Helpme1986
Posts: 121
Joined: Tue Jan 03, 2012 3:48 pm

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 8:17 pm

can you confirm that TP2 the hole in the board next to RCA Video and which side of the F3 chip it should be (edge or inside) ive just tried it and seemed to short the board

User avatar
jbeale
Posts: 3500
Joined: Tue Nov 22, 2011 11:51 pm
Contact: Website

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 8:35 pm

Through hole "TP2" is labelled as such on the top side silkscreen, between the GPIO pins and the RCA video jack.  According to forum admin JamesH earlier in this thread, TP1 is +5V and TP2 is ground.

Probing with the meter (in voltage mode, mind- not current mode!) should not short anything out, unless the meter probes tilt over by accident and contact a pin or component, which is very possible unless care is taken. Not sure what's on the nearest GPIO pins, but it probably doesn't want to be shorted to ground.

As it is a through-hole, you can also probe TP2 from the back side, which is where F3 is also. TP2 is not labelled on the back side, but should be easy to identify first looking from the front.

Helpme1986
Posts: 121
Joined: Tue Jan 03, 2012 3:48 pm

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 8:39 pm

Ignore that, me being stupid!!!

5v with all working dipping to 4.98v at the lowest but goes to 5.10v when stuff fail

User avatar
Gert van Loo
Posts: 2486
Joined: Tue Aug 02, 2011 7:27 am
Contact: Website

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 8:40 pm

Definitely no jump wire! Measuring the voltage is a problem if the processor is running. Best is to measure the voltage with no SD-card installed. That way you know the processor is running a tiny loop and you do not have fluctuations due to variable tasks running. Easiest place to measure the voltage is across the big black diode D17 (next to the green poly fuse).

It the voltage is slowly rising your power supply is not good. With no SD-card the voltage should be rock-solid the same value.

I have tried my PI with a 5.03V bench supply. With nothing connected . So no SD-card. It draws 100mA then after a few seconds it goes to 75mA. I can not find any voltage over the poly fuse. Even with the voltmeter set to a range of 200mV, where I should be able to see a variation of 0.1mV it is zero. I did a quick test of the meter measuring the voltage drop over the 5V cable from my bench supply to the Raspberry-Pi: 1.7mV. So the meter seems to be working fine.

Possible causes:

1/ If your supply is not stable and the voltage goes up the D17 protection diode will kick-in to protect your supply from overvoltage. It will conduct all over voltage to ground. (Not very professional phrased but understandable). All the current from that overvoltage will heat up your poly fuse and after a while it will stop working. Cooling down is a matter of minutes (I estimate 5-15), not hours.

2/ Something is drawing too much current: Same with the poly fuse. It will kick in. This should not be any USB device as each USB port is protected by a much smaller polyfuse: 140 mA. So that should kick-in before your main fuse. The HDMI port is the other 5V connection. There is a small possibility that your TV is faulty and drawing too much current from the 5V. (I know it should not be I am tracing all links).

3/ Or there is a fault on the board which makes it draw too much current. As it takes 30 seconds to happen this might not be spotted during production test.

4/ The poly fuse is faulty and kicks in too early.

Ideally you should measure the current going into your board. But measuring current is always difficult as you have to break the power connection, put the meter in series and set it to AMPS.

Don't forget to always, always immediately set your meter back to Volts after measuring current. You will one day regret it if you don't.

User avatar
jbeale
Posts: 3500
Joined: Tue Nov 22, 2011 11:51 pm
Contact: Website

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 8:46 pm

Gert:> I can not find any voltage over the poly fuse. Even with the voltmeter set to a range of 200mV, where I should be able to see a variation of 0.1mV it is zero.

Hmmm... Helpme1986 has reported 0.11 to 0.16 volts across his F3 polyfuse (with the system up and running, I guess). Very interesting to see what his number is with the SD card out.

Helpme1986
Posts: 121
Joined: Tue Jan 03, 2012 3:48 pm

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 8:53 pm

Without SD i get 0.02 over F3 and 5.24v from TP2 to F3

User avatar
jbeale
Posts: 3500
Joined: Tue Nov 22, 2011 11:51 pm
Contact: Website

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 9:11 pm

So, about 20 mV drop through F3.  Does your meter have a 2000 mV or 200mV range setting, that could get a more accurate number.  Apparently D17 is a zener which acts to crowbar an overvoltage supply (I wonder what the threshold voltage is, and how sharp a turn-on it has).  D17 isn't warm to the touch, by any chance?

User avatar
johnbeetem
Posts: 945
Joined: Mon Oct 17, 2011 11:18 pm
Location: The Mountains
Contact: Website

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 9:12 pm

Another easy test is to check the resistance of polyfuse F3 with RasPi power off.  Set your multimeter to the most sensitive resistance setting (200 Ohm on my cheapie).  Push the probes together to test the resistance of the probe cables.  Then check the resistance between power input S1 pin 1 and TP1 (+5V).  It should be almost the same as the resistance of the probe cables.  If the difference is more than 0.2 Ohm or so (I don't know what the exact value should be for F3) then F3 may be bad or there's a bad solder joint between S1 pin 1 and TP1.  If this is the case, it's easy to probe further to narrow it down.

Be sure to set the multimeter back to volts after measuring resistance, and always measure resistance with RasPi power off.

User avatar
Gert van Loo
Posts: 2486
Joined: Tue Aug 02, 2011 7:27 am
Contact: Website

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 9:20 pm

OK, my error! I was bypassing the fuse! (My bench supply was not using the USB plug!)  The fuse resistance should be about 0.4 Ohm and with 75mA that is indeed about 3 milli volts! So your Pi is not ill. In fact it looks like your fuse is ~0.26 Ohms. To find out what is happening I suggest you slowly start adding devices and see if/when the voltage over the fuse jumps up. So First only the SD-card, then keyboard, mouse etc.

Helpme1986
Posts: 121
Joined: Tue Jan 03, 2012 3:48 pm

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 9:21 pm

if ive done this right (S1 has some very small connectors!!!)

Ohms on 200, directly together is 00.2 and between TP1 and S1 (I HOPE!!!) is 00.6

Helpme1986
Posts: 121
Joined: Tue Jan 03, 2012 3:48 pm

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 9:25 pm

Gert said:


OK, my error! I was bypassing the fuse! (My bench supply was not using the USB plug!)  The fuse resistance should be about 0.4 Ohm and with 75mA that is indeed about 3 milli volts! So your Pi is not ill. In fact it looks like your fuse is ~0.26 Ohms. To find out what is happening I suggest you slowly start adding devices and see if/when the voltage over the fuse jumps up. So First only the SD-card, then keyboard, mouse etc.


Can you confirm how I should be measuring mA? (sorry for asking so many questions!)

For Volts im using 20 under the V ~ section and connecting from TP2 to F3

Helpme1986
Posts: 121
Joined: Tue Jan 03, 2012 3:48 pm

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 9:38 pm

Between TP2 and F3

Nothing: 5.43v

SDCard:5.35v

SD+HDMI: 5.34v

SD+HDMI+KB: 5.33v

SD+HDMI+KB+Mouse: 5.33v

SD+HDMI+KB+Mouse+Eth: 5.31v

GPIO +5 - 0

Nothing: 5.33v

SDCard:5.22v

SD+HDMI: 5.21v

SD+HDMI+KB: 5.21v

SD+HDMI+KB+Mouse: 5.20v

SD+HDMI+KB+Mouse+Eth: 5.16v

User avatar
jbeale
Posts: 3500
Joined: Tue Nov 22, 2011 11:51 pm
Contact: Website

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 10:00 pm

If we assume the "SD+HDMI+KB+Mouse+Eth" configuration draws 400 mA (as a rough guess) then your power supply just on its own has a total effective resistance of (5.43 – 5.31 V)/(0.4 A) = 0.3 ohms which should be OK.

Likewise, the F3 part has a resistance of (5.31 – 5.16 V) / (0.4 A) = 0.375 ohms which I gather is within spec.

Current is not such a casual measurement as voltage. To actually measure current, you would have to sacrifice a USB cable by peeling back the insulating cover and foil/metal shield if present, separating out the wires, cutting the red +5V wire in two, stripping the ends of that wire, and inserting your multimeter in "CURRENT" mode into the circuit, connecting the meter leads between the two wire ends.  By this point you probably want grabber probes or alligator clips to connect the wires, or strip enough back that you can wrap the wires around the probes (which is still not usually a solid connection). It would be useful to know, but that measurement costs a cable and some time.

User avatar
Gert van Loo
Posts: 2486
Joined: Tue Aug 02, 2011 7:27 am
Contact: Website

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 10:01 pm

Helpme1986 said:

....
Can you confirm how I should be measuring mA? (sorry for asking so many questions!)

For Volts im using 20 under the V ~ section and connecting from TP2 to F3


No, no, no, no! You should be using the V= settings. If you are measuring anything on V~ do NOT connect the supply. It is most likely a non-regulated supply!

User avatar
Gert van Loo
Posts: 2486
Joined: Tue Aug 02, 2011 7:27 am
Contact: Website

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 10:08 pm

So I had to get that reply out ASAP.

If your are measuring nothing on 20V= (I think it actually has dashes at the bottom:)

_______

__ __ __

Then don't use the supply. It may be 5V but is is not regulated!

For measuring (milli)-Amps you have to place the meter in the supply line which is normally rather difficult as you have to make a cut in the wire.

Like this:

+ -----(A)-----[ Device_under_test]---------- -

Helpme1986
Posts: 121
Joined: Tue Jan 03, 2012 3:48 pm

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 10:08 pm

My bad gert, it was =, no idea y I said ~, been a very long day!

User avatar
Gert van Loo
Posts: 2486
Joined: Tue Aug 02, 2011 7:27 am
Contact: Website

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 10:14 pm

Fair enough. I am running out of ideas. Your voltage drop seem reasonable. All I can think of is do some measurements when it stops working. I off to bed now.

User avatar
jbeale
Posts: 3500
Joined: Tue Nov 22, 2011 11:51 pm
Contact: Website

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 10:14 pm

Gert, I think he might be using Volts-DC mode despite claiming a "~" mark, I don't think he could have gotten the readings he did in an AC Volts mode, and there aren't any unregulated 5V USB supplies made- are there?  -Scary, if so.

Helpme1986
Posts: 121
Joined: Tue Jan 03, 2012 3:48 pm

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 10:19 pm

Fair enough. I am running out of ideas. Your voltage drop seem reasonable. All I can think of is do some measurements when it stops working. I off to bed now.

I"ve already gone to bed, thanks for your help, don"t know what to try next other than if someone else with a working setup lives around Birmingham so I could test it on theirs

User avatar
johnbeetem
Posts: 945
Joined: Mon Oct 17, 2011 11:18 pm
Location: The Mountains
Contact: Website

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 10:35 pm

Helpme1986 said:

Ohms on 200, directly together is 00.2 and between TP1 and S1 (I HOPE!!!) is 00.6
OK, that gives you 0.4 Ohms for F3, which is what Gert expects.  The latest GPIO +5V values look good -- if anything, they're a little high but within spec, so F3 seems to be working OK.  What was the CPU doing when you took these measurements?

Just for fun, I'd check the +3.3V regulator RG2.  There should be a pin that's precisely +3.3.V, or you can check C11 which should have precisely +3.3V across it.  In the unlikely event that it's not precisely 3.3V, RG2 may be bad or your meter may not be calibrated properly.  How hot is RG2 getting?  I'm not sure how hot to expect (I don't know when I'm getting my RasPi) but if it gets too hot it will warm up F3 and (if I understand polyfuse behaviour correctly) increase F3's resistance, and then GPIO +5V voltage will start to droop.

Also, there is a schematic of the power supply section at http://www.raspberrypi.org/wp-.....12/psu.png, which may be helpful.

colin B
Posts: 119
Joined: Sun Mar 04, 2012 12:23 pm
Contact: Website

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 10:48 pm

No one has wodnered if the main reservoir cap (electrolytic) might be leaky or high ESR. (even though it shouldn't be being SMD)
On a clear disk one can seek forever

User avatar
johnbeetem
Posts: 945
Joined: Mon Oct 17, 2011 11:18 pm
Location: The Mountains
Contact: Website

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 11:22 pm

bodgyuk said:


No one has wodnered if the main reservoir cap (electrolytic) might be leaky or high ESR. (even though it shouldn't be being SMD)


In this thread, a new Razzer accidentally broke off electrolytic cap C6.  A commenter said he had done the same thing and "it didn't hurt the running of it none" or words to that effect.  In the thread, Gert says C6 is mostly there as "good design practice".

Here's my opinion: you need bypass caps if load current changes rapidly, e.g., you have a CPU that needs a current spike when driving a 64-bit data bus from all 0 to all 1.  In RasPi, all that sort of mishegoss is on the +3.3V or lower rails, so it's really important that they have good bypass.  The load presented to the +3.3V regulator is thus pretty steady, so its +5V current draw is probably pretty steady as well.

The +5V also goes to the GPIO connector, external USB devices, and HDMI, but those devices should have their own local bypass caps.  So C6 normally doesn't have to do much and nobody would notice if it's a cheap Chinese substitute that will boil off its electrolyte in the first few months of operation.

plugwash
Forum Moderator
Forum Moderator
Posts: 3464
Joined: Wed Dec 28, 2011 11:45 pm

Re: USB bus seems to be shutting down after 30seconds

Wed Apr 18, 2012 11:59 pm

John Beetem said:

Here's my opinion: you need bypass caps if load current changes rapidly, e.g., you have a CPU that needs a current spike when driving a 64-bit data bus from all 0 to all 1.
Something else to remember is that big caps tend to have relatively high equivilent series resistance and equivilent series inductance. So they are fine for dealing with the relatively slow spikes of current coming from say turning LEDs on and off but they are useless for dealing with the very short spikes that a high speed logic chip creates. To deal with those you need lots of very small capacitors with low ESR and ESL located as close to the chip as possible (some vendors even put them on the IC package itself).

colin B
Posts: 119
Joined: Sun Mar 04, 2012 12:23 pm
Contact: Website

Re: USB bus seems to be shutting down after 30seconds

Thu Apr 19, 2012 12:24 am

plugwash said:


Something else to remember is that big caps tend to have relatively high equivilent series resistance and equivilent series inductance.

True, but SMD electrolytics have much lower ESR and ESL than through hole of the same value.

On most circuits I design I tend to have 1 - 10uF plus 100 + 10 or 22nF (for this type of circuit).

It was just a thought, dodgy solder joint or below spec cap being a problem child. Probably need to scope (as has been mentioned) the area to get a handle on the problem.

Pic chips can display similar behaviour if the MCLR or Vdd bypass caps are left out - hand capacitance or general parasitic capacitance.

On a clear disk one can seek forever

User avatar
johnbeetem
Posts: 945
Joined: Mon Oct 17, 2011 11:18 pm
Location: The Mountains
Contact: Website

Re: USB bus seems to be shutting down after 30seconds

Thu Apr 19, 2012 1:01 am

bodgyuk said:


plugwash said:


Something else to remember is that big caps tend to have relatively high equivalent series resistance and equivalent series inductance.


True, but SMD electrolytics have much lower ESR and ESL than through hole of the same value.

On most circuits I design I tend to have 1 - 10uF plus 100 + 10 or 22nF (for this type of circuit).


The great thing nowadays is that you can use multi-layer ceramic (MLC) SMT caps for 10uF and similar values.

Return to “Troubleshooting”