Some days ago I've reported about the new Raspberry Pi 2, a single board computer, that's also Windows 10 IoT capable. Now some Raspberry Pi 2 owners are facing a bad surprise: A Xenon flash light, available on some cameras, is causing the computer to freeze for a few seconds and forces then a reboot.
I introduced Raspberry Pi 2 here: (see New Raspberry Pi 2 Modell B – it's Windows 10 capable. Since a few hours, there are reports in Raspberry Pi-Forum, discussing a mysterious issue. A user tried to make photos of his Raspberry Pi 2. Suddenly he recognized, that his Pi 2 freeze for a few seconds and reboots, if he uses a camera flash light. Here are two YouTube videos demonstrating this effect, that has been confirmed by several users.
Neowin.net has written this article, where they state, LED flash light doesn't harm, while Xenon flash light causes a reboot.
My first idea was, that it could be an electromagnetic pulse that causes this effect, because Xenon flash lights use a capacitor. Searching the web for xenon lights radio interference results in many hits. My theory was, that an electromagnetic pulse induce some voltage glitches in Raspberry Pi cables and wiring. The first entry within this forum seems to proof my theory. But at the end of the day, I was wrong – it was a photo sensitive effect, that causes the issue.
Somebody packed the Raspberry Pi 2 into a card box, to shield it from any light. Afterward a Xenon flash light could not force a freez or reboot anymore. But which chip could be the trouble maker? Here is a picture of a Raspberry Pi 2 board.
(Source: Raspberry Pi-Foundation/The Register)
It seems that one of the ICs in the lower left corner causes the issue. Or more precise, a user has shielded chip U16 (power regulator) with a piece of bluetack – see my arrow in picture below.
(Source: Raspberry Pi Forum)
After shielding the IC in that way – or after turning the board upside down – the freeze and reboot caused by a Xenon flash light was gone. Searching the web, I found this FAQ, where a ADP160 1.8v regulator used in a gps design causes a similar behavior. User LucaV reported, that after extensive investigation, it would appear that the ADP160 regulators shutdown under direct sunlight, and reduce their output level when partially shaded. The user wrote: When the WLCSP package is assembled on the board with the bump side of the die facing the PCB, reflected light from the PCB surface is incident on active silicon circuit areas and results in the increased leakage currents. Far more discussion may be found within this (already linked) forum.
As a conclusion, with a little help of a piece of bluetack the problem can be solved. Isulating tape will not be too much helpful. Another solution will be hosing the Rapsberry Pi 2 in a box. Because I did not have a Raspberry Pi 2, comments from blog readers owning such a computer will be appreciated. BTW: All older Raspberry Pi models are not affected. I guess, we will soon a modified version of Raspberry Pi 2, that replaces either the chip or hose it with some non transparent material.
Postscriptum: I was wondering a bit about if an what quality control Raspberry Pi developers will use. Beside electromagnetic conformance tests (EMC) I have seen that a stroboscope flash light has been used in additional tests.
Cookies helps to fund this blog: Cookie settings
Pingback: Raspberry Pi 2 doesn’t smile at pictures | gitpull.gherzanandrei