Raspberry Pi doesn't stay powered up for long.


255 posts   Page 3 of 11   1, 2, 3, 4, 5, 6 ... 11
by gevero » Wed Oct 31, 2012 9:47 am
Hi guys

I have the same exact problem with an RS 512Mb raspberry Pi. The Pi stays on for about a minute, then randomly powers off, and I have to unplug the power to make it start again. I tried everything, 3 different PSU, two different sd cards, different operating systems, etc... None of these solved the problem. I also monitored, with some friend of mines, the voltage drop between tp1 and tp2, both with a multimeter and with a 300Mhz tektronix oscilloscope. I must say that in my case the voltage is nice and steady, first at 4.99V and then at 4.82V when the LAN kicks in. I see no voltage drop or spikes when the Pi suddenly turns off, just the voltage going back to 5.01V.

Only one thing solved the problem, i.e. a fan blowing cold air at the Pi. By doing this the Pi stays on without any problem, even with a usb keyboard and a usb wifi dongle attached. So, there must be somewhere an overheating problem, but i could not tell where.

I live in Italy, therefore for me returning the Pi is not an option, i might as well buy a new one. I wrote RS if i could get a replacement without sending back the old one, and I am still waiting for a reply. I guess that in the end, if they say I need to send back my Pi, I will buy a new one, but I guess I will go for farnell this time.
Posts: 6
Joined: Wed Oct 31, 2012 9:35 am
by gevero » Wed Oct 31, 2012 10:11 am
antoniomilano wrote:Hi,
I have the same problem.
I' ve bought a couple of RPi from RS and tried both with the same SD.
One of them randomly hangs after some minutes while the other is running fine from last thursday night.
I' ve phoned to the italian RS and they said me that the board can be easily substituted with no charges since it is covered by a one-year warranty, and there's no need to post the board back, but just pack it again and a RS man will pick my RPi at my home.
I read in this forum that the fuse could be the problem, I think that measuring the voltage drop across it could give some indication, and also to record the serial no. of the faulty RPi and to check if the production lot is the same.
I go to pack again my RPi.
Sincerely
Antonio



Hi Antonio

Could you please provide the phone for the italian RS?

Thanks

Giovanni
Posts: 6
Joined: Wed Oct 31, 2012 9:35 am
by strawberry » Wed Oct 31, 2012 10:45 am
Also have problems with the fuse. Before the fuse I have 5.02V coming from the 5V/1.2A adapter that came from RS. After the fuse (TP1-TP2) I get 4.3V but depending on CPU load, USB devices attached,... it sometimes goes below 3.8V and at that point the USB devices/LAN switch off/on/off/on :cry: (CPU keeps running).

So I'm thinking about removing the fuse or connecting the 5V power to the GPIO header.
Posts: 16
Joined: Tue Oct 30, 2012 10:44 am
by neo2810 » Wed Oct 31, 2012 8:25 pm
Just for the record, same problem with my 512Mb Pi delivered yesterday.....
Posts: 1
Joined: Wed Oct 31, 2012 8:22 pm
by lui_gough » Fri Nov 02, 2012 12:34 pm
Well, I couldn't wait any longer. RS didn't even bother a reply to me in about a week! I am mentally at my wits end - I've got projects to do, not time to sit and wait!

So I went ahead and desoldered F3, and replaced it with a plain wire. Guess what? Even that didn't solve the problem when using the Nexus 7's power adapter (which is now happy powering my Element14/Farnell sourced Pi serving my webpage at home) - up for 2 days until I took it down for a test.

This time, I went heftier. I replaced the power supply with my 20A lab benchtop switching supply, specced for 10mV ripple set at 5.03v plugged right into the GPIO header, watched the Pi boot up consuming average of about 0.5A ... and then suddenly black screen in raspi-config, and the draw dropped to suddenly dropped to 0.2A, ethernet lights off. Just like before. Voltage over the TP1-2 was rock solid of course.

So it looks like my Pi was "born" with a defect ... damn. And now I'm left with nothing.
Posts: 15
Joined: Fri Oct 26, 2012 5:31 am
by lui_gough » Fri Nov 02, 2012 1:10 pm
In my frustration, I have written a post at http://gough.ath.cx/?p=153 and I have also posted on RS Components' Facebook Page at http://www.facebook.com/RSComponents about this and hope to find some resolution.
Posts: 15
Joined: Fri Oct 26, 2012 5:31 am
by gevero » Fri Nov 02, 2012 2:33 pm
lui_gough wrote:In my frustration, I have written a post at http://gough.ath.cx/?p=153 and I have also posted on RS Components' Facebook Page at http://www.facebook.com/RSComponents about this and hope to find some resolution.


Hi

I have experienced the same exact situation. My diagnose was that i had somehow a temperature issue. I ran my Pi under cooling fan and, miracle, it worked indefinitely. Could you try this experiment for me, so that we may understand if we have a similar problem?

all the best

Giovanni
Posts: 6
Joined: Wed Oct 31, 2012 9:35 am
by lui_gough » Sat Nov 03, 2012 1:45 am
Hi Giovanni,

I tried your suggestion using an 80mm computer fan pointed at the top of the Pi, powered by external supply. Also, today is a cooler day (yesterday around 34 degrees C, today about 22 degrees C).

It seems that the Pi is able to work for a little longer (instead of 30 seconds - 1 minute, today it worked for about 30 -50 minutes), but is still not stable.

I suspect this may be suggesting there is a microcrack somewhere in the PCB or a bad joint somehow. I really don't want to build projects using it knowing that it would "crash" and need to be power cycled to recover.

- Gough
Posts: 15
Joined: Fri Oct 26, 2012 5:31 am
by Dilligaf » Sat Nov 03, 2012 2:33 am
If you've given up on the warranty you could always try baking it in the oven at 400 for 10 minutes, or heat it with a heat gun. This may have a detrimental affect on the polyfuse but you could always bypass it or power from the gpio. Baking/heating it should reflow the solder if it's a bad solder joint. You've really got nothing to lose if returning it isn't an option.
Posts: 283
Joined: Wed May 23, 2012 6:48 pm
by lui_gough » Sat Nov 03, 2012 2:51 am
Dilligaf wrote:If you've given up on the warranty you could always try baking it in the oven at 400 for 10 minutes, or heat it with a heat gun. This may have a detrimental affect on the polyfuse but you could always bypass it or power from the gpio. Baking/heating it should reflow the solder if it's a bad solder joint. You've really got nothing to lose if returning it isn't an option.


That's an idea ... I might give that a go after some thought, but I don't want to contaminate my own oven, so probably try a toaster oven, and I'm afraid it will probably destroy the connectors (so probably watch the duration somewhat). With any luck, components don't fall off the board as well and or the components on the bottom get stuck to something ... (i.e. what's the best orientation, need to prop it up, make sure oven isn't fan-forced or it might blow something off.)

Maybe later ...
Posts: 15
Joined: Fri Oct 26, 2012 5:31 am
by Dilligaf » Sat Nov 03, 2012 3:03 am
I've fixed plenty of boards from motherboards to video cards by baking them. Generally anything that cant be physically removed from the board will withstand the heat. I have never had anything fall off but thinking about it it could be a possibility. If it's worthless in it's current state then there's nothing to lose. Do a google search for circuit board baking, you'll be amazed.
Posts: 283
Joined: Wed May 23, 2012 6:48 pm
by itimpi » Sat Nov 03, 2012 9:54 am
I assume when you give 400 as the temperature to bake at I assume you mean Fahrenheit? I know if you do it at too high a temperature or too long you will find the plastic of the connectors starts to melt which does them no good at all! I know as I did it to one board and had to replace them :( Still it did seem to fix the underlying problem. :)
Posts: 1027
Joined: Sun Sep 25, 2011 11:44 am
Location: Potters Bar, United Kingdom
by lui_gough » Sun Nov 04, 2012 1:02 am
Thanks a lot Dilligaf! It worked! More info at my blog, served from Raspberry Pi:
http://gough.ath.cx/?p=160
Posts: 15
Joined: Fri Oct 26, 2012 5:31 am
by Dilligaf » Sun Nov 04, 2012 2:23 am
Just read your blog. You went a little bit higher tech then I ever bothered to (a simple oven thermometer is all I ever use) Glad it worked for you, incredibly it usually does.
Posts: 283
Joined: Wed May 23, 2012 6:48 pm
by drgeoff » Sun Nov 04, 2012 12:13 pm
This is beginning to look like a problem with the soldering of the BGA SoC to the PCB. Although baking may effect a cure, I would be concerned about the long term reliability of such a "repair".
Posts: 2045
Joined: Wed Jan 25, 2012 6:39 pm
by mahjongg » Sun Nov 04, 2012 12:36 pm
@Lui_gough.

You nailed it! seems you did a proper job. When I read something about "baking it at 400 degrees", and unless they meant Farenheit, I thought you were going to melt it into a puddle, or worse causing it to catch fire, but you were right, you need a bit over about half that temperature (260 to 270 degrees Celcius) for soldering! And you even used the right temperature ramps.

The trick is to first go up to 100 degrees, then wait a few seconds for all the water based solvents to cook off, then go up to 150 degrees slowly then to go up to about 260 degrees faster, then stop, and let it cool off naturally (not reinforced). The whole thing should last no more than 4 minutes and the peak above 200 degrees no more than ten seconds or so. Obviously your method doesn't completely do this as well as a real solder oven, but well enough it seems.

congratulations.

The only pity is that you needed to do this at all. :|
User avatar
Moderator
Moderator
Posts: 4506
Joined: Sun Mar 11, 2012 12:19 am
by lui_gough » Sun Nov 04, 2012 2:28 pm
mahjongg wrote:The only pity is that you needed to do this at all. :|


Thanks for the read and comment - I concur wholeheartedly, however, I guess the fact that it was ultimately "fixable" is much better than an entire waste. I've managed to actually get packages installed, configured, and the thing overclocked to 950Mhz too. A definite success compared to what it was before.

Of course, a new product from the manufacturer shouldn't need to be "baked" (unless it's food), but what makes me most frustrated is how RS has handled the whole thing. At least I could get on with my project, rather than further debug the hardware that was going to power it.

My main fear was destroying all the connectors, so I tried my best to try and keep to the reflow profile - knowing that underbaking is probably not going to solve the problem but won't break anything, whereas overbaking is a one-way-street. The headers, being through-hole were probably not intended for reflow anyway, so I'm not surprise the substrate melted a little. But the LAN jack, microUSB and HDMI were perfectly intact, which means I'm all gold as far as my project is concerned.

Lets hope my next Pi needs no further baking ...

[EDIT: Yes, the poster probably meant 400F because my oven doesn't go to 400C. 400F = 204C which is a bit low for lead-free, maybe for leaded it could do something as that starts melting at 183C...]
Posts: 15
Joined: Fri Oct 26, 2012 5:31 am
by Fray Bentos » Sun Nov 04, 2012 4:21 pm
Shame to hear you bought a load of USB transformers unnecessarily, can you return those or have you fubar'd the packaging. I'm sure u could ebay the ones u dont need.
How did you find the date of production? I can't find it on my Farnell board (this ones a 512mb).

I reckon your best bet is to find a 'local' number for RS rather than an 0845, that way you'll get to speak to a human, and you can get them to call you back when they've found out whats going on, and a solution - ie a return policy. Might be an idea to start by saying something along the lines of "I've got a faulty item, can you tell me how to send it in for replacement" rather than saying immediately that you don't know what the fault is, or whether there is a fault, leave that until you have got the procedure (for returns) off them.
This way you won't be fobbed off with the standard "we don't know if there is a fault please wait for us to get more complaints before we decide whether we're at fault or not", as long as you are fairly sure the product is at fault, as you and others on this forum seem to be, it should be returned and replaced.

They do testing at their end (RS and Farnell both) and RS even have videos of their test lab on Youtube. I think i read somewhere on here that the Farnell boards are also batch tested at the Sony factory (Pencoed) where they are made before packing.
Posts: 3
Joined: Fri Nov 02, 2012 2:33 pm
by Dilligaf » Sun Nov 04, 2012 8:09 pm
Sorry, I'm in the USA and unfortunately we choose not to participate in the metric system unlike the rest of the world. Typical pigheaded American stuff :) So yes I meant Fahrenheit. I don't think most ovens would reach 400 Celsius unless you set it to self cleaning which would likely vaporize the board.
Posts: 283
Joined: Wed May 23, 2012 6:48 pm
by lui_gough » Sun Nov 04, 2012 10:22 pm
drgeoff wrote:This is beginning to look like a problem with the soldering of the BGA SoC to the PCB. Although baking may effect a cure, I would be concerned about the long term reliability of such a "repair".


I agree, baking it may not necessarily make a long-term repair, especially with improper temperature profile which may make bad brittle joints, and reflowing solder which has already lost its flux and may oxidise further. But it's worth a shot in my case, and it did the trick, even if only in the short term. Better than nothing :).
Posts: 15
Joined: Fri Oct 26, 2012 5:31 am
by tass2001 » Mon Nov 05, 2012 12:18 am
Unfortunately I'm also experiencing the same issue with a Rev 2 RS Pi. I have an original Pi that I ordered from Farnell and It's doing just fine with the same peripherals that I've tried using the new Pi with.

I'm able to keep my Pi going indefinitely if I point a fan directly at it (in my case it's a floor fan on low speed). The funny part is that the SoC temp is only 30C with the fan pointing and 39C right before it shuts down (when the fan is not pointing at it).

If I don't use the fan, the board shuts down after ~3 minutes of operation. I'm very much considering reflowing the board as I ordered from RS in the UK and am located in the US and don't really feel like waiting a month+ for a replacement :/.
Posts: 3
Joined: Fri Jun 22, 2012 11:24 am
by carlmig » Mon Nov 05, 2012 12:17 pm
Just a heads up. I sent RS an email explaining the situation and they replied asking me to send the board back, and they would send me a new one within 10 days. So, fortunately, that's not too long, at least for someone who waited a few months!!
Packaging it right now.
Posts: 3
Joined: Thu Oct 25, 2012 10:46 pm
by dastasha » Wed Nov 07, 2012 8:43 am
Image

My solution. At least its stable but I have to replace the ice brick every hour or less.
Just a bit disappointed after 3 months wait. Using a 3200mA PSU and got 4.83V across TP1/TP2.
Tried several PSUs, hubs etc.
I bridged the polyfuse F3 to take that out of the equation. Its an RS revision 2 board.
I wont return it as I'm in Australia and I had poor communication from RS. It was too hard to get my hands on to start with.
I have ordered two more from au.element14 with a two week wait this time.
Hopefully I'll get one that works at room temperature and then I'll consider doing something nasty with this one - like reflowing it the kitchen oven
Posts: 12
Joined: Tue Nov 06, 2012 6:53 am
by M4T VW » Thu Nov 08, 2012 8:26 pm
Wow, Thats awesome! :lol:

I had a reply from RS yesterday, 10 days also so mine is going in the post tomorrow.
Not like i will miss it as i didnt get the chance to do anything with it anyway!
Posts: 3
Joined: Sat Oct 27, 2012 4:11 pm
by wallarug » Sun Nov 11, 2012 5:48 am
I have a different problem with a 512mb RPi:
I have discovered that there is a problem with the USB circuity. When-ever the device is touched on the USB controller (IC3), the device resets every thing attached to the USB circuit. When any device is removed or plugged into the device, the whole Raspberry Pi resets.


Could this be a contributing factor to the ' temperature' issues being experienced here?

We should also be telling each other what batch we are part of...
Mine is:
12/9 UK build

What date are yours?
Youtube Channel: http://www.youtube.com/user/CMDenterprises
Project page: http://code.google.com/p/cmd-robot/
~~ Now selling Apps on the Google Play, WP and Apple App Stores. ~~
Posts: 427
Joined: Mon May 14, 2012 8:21 am