Brand new rPi + bad F3 polyfuse


57 posts   Page 1 of 3   1, 2, 3
by thogue » Tue Oct 02, 2012 4:07 pm
By brand new, I mean spanking new, I just pulled it out of the static bag, plugged in an SD card.

Difference between both sides of F3 is > .3volts

my original pi has this problem and I thought this was due to me blowing the fuse..... now I suspect this was NOT me.... as these both came from the same place.

Also, my box looked like it was under a vise in the raid a few times.........
Posts: 131
Joined: Wed Sep 19, 2012 2:16 am
by RaTTuS » Tue Oct 02, 2012 4:12 pm
post pictures of box etx
shwo us exactly what your donig
how are you powering the RPI - show us urls of all comps
what is the power tp1 tp2
1QC43qbL5FySu2Pi51vGqKqxy3UiJgukSX - Prosliver FTW
"That's not right, the badgers have moved the goalposts."
User avatar
Posts: 4159
Joined: Tue Nov 29, 2011 11:12 am
Location: North West UK
by thogue » Tue Oct 02, 2012 4:23 pm
I am powering my Pi via a usb hub,

but no matter how i power it i see the polyfuses as bad, with my original pi I tested with 5+ adapters and ended up making an adapter via a PSU that had higher voltage (5.21). This increases voltage overall but does not solve the issue of the F3 fuse dropping all that power.

Currently

4.53 from first plug slowly rises to 4.6.

in front of the polyfuse measures 4.9, on the USB hub im using I get 5.1 on either side of its poly and 5.1 at the connector.

here are some pictures of the most recent pi I received moments ago

Here are pictures of my box I received, as you can see it was, there are more than few items of concern. I just heard back from a CS rep who was pretty appalled at the state of the box.
http://imgur.com/Y1Kbk,81lSQ,Nwdwp,9UqN ... a,VbNND#13

I am sure if I pull out my "better" PSU that has higher voltage that things will read better on T1 T2 but I will still have a drop of .3v or great on F3
Posts: 131
Joined: Wed Sep 19, 2012 2:16 am
by thogue » Tue Oct 02, 2012 4:29 pm
4.95 in front 4.6 in back of F3 with the USB hub (belkin rated 2 amps measures 5.1 at the connector)


with my 2.5 amp 5v home made psu, (which reads 5.21 no loads, 5.17 with the pi on it)
I am seeing 5,15 in front of f3 and 4.79 behind it (Tp1,tp2 also read 4.79)


As far as I see both these Pis my brandnew and my several week old one show the same things.

Note that even with the better PSU once i plug anything into the USB ports it drops into the unacceptable range.
Posts: 131
Joined: Wed Sep 19, 2012 2:16 am
by mahjongg » Tue Oct 02, 2012 4:32 pm
annoying. :(

The polyfuse that is used has an "off factory" resistance of about 0.11 Ohm, and after soldering becomes something like 0.2 Ohm, (because it becomes hot from soldering). Tyco the manufacturer claims that after (repeated) blowing and cooling off (say recovering, as it can take days to "cool off") the typical end resistance will be near 0.45 Ohm.

a thought experiment.
Lets say that you indeed have a "bad" fuse, (with a 0.45 Ohm resistance) then what amount of current would be necessary to cause a 0.3 Volt drop?
It would be I = V/R or 0.3/0.45 = 660 mA, which indeed sounds reasonable!

What kind of current is needed for 0.3 V to drop across the polyfuse, while it is still "fresh", that would be I = V/R or 0.3/0.2 = 1.5A which is obviously way more than a PI would normally draw.

A few questions:
* Did you indeed power the PI with nothing else plugged in, only the SD-card?
* Did perhaps the SD-card became warm after your test?
* What is the open voltage of your power supply?
its possible that it will cause the over-voltage protection on your PI to trigger, which will cause a massive current draw, and then 1.5A is indeed quite possible!

The state of the package should not affect the state of the polyfuse, so that is a bit of an irrelevant comment, unless its meant as a general negative comment about the distributor :lol: .

Unless you are unlucky enough to have received a "refurbished" PI that has a previously blow fuse, the logical explanation would be that something is drawing in excess of 1A, and that something would be something that both PI's share, the SD-card.

If you are extremely sure the PI had a blow fuse when you got it, I would return it again, and next time I would measure the resistance of the fuse before connecting anything! A resistance of almost half an ohm should be easily measurable.
User avatar
Moderator
Moderator
Posts: 4494
Joined: Sun Mar 11, 2012 12:19 am
by thogue » Tue Oct 02, 2012 4:42 pm
Thanks for your reply.

As stated, this was a brand new pi, Cold sd card, and I have NEVER plugged anything into it.
err Edit: the sd card is still cold, but I have not kept the pi on for more than a few minutes.





When I posted this the only powersupply it had seen is the Belkin Hub.

and yes the shipping this was a general comment...... figured Id shared... never know if water, smashing, squeezing is relevant

I have never plugged my pi into anything that reads over 5.25 with no load. edit: I read a lot of your posts about power issues previously and this was something you stressed heavily

Is it possible I received TWO pis with bad Fuses??

Also just tried with an Apple iphone charger (1amp usb) and I am seeing the same drop in voltage.
2nd edit:
The FIRST thing I did with my pi was plug it in and test the fuse. as the one I received a week or so ago fuse was blown and I could not pinpoint how that happened..

. I wanted to see the voltage on a good pi. ps this cd card is from my other pi, fully updated rasparian never noticed it get hot before. I have noticed the USB ports on that Pi get hot before though

EDIT4: also I have had some long edits on this post and my second post. just fyi. hope information doesnt become scattered and I become hard to understand.
Posts: 131
Joined: Wed Sep 19, 2012 2:16 am
by thogue » Tue Oct 02, 2012 6:17 pm
The SD card is a Wintec Filemate 16GB Class 10 same from the wiki list tests.

Plugging in the rPi WITHOUT an SD card I see .08 difference across f3.

However with micro 4gb sandisk class 2, back to >.3v diffence , just tried another 1GB card that is has moebius, same thing >.3v, edit also just tried a 512 card.


Is it really possible that all these cards are causing a high current draw? I had this problem with my original pi before I started using the 16gb filemate as my primary sd card.

hmmmm...frustrating to be going through this with 2 rpi's .... I was about to just bridge over F3 on my original PI but would really like to get to the bottom of the issue.

I had voltage problems on my original Pi when the only PSU I had used was the Apple1amp Charger, I only started experiencing with HUBS, other usb chargers when I realized I was having a voltage issue.
Posts: 131
Joined: Wed Sep 19, 2012 2:16 am
by thogue » Tue Oct 02, 2012 6:36 pm
was reading some other power threads, So i checked the fuses with the multimeter but this time checking the resistance vs checking the voltage difference.

my new pi is measureing 1.3~ ohms

my older pi is measuring 1.0~ ohms

Doesnt this mean that both my polys are 100 percent or better??

Im now confused as to why I am seeing the >.3v difference when I am running on an SD card (any of my 4 sd cards at my desk currently, SanDisk 1GB,512MB, Wintec 16GB, SanDisk 4GB micro)
Posts: 131
Joined: Wed Sep 19, 2012 2:16 am
by thogue » Tue Oct 02, 2012 7:34 pm
From my understanding the only thing that could cause the fuse to show the voltage drop is it failing or being overloaded ..... its just not making sense to me that all my sd cards could cause a draw of 1.5 amps even if im using a power adapter thats only capable of 1amp (apple iphone psu brick)
Posts: 131
Joined: Wed Sep 19, 2012 2:16 am
by mahjongg » Tue Oct 02, 2012 10:42 pm
thogue wrote:was reading some other power threads, So i checked the fuses with the multimeter but this time checking the resistance vs checking the voltage difference.

my new pi is measureing 1.3~ ohms

my older pi is measuring 1.0~ ohms

Doesnt this mean that both my polys are 100 percent or better??

Im now confused as to why I am seeing the >.3v difference when I am running on an SD card (any of my 4 sd cards at my desk currently, SanDisk 1GB,512MB, Wintec 16GB, SanDisk 4GB micro)


A normal polyfuse should have a resistance of about 0.2 Ohm, so your measurements of 1.0 and 1.3 Ohm are way off, five or even six times more than normal. In fact Tyco claims that even after several times having been blown, but given enough time to recover (days) the fuse will return to less than half an ohm (0.45 Ohm) !

If your fuses are getting warm while doing this, it means they are starting to blow already, which might explain the high resistance.

Are you sure you are measuring right? Its difficult to measure sub one ohm resistances, even with a good Fluke, and with excellent probes you will get resistances of 0.3 Ohm, just connecting the probes together, but if you hold the probes weakly together the resistance may go up to over one ohm! Also you need a fresh battery in the meter, as an old battery may lose voltage after a few seconds, and that may influence the measurement.

The way to do it is to press both probes to the same side of the fuse, wait a second or so for the reading to stabilize then write down the reading, then repeat but with one probe on the other side of the fuse. Subtract the result of the first measurement, and you will get the real value! Before doing the measurement be sure to have had the PI off for several minutes so all capacitor charges in it are discharged!
User avatar
Moderator
Moderator
Posts: 4494
Joined: Sun Mar 11, 2012 12:19 am
by thogue » Wed Oct 03, 2012 4:13 am
Deleted a long reply as I feel my replies was missing valuable info while saying way too much :roll:

I have a cheapo multimeter(edit:Reliance Controls DT830) reads .1 resistance when i touch the probes like you said it should. I have it on the lowest 200ohm setting.

It settles to .1 if I read both probes one side, and 1.1 if I read fuse fuse, probe either side(previously I was just doing the second measurement)

Thanks for all your help!!

... still sounds like I could be doing something very wrong....
Posts: 131
Joined: Wed Sep 19, 2012 2:16 am
by psergiu » Wed Oct 03, 2012 9:57 am
Will a cooled polyfuse recover quicker ?
In theory, it should ...
http://www.te.com/content/dam/te/global ... entals.pdf
Has anyone tried putting his RPi in the freezer for 5 minutes then measuring the resistance of the polyfuses afterwards ? I would but i have a el-cheapo multimeter.
User avatar
Posts: 212
Joined: Mon Nov 07, 2011 8:36 am
Location: Bucharest, Romania
by mahjongg » Wed Oct 03, 2012 12:36 pm
well at least it seems your meter can do fractional values (it has a "floating point"). Some very cheap multimeters only have three digits, without a point, and have a lowest resistance setting of 2000 Ohm, obviously these are indeed worthless for measuring the polyfuse. But it sounds as if your polyfuse really is 1.0 Ohm.

Its possible that polyfuses will recover faster when in a cold place, they consist of a nonconductive polymere, with conductive carbon particles interspersed if the ends of the carbon particles reconnect the polyfuse becomes conductive again. But I think the descriptive "cooling" is a bit wrong for what is really happening, obviously such a small mass will cool off immediately, but the proces of returning to a nominally normal value (after blowing) will take at least an hour, and furthering lowering continues over the time of days, weeks, even months!

A typical new polyfuse (soldered to the board) will be 0.2 Ohm, and one that has blown, but has had an hour to recover will typically be 0.45 Ohm (according to the datasheet of the device).

That means that for a typical current consumption of say 500mA a good fuse will have about 0.2 x 0.5 = 0.1 Volt across it so a 5.0 Volt PSU will only get 5.0 - 0.1 = 4.9 volt to the PI, well within specs. One hour after removing the power to a shorted/overloaded PI the drop will be 0.45 x 0.5 = 0.225 Volt over it, and so the same 5.0 Volt PSU will be able to deliver just 4.775 Volt to the PI, which means it is very very marginal, as the minimum will be 4.75 V.

At least, for many USB peripherals that will be the case, there are many that will do with 4.4Volt, but not all! Certainly not if they use more than 100mA (that is why the USB polyfuses were such a problem). But the PI itself is reasonably tolerant for lower voltages, that is because most of the PI is actually powered with 3.3 Volt, and the drop down regulator will still make 3.3 Volt, even when its input only get 4 Volt. That is why a PI may boot without any USB devices attached, but will crash when they are attached, not working and generating so many errors that the PI itself crashes. Lastly the PI does use some power directly from 5V, because the SoC uses a built in switching regulator to make its own core voltage, which is very low but programmable (as in "overvolting"). I don't think that this switcher is very intolerant for lower than 5V power, because the difference between its input and output voltage will still be quite large even if its input voltage drops half a volt or more most such switchers will still continue to operate.
User avatar
Moderator
Moderator
Posts: 4494
Joined: Sun Mar 11, 2012 12:19 am
by thogue » Wed Oct 03, 2012 4:31 pm
Well, I want to say I must of been shipped a bad poly because blaming my sd card for pulling so much power is just seeming far fetched to me, i could be so so wrong....

Guess I will contact them and RMA. I will be testing my next fuse before it sees power at all now that I feel I have a full understanding of the fuses role and how to test it.

The one I tested last night was only on for less than 60 seconds, 3x times and was off for 7+ hours when I tested the fuse again last night.... My readings are showing it didn't recover at all.... I guess I shall see if there is any recovery later today.

As for the Pi tolerating lower voltages it seems to be very evident but once it drops below 4.6 it seems it starts to crawl before showing any issues.

Going to RMA and order another pi or two on top of that, maybe ill have some luck and get a version two and read a good fuse and feel better about myself....
Posts: 131
Joined: Wed Sep 19, 2012 2:16 am
by thogue » Wed Oct 03, 2012 6:04 pm
Still zero recovery :-/

hmmm


However my older Pi which was measureing 1.3 across now reads 1.0 across ( .1 probes on the same side). This pi was running last night full speed, playing video, doing bluetooth/wireless testing.
Posts: 131
Joined: Wed Sep 19, 2012 2:16 am
by thogue » Wed Oct 03, 2012 6:49 pm
As I am waiting for an rma and temperature seems to be a sensitive thing.


I sealed the new pi in question inside its static bag, did my best to remove all air, and sealed it air tight. then I put this inside another enclosure that was air tight, threw it in the freezer for 30 minutes.

I just took it out, let it rest for 3 minutes inside everything, then removed the first enclosure letting it rest again for another 3 minutes, as to not create any massive temp change too fast and cause condensation.

the metal on my pi is cold, the F3 fuse feels cold....

still the same thing results when reading the poly.... no recovery.... :-/ This pi has been off for about 24 hours now.
Posts: 131
Joined: Wed Sep 19, 2012 2:16 am
by mahjongg » Thu Oct 04, 2012 8:52 am
psergiu wrote:Will a cooled polyfuse recover quicker ?
In theory, it should ...
http://www.te.com/content/dam/te/global ... entals.pdf
Has anyone tried putting his RPi in the freezer for 5 minutes then measuring the resistance of the polyfuses afterwards ? I would but i have a el-cheapo multimeter.

Unfortunately the quoted article doesn't have any info on whether cooling a blown fuse will help it to recover faster! The physical processes involved in "blowing" are very complex. Whether cooling will help in the recovery process is frankly in doubt.
User avatar
Moderator
Moderator
Posts: 4494
Joined: Sun Mar 11, 2012 12:19 am
by thsBavR10 » Thu Oct 04, 2012 11:36 am
Hello,

Today i've got a new Pi from RS (ordered june-21).
First thing i did was measuring the resistance of F3 (without starting the Pi!),
comparing it to the value for the other Pi (from Farnell).
F3 (Farnell): reading 0.2 - 0.3 Ohm
F3 (RS, fresh!): reading 0.9 Ohm
Both leads shorted: 0.1 Ohm

This leads me to the conclusion, that there are massive differences in resistance of
fresh polyfuses, maybe there are different processes in producing the Pi's.

ps: sorry for my bad english - i hope you are unterstanding the meaning ;-)
Posts: 216
Joined: Sat Jul 21, 2012 3:11 pm
by scorp » Thu Oct 04, 2012 12:02 pm
thsBavR10 wrote:This leads me to the conclusion, that there are massive differences in resistance of
fresh polyfuses, maybe there are different processes in producing the Pi's.

it is not very difficult to mess up with reflow process
Posts: 96
Joined: Tue May 29, 2012 6:21 pm
by thogue » Thu Oct 04, 2012 12:56 pm
At least I'm not the only one thinking something is up... Both these came from same place, same manufacturer date, believe the serial is about 1000 off from the other one... As its hard to believe my sd card could be drawing the kind of power to blow the fuse esp it didn't even get warm.

Going to check the "new" pi again shortly.EDIT: results at bottom

Maybe we will discover something new with time.


and yeah I figured the freezer thing would not be fruitful effort but I figured since I did not feel I had anything to lose and seen it mentioned several times on the forum what the heck? now I know freezing my pi for 30 minutes will not help the fuse recover at all.

Still no recovery on the new Pi. This is gonna get RMA'd most likely depending on what Newark has to say/turn around time etc.
Posts: 131
Joined: Wed Sep 19, 2012 2:16 am
by thsBavR10 » Thu Oct 04, 2012 3:48 pm
Did you get yourPi from RS or from Farnell?
I added a post in this thread: viewtopic.php?f=23&t=18576
asking for more values (resistance of F3).
Posts: 216
Joined: Sat Jul 21, 2012 3:11 pm
by thogue » Thu Oct 04, 2012 7:59 pm
maybe im off on how these are getting distributed. I am in the US I ordered from Newark/E14 (which is a Farnell company) So I guess thats farnell

my build build date is 1228
the stick on the back I assume its a serial FN12085XXXX, my older by being in the 2000s and by newest one (the one I still havnt used except for testing per this thread) is a 3000s

Dont know if this will become relevant or not we shall see.
Posts: 131
Joined: Wed Sep 19, 2012 2:16 am
by RaTTuS » Fri Oct 05, 2012 7:49 am
thogue wrote:maybe im off on how these are getting distributed. I am in the US I ordered from Newark/E14 (which is a Farnell company) So I guess thats farnell

my build build date is 1228
the stick on the back I assume its a serial FN12085XXXX, my older by being in the 2000s and by newest one (the one I still havnt used except for testing per this thread) is a 3000s

Dont know if this will become relevant or not we shall see.

no that will be date of manufacturing [probably]
serial is got from
cat /proc/cpuinfo
1QC43qbL5FySu2Pi51vGqKqxy3UiJgukSX - Prosliver FTW
"That's not right, the badgers have moved the goalposts."
User avatar
Posts: 4159
Joined: Tue Nov 29, 2011 11:12 am
Location: North West UK
by thogue » Fri Oct 05, 2012 2:19 pm
according to http://elinux.org/RaspberryPi_Boards

1228 is my build date but that might be just for the board , so I am not sure what the sticker is, I can only assume that it is put by the manufacturer, that put everything together, so It might contain another date, and more information, looking at pictures of other Pis it does not seem like there is any sort of standard format for the sticker.
Posts: 131
Joined: Wed Sep 19, 2012 2:16 am
by thogue » Fri Oct 05, 2012 4:12 pm
Also, my daily check of the newest rPi's poly, still zero recovery. ( still have not used this one since I originally opened the package to test.)

My current in use rPi (refered to as the older one in this thread) which was reading higher before had dropped down to about .9 ( .10-.01) and Ive been running this pretty much all day everyday.
Posts: 131
Joined: Wed Sep 19, 2012 2:16 am