omniboard
Posts: 22
Joined: Thu Sep 20, 2018 8:50 am

Re: The USB power appeal

Fri Sep 21, 2018 12:36 pm

jamesh wrote:
Fri Sep 21, 2018 12:11 pm
omniboard wrote:
Fri Sep 21, 2018 11:00 am
No - as I wrote it was just the Pi connected to the monitor + the dongle . Nothing else.
Most odd then. I suspect the dongle is broken and shorting something out, but difficult to tell. Clearly not in inherent issue with the Pi itself, if all other devices seem to work fine (ie for everyone else out there using USB devices).
The dongle is OK - works with another computers and SBCs. Actually the current limiting IC should not get damaged even if there is a short circuit. Otherwise what is it for? It is like burning someone's houses to prevent the home accidents ;)

BTW I do not nitpick - I am concerned about my USB product - it is powered from the USB, and it can drive up to 5 units. Just do not want to damage users PIs.

jamesh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 20923
Joined: Sat Jul 30, 2011 7:41 pm

Re: The USB power appeal

Fri Sep 21, 2018 2:30 pm

What model is the dongle you plug in that causes the problem? Can you provide a link to its specification?
Principal Software Engineer at Raspberry Pi (Trading) Ltd.
Please direct all questions to the forum, I do not do support via PM.

mattmiller
Posts: 1966
Joined: Thu Feb 05, 2015 11:25 pm

Re: The USB power appeal

Fri Sep 21, 2018 8:56 pm

Thank you :) But if you read my previous post - it is not the suspect.
It really, really is the prime suspect.

If it is the only thing plugged in at the time when the USB power fails - it is the No 1 suspect

Idahowalker
Posts: 399
Joined: Wed Jan 03, 2018 5:43 pm

Re: The USB power appeal

Fri Sep 21, 2018 10:29 pm

Get 2 NEW RPi's plug them in and power them up.

Plug into the Pis the mouse, kybd, and HDMI.
Plug the dongle into one of the Pi's .
See if the RPi without the dongle fails.
Without knowing why you are deleting my postings, I will not know how...

omniboard
Posts: 22
Joined: Thu Sep 20, 2018 8:50 am

Re: The USB power appeal

Sat Sep 22, 2018 10:19 am

mattmiller wrote:
Fri Sep 21, 2018 8:56 pm

It really, really is the prime suspect.

If it is the only thing plugged in at the time when the USB power fails - it is the No 1 suspect
It is not the suspect - but of course it kills the RPi? A paradox? Not - The dongle was tested and dynamically measured. And the current draw does not exceed 3 units. So there is a problem with the Pi as it dies when it has to supply max 3 units.
jamesh wrote: What model is the dongle you plug in that causes the problem? Can you provide a link to its specification?
HP mouse dongle model: DGRFEO

mattmiller
Posts: 1966
Joined: Thu Feb 05, 2015 11:25 pm

Re: The USB power appeal

Sat Sep 22, 2018 1:02 pm

I won't argue any further with you on the likely culprit :)

but
And the current draw does not exceed 3 units
If you stick to using standard mA instead of your "units", it will help others understand your posts a bit better

Good luck :)

User avatar
Burngate
Posts: 5513
Joined: Thu Sep 29, 2011 4:34 pm
Location: Berkshire UK
Contact: Website

Re: The USB power appeal

Sat Sep 22, 2018 5:26 pm

I've been keeping clear of this thread, so far, in part because others have said what I would have said.
But having again read through the whole thread, these are my thoughts.

So, some of your Pis - not 3B+, by the look of it - have died under similar circumstances.
But they haven't actually died - they're still working, but no power reaches the USB sockets.
Apparently, adding a wire from the power input socket to the USB power out restores things.

1) Does connecting pins 1 & 6 of U13 similarly restore power to the USB sockets? I ask, because connecting your wire at the power input socket bypasses both the polyfuse and the ideal diode.

2)I believe U13 shouldn't die, even if you connect a dead short across the USB power and ground - my understanding is that it's designed to cope with that, limiting the current to what the upstream circuit can handle.
So something else is happening.

omniboard
Posts: 22
Joined: Thu Sep 20, 2018 8:50 am

Re: The USB power appeal

Sat Sep 22, 2018 10:34 pm

mattmiller wrote:
Sat Sep 22, 2018 1:02 pm
If you stick to using standard mA instead of your "units", it will help others understand your posts a bit better
Sorry mate - but the units are not mine only defined in the USB standard. And I think, only pepole who know what does it mean can answer this question. If someone does not - just does not have enough knowledge. So please do not such a malignant comments. If you want to learn something about the USB - here is a good starting point : https://www.beyondlogic.org/usbnutshell/usb2.shtml
Burngate wrote: So, some of your Pis - not 3B+, by the look of it - have died under similar circumstances.
But they haven't actually died - they're still working, but no power reaches the USB sockets.
I think it is obvious if the workaround works :) As i see on internet this problem is quite common.
Burngate wrote: 1) Does connecting pins 1 & 6 of U13 similarly restore power to the USB sockets? I ask, because connecting your wire at the power input socket bypasses both the polyfuse and the ideal diode.
I did not try yet.
Burngate wrote: 2)I believe U13 shouldn't die, even if you connect a dead short across the USB power and ground - my understanding is that it's designed to cope with that, limiting the current to what the upstream circuit can handle.
So something else is happening.
It should not - so the question is what is causing this problem and my request to unveil this part of the schematics (I do not think that something secret and revolutionary is there). It may benefit the PI owners and the RPi foundation as well - maybe crowd of the decent electronics will discover the source of the problem, helping the RPiF in its efforts to improve the product.

W. H. Heydt
Posts: 9223
Joined: Fri Mar 09, 2012 7:36 pm
Location: Vallejo, CA (US)

Re: The USB power appeal

Sun Sep 23, 2018 4:30 am

Paeryn wrote:
Fri Sep 21, 2018 12:03 am
From memory the max current the RPi provides for USB is 1.2A shared between the 4 ports (used to be 600mA on the original RPis unless a config setting was enabled, though they only had 2 ports). I don't know if the RPi3B+ upped this to allow the full 2A (especially when powered by PoE). I've always assumed this limit was due to being powered via a USB connection.

I may be wrong, I've never had anything needing much power plugged in.
Just a bit of historical correction.... The original Model B had a 140mA polyfuse for each of the two USB ports (and in successive runs and revisions, they were replaced with zero ohm resistors and then eliminated altogether). The Pi2B was limited to 600mA combined for 4 ports, with a config.txt setting to change that to 1.2A. Eventually, 1.2A became the default.

I have Model B Pis with USB polyfuses and ones with zero ohm resistors as well as ones--Rev 2.0--without anything there at all.

jamesh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 20923
Joined: Sat Jul 30, 2011 7:41 pm

Re: The USB power appeal

Sun Sep 23, 2018 6:52 am

omniboard wrote:
Sat Sep 22, 2018 10:34 pm
mattmiller wrote:
Sat Sep 22, 2018 1:02 pm
If you stick to using standard mA instead of your "units", it will help others understand your posts a bit better
Sorry mate - but the units are not mine only defined in the USB standard. And I think, only pepole who know what does it mean can answer this question. If someone does not - just does not have enough knowledge. So please do not such a malignant comments. If you want to learn something about the USB - here is a good starting point : https://www.beyondlogic.org/usbnutshell/usb2.shtml
Burngate wrote: So, some of your Pis - not 3B+, by the look of it - have died under similar circumstances.
But they haven't actually died - they're still working, but no power reaches the USB sockets.
I think it is obvious if the workaround works :) As i see on internet this problem is quite common.
Burngate wrote: 1) Does connecting pins 1 & 6 of U13 similarly restore power to the USB sockets? I ask, because connecting your wire at the power input socket bypasses both the polyfuse and the ideal diode.
I did not try yet.
Burngate wrote: 2)I believe U13 shouldn't die, even if you connect a dead short across the USB power and ground - my understanding is that it's designed to cope with that, limiting the current to what the upstream circuit can handle.
So something else is happening.
It should not - so the question is what is causing this problem and my request to unveil this part of the schematics (I do not think that something secret and revolutionary is there). It may benefit the PI owners and the RPi foundation as well - maybe crowd of the decent electronics will discover the source of the problem, helping the RPiF in its efforts to improve the product.

This is NOT a common problem. You are the only person who has reported it here. This makes it one in 20million from our POV. Randomn internet claims without citations are useless.

We won't be releasing any more schematics.

And, the request to use mA is perfectly valid, using 'unit' makes no sense unless you say what a unit is. It may be in the USB spec somewhere, but to save everyone looking it up, just give values in mA. Its the standard SI unit for current after all.

Edit: USB spec defines a unit load as 100mA. Although the page linked about also uses 2mA as a unit at one point, so is suspect.
Principal Software Engineer at Raspberry Pi (Trading) Ltd.
Please direct all questions to the forum, I do not do support via PM.

omniboard
Posts: 22
Joined: Thu Sep 20, 2018 8:50 am

Re: The USB power appeal

Sun Sep 23, 2018 11:18 am

jamesh wrote:
Sun Sep 23, 2018 6:52 am
Edit: USB spec defines a unit load as 100mA. Although the page linked about also uses 2mA as a unit at one point, so is suspect.
It is not suspect. You just did not read it. The units of power are 100mA or 150mA.

The 2mA unit is only used in the device descriptors (data sent from the device to the USB host. It has to be smaller as the power consumption field in the descriptor has an integer value (fractional numbers cannot be represented by the integers)
jamesh wrote:
Sun Sep 23, 2018 6:52 am
This is NOT a common problem.You are the only person who has reported it here. This makes it one in 20million from our POV. Randomn internet claims without citations are useless.
Maybe I am the only one :)
https://www.raspberrypi.org/forums/view ... p?t=154343
and many more

But you did not even try to answer the question and are focused on me instead of the problem.

It should not happen. I know what I am doing. I have done many tests to eliminate my fault before posting here.

Please answer the question: is it safe to draw power from USB or users should use the powered hubs instead.

mattmiller
Posts: 1966
Joined: Thu Feb 05, 2015 11:25 pm

Re: The USB power appeal

Sun Sep 23, 2018 11:57 am

If you stick to using standard mA instead of your "units", it will help others understand your posts a bit better
Was a genuine post meant in a positive manner to help YOU get the best answers to your problem

It was neither malignant or intended to be anything else other than to be helpful to you

User avatar
Burngate
Posts: 5513
Joined: Thu Sep 29, 2011 4:34 pm
Location: Berkshire UK
Contact: Website

Re: The USB power appeal

Sun Sep 23, 2018 12:05 pm

omniboard wrote:
Sun Sep 23, 2018 11:18 am
... I know what I am doing. I have done many tests to eliminate my fault before posting here.
Could you answer my question?

If I had the device in front of me, I'd be able to -fault-find myself; as it is I have to rely on you performing the tests.
So measuring around the chip, shorting it out, etc. is something you will have to do for me

omniboard
Posts: 22
Joined: Thu Sep 20, 2018 8:50 am

Re: The USB power appeal

Sun Sep 23, 2018 5:14 pm

Burngate wrote:
Sun Sep 23, 2018 12:05 pm
Could you answer my question?
Not at the moment as I am not at work today :)

jamesh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 20923
Joined: Sat Jul 30, 2011 7:41 pm

Re: The USB power appeal

Sun Sep 23, 2018 5:49 pm

omniboard wrote:
Sun Sep 23, 2018 5:14 pm
Burngate wrote:
Sun Sep 23, 2018 12:05 pm
Could you answer my question?
Not at the moment as I am not at work today :)
To answer your question.

Yes, it is completely safe to pull current from the USB ports. Millions of people do it without problem with devices from low power to high power with no problems. If you start puling too much the device will likely show a warning icon (lightning bolt) and a log message will be entered saying there is an undervoltage.

I have no idea what your problem is,. but it's not the Pi.
Principal Software Engineer at Raspberry Pi (Trading) Ltd.
Please direct all questions to the forum, I do not do support via PM.

User avatar
Imperf3kt
Posts: 1472
Joined: Tue Jun 20, 2017 12:16 am
Location: Australia

Re: The USB power appeal

Mon Sep 24, 2018 12:08 am

I have a test for you.
You ask if three 'units' is too much for a Pi. A "unit" being 100mA
As you claim drawing more than 300mA is killing a Pi, then try a different device or multiple devices that add up to or more than '3 units'.
Do NOT plug in the device that causes it to fail.


If the power dies, you might have a point.
If it functions just fine, then your suspicious part is not the Pi.
Stop plugging your fan directly into the GPIO 5v
https://www.electronics-tutorials.ws/power/transient-suppression.html

omniboard
Posts: 22
Joined: Thu Sep 20, 2018 8:50 am

Re: The USB power appeal

Mon Sep 24, 2018 12:24 pm

jamesh wrote:
Sun Sep 23, 2018 5:49 pm


To answer your question.

Yes, it is completely safe to pull current from the USB ports.
Thank you.

So I am starting to port the software.

W. H. Heydt
Posts: 9223
Joined: Fri Mar 09, 2012 7:36 pm
Location: Vallejo, CA (US)

Re: The USB power appeal

Tue Sep 25, 2018 4:03 am

omniboard wrote:
Sun Sep 23, 2018 11:18 am
Please answer the question: is it safe to draw power from USB or users should use the powered hubs instead.
I have a Pi3B+ with a WD PiDrive powered from one of the Pi's USB port. I switch it out for other Pis fairly regularly, so it gets shut down, disconnected, reconnected, and rebooted reasonably often. It is also connected to a KVM switch--USB port for mouse and keyboard, HDMI for video.

So, yes, you can draw power from the USB ports on a Pi3B+ with no problems.

I have to agree with others that there is something peculiar going on with your dongle.

User avatar
HawaiianPi
Posts: 3031
Joined: Mon Apr 08, 2013 4:53 am
Location: Aloha, Oregon USA

Re: The USB power appeal

Tue Sep 25, 2018 6:14 am

omniboard wrote:
Sun Sep 23, 2018 11:18 am
Please answer the question: is it safe to draw power from USB or users should use the powered hubs instead.
I answered your question 4 days ago when I told you what you were claiming was complete nonsense, and then went on to explain that I regularly use power hungry USB devices on multiple Pi computers, even overloading my USB ports with multiple USB hard drives, but you completely ignored my post, just like you have rejected many other replies and other helpful advice in this thread.

You seem unable to accept any help or advice that disagrees with your distorted perception that there is a severe design flaw in the Pi, so no one here will be able to help you. And I honestly don't understand how you expected to get help with that attitude. There are people here who were genuinely trying to help you troubleshoot your problem (and yes, I know my reply wasn't one of them), but you respond by attacking them.

As jamesh said, I have no idea what your problem is, but it's not the Pi.
omniboard wrote:
Sun Sep 23, 2018 11:18 am
It is not suspect. You just did not read it. The units of power are 100mA or 150mA.
And you think this clarifies, or even justifies your improper use of "units" instead of milliamps? :roll:

If you actually want help, lose the attitude and stop attacking the people who are trying to help you.
My mind is like a browser. 27 tabs are open, 9 aren't responding,
lots of pop-ups...and where is that annoying music coming from?

Return to “Troubleshooting”