Avrdude and is this chip really dead?

[Rick Law]

I am not familiar with Avrdude's behavior or the ATTINY13A's behavior to know if I am getting false hope here.  With the vast experience base here, I appreciate some input.

This is the scenario:  By mistake, I connected the ATTINY13A to 12V.  So I think it is toasted.  I have some replacements from Mouser enroute.  So, I don't so much need the chip, but I hate to loose the week's "playtime" while I wait for the replacement.

Recalling that some has use 12V for fixing fuse issues when fuse is set wrong (disabling SPIEN), I thought perhaps 12V wouldn't kill the chip.   So I decided to give AvrDude a go and try to read the chip.  AvrDude seem to be reading something except it is always reading zeros including chip ID.   The fuse reads zero, so SPIEN is disabled assuming that is right - which could not be right because if it is right, it means SPIEN is disabled and AvrDude could not have communicated with it...  AvrDude is behaving as if it is communicating, and it seems to think the chip is communicating back.  Perhaps the chip is not stone dead... I thought...  So I do a little more looking into.

Other searches said a 000000 for chip ID is lack of an external clock.  So, I gave it a 1Mhz square wave at pin2, clk-in pin.  That doesn't help.

I hook up the "high volt fuse recuse" circuit found here:
http://www.instructables.com/id/Simple-and-cheap-Fuse-Doctor-for-Attiny/
And the circuit is:


I don't have a BC547 or an 2N3904.  I have a pn2222a and a pn3565.  It does seem to switch the 12V, I got 11V-ish at the collector.  I am not sure the either is fast enough for the circuit.  ATTINY13 does not communcate ( if(Serial.available()>0) is false readfuse() doesn't come back)  so the TINY13A could in fact be dead.  But AvrDude acts as if it is talking to it and it is talking back (with zeros).  AvrDude thinks it can read and write flash/fuse/eeprom (but it wont verify of course since it always read 0.)

Question:

(1) Does AvrDude do any communication check?  Or is it pretending the Tiny13A is talking back even when it is stone dead.

(2) Given the way the Tiny13A stopped working after connecting it to 12V, is what I described the behavior of a dead chip or a recoverable chip?

(3) For the "recuse circuit", is pn2222a or a pn3565 adequate for the job?

Mouser should be sending me the Tiny85V, but I sure like to know if the Tiny13A is recoverable so I can continue development while the 85V makes its way to me.


Thanks for sharing your experience and knowledge!

Rick

[wraper]

AVR won't survive 12V, you could just measure how much current it consumes without any other than power pins being disconnected. If it is more than a few mA it is 100% dead.

[Rick Law]

AVR won't survive 12V, you could just measure how much current it consumes without any other than power pins being disconnected. If it is more than a few mA it is 100% dead.

Ah!  Thanks.  "AVR won't survive 12V" leaves me with little hope.  Let me hook this thing up and see how many mA's it is eating up....

I guess I will just have to buckle down and wait for the parts.

[sleemanj]

Recalling that some has use 12V for fixing fuse issues when fuse is set wrong

She's dead Jim.  12V goes on the reset line to put into HVSP mode, not on VCC, Reset handles it, VCC notsomuch.

[Rick Law]

AVR won't survive 12V, you could just measure how much current it consumes without any other than power pins being disconnected. If it is more than a few mA it is 100% dead.

Thanks for the info!  I just finished re-hooking up and measured how much juice it sucks up (still in my Nanjg105c led driver, no led connected).   Just a day or so ago, I did the exact same measurement when it was working -  2.3mA at 3.5V so I even have a baseline - now it is taking 44mA!  Dead as a mummy in the museum....

I guess for my play time this week, I can chase my neighbor's dog and try to scare the daylight out of her whenever she dumps on my lawn.  I hate those dead grass patches because of that dog peeing/pooping there anyway.

Recalling that some has use 12V for fixing fuse issues when fuse is set wrong

She's dead Jim.  12V goes on the reset line to put into HVSP mode, not on VCC, Reset handles it, VCC notsomuch.

Rick, not Jim...  Thanks!  Interesting that reset is design to handle that and VCC isn't.  I've a lot to learn.

Once I get my 85V from mouser, I can free up the working 13A inside the flashlight.  I will experiment with that fuse-rescue more.  Mean time, since I can't really play with the dead TINY13A on my workbench anymore, I will get warm before I go entertain myself with "catch that dog pooping on my lawn."

[sleemanj]

Rick, not Jim... 

Please return your geek card, it's been cancelled.

https://youtu.be/MH7KYmGnj40

[jwm_]

I guess for my play time this week, I can chase my neighbor's dog and try to scare the daylight out of her whenever she dumps on my lawn.  I hate those dead grass patches because of that dog peeing/pooping there anyway.

If you still want to do something with the Attiny, get some acid and decap it, post some high res photos of the die, see if you can visually see the consequences of 12V.

   John

[Rick Law]

I guess for my play time this week, I can chase my neighbor's dog and try to scare the daylight out of her whenever she dumps on my lawn.  I hate those dead grass patches because of that dog peeing/pooping there anyway.

If you still want to do something with the Attiny, get some acid and decap it, post some high res photos of the die, see if you can visually see the consequences of 12V.

   John

That's an idea...  But I would need to decap a working one to compare.  I am a bit short on working Tiny13a's right now.  So that will have to wait.

[amyk]

The RESET pin can take 12V because it's designed with thicker gate oxide and has no upper ESD diode. The datasheet specifies 13V absolute maximum for it, while VCC max is 6V. The failure mode will likely be punctured gate oxide in many of the transistors. It might not be visible without stripping the metal layer.

[Mechatrommer]

Rick, not Jim... 

Please return your geek card, it's been cancelled.

the star trek?" no way! thats for retardeds..

[Rick Law]

Rick, not Jim... 

Please return your geek card, it's been cancelled.

the star trek?" no way! thats for retardeds..

Oh my God!  My geek card cancelled.  Now I know I am truly worthless.

That explained a lot.  Lacking a Tiny13A to play with, last night, I failed to overcome my kindness so as to go scare a dog...

So, I set by the creek near my house all night with full expectation.  But, no "watery tart" threw a sword to me.  None.  Nothing.  I froze my foot for nothing.  I was wondering where she found another worthier than me, as Arthur is already got his.  I didn't know my geek card was cancelled.  Now I understand.  After I dried my tears, I'll go play with the Dead Parrot till my t85A arrives, I guess.

[AlxDroidDev]

The other day I fried a Amtega328, along with several other chips on the board. The regulator I was using was a MAX603, with a MAX604 after it for 3.3V. I decide to power my board with a 13.5V walwart, and forgot that the MAX604 couldn't handle that much (11.5V max). That was enough to fry the 2 voltage regulators, the Atmega, a MAX412 OpAmp, an ACS712-05... and make a couple green leds turn yellowish!

On the upside, I am glad I fried the ACS712. After that, I found out about the INA219, and got it working beautifully, without having to resort to any opamp magic to get milliamp readings.

[Rick Law]

The RESET pin can take 12V because it's designed with thicker gate oxide and has no upper ESD diode. The datasheet specifies 13V absolute maximum for it, while VCC max is 6V. The failure mode will likely be punctured gate oxide in many of the transistors. It might not be visible without stripping the metal layer.

Thanks for the info!  Once I get the 85V, I will risk my one remaining TINY13A to see if those fuse-reset thing works and get some experience with that.

[amyk]

85V will certainly do enough damage to be visible on the die easily, and probably even without needing a decap.

[newbrain]

85V will certainly do enough damage to be visible on the die easily, and probably even without needing a decap.

/s missing? I think he was referring to the ATtiny85V, but as for high-ish voltage, see below.

When a chip is dead, then it can rest in peace...
It's a bit unsettling when it turns into some kind of vampire:
I managed yesterday evening early this morning to half-cook an STM32F411.
As a public service, I'll state here that it is now ascertained that a GPIO pin, configured as output, will not withstand unregulated 26V DC.

Strangely enough the MCU is still mostly OK, all other output and inputs working, including ADC.
I could use any other free pin instead of the fried one...were it not for the fact that Nosferatu is now sucking about 300mA (up from 13...)!

To give it eternal peace, I'll try zapping the same pin again...it'll either die or, who knows, I might get lucky and burn out just the part that's now crowbarring the supply. Not much to lose.

[mcinque]

Please return your geek card, it's been cancelled.

the star trek?" no way! thats for retardeds..

[Rick Law]

85V will certainly do enough damage to be visible on the die easily, and probably even without needing a decap.

/s missing? I think he was referring to the ATtiny85V, but as for high-ish voltage, see below.

When a chip is dead, then it can rest in peace...
It's a bit unsettling when it turns into some kind of vampire:
I managed yesterday evening early this morning to half-cook an STM32F411.
As a public service, I'll state here that it is now ascertained that a GPIO pin, configured as output, will not withstand unregulated 26V DC.

Strangely enough the MCU is still mostly OK, all other output and inputs working, including ADC.
I could use any other free pin instead of the fried one...were it not for the fact that Nosferatu is now sucking about 300mA (up from 13...)!

To give it eternal peace, I'll try zapping the same pin again...it'll either die or, who knows, I might get lucky and burn out just the part that's now crowbarring the supply. Not much to lose.

I considered this Tiny13A dead until AVRDUDE gave me some false hope when it was able to successfully read (as least it thought so) but reads all zeros.

So now I know.  When AVRDUDE thinks it successfully reads but reads all zeros, the chip may be dead.

[Rick Law]

...
When a chip is dead, then it can rest in peace...
It's a bit unsettling when it turns into some kind of vampire:
I managed yesterday evening early this morning to half-cook an STM32F411.
As a public service, I'll state here that it is now ascertained that a GPIO pin, configured as output, will not withstand unregulated 26V DC.

Strangely enough the MCU is still mostly OK, all other output and inputs working, including ADC.
I could use any other free pin instead of the fried one...were it not for the fact that Nosferatu is now sucking about 300mA (up from 13...)!

To give it eternal peace, I'll try zapping the same pin again...it'll either die or, who knows, I might get lucky and burn out just the part that's now crowbarring the supply. Not much to lose.

[RL: bold added in quote]

Well, there is always the option of wood stick right in the center of the chip.

I removed the ATTINY13A now, and I am more than ready for the 85V.

My wooden tooth-pick could not penetrate the plastic casing of the ATTINY13A.  So, I will pull out the legs, draw a big cross on the casing, exposed to direct sunlight, then bury with ample garlic.  A vampire with no legs should not be too scary.

I have a habit of taking a picture of the PCB whenever I removed chips.  When I trace, I can easily check what traces are under the chip with the picture.  Sharing with anyone interested.

[newbrain]

[...]
I considered this Tiny13A dead until AVRDUDE gave me some false hope when it was able to successfully read (as least it thought so) but reads all zeros.

So now I know.  When AVRDUDE thinks it successfully reads but reads all zeros, the chip may be dead.

I had the more or less the same with an ATmega 328, I was able to read ID and fuses, and (with a High Voltage programmer) to write some of them. No way to reliably program the chip with HVP or SPI, though, so in the electronic waste recycle bin it ended.

[...]
My wooden tooth-pick could not penetrate the plastic casing of the ATTINY13A.

A (very) hot soldering iron might be the weapon of choice, there!
But I'll spare my Hakko for better endeavours.

[Neverther]

By default tiny13a runs at 1.2Mhz so you need to slow avrdude down, atleast if you use usbasp. But as you can try reading chip ID, I think there is some kind of connection?

Pulled a fresh one for test:
avrdude -c usbasp -p t13 -U flash:w:AutoPWM.elf   ->  target doesn't answer

avrdude -B 10 -c usbasp -p t13 -U flash:w:AutoPWM.elf   ->  no problems

You run into this if you run any avr at low clock frequencies.

I've blown one pin on one of mine (buy a bunch theyre only 0.50€ a piece for so-8 when you order bunch from china), but it still worked drawing a lot of current. I think it was non ICSP pin, but that was couple years ago. Atleast programming it worked fine.

[Rick Law]

Yipee...

The Tiny85V arrived.  It is installed.  It is working.  I needed some changes made to the software to make it work, but it does now.  I am not sure what other impacts the change in CPU might have caused.  My PWM looks good on the scope, but worth while taking a pause and see if I have other time/frequency dependent codes that I forgot about.

I will give it some work-out.  After I found everything to be aok, I will take out the other Nanjg105c in my flashlight and replace that one with the Tiny85V also.