Raspberry Pi Forums

For those who think you have a power supply problem, I found one on digikey.ca with a slightly higher output voltage and it is a 10W. The Digikey part number is 993-1196-ND.

Make sure you order this part as they have three different numbers for what appears to be the same thing.

The difference to be concerned about is one version has the slightly higher output voltage.

Looking at the data sheet it is the "D" model which "compensates for DC drops in the USB cable by 400mV at full load".

If you don't have access to Digikey try searching as I am sure there are other power supply manufactures who adjust the output voltage to compensate for USB cable voltage drops.

The answer is NOT to compensate for Volt Drop in the cable by raising the source voltage. This still leaves the problem that the load voltage varies with load current.

The answer is to use a heavier guage cable with less volt drop.

The biggest issue is not the cable loss, it's the loss across the polyfuse. That loss increases as the current load of the peripherals increases. It can take an in-spec voltage input and reduce it to out-of-spec.

Again, the solution is NOT to raise the input Voltage but to fix the volt drop across the Polyfuse.

It would appear from other posts here, that there is a manufacturing problem with Polyfuse resistance on the Chinese manufactured Rev 2 boards. I suspect that the Polyfuse F3 is being overheated during the Solder re-flow process. ... but I am only speculating.

Wheel_nut wrote:It would appear from other posts here, that there is a manufacturing problem with Polyfuse resistance on the Chinese manufactured Rev 2 boards.

Really? I do not think I have ever seen a case that was proven to be caused by a faulty polyfuse.

Wheel_nut wrote:http://www.raspberrypi.org/phpBB3/viewt ... 26&t=23901

Excessive drop on F3 fixed by replacing the board, not by replacing F3. No proof that F3 was faulty, as opposed to a fault elsewhere on the board and F3 tripping as intended.

jojopi wrote:

Wheel_nut wrote:http://www.raspberrypi.org/phpBB3/viewt ... 26&t=23901

Excessive drop on F3 fixed by replacing the board, not by replacing F3. No proof that F3 was faulty, as opposed to a fault elsewhere on the board and F3 tripping as intended.

You can measure the F3 resistance is way out of spec with no power through the board and people have replaced F3 and found it to correct their power problems as well.

Someone needs to specify exactly what constitutes "excessive drop". Polyfuses, according to what I have read, are by nature a device with an internal resistance which means some sort of voltage loss. I apparently have a version 1.0 board even though I bought it in September (here in the states). I say that because it has polyfuses in the USB outputs as well and my understanding is that 1.1 replaced those with shorting wires and 2.0 did away with them altogether. As for me, I'm going to replace the input polyfuse with a standard fuse. End of voltage drop.

Mobius wrote:Someone needs to specify exactly what constitutes "excessive drop". Polyfuses, according to what I have read, are by nature a device with an internal resistance which means some sort of voltage loss. I apparently have a version 1.0 board even though I bought it in September (here in the states). I say that because it has polyfuses in the USB outputs as well and my understanding is that 1.1 replaced those with shorting wires and 2.0 did away with them altogether. As for me, I'm going to replace the input polyfuse with a standard fuse. End of voltage drop.

The below post with a link to the datasheet for the fuse also shows what the in spec resistance should translate to as far as voltage drop goes, short version: not more than 0.3v

http://www.raspberrypi.org/phpBB3/viewt ... 26#p222526

bgirardot wrote:

Mobius wrote:Someone needs to specify exactly what constitutes "excessive drop". Polyfuses, according to what I have read, are by nature a device with an internal resistance which means some sort of voltage loss. I apparently have a version 1.0 board even though I bought it in September (here in the states). I say that because it has polyfuses in the USB outputs as well and my understanding is that 1.1 replaced those with shorting wires and 2.0 did away with them altogether. As for me, I'm going to replace the input polyfuse with a standard fuse. End of voltage drop.

The below post with a link to the datasheet for the fuse also shows what the in spec resistance should translate to as far as voltage drop goes, short version: not more than 0.3v

http://www.raspberrypi.org/phpBB3/viewt ... 26#p222526

Thanks. That's about what I'm seeing but the caveat is that the more current drawn, the larger the voltage drop. Even at 300 millivolts, that puts the operating voltage of the Pi outside of the 5% range if you start with a 5 volt supply. It's even worse at the USB ports if, like me, you have an early model that has the USB polyfuses.

If you have a Rev1 with USB Polyfuses, you should resort to Powering the RPi on the GPIO Header.

You will still have the problem of the USB Polyfuses limiting the current and consequently Voltage to the peripherals. Only solution there is to bridge the Polyfuses.

What if you don't have a revision 1 board but one of the new ones purchased a week ago and are having the same issues? Powering through the GPIO whilst being an acceptable solution is not something everyone should have to resort to if there is an issue with the main polyfuse

If you are having problems with volt drop across the F3 Polyfuse on a Rev2 board then your best solution is to "backfeed" it from a Powered USB Hub fed fron a 5V, 2A Power Supply.

Is your Board made in China by any chance?

I'll bet the foundation are cursing the day they decided on micro-usb as a power solution......

Feed the damn thing with 7-9V and regulate it on board. Half a volt or more drop in the cable, polyfuse or anywhere else is then irrelevant. It wouldn't have been that expensive to supply the Pi with a power supply. Or do it ala Arduino and make it so you can feed it anywhere between 7 and 20 volts so practically any power supply with the right connector will work.

Yep ive got one of the made in China ones.

Correct me if I'm wrong but isn't backfeeding it through the USB port effectively the same as powering it through the GPIO?