jdb
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 2075
Joined: Thu Jul 11, 2013 2:37 pm

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 2:30 pm

Ravenous wrote:Trivia question: is it a flash (ie a brief transient) or can a constant but very high intensity light cause this? I'm thinking of the step response in the control circuits of the regulator chip - will it be stable under a sustained light level? If so, sunlight is less likely to be a serious problem.

(Just a thought. Not that it matters much to most of us.)
Of all the constant light sources, direct sunlight produces the most pronounced change in 1v2 core voltage (which is to be expected with a massively broad-band, high intensity source) but this is far more benign than the instantaneous change caused by a xenon lamp.

The discharge time for a lamp is of the order of 10s of microseconds which interferes with the control loop response in addition to the output voltage. I can imagine the same result with a narrow point source - lasers reflecting off various surfaces near the device causing instantaneous fluctuations.

The resulting dip in the 1v2 line (as displayed prominently in the blog post) is what causes the Pi to crash/brownout/reboot.
Rockets are loud.
https://astro-pi.org

simpex
Posts: 45
Joined: Tue Feb 03, 2015 8:40 am

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 2:33 pm

jamesh wrote:The sun is VERY bright in Cambridge at the moment...
And have you tried to suddenly (not little by little) expose Raspberry Pi 2 to direct sunlight? (Windows reflect the infrared radiation so they have to be opened if you are indoors.)
Last edited by simpex on Mon Feb 09, 2015 3:12 pm, edited 2 times in total.

Die Schatzjaeger
Posts: 3
Joined: Mon Feb 09, 2015 12:43 pm

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 2:34 pm

As far as I tested the (hardware) Watchdog does it's job.
It even does recognize that something is wrong...

Aktivated the WDT:

Code: Select all

[email protected]:~# modprobe bcm2708_wdog heartbeat=14
[email protected]:~# echo 0 > /dev/watchdog
Flashed the Pi:

Code: Select all

[  753.759568] wdt: WDT device closed unexpectedly.  WDT will not stop!
About 14 Seconds later:

Code: Select all

[email protected]:~# Uncompressing Linux... done, booting the kernel.
[    0.000000] Booting Linux on physical CPU 0xf00
Output is via UART of the Pi.

Thomas

Ravenous
Posts: 1956
Joined: Fri Feb 24, 2012 1:01 pm
Location: UK

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 2:42 pm

jdb wrote:The discharge time for a lamp is of the order of 10s of microseconds which interferes with the control loop response in addition to the output voltage.
Thanks, that's interesting (it was purely academic interest in my case) - I think this means there's both a steady state offset from bright light (probably insignificant) and a destabilising effect from very sudden transients (which is the main problem here).

Anyway no need to worry about sunlight, I'm sure now.

Not that it matters to most of us, but it's a fun one for Control Theory students. :)

simpex
Posts: 45
Joined: Tue Feb 03, 2015 8:40 am

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 2:44 pm

Die Schatzjaeger wrote:As far as I tested the (hardware) Watchdog does it's job.
Then why does Raspberry Pi 2 freeze, after being illuminated with a Xenon Flash or Laser Pointer, without resetting and starting to run its code again?

aoakley
Posts: 66
Joined: Fri May 11, 2012 10:27 pm
Contact: Website

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 2:51 pm

So, in short: small lump of wall tack or snippet of black electrical insulating tape over the offending chip, identified in mikerrr's photo, and we're done, right?

Image

My experiment for when I get home, is to discover whether using Kapton tape, pulled from from a 1B/B+ HDMI socket, is sufficiently opaque as to prevent the problem too.

Because heaven knows how much we miss that Kapton tape. Our two B2s arrived from RS on Tuesday without any; the HDMI sockets did not have the cutaways, so presumably the little sucking component placement robot didn't need it any more. Peeling it off the HDMI port was a job my little 5-year-old goblin... er,daughter Beatrice considered to be a crucial part of setting up a new Raspberry Pi. Poor little mite was distraught that her key role in computing was no longer required with the Raspberry Pi 2. I had to assign her a new role as third Pibow hexnut tightener.

It would be *hilarious* if the solution was to start shipping with a patch of Kapton tape again... ;-)

Also we totally missed the opportunity to name this XENON (pi) 2 MEGABLAST . Remember, it's not retro if it's just been sitting in a box in your attic for thirty years, it's just old.
Last edited by aoakley on Mon Feb 09, 2015 3:09 pm, edited 2 times in total.
Andrew Oakley - Cotswold Raspberry Jam, Gloucestershire - http://www.cotswoldjam.org

jdb
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 2075
Joined: Thu Jul 11, 2013 2:37 pm

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 2:52 pm

simpex wrote:
Die Schatzjaeger wrote:As far as I tested the (hardware) Watchdog does it's job.
Then why does Raspberry Pi 2 freeze, after being illuminated with a Xenon Flash or Laser Pointer, without resetting and starting to run its code again?
The watchdog driver isn't modprobed by default, because a stack trace left on-screen is far more desirable than a constantly rebooting Pi.

It's available as an option if people want to use it.
Rockets are loud.
https://astro-pi.org

Die Schatzjaeger
Posts: 3
Joined: Mon Feb 09, 2015 12:43 pm

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 2:52 pm

@simpex:
The hardware Watchdog needs
a) to be enabled, and
b) constantly feed with 'don't reset me' signs.

I think no Raspberry Pi Distribution has the Watchdog enabled by default.

Thomas

Ravenous
Posts: 1956
Joined: Fri Feb 24, 2012 1:01 pm
Location: UK

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 2:58 pm

aoakley wrote:Peeling it off the HDMI port was a job my little 5-year-old goblin... er,daughter Beatrice considered to be a crucial part of setting up a new Raspberry Pi.
Now give her the job of painting a small blob of dark coloured nail varnish on the offending component! :)

(Or perhaps the job of soldering a good capacitor very close to that regulator, if you trust her soldering skills!!)

simpex
Posts: 45
Joined: Tue Feb 03, 2015 8:40 am

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 3:10 pm

If the watchdog is active, does Raspberry Pi 2 always reset after being illuminated with strong light sources that cause glitches to that U16?

simpex
Posts: 45
Joined: Tue Feb 03, 2015 8:40 am

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 3:25 pm

jdb wrote:Of all the constant light sources, direct sunlight produces the most pronounced change in 1v2 core voltage (which is to be expected with a massively broad-band, high intensity source) but this is far more benign than the instantaneous change caused by a xenon lamp.
The discharge time for a lamp is of the order of 10s of microseconds which interferes with the control loop response in addition to the output voltage. I can imagine the same result with a narrow point source - lasers reflecting off various surfaces near the device causing instantaneous fluctuations.
From what I have seen in a video clip, somebody was relatively slowly moving a red spot, generated by a laser pointer, along the Raspberry Pi 2 PCB and when the spot reached a certain place the board malfunctioned.

Definitely, the laser light "slowly" covered the problematic IC (possible U16) in likely milliseconds or more not hundreds on nanoseconds (10s of microseconds).

agprimatic
Posts: 2
Joined: Mon Feb 09, 2015 4:13 pm

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 4:15 pm

In the comments section of the Hackaday post, tekkieneet posted this informative link about light sensitivity and CSP packages:

http://www.silabs.com/Support%20Documen ... AN0878.pdf

jamesh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 23369
Joined: Sat Jul 30, 2011 7:41 pm

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 4:34 pm

simpex wrote:
jdb wrote:Of all the constant light sources, direct sunlight produces the most pronounced change in 1v2 core voltage (which is to be expected with a massively broad-band, high intensity source) but this is far more benign than the instantaneous change caused by a xenon lamp.
The discharge time for a lamp is of the order of 10s of microseconds which interferes with the control loop response in addition to the output voltage. I can imagine the same result with a narrow point source - lasers reflecting off various surfaces near the device causing instantaneous fluctuations.
From what I have seen in a video clip, somebody was relatively slowly moving a red spot, generated by a laser pointer, along the Raspberry Pi 2 PCB and when the spot reached a certain place the board malfunctioned.

Definitely, the laser light "slowly" covered the problematic IC (possible U16) in likely milliseconds or more not hundreds on nanoseconds (10s of microseconds).
Speed of moving the dot is irrelevant - the point is that as the dot hits the edge, there is an instant increase in light (providing it's a decent laser). One moment unlit, next moment lit.
Principal Software Engineer at Raspberry Pi (Trading) Ltd.
Contrary to popular belief, humorous signatures are allowed. Here's an example...
"My grief counseller just died, luckily, he was so good, I didn't care."

User avatar
PeterO
Posts: 4942
Joined: Sun Jul 22, 2012 4:14 pm

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 4:51 pm

agprimatic wrote:In the comments section of the Hackaday post, tekkieneet posted this informative link about light sensitivity and CSP packages:

http://www.silabs.com/Support%20Documen ... AN0878.pdf
I vote for the "Glob-Top Epoxy" solution :D
PeterO
Discoverer of the PI2 XENON DEATH FLASH!
Interests: C,Python,PIC,Electronics,Ham Radio (G0DZB),1960s British Computers.
"The primary requirement (as we've always seen in your examples) is that the code is readable. " Dougie Lawson

User avatar
aTao
Posts: 1087
Joined: Wed Dec 12, 2012 10:41 am
Location: Howlin Eigg

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 4:54 pm

Dimmable very bright lights are tricky to come by, a lens and the sun would do, start with a wide dot and stop before you set fire to the board.

Next test is how bright does direct sunlight need to be?
Winter's day?
Spring morning?
Outside a pub in the summer?
Caribbean beach?

Testers need right now....
>)))'><'(((<

User avatar
RaTTuS
Posts: 10412
Joined: Tue Nov 29, 2011 11:12 am
Location: North West UK

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 4:56 pm

aTao wrote:Dimmable very bright lights are tricky to come by, a lens and the sun would do, start with a wide dot and stop before you set fire to the board.

Next test is how bright does direct sunlight need to be?
Winter's day?
Spring morning?
Outside a pub in the summer?
Caribbean beach?

Testers need right now....
I volunteer for the Caribbean send me ticket to my email.... ;)
How To ask Questions :- http://www.catb.org/esr/faqs/smart-questions.html
WARNING - some parts of this post may be erroneous YMMV

1QC43qbL5FySu2Pi51vGqKqxy3UiJgukSX
Covfefe

simpex
Posts: 45
Joined: Tue Feb 03, 2015 8:40 am

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 5:25 pm

jamesh wrote:Speed of moving the dot is irrelevant - the point is that as the dot hits the edge, there is an instant increase in light (providing it's a decent laser). One moment unlit, next moment lit.
Speed is relevant because the laser dot does not present a constant intensity across its entire surface. The edges of the spot have close to zero intensities while the center reaches the maximum. As the red dot comes closer to the IC that causes the glitch, more and more light illuminates the circuit. So, that U16 (I hope U16 is the real culprit) is gradually illuminated by the laser dot.

What exactly do you see on the oscilloscope when a laser dot is switched on, pointed a few inches away from U16 and then moved above U16?

User avatar
jbeale
Posts: 3476
Joined: Tue Nov 22, 2011 11:51 pm
Contact: Website

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 6:03 pm

My bottle of black nail polish says: Ethyl Acetate, Butyl Acetate, Nitrocellulose, Acetyl Tributyl Citrate, Isopropyl Alcohol, Adipic Acid/Neopentyl Glycol/Trimellitic Anhydride Copolymer, Acrylates Copolymer, Stearalkonium Bentonite, Styrene/Acrylates Copolymer, n-Butyl Alcohol, Benzophenone-1, Silica and MAY CONTAIN Bismuth Oxychloride/CI 77163, Black 2/CI 77266, Ferric Ammonium Ferrocyanide/CI 77510 and a long list of various other coloring agents. Wow.

Any fluid you paint over a BGA package is going to get wicked underneath around the solder balls and die surface, and even a little bit of conductivity or corrosiveness could be an issue. I guess the only materials actually tested for that application are the plastics (epoxy?) used for chip packaging and "glob-top" chip-on-board mounting.

I’ve used black nail polish for anti-reflection coating various optoelectronic prototypes without ever seeing conductivity or corrosion issues, but I never tried it on a bare die. I have also used black acrylic paint but found it slightly conductive; maybe it uses carbon black as the coloring agent.

EDIT: google reports that "Black 2 CI 77266" is indeed carbon black. So the dried film may be expected to have some conductivity, and may even be microphonic like the early telephones using packed carbon granules as a microphone.
Last edited by jbeale on Mon Feb 09, 2015 6:12 pm, edited 3 times in total.

Ravenous
Posts: 1956
Joined: Fri Feb 24, 2012 1:01 pm
Location: UK

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 6:05 pm

jbeale wrote:Any fluid you paint over a BGA package is going to get wicked underneath around the solder balls and die surface, and even a little bit of conductivity or corrosiveness could be an issue. I guess the only materials actually tested for that application are the plastics (epoxy?) used for chip packaging and "glob-top" chip-on-board mounting.
I'd suggest differential thermal expansion too. It may cause the solder balls to fail eventually. I always wondered if this was the reason why people recommended against embedding the raspberry pi in potting compound, acrylic, etc...

simpex
Posts: 45
Joined: Tue Feb 03, 2015 8:40 am

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 6:11 pm

It appears that circuits inside CSP packages are upset by irradiances as low as 4 mW/cm2 (direct sunlight > 100 mW/cm2)

Not only Xenon flashes or Laser Pointers can cause ICs inside CSP packages to malfunction but continuous halogen light with irradiances ranging from 4 to 30 mW/cm2 (well below the Sun power density which is 1350 W/m2).
Raspberry Pi 2 can easily come close to a halogen lamp (http://www.homedepot.ca/catalog/halogen-bulbs/172459) that likely will cause it to either freeze or reset.
Image

Citation from:
"The opaque backside material on top of the CSP package blocks most of the light. However, it is a thin layer of material and does not block all light. Experiments using a bight halogen light show that an irradiance of 30 mW/cm2 directly on the top of the EFM32WG360F256-A-CSP81 will cause it to reset. In comparison, bright sunlight is about 100 mW/cm2. While your fingers may be opaque, a light that is bright enough will also pass through flesh.

The EFM32WG360F256-A-CSP81 is most sensitive to a light source at a 45° angle to the left side of the die, which is the side with the voltage regulator decouple (DEC_0) ball. A light level of about 4 mW/cm2 at this angle will cause a reset, which indicates that the edges of the die and the underside of the package are more susceptible to light than the top with the black coating."

jdb
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 2075
Joined: Thu Jul 11, 2013 2:37 pm

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 6:19 pm

You've confused your units.

1350W/m² is equivalent to 135mW/cm². You need an intensely bright halogen lamp (or a lesser wattage one at extremely close range) to match the incident flux that sunlight delivers.

That linked page is for a specific, complex integrated circuit (microcontroller). There are many factors that will affect the sensitivity of a CSP package, the effects of which have been investigated and categorised in this thread.

The fact that direct sunlight (even "flashing" it on and off with e.g. piece of card over the chip) doesn't noticeably upset the Pi implies that U16 is far less sensitive than the example in that PDF.
Rockets are loud.
https://astro-pi.org

User avatar
clive
Forum Moderator
Forum Moderator
Posts: 1013
Joined: Tue Feb 07, 2012 8:19 pm

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 6:59 pm

simpex wrote:Raspberry Pi 2 can easily come close to a halogen lamp (http://www.homedepot.ca/catalog/halogen-bulbs/172459) that likely will cause it to either freeze or reset...
You might want to test your data-surmised hypothesis. By shining one of those lights on a Pi 2 for example :) Empiricism FTW :ugeek:

User avatar
joan
Posts: 14196
Joined: Thu Jul 05, 2012 5:09 pm
Location: UK

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 8:30 pm

I skimmed the thread. Apologies if this has already been posted a dozen times.

http://www.bbc.co.uk/news/technology-31294745

ktb
Posts: 1380
Joined: Fri Dec 26, 2014 7:53 pm

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 8:39 pm

joan wrote:I skimmed the thread. Apologies if this has already been posted a dozen times.

http://www.bbc.co.uk/news/technology-31294745
Albert Einstein won a Nobel Prize for his discovery that if a light hits a component, it generates a charge. In the Pi this charge causes a roboot.
:) How do I make my Pi roboot?

simpex
Posts: 45
Joined: Tue Feb 03, 2015 8:40 am

Re: Why is the PI2 camera-shy ?

Mon Feb 09, 2015 8:41 pm

I made no confusion regarding the solar irradiance. I simply used both units, mW/cm2 and W/m2.
135 mW/cm2 (sunlight) is well above 4 mW/cm2 (the reset level)!

Secondly, http://www.silabs.com/Support%20Documen ... AN0878.pdf does not say the CSP packed circuit, reset there under halogen light, was also illuminated with sunlight. We simply do not know whether it is sensitive to sun rays or not.

A short evaluation shows that a 100 W tungsten incandescent ordinary bulb with a typical 2.5% efficiency (electric power to visible light power conversion) generates an irradiance of 6.25 mW/cm2 on a 20 cm x 20 cm surface.
However, the total radiated power, taking into account the IR an UV spectra emitted by the filament, is much higher and Integrated Circuits are also sensitive to this kinds of invisible light.
If you concentrate the light of the bulb on only 10 cm x 10 cm you already get 25 mW/cm2 (visible light).

I believe that a lot o light sources, that people use at home or in labs, can reset Raspberry Pi 2 if they are positioned less than 20 cm away from the board.

Return to “Troubleshooting”