What I hear is everytime alpha particle(cosmic ray) hits memory, it flips a bit or two around. Elledan or somebody said chip manufactures are VERY CAREFUL about possible radioisotope found in raw material for making chip packaging. ECC is supposed to survive this, where regular memory crashes computer. How can I reproduce the situation? Holding a 33.3kBq Americium-241 radio isotope by it hasn't been able to create the problem yet.
Reproducing cosmic ray induced memory error
- Thread starter Jerboy
- Start date
I don't think it are alpha particles which cause the problem when talking about sources outside the packaging of the memory chips, since those particles are already stopped by the packaging of the chips.
Beta particles and gamma radiation might do the trick, where gamma radiation will take a while, since the half-life of 241 Am is 432 years, resulting in relatively little gamma radiation.
In short: why not try an isotope which produces Beta particles?
Beta particles and gamma radiation might do the trick, where gamma radiation will take a while, since the half-life of 241 Am is 432 years, resulting in relatively little gamma radiation.
In short: why not try an isotope which produces Beta particles?
Alpha particles can only travel a couple of cm in air and would probably be blocked by the ceramic of the memroy modules (it can be blocked by skin). The only real danger from alpha particles would be inside the memory module, not outside.
<< Alpha particles can only travel a couple of cm in air and would probably be blocked by the ceramic of the memroy modules (it can be blocked by skin). The only real danger from alpha particles would be inside the memory module, not outside. >>
On the net, I found out Americium 241 emits a little bit of gamma ray in addition to alpha particles. I still couldn't get the computer to crash :|
<<
<< Alpha particles can only travel a couple of cm in air and would probably be blocked by the ceramic of the memroy modules (it can be blocked by skin). The only real danger from alpha particles would be inside the memory module, not outside. >>
On the net, I found out Americium 241 emits a little bit of gamma ray in addition to alpha particles. I still couldn't get the computer to crash :| >>
Read my post again.
I explained (attempted to) that the Gamma radiation emitted by 241 Am will have little effect, since it's so little due to the long half-life of this isotope, and the fact that gamma radiation does not ionize every molecule it approaches, like Alpha particles do. Otherwise one could stop Gamma radiation by a dense material, like Pb, which clearly isn't the case.
What you need are Alpha particles in (very) close proximity to the actual RAM, or an isotope which emits Beta particles.
You can stop Alpha radiation with a brown paper bag. In fact the base of the Alpha radiation measuring device I used was made of Mylar that was so thin a blade of grass would poke a hole in it. You can imagine how laborious it was to find a small sample that was dropped on a lawn. You had to carefully smooth the grass with your hand one way and test and then bend it all over the other way and test again. For one exercise we found ourselves looking at a forty acre field and mentally projecting a lifetime task.
<< Why do you need to crash you computer for
A simpler solution would be installing Win me >>
lol
take a walk with a laptop around chernobyl's reactor 4 and let us know
dude, please tell me your not holding abunch of this crap in your hand and trying to crash your system?
Also, cosmic radiation isn't what causes the problems, its solar radiation which is made up of just about every freakin thing U can imagine.
What R U trying to accomplish here anyways?
Also, cosmic radiation isn't what causes the problems, its solar radiation which is made up of just about every freakin thing U can imagine.
What R U trying to accomplish here anyways?
Elledan's right, you need a beta generator, or you need to get your alpha source very, very close. You'd need very, very high energy gamma radiation to have that cause the effect.
Cosmic rays, and not solar radiation. are the cause of most bit flips in modern RAM.
Cosmic rays, and not solar radiation. are the cause of most bit flips in modern RAM.
<< Elledan's right, you need a beta generator, or you need to get your alpha source very, very close. You'd need very, very high energy gamma radiation to have that cause the effect.
Cosmic rays, and not solar radiation. are the cause of most bit flips in modern RAM. >>
Where do I get a ß emitting source?
<< Cosmic rays, and not solar radiation. are the cause of most bit flips in modern RAM. >>
uhm, whatever U define cosmic rays are... the primary source of ionizing radiation while in orbit is the sun. Ionizing radiation is why we need rad hard chips.
cosmic rays is a very broad term that can mean alot of different things, solar radiation is more specific and IS correct.
JonB
Platinum Member
The type of memory errors you are trying to reproduce would be caused eventually by any high energy gamma ray, regardless of source. That energy is measure in MeV (mega-electron volts) or GeV (giga) rather than the gamma from Americium, which is measured in KeV (kilo-electron volts).
The memory bit-flipping error could be caused by a Compton Scattering or Pair-Production event but would have to hit the memory module in very precise locations to cause a change. Probably would have to right in the P-N junction of a transistor. Plus, there are times in the RAM refresh cycle where it will be more or less sensitive. All in all, it becomes a probability question.
High energy gammas of sufficient quantity to hit a precise location at a sensitive time.
Sorry, but Beta and Alpha particles aren't going to be big players in this. They won't make it through the plastic case. There are no thin, effective gamma or neutron shields though, so they will be your culprits.
For manufacturers of circuitry to be used in space or at nuclear facilities, the bigger question is how to use smaller and smaller die sizes and fabrication methods but maintain some stability in radiation flux. If you've ever wondered why they still use the equivalent of older 486 and 386 processors in spaceborne craft, its because the wider silicon data path is much less sensitive to disruption by an occasional gamma collision. Make the path too skinny, and a single high energy gamma can actually damage the conductor. All materials are damaged by radiation flux. Plastics harden, metals get brittle, PVC starts to give off chlorides to cause corrosion, etc...
The memory bit-flipping error could be caused by a Compton Scattering or Pair-Production event but would have to hit the memory module in very precise locations to cause a change. Probably would have to right in the P-N junction of a transistor. Plus, there are times in the RAM refresh cycle where it will be more or less sensitive. All in all, it becomes a probability question.
High energy gammas of sufficient quantity to hit a precise location at a sensitive time.
Sorry, but Beta and Alpha particles aren't going to be big players in this. They won't make it through the plastic case. There are no thin, effective gamma or neutron shields though, so they will be your culprits.
For manufacturers of circuitry to be used in space or at nuclear facilities, the bigger question is how to use smaller and smaller die sizes and fabrication methods but maintain some stability in radiation flux. If you've ever wondered why they still use the equivalent of older 486 and 386 processors in spaceborne craft, its because the wider silicon data path is much less sensitive to disruption by an occasional gamma collision. Make the path too skinny, and a single high energy gamma can actually damage the conductor. All materials are damaged by radiation flux. Plastics harden, metals get brittle, PVC starts to give off chlorides to cause corrosion, etc...
<< Sorry, but Beta and Alpha particles aren't going to be big players in this. They won't make it through the plastic case. >>
There will still a lot of Beta particles penetrate the plastic packaging. Don't believe it? Try it with a piece of plastic, a Beta-emitting source and measure the particles which come through. You'll find that Beta-particles still penetrate the piece of plastic.
JonB
Platinum Member
Elleden, it would probably depend on the plastic (type and thickness) as to how penetrable it is. A Beta particle is just a free proton or anti-proton. It has high mass and a super high positive charge so any high-density plastic will stop all or most particles.
Saran wrap or thin polyethylene won't stop much, but a hard plastic case, even a few mils thick, with carbon black as the coloring agent will be very effective (like on a RAM module or a CPU).
Saran wrap or thin polyethylene won't stop much, but a hard plastic case, even a few mils thick, with carbon black as the coloring agent will be very effective (like on a RAM module or a CPU).
<< Elleden, it would probably depend on the plastic (type and thickness) as to how penetrable it is. A Beta particle is just a free proton or anti-proton. It has high mass and a super high positive charge so any high-density plastic will stop all or most particles.
Saran wrap or thin polyethylene won't stop much, but a hard plastic case, even a few mils thick, with carbon black as the coloring agent will be very effective (like on a RAM module or a CPU). >>
Hmm... you're probably right. Any idea which kind of plastic is commonly used for the packaging of ICs and what the density of it is?
A Beta particle is just a free proton or anti-proton.
Nope.
It's and electron, or a positron (beta plus)
The difference is in the mass, if someone asks.
Chances of beta radiation from space to go through atmosphere, house, and computer case = 0.
Chances of beta radiation from within the computer doing something to the memory > 0, but not very high.
Honestly, gamma rays are really the bad guys, as they still pack quite a punch and penetrate a lot of things. Not sure if I am concerned though.
BTW, anyone who has an aluminum case does not have to worry about beta radiation. Most metals, even thin, shield alpha and beta highly effectively. With gamma radiation, you need a material with high Z, such as lead - anyone out there have a lead case????
Nope.
It's and electron, or a positron (beta plus)
The difference is in the mass, if someone asks.
Chances of beta radiation from space to go through atmosphere, house, and computer case = 0.
Chances of beta radiation from within the computer doing something to the memory > 0, but not very high.
Honestly, gamma rays are really the bad guys, as they still pack quite a punch and penetrate a lot of things. Not sure if I am concerned though.
BTW, anyone who has an aluminum case does not have to worry about beta radiation. Most metals, even thin, shield alpha and beta highly effectively. With gamma radiation, you need a material with high Z, such as lead - anyone out there have a lead case????
JonB
Platinum Member
Ooops, I missed that one. Mixing protons with electrons.
The plastic that Intel uses in their PPGA chips is called Bismaleimide Triazine. I'm clueless to its properties other than it is thermally conductive. It must work, they've made enough of them.
I didn't find anything about the plastic they encase RAM chips in. Someone else may have better luck.
The plastic that Intel uses in their PPGA chips is called Bismaleimide Triazine. I'm clueless to its properties other than it is thermally conductive. It must work, they've made enough of them.
I didn't find anything about the plastic they encase RAM chips in. Someone else may have better luck.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 24K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter