er.

On and on it goes, everything is inefficient. Which is fine, sort of. But i t opens the door for competitors selling stuff at a tenth the price.

The biggest issue I had with it all wasn't just the costs of the product pa rts, those are all justifiable for the target market. A similarly inefficie nt staffing setup was what bugged me most. Despite being very successful I became convinced that a new competitor would spring up & wipe them out. It happened. Success breeds inefficiency & cruft which leads to downfall.

NT

You keep harping on about that statement of mine without acknowledging the context. This is very different from what you said in another thread. Let's see -

"Probably...." "I had one fixture that seemed to work, but...." "Plug one in and try it" "I've had some success...."

doesn't sound like you need to be 300% sure of everything before actually doing it.

You guys crack me up. Golly gosh and gadzooks! You scoundrels are being unpleasant to another person. :-)

This reminds me of a story I read in Archie comics some time ago (yes, I still read them). Principal Weatherbee caught Archie using the word 'lousy' and chided him for using a crude expression. Archie watched his words for awhile and was tormented by his friends who bombarded him with excessively staid language until Weatherbee relented.

Maybe I should start cussing a bit.

It's hole-in-the-ground stuff and it's the temperature specs that are routinely exceeded, because otherwise you couldn't do the job and short equipment lifetimes are the norm. Most other specs are treated very conservatively.

Previously unused devices are tested of course, and each designer or team has an inventory of parts which are known to work and even parts of parts which don't work. That's valuable information. The fun starts when a manufacturer moves, for example, to a different process and doesn't mention it because it doesn't change the spec.

Cheers

I know. I'll use FR4 and Molex from now on :-)

Cheers

The way that you tell if an FPGA is fast enough is to crank up the speed until it breaks, and then back off some.

Are you not using FR4 now?

Polyimide, 200'C rated. FR4 will last a short while for prototypes.

Cheers

A long time ago I needed to use a TMS320C52 DSP outside its rated specs. TI were kind enough to give me some test code that exercised the known critical timing paths so that I could test the devices myself. I put the code in the boot path, so that each device was auto-tested every time it was used. This worked very nicely.

John

THAT'S HOW IT'S DONE! You GUARANTEE that each device works in the application.

The marginal cost of that test is almost zero.

YOU take responsibility for the vendor's ability to supply parts that work. That calculated risk may cost you in the long run, but the customer gets a quality product.

You consider the consequences early in the design phase and ENGINEER ways to make it work. That's the mindset you want in all of your design engineers.

That's a far cry from the context of this thread as detailed in the subject line: asking random internet denizens if it's OK to exceed some voltage breakdown spec on some unspecified component by some nondescript amount.

I stand by my original statements in this context, "NO, it ain't OK!"

"Another thread" is the definition of different context.

I can't tell which thread/context to which you refer. I will say that diagnostic experiments aimed to determine what features are compatible or what has failed are not the same as ENGINEERING a product.

When you run out of standard diagnostic options, random non-destructive testing may increase insight.

In the newsgroups, you often see a lot of arguments over whether this USB device that I have in my hand might do what I want when just plugging in the device would answer the question.

If I ever stated that it was OK to exceed specified breakdown voltages in a production device, I hereby retract that statement, offer sincere apologies for my error and vow to do my best to prevent such an irresponsible act in the future.

Well, arrest the OP. I think he's in India.