Is that how the Ti bsp is setup?
I was considering doing a yocto build for it. Only tried it on scarthgap and it had some issue and would not build(my fault for getting ahead of them) and will need to make some changes. Is that due to lack of emmc or just simple code tweak?
@foxsquirrel - unfortunately no. The board that I have tends to get quite hot after a few minutes of trying to boot (flashing green led) so I guess I’ll consider a RMA/exchange today as well after trying a few more times.
If memory is correct it was about 60 or so mA when intramfs/green heart beat are up. Then on the board that boots it was running around 90 or so mA, making an assumtion the additional current is from the hdmi chip. Its hard to even begin to guess at what might be wrong, what ever it is both our boards have the same problem. To generate that much thermal energy all the cores must lite up, at least that is the first thought in my mind. The running board is not too bad, I did place some small heatsinks on all the chips to bring down the avg temp. NVMe is running around 750 mbs and it might not need a heatsink at that speed.
Just had another guess, might be their firmware/software for the secure boot using t3boot is in race condition. That would explain the high thermal. Ti is cloak and dagger with that secure boot, at least I have not been able to find any source code for the ti3boot.
@foxsquirrel from my tests yesterday, after about an hour of the flashing-led situation, the entire board (including the usb and ethernet ports) had gotten quite hot - I was trying to hold the power button to shut down the board and that was getting difficult (edit) because of the board temperature.
I am trying another run now after manually setting the ‘user_password’ in the sysconfig.txt (?) file and notice that the cpu, the kingston 4086414, and the tusb8041 are currently the hottest, in this order; on shutdown just now, the ports were also getting warm even though they don’t seem physically close to the above 3 chips.
I will go ahead and request an rma from the link that @RobertCNelson posted and hope this works out. Thanks for the help.
This is the same board with the funky pink and green screen out of the box. Next day it was fine, it was up and running. Then progressively has gotten worse and would not boot. I assumed the monitor issue was just a connector problem, wiggled it and it worked. However, after that the boot issue started. Are they related, ???.
Having the same issue here, along with now the beagley imager.exe is not opening
edit: I’m glad we are discussing and talking about this issue today, because I have the BeagleyAI, The Beagley Imager is not launching for me, and when I get the OS onto a MicroSD. No booting, but two rapid green flash, and pause. Then a repeat
You can use the raspberry pi imager. The code for the imager is difficult to build for a windows box. If you have Debian or Ubuntu 22.04 it will work perfectly.
@RisingStar So the rpi-imager doesn’t know how to edit sysconf.txt… Please open the microSD (first fat partition) in windows/mac/linux and please change at-least your password.
(otherwise it’s the same flashing software underneath)
So couple reasons, this am67a soc is brand new, so while it shares the same microSD controller of previous TI k3 devices, it’s a new physical layout, (same ip, new routing, maybe tweaked fab etc)… One of the big things in the Linux sdhci driver, are otap values and other variables that are used to tune the controller… In the old days when microSD was really just spi this wasn’t an issue, but with SDR and DDR high speed signaling… Now things need to be tuned, and sadly the first few 100 users always have new never tested microSD’s!
I’ve pinged our TI contact, and also ordered that Lexar microSD for local testing…