Polyfuse reduces voltage input

[Mobius]

I had a previous issue where the ethernet seemed to connect but would not work when I brought up a browset in the graphical interface. I traced that to a particular USB power adapter I was using. Since that time I've actually measured working voltages with the two different adapters and found that the voltage on the output side of the poly fuse is significantly lower than the input. My input is right at 5 volts (with the known good adapter) but the operating voltage is 120 millivolts lower. If I plug in my wifi adapter, the operating voltage drops about 350 millivolts which explains the problems I'm seeing with getting wifi to work. I did some online research and apparently that is normal behavior for polyfuses. Sounds like adapters need to be 5.1 or 5.2 volts output to ensure an operating voltage around 5 volts and even more if a wifi adapter is used. I do have a powered USB hub on order which will hopefully fix the wifi issue but the loss across the polyfuse is still about 112 millivolts without peripherals. That means an in spec adpater may end up in an out of spec operating voltage. Did I miss this caveat somewhere?

[tomexx]

Hi,
I had the same issue. Tried different power supplies. Turned out to be a faulty F3 polyfuse.
My polyfuse measured 0.9 ohm, when it should normally measure 0-0.3 ohm.

I ordered new polyfuses from Newark at $0.30 each and replaced the F3. My TP1-TP2 voltage went from 4.6V (with the original fuse) to 5.1V (with the new fuse) using a 5.25V Adafruit power supply.

Tom

[thekingcat]

I am new to electronics, can you advise how I measure my polyfuse correctly to determine its reading?

I have attached the photo of how I currently have my multi-metre setup for it, not sure if it is correct?

I unplugged the Pi and read the F3 fuse and with the current settings of the MM i get a reading of 2.0.

Havent a clue what I am looking at really. This is the replacement Pi due to an issue with losing voltage over the F3 fuse, didnt really want to have to send it back for another if I can help it.


EDIT: Forgot to attach the photo of my multi metre.

[chipv12]

with meter in ohms as you have in the pic, put your leads together and it should read "0". however 0.1 or 0.2 in normal. if this is the case, then u have a bad fuse. I have found that my pi's (i just bought 2) wont carry much load. not even ethernet. prob because of this fuse. I havent read up on poly fuses yet but i understand that they will recover, but it takes days. 25 years in electronics. new to pi.

[thekingcat]

Set up as per the picture, the reading with the probes not touching is 1. with them touching it is 1.2(ish) - it fluctuates


But the voltage is definately dropping over the F3 fuse by at least 0.45v

Another return to where I bought it?

[newdendrite]

Is there a problem with just shorting the fuses to remove them from the circuit? I was under the impression that the newest revision eliminated them entirely.

Thanks,
Michael

[cnt]

Is there a problem with just shorting the fuses to remove them from the circuit? I was under the impression that the newest revision eliminated them entirely.

Thanks,
Michael

The USB poly fuses were removed. The main poly fuse was not. Mine (new as of 2 days ago; made in china, not UK) is currently running at 4.47v (TP1-TP2) with .56v dropped across the poly fuse. Its kind of annoying actually. I actually just took a 2pin header, bent the pins and soldered them across the poly fuse in my other r-pi. That way I can install a jumper when i'm using USB devices that need extra current. So far its worked fine, though I'm sure the foundation doesn't recommend doing what I've done.

[thekingcat]

Mine is also a Made in China model...

This is the 2nd unit I have had now, sent the first back to NewIt.co.uk and got this replacement.

The first one was dropping to 4.3v over the F3 polyfuse. This one is dropping to 4.5v.

I have tried just powering the Pi through the USB port, bypassing the F3, but my reading is only 4.4v backfed from my Belkin USB hub.

[Mobius]

As I mentioned in my original post, without the wireless adapter plugged into the USB I got about 120 millivolts drop across the polyfuse. That setup has a USB keyboard, USB mouse, ethernet cable, and HDMI. I'm thinking that's probably the normal range. Just measure the voltage prior to the polyfuse and then measure it at TP1 and TP2 to find the difference. The totally unacceptable voltage drop came when I unlpugged the mouse and plugged in the wireless adpater. I just got my powered USB hub so I'll try the wireless again. One poster is correct about the polyfuses being gone for the two USB ports on later model Pi's. Since I'm hooking up a powerd hub anyway, I probably won't short across them on my board. Just shorting across the main polyfuse is a bit risky. A better solution would be to put a 1 amp glass fuse inline with the plus side of your power source and then short across the polyfuse. I'll probably dig around in my junk box to see if I have the parts to do that on mine.

[shpank]

Hi!

I recently got my Raspi too and ran into exactly the same Problem.

The first two days, everything worked like a charm. No hangups whatsoever.

Today I got a new (bigger and faster) SD-card and a HDMI to DVI adapter cable and the problems began. I installed Raspbian on the SD and booted it. Still no problems. Then i connected a Mouse to my setup and things started going south. In the end it didn't even recognize the keyboard so I started updating the firmware and kernel.

Nothing got better. So I fired up my oscilloscope and measured the 3V3 and 5V rails. 3V3 seemed completely normal and 5V fluctuated up to 200mV depending on the AC adapter I used. After I flipped it over, i nearly burned my finger on the damn polyfuse! It was boiling hot!

So I measured across it and got 100 to 200mV of voltage drop across this sucker.

I then bridged it by clipping an alligator clip over it and reapplied power to the system.
No problems since then. It's running for about half an hour now and finally does what it's supposed to.

I'm thinking of just bridging it with a piece of wire. Then it will (if it ever will) fail spectacularly instead of quietly

This certainly dampened my first impression on the Raspi, but nonetheless I'm thrilled to finally start working on my first real projects with this little beast!

[newdendrite]

Thanks for the info on the polyfuses. I didn't realize that the newest version still had one in the circuit.

I have an early v1 RP, and I've been getting about 4.4v at the onboard testpoints from several power sources. All the supplies claimed 1A or better, but still showed about 4.4 v. The low voltage has caused a problem with at least one USB dongle operating improperly

This must be the same issue with the polyfuse that you are describing. If this is the case, then the specs from the foundation about suitable power sources is very misleading and need to be re-written.

Best regards,
Michael

[Lob0426]

I replaced F3 with a 1.1A unit. I had a lot less problems afterwards. The higher the amp rating on polyfuses, the lower the drop. Another way would be to bridge F3 then use an inline fuse holder on your power supply.

I just received some 512MB RasPii. They still will not power a USB HDD, because of the drop through F3. I am going to run them for a while to make sure they are good. Then I am going to modify one. I have a heavily modified ver1.0. It still will not power a HDD on a single port. Even with the 1.1A F3. It will work with a "Y" adapter in both ports, so I do not know what is going on there.

[JCT250]

Ive been having the same problem, 512mb board running Raspbian.
When sitting there powered on but not running any videos or audio etc Ill get approx 4.72v across the test points. When I load a video or anything that is slightly intensive the voltage across the test points will drop to as low as 4.62v and the board will browns out I'm assuming.

This got really annoying to start with as even installing Raspbian (which at points I'm guessing used most of the processing power) the board would crash out. Managed to solve it by hooking the 5v rail from an ATX power supply up to the pi through the GPIO and that gave me a nice 5.1v across the test points.

As well as narrowing the issue down by powering through the GPIO I've also tried 3 different 5v power supplies, one 2.0A, one 2.1A and one 1.5A. Ive also tried 3 different leads with each so Im PRETTY certain that this is not where the issue lies...

Across the Polyfuse I'm dropping around .25V which I'm guessing is the difference from the test points to make 5v.

I would prefer to not mess around with the Polyfuse and would also prefer to not have to power through the GPIO but I can't really see any other option? Its also rather annoying that when issues with the boards come up the general answer is "get a more powerful supply, that's your problem" Whilst I accept that yes in a number of cases it bay be the issue there is no denying that there is a deeper issue here.

There must be some form of solution?

[thekingcat]

Are you boards Made in China or Made in UK?

My problems were
Both with made in china boards.

My replacement made in UK board is 100% ok.

[Dilligaf]

All this talk of bridging polyfuses, just power through the gpio and keep your warranty intact. The issue with the excessive drop over the input polyfuse is a recent "enhancement" earlier boards didn't have problems with the fuse but had and still do have lots of problems caused by power supplies not supplying their rated output hence the "check your power supply" which is solid advise as low power probably causes 90% of the problems people are having with their PIs

[JCT250]

Yea I've got one of the made in china boards. Hoping to grab one of the UK made ones though if they ever make it down to NZ.

[cyglor]

Having stability woes that seem polyfuse related.
My tp2 to tp1 voltage is always lower than 4.7, from several power supplies, cords, etc.
Thankful to find this post after wading though many "it must be your PSU" posts.

I seem to have a quite a large voltage drop across the F3 polyfuse under load.

Bare board, power over micro usb
5.20V = tp2 to input of polyfuse F3
5.10V = tp2 to output of polyfuse F3

With HDMI + SD Card + Ethernet, power over micro usb (nothing connected to USB)
5.07V = tp2 to input of polyfuse F3
4.60V = tp2 to output of polyfuse F3

This is a huge drop: ~.47V

These #'s are with the adafruit supply power (5.25V 1A) - which has given me the most luck (not a surprise since my woes seem voltage related)

I have also tried several PSUs: 5V 1A, 5V 2.1A, 5V 700ma, etc

I'll check where the board was manufactured later today.

I have a GPIO breakout - perhaps i'll to power my pi that way, maybe I'll try to replace the F3.

[Mobius]

I started this post and since then have made the following mods to my first gen board:

- Shorted across the USB polyfuses
- Upgraded to a 5 volt, 2.5 amp power supply
- Added a standard fuse (1.25 amp fast blow) to bypass the input polyfuse

I now have a relatively stable 4.94 volts but I'm also thinking of cutting out the 5 volt zener diode (D17). I'm using a USB hub just because I have three USB devices and can now easily power it directly from the Pi with those polyfuses out of the way. I'm using a USB wireless device that drew down the power before the mods as well as a HDMI to VGA adapter that also draws a bit of current.

[cyglor]

Long delayed followup:

I replaced the bad F3 polyfuse with a new fuse and my power woes are gone.