Cosmic rays and transistors

Thread Starter

Nathan Hale

Joined Oct 28, 2011
159
Hi folks! Hope all is well! I was just reading an article about cosmic rays.
The article said "Cosmic rays have sufficient energy to alter the states of circuit components in electronic integrated circuits, causing transient errors to occur, such as corrupted data in electronic memory devices, or incorrect performance of CPUs, often referred to as "soft errors" (not to be confused with software errors caused by programming mistakes/bugs). This has been a problem in electronics at extremely high-altitude, such as in satellites, but with transistors becoming smaller and smaller, this is becoming an increasing concern in ground-level electronics as well."

Can some one please let me know why smaller transistors have a higher probability of being affected by cosmic waves? Why would the size of a transistor matter?
Thanks for the replies!
 

wayneh

Joined Sep 9, 2010
17,496
It's less likely to be hit, by virtue of being smaller, but I'm guessing the energy is relatively larger and has a greater effect on the smaller device. Is an elephant affected? No. Is a virus? Maybe.
 

Aleph(0)

Joined Mar 14, 2015
597
Hi folks! Hope all is well! I was just reading an article about cosmic rays.
The article said "Cosmic rays have sufficient energy to alter the states of circuit components in electronic integrated circuits,
I can just say that 100kev gamma dose rate of like 500 Gy per minute (according ionization chamber survey meter) totally crashes tablets and phones! Also there is app that makes phone work like GM counter by counting flipped camera sensor pixels. Is not accurate but fact that it works at all to super low level says something really unkind about cmos:eek:!
 
Last edited:

dl324

Joined Mar 30, 2015
16,845
This has been a problem in electronics at extremely high-altitude, such as in satellites, but with transistors becoming smaller and smaller, this is becoming an increasing concern in ground-level electronics as well."
Can you post a link to the article? The problem has affected devices at sea level for decades.
Can some one please let me know why smaller transistors have a higher probability of being affected by cosmic waves? Why would the size of a transistor matter?
Each node in an integrated circuit has what's known as a "critical charge" which is the amount of charge required for a node to change states. Reducing capacitance and supply voltage has had more impact than transistor size.

In the 70's Intel famously had a problem with soft errors in a DRAM. Gordon Moore (and likely others) thought it was from cosmic rays and Intel purchased some lead bricks to build an enclosure to block them. It turns out that the problem was alpha particles coming from the ceramic package. Moore is a fishing aficionado and used some of the lead to make his own sinkers.
 
Last edited:

Thread Starter

Nathan Hale

Joined Oct 28, 2011
159
Can you post a link to the article? The problem has affected devices at sea level for decades.
Each node in an integrated circuit has what's known as a "critical charge" which is the amount of charge required for a node to change states. Reducing capacitance and supply voltage has had more impact than transistor size.

In the 70's Intel famously had a problem with soft errors in a DRAM. Gordon Moore (and likely others) thought it was from cosmic rays and Intel purchased some lead bricks to build an enclosure to block them. It turns out that the problem was alpha particles coming from the ceramic package. Moore is a fishing aficionado and used some of the lead to make his own sinkers.
https://en.wikipedia.org/wiki/Cosmic_ray#Effect_on_electronics

https://en.wikipedia.org/wiki/Cosmic_ray#cite_note-77
 

RichardO

Joined May 4, 2013
2,270
In the 70's Intel famously had a problem with soft errors in a DRAM. Gordon Moore (and likely others) thought it was from cosmic rays and Intel purchased some lead bricks to build an enclosure to block them.
One fix was to coat the DRAM die with polyimide (Kapton).
There were also problems with the ceramic packaging of the DRAM's emitting alpha rays. This was solved by making sure the clay used to make the ceramic did not contain decaying radioactive material.
 

dl324

Joined Mar 30, 2015
16,845
One fix was to coat the DRAM die with polyimide (Kapton).
Not likely. Neutrons from cosmic rays can penetrate tens of feet of rock. The usual way to make integrated circuits less susceptible is to increase the critical charge.
There were also problems with the ceramic packaging of the DRAM's emitting alpha rays. This was solved by making sure the clay used to make the ceramic did not contain decaying radioactive material.
No one knew that before Intel's investigation. In Intel's case, the problem was caused by using water downstream from an old uranium mine.
 

RichardO

Joined May 4, 2013
2,270
Not likely. Neutrons from cosmic rays can penetrate tens of feet of rock. The usual way to make integrated circuits less susceptible is to increase the critical charge.
Not neutrons but alpha particles -- a byproduct of cosmic rays hitting the atmosphere. And, yes increasing the number of electrons that store the memory state helps.
http://eesemi.com/soft-error.htm

No one knew that before Intel's investigation. In Intel's case, the problem was caused by using water downstream from an old uranium mine.
That I did not know. The above reference shows that the package also emits alpha particles.
 
Top