Raspberry Pi: Crash, then solid red light

[scotoma]

Hi, this is my first post. I was encouraged by adafruit technical support to post my query to the forums as well as through email to them. Here is a summary of my problem:

My raspberry pi recently stopped working. It worked for about 1 week with two different SD cards using two different linux distros (wheezy and occidentalis), but last Friday it just turned off randomly. I wasn't even typing or updating or doing anything at the time, though I had just (5 minutes earlier) finished installing the python-imaging software per these (http://learn.adafruit.com/light-paintin ... i/software) instructions. Now when I plug it in, I get a red LED and nothing else. I don't get any video output and I don't see the raspberry pi on the network. I have tried unplugging everything except for the power, and I still just get a solid red LED. I've tried using both SD cards (which again, worked previously). I've also tried using two different power sources. One is a 10A 5V power supply that I am using through the +5V and GND GPIO pins, and the other is a 5V microusb drive. The 10A 5V supply is outputting a steady 4.95V-5V according to my DMM. Both worked previously to power the RPi. I also gave it one day rest, and then 3 days rest, since I read that sometimes RPi "timer trip fuses" take a while to reset if there is a surge or something (this is unlikely as all other devices connected to the surge protector were fine). The SD card pins look fine on close inspection (and the shutoff occurred during use and the RPi wasn't physically moving when it shut off). The shut off did occur after a longer period of use then I had put it through before (roughly 30 minutes). It is in a Ninja Pibow case and it was 65 degrees in my room, so I didn't think overheating would be a big issue, but I don't know. I was just wondering if there was anything else I can do or if it's time to get a replacement. Thank you for your help!

[adafruit_support_mike]

The red LED indicates a power connection, but if you don't see anything from the green LED it means the bootloader isn't executing. It sounds like you've checked most of the usual sources of trouble, but let's dig into it in a little more detail.

First, grab a multimeter and look for the two pads on the front of the PCB marked TP1 and TP2. TP1 should be below the Raspberry Pi logo, and TP2 should be next to the NTSC video output jack. Measure the voltage between those to make sure the board is getting 5v or something close to it. TP1 is connected to VCC, TP2 is connected to GND.

Second, flip the board over and find the TP2 pad on the opposite side (it should be directly below the one on top). Then check the voltage at either side of the polyfuse F3 (just below the USB connector, usually pale green, about 1/8" x 1/4"). If the difference between those two voltages is more than about 300mV, your fuse may be dead.

[tldr]

Second, flip the board over and find the TP2 pad on the opposite side (it should be directly below the one on top). Then check the voltage at either side of the polyfuse F3 (just below the USB connector, usually pale green, about 1/8" x 1/4"). If the difference between those two voltages is more than about 300mV, your fuse may be dead.

sorry to crash the party, but i, too, have a dead pi. in addition to the fuse quickly getting painfully hot to the touch, i measure close to five volts on the end of the fuse nearer the board edge and under two volts at the other end. the voltage between tp1 and tp2 is about 2.2v. the thing doesn't really weigh enough to make a decent paper weight and i'd go broke destroying enough of them to tile the bathroom. is the fuse replaceable? got a part number?

[adafruit_support_mike]

Yep.. that's a dead fuse.

There's no special magic if you're already comfortable doing rework on SMT components. I don't have a part number off the top of my head because I can't find anything that big with specs low enough to match what's listed on the schematic. Officially, it's a 6v 750mA polyfuse with a 1.1A trip current. Thing is, the one I just measured is an 1812 package, and I can't find anything at Mouser or Digikey in 1812 with a holding current less than 1.1A.

The Littelfuse 1812L110 is probably the closest match you'll find without going to Element14 and asking for the specific device.

[tldr]

how about i just solder a penny across the pads?

[adafruit_support_mike]

It'd be a bit tight, but the theory is sound. People have replaced the fuse with a jumper.

The part's main job is to kill spikes from the supply that could fry the CPU or other delicate/expensive/hard-to-solder components.. basically the job a low-dropout regulator would do. At a guess, I'd say they chose the polyfuse for reasons of cost, space, and simplicity.

If you know you have a nice, quiet 5v supply, you can probably live without the fuse.

[scotoma]

Thanks mstone,

I just measured the voltages you suggested:

TP1-TP2: 3.98V
TP2-edge side polyfuse F3: 3.98V
TP2-inner side polyfuse F3: 3.98V

It seems like the fuse may be ok, though it's strange that I'm now getting 4.5V across the power supply output wires and 4V across TP1 and TP2. Also, the red LED seems to be dimmer now than it was before. Any other ideas? Thanks again.

The red LED indicates a power connection, but if you don't see anything from the green LED it means the bootloader isn't executing. It sounds like you've checked most of the usual sources of trouble, but let's dig into it in a little more detail.

First, grab a multimeter and look for the two pads on the front of the PCB marked TP1 and TP2. TP1 should be below the Raspberry Pi logo, and TP2 should be next to the NTSC video output jack. Measure the voltage between those to make sure the board is getting 5v or something close to it. TP1 is connected to VCC, TP2 is connected to GND.

Second, flip the board over and find the TP2 pad on the opposite side (it should be directly below the one on top). Then check the voltage at either side of the polyfuse F3 (just below the USB connector, usually pale green, about 1/8" x 1/4"). If the difference between those two voltages is more than about 300mV, your fuse may be dead.

[adafruit_support_mike]

Hmm.. that explains why it won't boot. The supply voltage needs to be between 4.75v and 5.25v. The searching that I've done suggests that the microprocessor will boot down to about 4.5v, but performance will be marginal. It will have trouble recognizing USB devices like the keyboard and mouse at around 4.6v, and the system will lock up during CPU-intensive operations below 4.4v.

The lack of voltage drop across F3 would be unusual in normal operating conditions.. I just measured mine and saw about 150mV across it, with a TP1-TP2 voltage of 4.89v. The fact that you aren't seeing any drop at all probably means the current flowing through the thing is small.. maybe a milliamp. That would make sense it the power system was refusing to start anything up.

You mentioned getting a 4.5v reading from the power supply when it was disconnected. The USB spec voltage is the same as the RasPi's requirements: 4.75v to 5.25v. If you're seeing less than that, your supply has probably gone bad. It's possible that it was rated to supply 500mA (the maximum load for a single USB device) and the RasPi was drawing a bit more than that.. the rated load is 700mA.

You mentioned using a 5v/10A supply to power the RasPi through the GPIO pins, which is a sensitive issue in its own right. The internal power rails don't provide any current protection for D17 (a transient voltage supression diode), so you need to be very sure your supply is smooth and spike-free.

Do you have a powered USB hub that you could use as an alternate through-the-USB-port supply?