AM3358: SD card boot fails

Lars Baltzer

Part Number: AM3358

I am trying to get my custom board up and running but it fails to start from the SD-card. My custom board is based on the BBB board with only minor modifications. Boot configuration is the same as the BBB.

I have followed the AM335x bringup tips and all voltage and frq (also CLKOUT1 after enabling it) is ok as well as PORZ.

When powering up the board I can measure data on the CLK pin of the SDcard and if a valid sdcard is inserted there is also data on data0 but it dies out short time after and nothing more happens. Uart0 is not available for measureing so I can't check if there is a "C" output. The SD-card has been tested on a BBB so I know for sure that it is working.

over 6 years ago

0 Biser Gatchev-XID over 6 years ago

TI__Guru**** 393215 points

Hi,

Can you change the boot sequence so that MMC0 boot is first? What software are you trying to boot? Do you have any debug interface on the board?

0 Lars Baltzer over 6 years ago in reply to Biser Gatchev-XID

Prodigy 110 points

Thanks for your feedback.

I have tried to change the boot sequence now with same result.

There is a Jtag interface and I have access to a J-link and have tried with J-link commander but get the following error message. TDO is constant high. I have measured the pin and it is indeed high all the time. As the tdo pin goes directly into the processor I am a bit puzzled what the problem could be here.

-Lars

0 Biser Gatchev-XID over 6 years ago in reply to Lars Baltzer

TI__Guru**** 393215 points

One important question here: Do you have external 4.7kOhm pullups to VDDSHV6 on EMU0 and EMU1 signals?

0 Lars Baltzer over 6 years ago in reply to Biser Gatchev-XID

Prodigy 110 points

Yes buth EMU0 and EMU1 has a 4k7K ohm pullup resistor.

0 Biser Gatchev-XID over 6 years ago in reply to Lars Baltzer

TI__Guru**** 393215 points

Does your J-link control EMU0 and EMU1? These are the two signals that the emulator uses to enable the device internal debug circuitry.

0 Biser Gatchev-XID over 6 years ago in reply to Biser Gatchev-XID

TI__Guru**** 393215 points

For normal operation both EMU0 and EMU1 must be high at reset release time.

0 Lars Baltzer over 6 years ago in reply to Biser Gatchev-XID

Prodigy 110 points

EMU1 is pulled low when connecting the J-link but I dont have much experience with the J-link debugger. My software guy is not available at the moment so I have to figure it out somehow.

0 Biser Gatchev-XID over 6 years ago in reply to Lars Baltzer

TI__Guru**** 393215 points

See section 27.1.2.2.2 of the AM335x TRM Rev. P. EMU1 = 0 is a reserved situation and should not be used.

0 Lars Baltzer over 6 years ago in reply to Biser Gatchev-XID

Prodigy 110 points

Hi Again

Got the Jtag up and running and I can connect via J-link commander but this only proves that I can connect.

What would be the next step to check why the processor is not booting from the SD-card.

-Lars

0 Biser Gatchev-XID over 6 years ago in reply to Lars Baltzer

TI__Guru**** 393215 points

What software are you trying to boot?

0 Lars Baltzer over 6 years ago in reply to Biser Gatchev-XID

Prodigy 110 points

Debian for BBB and the sw is starrting on the original BBB board

0 Biser Gatchev-XID over 6 years ago in reply to Lars Baltzer

TI__Guru**** 393215 points

I'm sorry, Debian is not supported by TI. We recommend using the Linux SDK, which can be downloaded here: www.ti.com/.../PROCESSOR-SDK-AM335X

0 Biser Gatchev-XID over 6 years ago in reply to Biser Gatchev-XID

TI__Guru**** 393215 points

For Debian support you should ask on www.beagleboard.org

0 Lars Baltzer over 6 years ago in reply to Biser Gatchev-XID

Prodigy 110 points

Used your link and downloaded the Linux SDK, transferred it to the SD-card and tried with same result. The board is still not booting but clk pulses to the SD-card is there for brief second as before.

0 Biser Gatchev-XID over 6 years ago in reply to Lars Baltzer

TI__Guru**** 393215 points

OK, follow these steps: processors.wiki.ti.com/.../AM335x_board_bringup_tips

0 JJD over 6 years ago in reply to Biser Gatchev-XID

TI__Guru 68880 points

Lars, here are a few things to try:

-after a failed boot, connect with JTAG and record the PC (program counter). Do this several times and see if you are ending up in the same location (maybe a exception fault) or at different locations (ie, it is re-attempting to boot because it doesn't find a valid boot file).

-when you probe MMC_CLK, do you see it attempt to boot multiple times, or it tries just once and then you don't see any more bursts?

-do you have a series resistor on the MMC_CLK signal, close to the processor? This is required to reduce reflections because this clock is looped back to the MMCSD controller to retime the read data

Regards,

James

0 Lars Baltzer over 6 years ago in reply to Biser Gatchev-XID

Prodigy 110 points

I have tried to follow the guide but get the following error message in the console window:

Could not open session. No matching devices found. (C:\ti\am335x-boot.dss#60)

Picture of CCS window. Any Ideas where to look?

0 Lars Baltzer over 6 years ago in reply to JJD

Prodigy 110 points

Hi James
As soon as I understand why I can't get CCS to connect via J-link and will try the PC count.

For the MMC_CLK I only see one burst and then it dies out.
I dont have a series resistor on the MMC_CLK signal but I will look into it to see if I can insert one.

-Lars

0 Biser Gatchev-XID over 6 years ago in reply to Lars Baltzer

TI__Guru**** 393215 points

Please check what JJD suggested above.

0 Lars Baltzer over 6 years ago in reply to Biser Gatchev-XID

Prodigy 110 points

Got the boot analycis up and running and the PC counter gets stuck at the same place at 0x402f40e4. And from the TRM I can see that it then get stuck at the internal SRam

Anyone knows whats it looking at here?

-Lars

0 Biser Gatchev-XID over 6 years ago in reply to Lars Baltzer

TI__Guru**** 393215 points

This should mean that the MLO is loaded, started executing and stopped in there for some reason.

0 JJD over 6 years ago in reply to Biser Gatchev-XID

TI__Guru 68880 points

Yes, that means at least the ROM is executing properly, MLO is getting downloaded to internal SRAM, and something is going wrong and branching to an exception vector.

This is actually good because it will make debugging easy. In CCS, you can use the file u-boot-spl as the symbol file to perform source code correlation. When the code is stuck at 0x402f40e4, load up the symbol file to give you better visibility as to where the code is stuck. I haven't used the specific emulator you are using, but you should be able to see the symbols in the disassembly window, and also the source code if you point it to the source directory.

Most likely, you will have to reset and step through the first part of the bootloader to find out why it is getting there. Be sure to use hardware breakpoints, since the code you will be debugging is relocated into internal SRAM.

Regards,

James

0 Biser Gatchev-XID over 6 years ago in reply to JJD

TI__Guru**** 393215 points

Is this resolved? If so, please verify, so we can close the thread.

0 Lars Baltzer over 6 years ago in reply to Biser Gatchev-XID

Prodigy 110 points

It has not been resolved but I have not had the time to investigate and it will be 2-3 weeks before I can look at it again.

I will try the suggested things but as this will take time I suggest that you close this thread and I can open a new when I have time again.

-Lars

Processors

Processors forum

AM3358: SD card boot fails