Author Topic: Getting Inside A Microcontroller  (Read 7370 times)

0 Members and 1 Guest are viewing this topic.

Offline Lance

  • Frequent Contributor
  • **
  • Posts: 317
  • Country: 00
  • Resistance if futile if R<1Ohm
Getting Inside A Microcontroller
« on: February 03, 2011, 02:22:53 am »
Okay, so this was my second attempt at getting inside a chip, specifically a dead PIC18f4525. It didn't go well. I tried using a handheld propane torch to burn the plastic off, as opposed to sanding this time. It failed. The plastic just bubbled a little. I got green fire, which I know is copper, so I'm fairly certain I trashed whatever is in there.

Does anyone know of a method I could do at home which would let me get at the die?
#include "main.h"
//#include <killallhumans.h>
 

Offline Hypernova

  • Supporter
  • ****
  • Posts: 655
  • Country: tw
Re: Getting Inside A Microcontroller
« Reply #1 on: February 03, 2011, 03:54:51 am »
Acid? Forgot which type that only eat away at the plastic.

Might wanna do it out side of course.
 

Offline Jon Chandler

  • Frequent Contributor
  • **
  • Posts: 545
    • Throw Away PIC
Re: Getting Inside A Microcontroller
« Reply #2 on: February 03, 2011, 04:03:08 am »
Sparkfun did a series of posts on this subject while investigating some fake chips.  There are at least 3 posts in the series:

http://www.sparkfun.com/news/350

http://www.sparkfun.com/news/364

http://www.sparkfun.com/news/395


There are also some additional columns on dissecting fake SD cards on Sparkfun.



Jon


 

Offline Lance

  • Frequent Contributor
  • **
  • Posts: 317
  • Country: 00
  • Resistance if futile if R<1Ohm
Re: Getting Inside A Microcontroller
« Reply #3 on: February 03, 2011, 05:45:55 am »
Wait, so why did fire not burn off the plastic?
#include "main.h"
//#include <killallhumans.h>
 

Offline Simon

  • Global Moderator
  • *****
  • Posts: 16847
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: Getting Inside A Microcontroller
« Reply #4 on: February 03, 2011, 06:22:55 am »
must be fairly heat resistant but plastic will melt before burning at which point you have already made a mess. Who said the chip inside is fire proof ?
 

Offline DJPhil

  • Frequent Contributor
  • **
  • Posts: 511
Re: Getting Inside A Microcontroller
« Reply #5 on: February 03, 2011, 08:25:15 am »
Acid? Forgot which type that only eat away at the plastic.

Might wanna do it out side of course.
Nitric or maybe sulfuric should work. It'll need to be strong enough to require lots of protective gear and a fume hood. I might try this with a few old chips I have next summer.
 

Offline Hypernova

  • Supporter
  • ****
  • Posts: 655
  • Country: tw
Re: Getting Inside A Microcontroller
« Reply #6 on: February 03, 2011, 02:12:46 pm »
must be fairly heat resistant but plastic will melt before burning at which point you have already made a mess. Who said the chip inside is fire proof ?

Not all plastic melt, some will simply burn. That's why not all of them are recyclable as you can't melt them down to make new things.
 

Offline Lance

  • Frequent Contributor
  • **
  • Posts: 317
  • Country: 00
  • Resistance if futile if R<1Ohm
Re: Getting Inside A Microcontroller
« Reply #7 on: February 03, 2011, 05:48:02 pm »
must be fairly heat resistant but plastic will melt before burning at which point you have already made a mess. Who said the chip inside is fire proof ?
I was getting green flames. I'm fairly sure I trashed whatever was inside that case. There was a paper sticker that had pin labels on it which I couldn't get fully off. The fire stripped the remaining paper right off! It did form a really cool effect though. Imagine a microchip, reentering the atmosphere. It looked awesome.

So lesson learned, they don't pick cheap plastic for these things. Is acid really the only way to go here?
« Last Edit: February 03, 2011, 05:54:47 pm by Lance »
#include "main.h"
//#include <killallhumans.h>
 

Offline armandas

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: jp
    • My projects
Re: Getting Inside A Microcontroller
« Reply #8 on: February 03, 2011, 09:07:26 pm »
must be fairly heat resistant but plastic will melt before burning at which point you have already made a mess. Who said the chip inside is fire proof ?
I was getting green flames. I'm fairly sure I trashed whatever was inside that case. There was a paper sticker that had pin labels on it which I couldn't get fully off. The fire stripped the remaining paper right off! It did form a really cool effect though. Imagine a microchip, reentering the atmosphere. It looked awesome.

So lesson learned, they don't pick cheap plastic for these things. Is acid really the only way to go here?

I've said it before, and I'll say it again: if you just want to extract the die, burning is the easiest way. I've never seen a chip package that melts instead of burning to ashes, can you post pics of what it looked like?

When the chip is sufficiently charred, you should be able to use a hammer or pliers to break the package and find a little piece of silicon inside.
 

Offline Lance

  • Frequent Contributor
  • **
  • Posts: 317
  • Country: 00
  • Resistance if futile if R<1Ohm
Re: Getting Inside A Microcontroller
« Reply #9 on: February 03, 2011, 10:15:18 pm »
Unfourtinately I binned it. The plastic literally just sat there. It bent and expanded a little in the middle, and my attempts at removing the plastic failed. Is there a specific type of fire source that works better? I used a handheld propane torch.
#include "main.h"
//#include <killallhumans.h>
 

Offline armandas

  • Frequent Contributor
  • **
  • Posts: 336
  • Country: jp
    • My projects
Re: Getting Inside A Microcontroller
« Reply #10 on: February 03, 2011, 11:08:17 pm »
I've got a cheapy torch from ebay, so nothing special. You could try your luck with different chips you have any. IC's are not hard to come by.
 

Offline TopherTheME

  • Regular Contributor
  • *
  • Posts: 196
Re: Getting Inside A Microcontroller
« Reply #11 on: February 03, 2011, 11:20:52 pm »
You need to use a high molar nitric acid to remove the encasing. I don't know of any other method besides sanding or sandblasting that will give you any real results.
Don't blame me. I'm the mechanical engineer.
 

Offline Lance

  • Frequent Contributor
  • **
  • Posts: 317
  • Country: 00
  • Resistance if futile if R<1Ohm
Re: Getting Inside A Microcontroller
« Reply #12 on: February 04, 2011, 12:03:03 am »
I've got a cheapy torch from ebay, so nothing special. You could try your luck with different chips you have any. IC's are not hard to come by.
True enough, but I'd prefer to not destroy any working technology, at least until I get a viable technique going. So far I've been experimenting with dead chips.
You need to use a high molar nitric acid to remove the encasing. I don't know of any other method besides sanding or sandblasting that will give you any real results.
I tried sanding already, it trashed the die.
#include "main.h"
//#include <killallhumans.h>
 

Offline Wim_L

  • Regular Contributor
  • *
  • Posts: 211
  • Country: be
Re: Getting Inside A Microcontroller
« Reply #13 on: February 04, 2011, 02:34:57 am »
It's very unlikely that a small blowtorch would reach sufficient temperatures to really damage the silicon. Worst case you'd oxidise the surface a bit further. Do make sure you have a plastic and not a ceramic DIP. Ceramic won't burn, though the glue holding the parts together probably would.

For acid, nitric is certainly more versatile. Sulfuric has a reputation for burning very hard, but that's mostly for biological matter (it certainly damages humans quickly). Nitric usually is the better bet if you want to dissolve something. Or in some cases specialised mixtures like aqua regia.

Oh, and ventilate well. If concentrated nitric acid gets too warm or digests a lot of metal, it will form NO2, which is a nasty poison with effects (lung edema) that can be delayed for hours. Something to be careful with if you value your lungs.
 

Offline Lance

  • Frequent Contributor
  • **
  • Posts: 317
  • Country: 00
  • Resistance if futile if R<1Ohm
Re: Getting Inside A Microcontroller
« Reply #14 on: February 04, 2011, 06:34:47 am »
It's very unlikely that a small blowtorch would reach sufficient temperatures to really damage the silicon. Worst case you'd oxidise the surface a bit further. Do make sure you have a plastic and not a ceramic DIP. Ceramic won't burn, though the glue holding the parts together probably would.

For acid, nitric is certainly more versatile. Sulfuric has a reputation for burning very hard, but that's mostly for biological matter (it certainly damages humans quickly). Nitric usually is the better bet if you want to dissolve something. Or in some cases specialised mixtures like aqua regia.

Oh, and ventilate well. If concentrated nitric acid gets too warm or digests a lot of metal, it will form NO2, which is a nasty poison with effects (lung edema) that can be delayed for hours. Something to be careful with if you value your lungs.
My lungs are kind of important. The more I learn, the more I see I'm NOT equipped to deal with this at home. It's definitely plastic, unless Microchip's PIC18 devices are made out of some ceramic that really feels a lot like plastic.

It wasn't a small torch, but it wasn't particularly huge. How hot do you need to get before you damage the silicon? Also, if I remember correctly aren't these semiconductors encased in a thin layer of glass?
#include "main.h"
//#include <killallhumans.h>
 

Offline RayJones

  • Frequent Contributor
  • **
  • Posts: 490
    • Personal Website
Re: Getting Inside A Microcontroller
« Reply #15 on: February 04, 2011, 07:28:04 am »
Many ceramic chips simply had a lid soldered on following wafer installation and lead bonding. Lead as in the fine gold wires, not the heavy metal!
They would be real easy to expose.
 

Offline Lance

  • Frequent Contributor
  • **
  • Posts: 317
  • Country: 00
  • Resistance if futile if R<1Ohm
Re: Getting Inside A Microcontroller
« Reply #16 on: February 05, 2011, 02:33:58 am »
Okay, so it turns out that I have a readily available source of Z80's available. I wonder if getting inside those would be any easier.
#include "main.h"
//#include <killallhumans.h>
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 17330
  • Country: gb
  • 0999
Re: Getting Inside A Microcontroller
« Reply #17 on: February 05, 2011, 10:20:13 am »
Careful removal of the package can be used to disable code protection and enable reverse engineering.

http://www.cl.cam.ac.uk/~sps32/mcu_lock.html
 

Offline Lance

  • Frequent Contributor
  • **
  • Posts: 317
  • Country: 00
  • Resistance if futile if R<1Ohm
Re: Getting Inside A Microcontroller
« Reply #18 on: February 05, 2011, 06:52:13 pm »
Be that as it may, I'm not interested in reverse engineering. I just wanna get at some dies. So it's looking like nitric acid is the only way to go. Acetone apparently won't work on this stuff, and they're apparently built to survive some really stupidly high temperatures.
#include "main.h"
//#include <killallhumans.h>
 

Offline Wim_L

  • Regular Contributor
  • *
  • Posts: 211
  • Country: be
Re: Getting Inside A Microcontroller
« Reply #19 on: February 05, 2011, 09:54:11 pm »
It wasn't a small torch, but it wasn't particularly huge. How hot do you need to get before you damage the silicon? Also, if I remember correctly aren't these semiconductors encased in a thin layer of glass?

Very hot. Melting point of pure silicon is over 1400 degrees, so if it isn't glowing bright red, it'll keep its shape. But it will be chemicall altered somewhat at the surface.

And yes, just about every silicon semicondictor is encased in "glass", if you can call it that. It's really surface oxidation of the silicon, through a high temperature with oxygen and/or water, forming silicon dioxide, the main component of glass. When heating the chip with a burner, you'll grow this layer some more, making it thicker. This will affect the visual appearance somewhat, mostly by making the colours different, because the colour of reflected light depends on the thickness of that thin layer. The effect is rather obvious when heating a silicon wafer over a small area. It discolours visibly because the oxide layer becomes thicker.
 

Offline Lance

  • Frequent Contributor
  • **
  • Posts: 317
  • Country: 00
  • Resistance if futile if R<1Ohm
Re: Getting Inside A Microcontroller
« Reply #20 on: February 07, 2011, 08:55:40 pm »
Good news everyone! I managed to crack open a ceramic intel chip by driving a flathead screwdriver into the side with a hammer. It broke the top right off, and the IC is undamaged! There's a few microscopic bits of ceramic on it, but other then that it looks fine. I'm going to see if a similar method will work for a plastic device.
#include "main.h"
//#include <killallhumans.h>
 

Offline osmosis321

  • Regular Contributor
  • *
  • Posts: 64
Re: Getting Inside A Microcontroller
« Reply #21 on: February 08, 2011, 06:08:07 pm »
Quote
My lungs are kind of important. The more I learn, the more I see I'm NOT equipped to deal with this at home.

Words of exceptional wisdom.
 

Offline Lance

  • Frequent Contributor
  • **
  • Posts: 317
  • Country: 00
  • Resistance if futile if R<1Ohm
Re: Getting Inside A Microcontroller
« Reply #22 on: February 08, 2011, 06:22:47 pm »
Quite. I would love to take some pictures of this thing with my microscope, but I don't have the hardware for it.
#include "main.h"
//#include <killallhumans.h>
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf