High Voltage drop across polyfuse causes FREEZE


3 posts
by skynet » Sat Jun 23, 2012 2:44 pm
As my Pi (from RS) never ran more than 2 minutes before freezing :( I did some research and came across the power supply topic.
I started measuring the voltage (in idle mode, no cpu load) between TP1-TP2 and the voltage at the USB-plug
on a bare Pi then with a Keyboard then with Mouse and Keyboard and then with everything (hdmi+network) connected
In short: The power supply always delivered 5.04 V
BUT the measurement TP1-TP2 fell from 4.9 V (bare) down to 4.8 V
and at that level (no cpu load) my Pi becomes unstable at ANY activity (like access to the sd-card, etc.)

looking at the schematic (thanks for supplying, this helped a lot) I found the missing 200 mV across the polyfuse F3

this brings me to my question: Is the Polyfuse out of specs or is this an expected Voltage drop?

If yes, the 5 V are not enough for a stable operation.
Posts: 4
Joined: Wed Nov 16, 2011 11:15 am
by jbeale » Sat Jun 23, 2012 3:43 pm
Sounds like your input polyfuse is about 0.4 ohms (500 mA * 0.4 ohms = 200 mV) which is on the high side for resistance, I've seen 0.2 to 0.3 ohms mentioned. Nominally 4.8V on the board should be ok, actually the Pi itself has been reported to work as low as 4.6 V, but some USB peripherals can be more picky, especially if you are trying to use anything wireless. Also, if there are brief downward spikes from brief current pulses, that could cause the stability issues. You'd need a scope to check that; a multimeter does not show fast spikes.

There is also the possibility of some software issues, SD card issues, etc. as you will find documented at http://elinux.org/index.php?title=R-Pi_ ... did=141722
User avatar
Posts: 2076
Joined: Tue Nov 22, 2011 11:51 pm
by skynet » Sat Jun 23, 2012 4:38 pm
well, in theory it migth be some other Hardware ...
but I also tried ssh only (nothing on the USB Ports) with 5 different SD cards (from the whitelist).
but also here it crashed even idle if you were willing to wait 5 minutes, just watching the cursor blinking....

all my observations led to the conclusion: it does not matter what you have connected or what you're doing, it will freeze in a few minutes

I just picked randomly a datasheet for a comparable polyfuse (1A) which is talking of 0.04 ohms (no load) therefore 140 mV with nothing connected and idle looks like a QA issue in manufacturing. But it would be necessary to get data from other users on the voltage drop on F3

Now by ensuring 5V between TP1-TP2 (not on the micro usb jack) it looks stable as far as I see ...
Posts: 4
Joined: Wed Nov 16, 2011 11:15 am