Stupid question time: what exactly happens when you (far) exceed pin current?

Go To Last Post
40 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If a GPIO has a max rated current of 40mA... what exactly happens if it is exceeded?  By a little?  By a lot?

 

I understand that it gets "damaged", but how?  I figure it is one of the 3 options below:

 

1) The pin "breaks" and is stuck high, sort of like a shunt, allowing the current to flow (eventually causing a fire, which is manageable)

2) The pin "burns out" and is high Z (physically disconnected) - something like a fuse

3) The pin is in an indeterminate state... could be high, could be low, etc

 

Anyone know which?  Is it consistent and predictable?  I just hope someone's experience can give me a preview of what to expect.

 

The application calls for 3.3V micro powered by a 1S LiPo doing some job (not important what) and IF a certain terminal condition occurs, to briefly power a special bulb which will momentarily draw roughly 2A before burning out.  At that point, it is perfectly OK for the device to be damaged...  which is why I'm hoping it is #1 above and just becomes a shunt.  If it is #2, then how quickly?  Can I expect a few milliseconds of high draw to get through?  A few hundred milliseconds?

 

If the answer is #2, then I need to handle the current more "normally"...  a MOSFET switch.  That's not so easy because I haven't found a MOSFET that can be switched at 3.3V and handle 2A/8W and is reasonably price AND is available :).   Relay is just too big or expensive. 

 

 

Last Edited: Thu. Mar 16, 2017 - 01:39 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

1,2 and 3.

John Samperi

Ampertronics Pty. Ltd.

https://www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

2 Amps!

 

You'll need an O'scope to see the brief "flash".

 

Recall that inside the microcontroller "chip" there are dinky tiny little wires that connect the die to the pins.

 

2 Amps is 50 times the Maximum Limit, so you can expect the dinky tiny little wire to act like a fuse and go up in a puff of smoke.

(Obviously the Vcc bond wire might fry, also, and other internal current carrying structures as well, so the above explanation is a bit of a simplification and perhaps viewed as more of a conceptual answer...)

 

If the output driver is fried before the wire I'd think the exact failure mode would be a bit difficult to reliably predict, and hence John's answer, above.

 

JC

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The pin buffers have an equivalent series resistance that will limit the current to some level (at least, before it fails.)  (Someone has measured that resistance, way back in time, on these forums.)

Either the chip/buffer heats up to the point where it breaks (or "becomes damaged"), or the inherent resistance is enough to prevent that and you just get a voltage level that creeps into a range where digital outputs aren't supposed to go.

(so, say the resistance is 50 ohms when sinking current.  You are connected to a 5V rail (which is "ideal", of course :-)), so you're putting 100mA through the pin, and it will rise to within epsilon of 5V instead of staying down near 0V.  Meanwhile, things inside the chip are dissipating 0.5W and starting to get warm.  Various parts that expect to have ~0 or ~5V are probably biased somewhere in the middle, interfering with their low-power "off" and "full on" states, causing excess power dissipation in transistors that weren't designed for it.  Maybe something gives.  Maybe not.  Maybe you can't tell until some other circumstance happens.)

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

There is no way 2A can actually be supplied by a MCU pin, right? For example, typical AVRs have at least 25 ohm output impedance, they can supply no more than 100-150 mA. Sure, this is way over the limit but it's not 2 amps...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I'm with El Guapo. If sufficient current did flow, it would probably blast a crater in the chip. For those that have grossly overloaded a transistor or a ULN2003, you know what I mean.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

El Guapo... yeah... excessive "current" can do this...

 

https://www.youtube.com/watch?v=...

 

 

Ross McKenzie ValuSoft Melbourne Australia

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It is in the nature of semiconductors that there are many failure modes, each with their own cause and each dependent upon the construction of the device.


You will not get 2A directly with an I/O pin.


Picking the Atmega48PA as an example ;
From the 'Pin Driver Strength' in the datasheet you can calculate that the internal resistance of an I/O pin is somewhere in the region of 27 to 55 ohms at a supply voltage of 3V.
That internal resistance will produce a short-circuit current in the vicinity of 80 or so mA, so you will need to parallel about 25 I/O pins to get 2A.


Conclusion : Use an external mosfet to switch the high-current.

See all the previous posts

Last Edited: Thu. Mar 16, 2017 - 03:08 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I told you it was a stupid question.

 

 

Figured I'd explore it before dismissing.  Consider it dismissed, thank you.

 

If anyone knows of a 3V suitable MOSFET that can handle said power requirements... I'd really appreciate it.  I searched for 7 hours before my brain went to mush and I asked that question.

Requirements = 3V activation, 2A, 8W (however brief), AVAILABLE, under a buck would be nice, smaller the better but I'm not too picky.  SMD preferred.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I used 100 of the BUK138-50DL in a recent project, but gee... there have to be literally hundreds of choices. Try the DigiKey parametric search engine for starters.

 

 

Ross McKenzie ValuSoft Melbourne Australia

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You'd think, right?  Alas... not so.  Take the BUK138-50DL for example... if it were available (it is not), the datasheet makes no indication for anything < 4.4V to suggest it would reliably switch at ~3V

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If it is not available... where do you think I was able to buy 150 pieces at around A$0.50 each last October?

 

How many units of whatever do you need? 10, 1000, 1 million?

 

Ross McKenzie ValuSoft Melbourne Australia

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Try the DigiKey parametric search engine

The trick is to pick RANGES of parameters rather than specifics.   Id(cont) 1.8 to 14A, drive voltage 0.9 to 2.5V, N-Channel, sort by price...  There are about 400 choices.  Cheapest looks like a BSS806n in SOT23 for about $0.36...

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

valusoft wrote:

If it is not available... where do you think I was able to buy 150 pieces at around A$0.50 each last October?

 

How many units of whatever do you need? 10, 1000, 1 million?

 

 

OK, let me rephrase... reasonably available with a foreseeable supply going forward from reputable dealers in good supply :)

 

Basically, if DigiKey lists it as "obsolete", I don't design with it.  Call it a rough litmus test, but...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

AlphaOmega should have some suitable parts - or rat some mosfets off a dud motherboard or dud laptop battery.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

westfw wrote:

Try the DigiKey parametric search engine

The trick is to pick RANGES of parameters rather than specifics.   Id(cont) 1.8 to 14A, drive voltage 0.9 to 2.5V, N-Channel, sort by price...  There are about 400 choices.  Cheapest looks like a BSS806n in SOT23 for about $0.36...

 

 

Thanks.  I do know how to use their search though.  BSS806n is only suitable for half a watt, which is why it was not on my results.

 

Mind you, if I'm willing to blow up a micro I should be willing to cook a FET, right?  So perhaps I should be backing off that power dissipation requirement.

 

So... new question:  can I expect a FET that is rated for 0.5W to handle 8W for more than say 100ms before the circuit craps out?  I think I need to do some research on how exactly a FET fails when it fails.

 

This is new to me, come to think about it.  I've never really concerned myself with HOW things fail.  The job was to always spec things so they don't fail in the first place.  lol   Interesting project.  Since the data I require can't possibly be in the datasheet, I suppose its time to just order some up and start playing.

 

Now where's my fire extinguisher...

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

OK... apologies from me. I didn't mean to offend.

 

Only you know all of your selection criteria. Mine were met by my choice.

 

Best of luck with your search.

 

Ross

 

Ross McKenzie ValuSoft Melbourne Australia

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Kartman wrote:

AlphaOmega should have some suitable parts - or rat some mosfets off a dud motherboard or dud laptop battery.

 

Thanks for the suggestion.  I forgot about them.  I've used their stuff in the past when I couldn't find anything more mainstream, so certainly worth looking at.

 

Their parametric search is buggy, but thankfully it actually didn't take that long to scan through their 60 pages of N-channel mosfet listings.  Out of that, there was exactly 3 units that met the stated requirements, not one of which is in production!!  One is listed as obsolete and the other 2 are "last time buy".  I haven't gone through the datasheets to see if they tick off all the boxes, but given their production status... no point anyway.

 

I suspect there are plenty that will work, they just aren't committing the data for low voltage RDS(on) values.  Without that, there's no way to know (aside from experimentation) which will reliably switch.

 

 

I think what I'll have to do is pick the closest one that will work with the voltage and can handle the amps and then not worry about the power dissipation.  I'll try it and see how long it takes to cook. 

 

Thanks for help everyone.  Consider this solved.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

valusoft wrote:

OK... apologies from me. I didn't mean to offend.

 

Only you know all of your selection criteria. Mine were met by my choice.

 

Best of luck with your search.

 

Ross

 

 

Certainly no offense Ross!  I can't start with a vague question and expect people to know every single criteria... nor can I possibly think of every criteria to present.  I did my best, but I made no mention of intent w.r.t. production, etc.  You couldn't know that I can't be buying grey market due to future instability of supply.

 

To be clear, I appreciate your efforts.  Thank you.

Last Edited: Thu. Mar 16, 2017 - 06:53 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

s_mack wrote:
... to briefly power a special bulb which will momentarily draw roughly 2A before burning out.
The resistance of a bulb's cold filament will result in well over 2A for a "brief" duration.

s_mack wrote:
That's not so easy because I haven't found a MOSFET that can be switched at 3.3V and handle 2A/8W and is reasonably price AND is available :).
A large power FET might handle the bulb's surge current.

If the Vds is low enough then Vgs-th will likely be low but the Cgs will be large resulting in a brief AVR over-current.

The Vds is low so could consider a bipolar power transistor (no secondary breakdown) with its effectively lower Vgs-th equivalent.

The following is a sequence of current sinks and sources from the bulb back to the AVR with the assumption that the bulb is powered from the 1S Li-Po :

Switch a 2A bulb thru current sink, current source, current sink, AVR

All electronics survive the system's terminal condition; recover from the terminal condition by replacing the bulb.

Some Li-Po can be damaged by excessive discharge current; might need to replace it.

Q1 should have a small or mid-size pad on its collector tab if the bulb takes "a while" to open.

 


 

Attachment(s): 

"Dare to be naïve." - Buckminster Fuller

Last Edited: Thu. Mar 16, 2017 - 07:18 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

My first reaction on just reading the thread title:

 

https://i.ytimg.com/vi/HqGpKbCAt70/hqdefault.jpg

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 BSS806n is only suitable for half a watt

 Remember that that rating is power dissipation IN THE MOSFET, not in the load.  If you have 57mOhm On resistance (Vgs=2.5V) and 2A, the mosfet is dissipating a quarter-watt or so...

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

s_mack wrote:
The application calls for 3.3V micro powered by a 1S LiPo doing some job (not important what) and IF a certain terminal condition occurs, to briefly power a special bulb which will momentarily draw roughly 2A before burning out. At that point, it is perfectly OK for the device to be damaged... which is why I'm hoping it is #1 above and just becomes a shunt. If it is #2, then how quickly? Can I expect a few milliseconds of high draw to get through? A few hundred milliseconds?

 

This sounds like an IED ignition device? 

 

(Possum Lodge oath) Quando omni flunkus, moritati.

"I thought growing old would take longer"

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I agree with #20, maybe this is a job for BJTs. 3V may be too low to switch a power MOSFET on.

 

Here are some candidates. Find one that is cheap and has low VCE saturation drop. If the gain is not enough so that the MCU can drive it with 20 mA or so, drive the base with another (small signal) BJT or MOSFET. Example: 2SB1202T.

 

 

edit:

ki0bk wrote:

This sounds like an IED ignition device? 

 

It's low voltage, remotely controlled suicide vest most likely.

Last Edited: Thu. Mar 16, 2017 - 02:08 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

I wonder how many ears at NSA and GCHQ just popped up ?? smiley

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Recommend at least one moderator apply their intuition.

If the moderator feels that this thread by s_mack should be locked then lock it until the PM between the two reaches a mutually agreeable resolution.

 

"Dare to be naïve." - Buckminster Fuller

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ki0bk wrote:
This sounds like an IED ignition device?

Or pyrotechnics?

 

 

Top Tips:

  1. How to properly post source code - see: https://www.avrfreaks.net/comment... - also how to properly include images/pictures
  2. "Garbage" characters on a serial terminal are (almost?) invariably due to wrong baud rate - see: https://learn.sparkfun.com/tutorials/serial-communication
  3. Wrong baud rate is usually due to not running at the speed you thought; check by blinking a LED to see if you get the speed you expected
  4. Difference between a crystal, and a crystal oscillatorhttps://www.avrfreaks.net/comment...
  5. When your question is resolved, mark the solution: https://www.avrfreaks.net/comment...
  6. Beginner's "Getting Started" tips: https://www.avrfreaks.net/comment...
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Won't a lot depend on what the "bulb" requirements are for voltage and current?  When I hear "bulb", I think incandescent.  Are there e.g. 1V "bulbs" that are going to quickly fry at 3V?

 

Anyway, my first thought was to charge a cap during the idle time, and then release it at the critical moment.

 

Indeed knowing the end need would be most interesting.  Is it the flash of light that is important?  Or the creation of the open circuit?  If the latter, then a solid-state fuse?

 

 

You can put lipstick on a pig, but it is still a pig.

I've never met a pig I didn't like, as long as you have some salt and pepper.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have used the IRLB8743PBF with excellent results.

Another way to narrow your search is to look for "Logic Level MOSFET". Google is your friend for this.

There are many out there that fit your needs.

Jim

I would rather attempt something great and fail, than attempt nothing and succeed - Fortune Cookie

 

"The critical shortage here is not stuff, but time." - Johan Ekdahl

 

"Step N is required before you can do step N+1!" - ka7ehk

 

"If you want a career with a known path - become an undertaker. Dead people don't sue!" - Kartman

"Why is there a "Highway to Hell" and only a "Stairway to Heaven"? A prediction of the expected traffic load?"  - Lee "theusch"

 

Speak sweetly. It makes your words easier to digest when at a later date you have to eat them ;-)  - Source Unknown

Please Read: Code-of-Conduct

Atmel Studio6.2/AS7, DipTrace, Quartus, MPLAB, RSLogix user

Last Edited: Thu. Mar 16, 2017 - 10:32 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

theusch wrote:
Anyway, my first thought was to charge a cap during the idle time, and then release it at the critical moment.

Previously, on AVRFreaks: https://www.avrfreaks.net/forum/c...

Top Tips:

  1. How to properly post source code - see: https://www.avrfreaks.net/comment... - also how to properly include images/pictures
  2. "Garbage" characters on a serial terminal are (almost?) invariably due to wrong baud rate - see: https://learn.sparkfun.com/tutorials/serial-communication
  3. Wrong baud rate is usually due to not running at the speed you thought; check by blinking a LED to see if you get the speed you expected
  4. Difference between a crystal, and a crystal oscillatorhttps://www.avrfreaks.net/comment...
  5. When your question is resolved, mark the solution: https://www.avrfreaks.net/comment...
  6. Beginner's "Getting Started" tips: https://www.avrfreaks.net/comment...
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

gchapman wrote:

Recommend at least one moderator apply their intuition.

If the moderator feels that this thread by s_mack should be locked then lock it until the PM between the two reaches a mutually agreeable resolution.

 

 

Huh?  Why in the world would this discussion need locking???

 

edit:  oh, the IED suggestion?   Lol.  I was sure that was a joke, but just in case... if I were arming a vest, surely I wouldn't care about a 3V requirement or how much the FETs cost, right? 

Last Edited: Sat. Mar 18, 2017 - 09:49 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jgmdesign wrote:

I have used the IRLB8743PBF with excellent results.

Another way to narrow your search is to look for "Logic Level MOSFET". Google is your friend for this.

There are many out there that fit your needs.

Jim

 

From the datasheet, how can I know it would work at ~3V?  There's no data < 4.5V.  That's the problem.  And a lot of you say "there are lots"... but there aren't!

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

westfw wrote:

 BSS806n is only suitable for half a watt

 Remember that that rating is power dissipation IN THE MOSFET, not in the load.  If you have 57mOhm On resistance (Vgs=2.5V) and 2A, the mosfet is dissipating a quarter-watt or so...

 

 

I just came to that realization this morning.  You're absolutely right, of course.  I was thinking the bulb says it pulls 8W so I need a FET rated for 8W.  Duh.   THAT changes everything, and now I can re-start my search.

Last Edited: Sat. Mar 18, 2017 - 09:36 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

And with that.... seeing that perhaps power dissipation may not be the concern I thought it was... I did rig up a test this morning using a random N-ch Mosfet I found in the bin.  I don't even know what it is, but it isn't very large (SOT-23-3 style package).  Someone forgot to label them and I don't recognize it from previous projects so I'm not sure how useful this test is, but I have a few dozen of them so whatever.  Nevertheless, it worked how I was hoping:  it switched on, stayed on longer than I expected (between 4 and 5 seconds), and then died an uneventful death.  The FET was warm, but not hot.  I put in a new flash bulb and tried it again but it didn't work.  So I figure I did blow something.  I soldered up a new FET and it still won't turn on.  Could I have damaged the micro?  Investigating.

Last Edited: Sat. Mar 18, 2017 - 10:06 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

theusch wrote:

Won't a lot depend on what the "bulb" requirements are for voltage and current?

Yes, absolutely.  Everything depends on that, which is why I'm here :)

 

theusch wrote:
Indeed knowing the end need would be most interesting.  Is it the flash of light that is important?

 

The end result is switching on the bulb.  Yes, the flash is important.

 

It seems curiosity is calling for me to explain my project.  If anyone cares, sure I could elaborate.  But not now... my wife is beckoning.  Post back if you're genuinely interested.  If those "IED" posts are the extent of the curiosity... don't bother, it is nothing so exciting.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

A Pest/Burglar frightener ?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

s_mack wrote:
If those "IED" posts are the extent of the curiosity... don't bother, it is nothing so exciting.

I am going to ask that the IED references stop as I will have no other choice but to move this thread to the Administrators forum like I did with another thread earlier today.

Jim - Moderator

 

s_mack wrote:
It seems curiosity is calling for me to explain my project. If anyone cares, sure I could elaborate.

Please do tell us what you are making.  the suspense has me intrigued.

 

Jim - the Nerd

I would rather attempt something great and fail, than attempt nothing and succeed - Fortune Cookie

 

"The critical shortage here is not stuff, but time." - Johan Ekdahl

 

"Step N is required before you can do step N+1!" - ka7ehk

 

"If you want a career with a known path - become an undertaker. Dead people don't sue!" - Kartman

"Why is there a "Highway to Hell" and only a "Stairway to Heaven"? A prediction of the expected traffic load?"  - Lee "theusch"

 

Speak sweetly. It makes your words easier to digest when at a later date you have to eat them ;-)  - Source Unknown

Please Read: Code-of-Conduct

Atmel Studio6.2/AS7, DipTrace, Quartus, MPLAB, RSLogix user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 

I certainly didn't start that ridiculous commentary, but sure I can refrain from the reference.  I don't understand the paranoia, but I can comply no problem.

 

The project...  photography capturing events in nature.  Specifically, destructive events such as an earthquake where the equipment is less likely to survive, which is why I want it as disposable (ie. inexpensive) as possible.  Other projects exist that are extremely robust, but they are expensive and that prohibits having numerous collection points.  My hope is to have hundreds of these, if not more.  The camera we're considering (due to specific attributes and cost) doesn't work well in low-light conditions.  External digital flash units are very expensive.  I found a source of cheap old school flash bulbs for next to nothing.  The idea was that if the event happened in low light (easy enough to detect), to trigger the flash and take the picture.  But I was having trouble marrying the digital aspects of the controls with the distinctly non-digital requirements of the bulb. 

 

Anyway, it may be a moot point.  The experiments I ran earlier seemed promising at first but I now see I have to go back to the drawing board.  The flash is nowhere near as bright or fast as it needs to be.  The bulbs may have been cheap for a reason... perhaps they are simply past their shelf life.  Or I'm not giving it enough power... though I'm limited there for other reasons.  I'll probably go back to looking for sufficiently bright LEDs.  I was on that avenue when I came across the bulbs.

 

Once again, feel free to consider this closed.  I'm sure I'll start another thread when I run into my next problem :)

 

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Post a picture of the bulbs.  If they are of the old school style with what looks like steel wool in them, yes they can be rather slow.

 

LED's are not very fast compared to xenon strobes like in standard flash units.  You can probably get those on ali-whatever for a couple of bucks.

 

Worth looking into

 

JIm

I would rather attempt something great and fail, than attempt nothing and succeed - Fortune Cookie

 

"The critical shortage here is not stuff, but time." - Johan Ekdahl

 

"Step N is required before you can do step N+1!" - ka7ehk

 

"If you want a career with a known path - become an undertaker. Dead people don't sue!" - Kartman

"Why is there a "Highway to Hell" and only a "Stairway to Heaven"? A prediction of the expected traffic load?"  - Lee "theusch"

 

Speak sweetly. It makes your words easier to digest when at a later date you have to eat them ;-)  - Source Unknown

Please Read: Code-of-Conduct

Atmel Studio6.2/AS7, DipTrace, Quartus, MPLAB, RSLogix user

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

My guess it that when the bulb fires, the gas is ionised and the impedance of the bulb is near zero, so large current will flow especially since a lipo is involved. I'd suggest some small series resistance - low enough to not affect the ignition but high enough to limit the peak currents. Or an inductor.