skyluke
Posts: 13
Joined: Mon May 14, 2018 1:37 pm

USB voltage slowly dropping

Mon May 14, 2018 2:06 pm

Hi,

I do have a Raspi 3B that slowly drops USB voltage when running and I can measure that at the USB ports.
The Raspi is powered by an industrial/lab PSU with exactly 5V. All time when the Raspi is running this PSU provides 5V.
The PSU is connected using a cable (25cm, 0,75mm^2) that is soldered to the GPIO pins of the Raspi (so no fuse problem).
The Raspi has nothing connected except a HDMI monitor. But it has nearly no influence what devices are connected.

The Raspi is turned off for about 60 minutes and now turned on. The voltage at the USB ports is then between 4,90V and 4,92V.
(While booting the USB power is turned off and on again by raspbian (seems to be normal)).
After booting the running system's USB voltage is at about 4,86V-4,88V. Within the next hours the voltage is pretty constant, it goes down to 4,85V but not less. After 5-8 hours something strange happens and the voltage starts to drop down to 4,35V and the maximum is not more than 4,70V.

Now I turn off the Raspi and expect it to start with 4,9V again. But that does not happen. It starts with about 4,75V-4,78V and after boot it goes down to less than 4,5V. I have to turn it off for about 60 minutes to start again with 4,9V. Sometimes when I first boot I get a kernel panic telling me something about being unable to initialise CPU Core 2 or whatever. This is exactly at the moment when the USB ports are turned off and on while booting. I turn it off and on again and it starts fine with 4,9V.

Any ideas anyone? Maybe the Raspi is simply broken. But is there any explanation for this strange voltage behaviour?
I would like to understand what may have caused that because I tried several hours (days) now. I can reproduce but cannot understand.

Thanks a lot!
skyluke


Sorry, I know under-voltage is boring... ;-)

User avatar
Imperf3kt
Posts: 4018
Joined: Tue Jun 20, 2017 12:16 am
Location: Australia

Re: USB voltage slowly dropping

Mon May 14, 2018 10:40 pm

What are you using to power the Pi? It sounds like it is getting thermally stressed.
55:55:44:44:4C
52:4C:52:42:41

Rose coloured glasses are difficult to see through.

MaxK1
Posts: 1043
Joined: Sun Aug 26, 2012 11:34 pm

Re: USB voltage slowly dropping

Tue May 15, 2018 8:34 am

Can we assume latest Raspbian, fully updated and no overclocking? Does it get VERY hot at any point? And just to be sure, this is a 3B not a 3B+?
Could you post your /boot/config.txt?

The kernel crapping out on CPU 2 is probably reason enough to RMA it but the fact that you soldered the power connection directly to the board may complicate that process.

skyluke
Posts: 13
Joined: Mon May 14, 2018 1:37 pm

Re: USB voltage slowly dropping

Tue May 15, 2018 9:37 am

@Imperf3kt:
Temperature is 53-54°C


@MaxK1:
Yes, latest Raspbian (light, no GUI), latest firmware, no overclocking, not very hot at any point and model 3B.
Btw, just ordered a 3B+, will try that.

Here is the config.txt:

disable_overscan=1
hdmi_force_hotplug=1
dtparam=audio=off
dtparam=i2c_arm=on
dtparam=i2c=on
dtparam=spi=on
start_x=0
gpu_mem=16
dtoverlay=pi3-disable-bt
dtoverlay=pi3-disable-wifi
hdmi_blanking=1

User avatar
PeterO
Posts: 5968
Joined: Sun Jul 22, 2012 4:14 pm

Re: USB voltage slowly dropping

Tue May 15, 2018 9:43 am

Imperf3kt wrote:
Mon May 14, 2018 10:40 pm
What are you using to power the Pi? It sounds like it is getting thermally stressed.
It seems you didn't even get to Line 2 of the original post... :roll:
The Raspi is powered by an industrial/lab PSU with exactly 5V. All time when the Raspi is running this PSU provides 5V.
Are you getting any low voltage warnings on screen ?

Do you have another PI you can use just to check it doesn't behave the same and that it isn't something to do with your PSU ?

It sounds to me like a poor solder joint somewhere which increases in resistance with thermal stresses causing slight movement. The increased resistance could either be directly causing the 5V to drop, or is having an effect on the USB current limiting circuitry (but that would not explain the CPU core problem),

PeterO
Discoverer of the PI2 XENON DEATH FLASH!
Interests: C,Python,PIC,Electronics,Ham Radio (G0DZB),1960s British Computers.
"The primary requirement (as we've always seen in your examples) is that the code is readable. " Dougie Lawson

MaxK1
Posts: 1043
Joined: Sun Aug 26, 2012 11:34 pm

Re: USB voltage slowly dropping

Tue May 15, 2018 12:06 pm

It does sound like a bad solder joint on the board (maybe more than 1, which could account for the failure of 1 core I suppose)
You are in a maze of twisty little passages, all alike.
When General Failure and Major Disaster get together, Private Parts usually suffers.

skyluke
Posts: 13
Joined: Mon May 14, 2018 1:37 pm

Re: USB voltage slowly dropping

Tue May 15, 2018 3:58 pm

Yes, of course - I believe I can't see anything else except under-voltage warnings on screen if I reach the critical threshold. :-D
Btw, it started after upgrade to 4.14 or I started to see. Yellow bolt & console messages/dmesg was how I became aware of this. (Who sees the red LED?)

This is the PSU: https://de.rs-online.com/web/p/labor-ne ... e/7568821/
It's my faithful companion. It provides exactly 5V (measured) if I say "5V please".
(@PeterO: DB5VW since 1982. So, you can be pretty sure that I have the right equipment and know how to handle it. ;-))

The new 3B+ just arrived and I got another 3B. I will test them.

When does thermal stress start for a Raspi?

User avatar
PeterO
Posts: 5968
Joined: Sun Jul 22, 2012 4:14 pm

Re: USB voltage slowly dropping

Tue May 15, 2018 4:23 pm

skyluke wrote:
Tue May 15, 2018 3:58 pm
Yes, of course - I believe I can't see anything else except under-voltage warnings on screen if I reach the critical threshold. :-D
Btw, it started after upgrade to 4.14 or I started to see. Yellow bolt & console messages/dmesg was how I became aware of this. (Who sees the red LED?)

This is the PSU: https://de.rs-online.com/web/p/labor-ne ... e/7568821/
It's my faithful companion. It provides exactly 5V (measured) if I say "5V please".
(@PeterO: DB5VW since 1982. So, you can be pretty sure that I have the right equipment and know how to handle it. ;-))

The new 3B+ just arrived and I got another 3B. I will test them.

When does thermal stress start for a Raspi?
Ok, but because you specifically mentioned the USB voltage, and didn't mention the under voltge warning I assumed you were NOT seeing the low voltage warnings. I would now suggest trying a different PSU and different cables.

PeterO
Discoverer of the PI2 XENON DEATH FLASH!
Interests: C,Python,PIC,Electronics,Ham Radio (G0DZB),1960s British Computers.
"The primary requirement (as we've always seen in your examples) is that the code is readable. " Dougie Lawson

skyluke
Posts: 13
Joined: Mon May 14, 2018 1:37 pm

Re: USB voltage slowly dropping

Tue May 15, 2018 5:20 pm

Just did that. Tried a 5V 2.5A PSU using usb cable. Did not wait for an hour. And my Raspi started again with about 4.7V.

Will try a new Raspi this evening or try a Meanwell PSU.

Btw, I came from another smaller PSU. This was why I changed to the lab PSU. 20A and 0,75mm^2 cable should be enough for such a small device. I know it delivers 5V. If I tune up for 5.2V I will get 0.2V more until voltage drops. But the PSU that should be used can deliver 5V only.

I'm sure a new Raspi will help.

Are there any design differences regarding power between 3B and 3B+?

And when does thermal stress start? At 54°C ??

User avatar
FTrevorGowen
Forum Moderator
Forum Moderator
Posts: 5901
Joined: Mon Mar 04, 2013 6:12 pm
Location: Bristol, U.K.
Contact: Website

Re: USB voltage slowly dropping

Tue May 15, 2018 8:53 pm

Just to clarify further - what, if anything, is plugged into the USB port(s)? (Just wondering if the total current through the Pi is close to its polyfuse limit and that's "soft-triggering". In that situation, AIUI, keeping the Pi unused and unpowered for a day or so should lead to some "voltage recovery")
Trev.
Still running Raspbian Jessie or Stretch on some older Pi's (an A, B1, 2xB2, B+, P2B, 3xP0, P0W, 2xP3A+, P3B+, P3B, B+, and a A+) but Buster on the P4B's. See: https://www.cpmspectrepi.uk/raspberry_pi/raspiidx.htm

User avatar
Imperf3kt
Posts: 4018
Joined: Tue Jun 20, 2017 12:16 am
Location: Australia

Re: USB voltage slowly dropping

Tue May 15, 2018 9:14 pm

When I said thermally stressed, I meant the PSU, not the Pi.

@PeterO:
What is the make / model of his/her PSU?
An "industrial PSU" tells me nothing useful about the power supply.
55:55:44:44:4C
52:4C:52:42:41

Rose coloured glasses are difficult to see through.

User avatar
mahjongg
Forum Moderator
Forum Moderator
Posts: 13370
Joined: Sun Mar 11, 2012 12:19 am
Location: South Holland, The Netherlands

Re: USB voltage slowly dropping

Tue May 15, 2018 9:50 pm

sounds like the PI's polyfuse is slowly giving up the ghost.

54 degrees celsius is a reasonable temperature for the SoC.

User avatar
PeterO
Posts: 5968
Joined: Sun Jul 22, 2012 4:14 pm

Re: USB voltage slowly dropping

Tue May 15, 2018 10:28 pm

mahjongg wrote:
Tue May 15, 2018 9:50 pm
sounds like the PI's polyfuse is slowly giving up the ghost.

54 degrees celsius is a reasonable temperature for the SoC.
I though powering via GPIO bypassed the ployfuse ?

PeterO
Discoverer of the PI2 XENON DEATH FLASH!
Interests: C,Python,PIC,Electronics,Ham Radio (G0DZB),1960s British Computers.
"The primary requirement (as we've always seen in your examples) is that the code is readable. " Dougie Lawson

User avatar
PeterO
Posts: 5968
Joined: Sun Jul 22, 2012 4:14 pm

Re: USB voltage slowly dropping

Tue May 15, 2018 10:30 pm

Imperf3kt wrote:
Tue May 15, 2018 9:14 pm
@PeterO:
What is the make / model of his/her PSU?
An "industrial PSU" tells me nothing useful about the power supply.
Then ask "what make and model" not some generic "what are you using".
PeterO
Discoverer of the PI2 XENON DEATH FLASH!
Interests: C,Python,PIC,Electronics,Ham Radio (G0DZB),1960s British Computers.
"The primary requirement (as we've always seen in your examples) is that the code is readable. " Dougie Lawson

User avatar
Imperf3kt
Posts: 4018
Joined: Tue Jun 20, 2017 12:16 am
Location: Australia

Re: USB voltage slowly dropping

Tue May 15, 2018 11:49 pm

Fair enough.
However it seems I am blind after all, OP later posted this link:
https://de.rs-online.com/web/p/labor-ne ... e/7568821/


I am not using a device that can translate this on the fly, but it looks like the current is adjustable?
55:55:44:44:4C
52:4C:52:42:41

Rose coloured glasses are difficult to see through.

skyluke
Posts: 13
Joined: Mon May 14, 2018 1:37 pm

Re: USB voltage slowly dropping

Wed May 16, 2018 7:16 am

It was the Raspi. It could have been so easy if I just replaced the Raspi and did not deep dive into this strange behavior.

I changed the Raspi yesterday night and let it run until this morning. Voltage at USB ports was still at 4.91V. Please note PSU voltage was set to 5.0V as the final PSU can supply 5V only.

The lab PSU is the one in the link, can adjust V and A and is the ISO version of the PSU shown. I use it for a pretty long time now and as a certified PSU I believe its settings and display.

Btw, I use the new 3B+ as replacement. Until this morning no problems with voltage and no boot problems.

I added piface digital 2 and piface rtc again and let it run again. I will see what happens this evening.

Question. What do you think of the Meanwell IRM-20-5? Do you have any experience with that? This model is planned as the PSU to be used in a DIN rail housing.

Thank you very much for your help.

User avatar
Burngate
Posts: 6371
Joined: Thu Sep 29, 2011 4:34 pm
Location: Berkshire UK Tralfamadore
Contact: Website

Re: USB voltage slowly dropping

Wed May 16, 2018 10:46 am

Imperf3kt wrote:
Tue May 15, 2018 11:49 pm
... I am not using a device that can translate this on the fly, but it looks like the current is adjustable?
https://docs-emea.rs-online.com/webdocs ... 38cf36.pdf
Meanwell IRM-20-5
No direct experience, but it seems from its spec it should work.

Can't remember the details, but the USB 5v goes through a chip that features over-current protection ... I wonder if it's that that's failing ...

Return to “Troubleshooting”