grp62
Posts: 32
Joined: Fri May 23, 2014 7:41 am
Location: Lyon, FR

Compute module boot issue (solved)

Mon Aug 11, 2014 9:13 am

Hi all,
We designed a carrier board for the CM and we noticed that the CM boots raspbian after few minutes.
We checked all the power supplies, power sequencing, RUN signal state...
Our schematic is very similar to the CMIO board about the EMMC_DISABLE_N and USB flashing circuitry.
However all the GPIOs are used in our design and we are wondering if some GPIOs are latched on reset and which could modify the boot sequence.
greg

Briarios
Posts: 119
Joined: Thu Feb 21, 2013 11:35 am
Location: Perth, Western Australia
Contact: Website

Re: Compute module boot issue

Tue Aug 12, 2014 11:44 pm

I think there are 4 GPIO's that you can't really use during boot up, can't think of which ones they are atm should be in the CMIO schematic.
http://rpi.pmfg.me -- Test board Eagle Cad Files.

grp62
Posts: 32
Joined: Fri May 23, 2014 7:41 am
Location: Lyon, FR

Re: Compute module boot issue

Wed Aug 13, 2014 8:09 am

Hi Briarios,
Thanks for your feedback.
Do you mean GPIO28/29 and GPIO44/45 ?
According to the datasheet there's no internal pull U/D on these GPIO that's the reason we find pull-down resistors on the CMIO schematic. For info these pull down resistors are populated in our design.
Greg

Briarios
Posts: 119
Joined: Thu Feb 21, 2013 11:35 am
Location: Perth, Western Australia
Contact: Website

Re: Compute module boot issue

Wed Aug 13, 2014 10:29 am

grp62 wrote:Hi Briarios,
Thanks for your feedback.
Do you mean GPIO28/29 and GPIO44/45 ?
According to the datasheet there's no internal pull U/D on these GPIO that's the reason we find pull-down resistors on the CMIO schematic. For info these pull down resistors are populated in our design.
Greg

ahh yes those are the ones, and it looks like you've covered it so... sorry can't be much help for you atm, still waiting on my design to be made up. so i will check that when they come.
http://rpi.pmfg.me -- Test board Eagle Cad Files.

gsh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 1454
Joined: Sat Sep 10, 2011 11:43 am

Re: Compute module boot issue

Wed Aug 13, 2014 11:23 am

grp62 wrote:Hi all,
We designed a carrier board for the CM and we noticed that the CM boots raspbian after few minutes.
We checked all the power supplies, power sequencing, RUN signal state...
Our schematic is very similar to the CMIO board about the EMMC_DISABLE_N and USB flashing circuitry.
However all the GPIOs are used in our design and we are wondering if some GPIOs are latched on reset and which could modify the boot sequence.
greg

Can you put a video on dropbox? Do you actually mean minutes?

Have you tried changing the state of the eMMC enable bit? Does it go into USB booting mode? (This will normally take an extra 5 seconds)

Gordon
--
Gordon Hollingworth PhD
Raspberry Pi - Director of Software Engineering

User avatar
Cancelor
Posts: 759
Joined: Wed Aug 28, 2013 4:09 pm
Location: UK

Re: Compute module boot issue

Wed Aug 13, 2014 11:31 am

I don't understand what you are asking.

Do you want it to boot raspbian?
Are you saying it is slow booting?
Or are you saying it boots when you don't want it to boot?
Or are you saying it doesn't boot even if you wait a few minutes?
Can't find the thread you want? Try googling : YourSearchHere site:raspberrypi.org

grp62
Posts: 32
Joined: Fri May 23, 2014 7:41 am
Location: Lyon, FR

Re: Compute module boot issue

Wed Aug 13, 2014 4:37 pm

Hi,
The compute module has been flashed from a CMIO board with the lastest and greatest raspbian image available from the rpi website.
When i plug this pre-flashed CM on the CMIO board it boots normally, i mean after few seconds.
Now when i plug the same pre-flashed CM on my own carrier board, it boots after few minutes (about 7mins).
Then we ran an rpi-update on the pre-flashed CM plugged into our carrier board and after a reboot we noticed that the board starts immediatly (i mean 35s) and boots like the pre-flashed CM plugged into the CMIO board.
Gordon, what do you mean by changing the state of the eMMC enable bit ?
Greg

gsh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 1454
Joined: Sat Sep 10, 2011 11:43 am

Re: Compute module boot issue

Wed Aug 13, 2014 4:43 pm

We've changed the default pulls / alternate settings recently so it could be that...

The eMMC disable input to the compute module is used to disable the clock to the eMMC so it will (on purpose) fail to respond to SD requests at boot time and therefore fall back to USB boot mode

Gordon
--
Gordon Hollingworth PhD
Raspberry Pi - Director of Software Engineering

grp62
Posts: 32
Joined: Fri May 23, 2014 7:41 am
Location: Lyon, FR

Re: Compute module boot issue

Wed Aug 13, 2014 5:02 pm

Gordon,
According to the schematic, the eMMC disable input of the CM allows to isolate the eMMC CMD signal when we enable the USB BOOT mode.
I'm not sure to understand what you mean.
Greg

gsh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 1454
Joined: Sat Sep 10, 2011 11:43 am

Re: Compute module boot issue

Wed Aug 13, 2014 6:10 pm

I don't think it matters, if that was the problem then you'd see it drop into USB boot mode and stay there forever.

Try leaving it in the 7 minute booting state and go around all the pins with your finger to see if you can trigger the boot to occur...

I think the problem is either your power supply sequencing causing a latch up that is taking time to get out of or the RUN signal is floating low and is taking 7 minutes to charge up to a 1 (which a finger should make happen quicker)
--
Gordon Hollingworth PhD
Raspberry Pi - Director of Software Engineering

grp62
Posts: 32
Joined: Fri May 23, 2014 7:41 am
Location: Lyon, FR

Re: Compute module boot issue

Thu Aug 14, 2014 7:23 am

Gordon,
Thanks for your feedback.
But why after issuing an rpi-update then our board boots normally everytime ?
The RUN pin doesn't float (CM schematic) and all the GPIOs have internal pull resistors (except gpio28/29 and gpio44/45 but we have populated PD resistors on our carrier board)
I will also check the power sequencing but i don't understand why the behaviour will be different if it was a power sequencing and/or gpio floating.
Greg

grp62
Posts: 32
Joined: Fri May 23, 2014 7:41 am
Location: Lyon, FR

Re: Compute module boot issue

Mon Aug 18, 2014 7:36 am

Gordon,
Where can i find some information about the modification on the default pulls / alternate settings.
Greg

gsh
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 1454
Joined: Sat Sep 10, 2011 11:43 am

Re: Compute module boot issue

Mon Aug 18, 2014 7:42 am

--
Gordon Hollingworth PhD
Raspberry Pi - Director of Software Engineering

ShiftPlusOne
Raspberry Pi Engineer & Forum Moderator
Raspberry Pi Engineer & Forum Moderator
Posts: 6124
Joined: Fri Jul 29, 2011 5:36 pm
Location: The unfashionable end of the western spiral arm of the Galaxy

Re: Compute module boot issue

Mon Aug 18, 2014 8:20 am

Please note that the documentation is slightly wrong right now.
https://github.com/raspberrypi/documentation/issues/112

db454
Posts: 2
Joined: Thu Jan 08, 2015 4:45 pm

Re: Compute module boot issue

Thu Jan 08, 2015 5:04 pm

Greg,

Did you ever solve your boot issue?

I too have built a PCB for the CM and have the exact same issue with the ~7 minute boot.

Briarios
Posts: 119
Joined: Thu Feb 21, 2013 11:35 am
Location: Perth, Western Australia
Contact: Website

Re: Compute module boot issue

Thu Jan 08, 2015 11:58 pm

do an rpi-update( i think its that) command after it eventually boots, i had to do this last night and now it boots straight away. :D
http://rpi.pmfg.me -- Test board Eagle Cad Files.

db454
Posts: 2
Joined: Thu Jan 08, 2015 4:45 pm

Re: Compute module boot issue (solved)

Fri Jan 09, 2015 3:18 pm

I did do the update that works, I was concerned why it does not work before the update
David

Briarios
Posts: 119
Joined: Thu Feb 21, 2013 11:35 am
Location: Perth, Western Australia
Contact: Website

Re: Compute module boot issue (solved)

Sat Jan 10, 2015 2:30 pm

Crazy I know but I'm sure Gordon can Explain it better when he has chance
http://rpi.pmfg.me -- Test board Eagle Cad Files.

Return to “Compute Module”