Is there a known effect in microcontrollers for working a 'little' beyond the absolute maximum voltage as power supply?

Log in or register to post comments
Go To Last Post
16 posts / 0 new
Author
Message
tglaria
tglaria's picture
Level: Hangaround
Joined: Tue. Apr 9, 2013
Posts: 151 View posts
#1
Posted by tglaria: Fri. Oct 10, 2014 - 01:59 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yes, the question is a little too wide.

right now I'm using an ATmega128RFA1 microcontroller whose power supply is rated to be 3.6V max.

 

I did some test with some boards powering them through USB (5V) just to test if they whould burn, but they didn't.

After that I've been powering them with a LiPo battery (4.2V max) and they're working fine for now.

 

Should the chip be damaged for using it at 4.2V?

Is that too much of an overvoltage?

Is there any symptom that I should/could check?

 

Since these are for prototipes, I don't care if they burn, in fact I'd like to know how much voltage they can tolerate without damaging them.

Tags:

AVR Microcontrollers, megaAVR and tinyAVR
Top
Torby
Torby's picture
Level: 10k+ Postman
Joined: Tue. Nov 11, 2003
Posts: 10179 View posts
Location: Bartlesville Oklahoma USA
Posted by Torby: Fri. Oct 10, 2014 - 02:26 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Will it work at 4.2? Likely.

 

Will it endure at 4.2? Not Likely.

 

If I made a hundred of them would they all work at 4.2? You're REALLY pushing your luck here.

The largest known prime number: 282589933-1

It's easy to stop breaking the 10th commandment! Break the 8th instead. 

Top
tglaria
tglaria's picture
Level: Hangaround
Joined: Tue. Apr 9, 2013
Posts: 151 View posts
Posted by tglaria: Fri. Oct 10, 2014 - 03:23 PM (Reply to #2)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Torby wrote:

Will it work at 4.2? Likely.

 

Will it endure at 4.2? Not Likely.

 

If I made a hundred of them would they all work at 4.2? You're REALLY pushing your luck here.

The idea is just to prototype using 4.2V since it's what I have at hand, the definitive solution should work with 3.7V max battery.

Top
dalpilot
dalpilot's picture
Level: Resident
Joined: Mon. Sep 1, 2008
Posts: 516 View posts
Location: Atlanta, GA
Posted by dalpilot: Fri. Oct 10, 2014 - 03:37 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Why not just put a diode, or 2,  between your battery and the processor?

 

Greg

Top
david.prentice
david.prentice's picture
Level: 10k+ Postman
Joined: Sat. Feb 12, 2005
Posts: 29269 View posts
Location: Wormshill, England
Posted by david.prentice: Fri. Oct 10, 2014 - 04:35 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

AFIK,    stress testing of any device is normally done at 'higher values'.      Then some maths is applied to relate the results to the design values.

 

After all,   if a product is supposed to have a MTBF of 20 years,    you are going to be dead before 20 years is up.     And better designs, production have appeared.     So they will use high voltages, extreme temperatures,  ... That way,   they might get the statistical results within hours, days, months, ...      At least within a sensible time.

 

I have no idea or knowledge of the 'voltage' relationship.    But the electronics industry statisticians will know it intimately.     It is your job to ask Google.

 

I would guess that you will be able to expect several hours of use at 5V.

And if you do the maths on 3.7V vs 3.6V,   I guess you will get one or two years instead of 20 years life.

 

If you have changed your job before they fail,   you are golden!

 

David.

Top
John_A_Brown
John_A_Brown's picture
Level: Raving Lunatic
Joined: Sat. Sep 20, 2003
Posts: 6054 View posts
Location: Surrey, England
Posted by John_A_Brown: Fri. Oct 10, 2014 - 05:31 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It all depends on what you take the words "absolute" and "maximum" to mean.

The way I understand the word "maximum", the word "absolute" seems superfluous, so the very fact that the manufacturers have chosen to use the term "absolute maximum" suggests to me that you should take notice.

Four legs good, two legs bad, three legs stable.

Top
alexru
alexru's picture
Level: Raving Lunatic
Joined: Wed. Apr 15, 2009
Posts: 8074 View posts
Location: San Jose, CA
Posted by alexru: Fri. Oct 10, 2014 - 05:54 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Remember that RF part of the RFA1 is a very delicate analog device. Digital parts will likely work at higher voltages, although they might cause permanent damage that will result in higher current consumption in the sleep modes.

 

So never do power consumption tests on devices exposed to high voltages or static shock. Those devices might appear to be working fine, but they are not.

NOTE: I no longer actively read this forum. Please ask your question on www.eevblog.com/forum if you want my answer.

Top
gchapman
gchapman's picture
Level: 10k+ Postman
Joined: Tue. Jan 9, 2007
Posts: 11072 View posts
Location: Arlington, Texas, u.S.A.
Posted by gchapman: Fri. Oct 10, 2014 - 06:23 PM (Reply to #6)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

John_A_Brown wrote:
It all depends on what you take the words "absolute" and "maximum" to mean.

Go beyond the datasheet – Part 1: Dependencies and guardbands

by Gene Frantz, Chau Mai, Ivan Garcia (Texas Instruments)

9/21/2010

http://www.eetimes.com/document.asp?doc_id=1278301

"Dare to be naïve." - Buckminster Fuller

Last Edited: Fri. Oct 10, 2014 - 06:24 PM
Top
gchapman
gchapman's picture
Level: 10k+ Postman
Joined: Tue. Jan 9, 2007
Posts: 11072 View posts
Location: Arlington, Texas, u.S.A.
Posted by gchapman: Fri. Oct 10, 2014 - 06:36 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

tglaria wrote:
Should the chip be damaged for using it at 4.2V?

Is that too much of an overvoltage?

Depends on the voltage, temperature, and process variation.

Some ICs contain a voltage shunt to limit an over-voltage on the IC's power.

4.2V -

If the volume and weight limits aren't too low could consider 3.6V with lithium phosphate or 3.8V with two series connected nickel zinc.

Or, limit the cell's voltage to 4.0V or 4.1V and quadruple or double the cell's lifetime.

With a 1mm thick cell LiPo sure are appealing.

"Dare to be naïve." - Buckminster Fuller

Top
ralphd
ralphd's picture
Level: Resident
Joined: Tue. Feb 11, 2014
Posts: 653 View posts
Posted by ralphd: Sun. Oct 12, 2014 - 03:29 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Look at the various overclocking sites, and you'll see people have been pushing voltage specs for decades.

 

In my experience, absolute maximum ratings in datasheets are 30-50% below what it will take to damage the chips.  And as for any long-term damage, the only thing I know that higher voltage will cause is faster silicon doping migration, which is extremely slow anyway.

 

While moderate over-voltage is highly unlikely to damage a chip, it might not perform within spec.   For the voltage differences you're talking about, I'd take a bet on 100:1 odds that you're safe.

 

However I'd second the suggestion to charge your lithium cell to 4.0 instead of 4.2V.

http://www.powerstream.com/lithu...

 

I have no special talents.  I am only passionately curious. - Albert Einstein

 

Top
John_A_Brown
John_A_Brown's picture
Level: Raving Lunatic
Joined: Sat. Sep 20, 2003
Posts: 6054 View posts
Location: Surrey, England
Posted by John_A_Brown: Mon. Oct 13, 2014 - 08:00 AM (Reply to #8)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

gchapman wrote:

 

John_A_Brown wrote:

It all depends on what you take the words "absolute" and "maximum" to mean.

 

Go beyond the datasheet – Part 1: Dependencies and guardbands

by Gene Frantz, Chau Mai, Ivan Garcia (Texas Instruments)

9/21/2010

http://www.eetimes.com/document.asp?doc_id=1278301

While that is an interesting paper, it would be more relevant(in my opinion) if the OP was raising the voltage in an attempt to squeeze more performance out of the chip. That doesn't seem to be the case, they just don't want the hassle of a regulated supply.

Four legs good, two legs bad, three legs stable.

Top
tglaria
tglaria's picture
Level: Hangaround
Joined: Tue. Apr 9, 2013
Posts: 151 View posts
Posted by tglaria: Mon. Oct 13, 2014 - 09:05 PM (Reply to #10)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

dalpilot wrote:

Why not just put a diode, or 2,  between your battery and the processor?

 

Greg

Just space and consumption/lifetime constraints.

I can get more juice out of the battery if I connect it directly to the microcontroller.

david.prentice wrote:
AFIK,    stress testing of any device is normally done at 'higher values'.      Then some maths is applied to relate the results to the design values.

 

After all,   if a product is supposed to have a MTBF of 20 years,    you are going to be dead before 20 years is up.     And better designs, production have appeared.     So they will use high voltages, extreme temperatures,  ... That way,   they might get the statistical results within hours, days, months, ...      At least within a sensible time.

 

I have no idea or knowledge of the 'voltage' relationship.    But the electronics industry statisticians will know it intimately.     It is your job to ask Google.

 

I would guess that you will be able to expect several hours of use at 5V.

And if you do the maths on 3.7V vs 3.6V,   I guess you will get one or two years instead of 20 years life.

 

If you have changed your job before they fail,   you are golden!

 

David.

As far as I knew, most devices should work a bit over their rated values. 

3.7 over 3.6V isn't too much according to my 'thinking', but just want to be sure.

 

alexru wrote:

Remember that RF part of the RFA1 is a very delicate analog device. Digital parts will likely work at higher voltages, although they might cause permanent damage that will result in higher current consumption in the sleep modes.

 

So never do power consumption tests on devices exposed to high voltages or static shock. Those devices might appear to be working fine, but they are not.

That's the part I'm more concerned about, the analog part.

I'm trying to find what could happen when powering with overvoltage, but haven't found what tests were done.

 

gchapman wrote:

 

John_A_Brown wrote:

It all depends on what you take the words "absolute" and "maximum" to mean.

 

Go beyond the datasheet – Part 1: Dependencies and guardbands

by Gene Frantz, Chau Mai, Ivan Garcia (Texas Instruments)

9/21/2010

http://www.eetimes.com/document.asp?doc_id=1278301

That gives me hope.

 

ralphd wrote:

Look at the various overclocking sites, and you'll see people have been pushing voltage specs for decades.

 

In my experience, absolute maximum ratings in datasheets are 30-50% below what it will take to damage the chips.  And as for any long-term damage, the only thing I know that higher voltage will cause is faster silicon doping migration, which is extremely slow anyway.

 

While moderate over-voltage is highly unlikely to damage a chip, it might not perform within spec.   For the voltage differences you're talking about, I'd take a bet on 100:1 odds that you're safe.

 

However I'd second the suggestion to charge your lithium cell to 4.0 instead of 4.2V.

http://www.powerstream.com/lithuim-ion-charge-voltage.htm

 

The whole idea is to power the chips with 3.7V max (the battery datasheet says it's a 3.6V one, but surely it gets a bit more)

Top
awneil
awneil's picture
Level: 10k+ Postman
Joined: Fri. Jul 1, 2005
Posts: 16433 View posts
Location: Basingstoke, Hampshire, UK
Posted by awneil: Mon. Oct 13, 2014 - 09:25 PM (Reply to #10)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ralphd wrote:
In my experience, absolute maximum ratings in datasheets are 30-50% below what it will take to damage the chips.

The Absolute Maximum ratings are those within which the thing is guaranteed not to get damaged.

 

That, clearly, does not mean that one step over them will cause damage - just that you have no guarantee.

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 oscillator: https://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...
Top
awneil
awneil's picture
Level: 10k+ Postman
Joined: Fri. Jul 1, 2005
Posts: 16433 View posts
Location: Basingstoke, Hampshire, UK
Posted by awneil: Mon. Oct 13, 2014 - 09:26 PM (Reply to #3)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

tglaria wrote:
The idea is just to prototype using 4.2V since it's what I have at hand, the definitive solution should work with 3.7V max battery.

 

Not a believer in the, "Fly What You Test; Test What You Fly" maxim - then?

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 oscillator: https://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...
Top
awneil
awneil's picture
Level: 10k+ Postman
Joined: Fri. Jul 1, 2005
Posts: 16433 View posts
Location: Basingstoke, Hampshire, UK
Posted by awneil: Mon. Oct 13, 2014 - 09:30 PM (Reply to #2)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Torby wrote:
If I made a hundred of them would they all work at 4.2? You're REALLY pushing your luck here.

And would they all continue to work on a hot summer's day? or a cold winter's night? etc, etc,...?

 

This is the difference between a hobbyist one-off and a professional, fully-specified design fit for long-term production & deployment at scale...

 

 

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 oscillator: https://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...
Top
tglaria
tglaria's picture
Level: Hangaround
Joined: Tue. Apr 9, 2013
Posts: 151 View posts
Posted by tglaria: Tue. Oct 14, 2014 - 12:25 PM (Reply to #15)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

awneil wrote:

 

ralphd wrote:

In my experience, absolute maximum ratings in datasheets are 30-50% below what it will take to damage the chips.

 

The Absolute Maximum ratings are those within which the thing is guaranteed not to get damaged.

 

That, clearly, does not mean that one step over them will cause damage - just that you have no guarantee.

I guess this confirms my suspitions then.

I'll just have to test them once I get to a 'beta' prototype to test if I can afford the smaller MTBF because of using a 3.7V power supply instead of a 3.3V regulated one.

 

awneil wrote:

 

tglaria wrote:

The idea is just to prototype using 4.2V since it's what I have at hand, the definitive solution should work with 3.7V max battery.

Not a believer in the, "Fly What You Test; Test What You Fly" maxim - then?

I'm not sure if I understood this maxim.

 

awneil wrote:
Torby wrote:

If I made a hundred of them would they all work at 4.2? You're REALLY pushing your luck here.

And would they all continue to work on a hot summer's day? or a cold winter's night? etc, etc,...?

 

This is the difference between a hobbyist one-off and a professional, fully-specified design fit for long-term production & deployment at scale...

And that's when testing starts to get complicated.

Top