DM365 DVEVM - No RS-232 Console via DB9 - running blind!

NickAllen said:

My question is thus as follows, should I be seeing anything from the DB9 even if the DM365 cannot find the NFS server? Surely I will still see some uBoot output?

Yes, you are correct, you should be seeing UBL and U-Boot output regardless of the status of the network, assuming they are properly burned into flash as they should be, and that you are setup in the right boot mode to boot them.

NickAllen said:

If I should be seeing some comms, I am concerned - should the next logical thought be that something has gone wrong with the HW? The DVEVM accompanied me home a number of times (protected) during optimistic weekends when I was hoping to get a chance to continue playing. I am now forced to consider whether it may have been damaged. How could I test this? What is my next "best bet" step?

From here there a few steps you can take to try to fix this or at least figure out what is wrong:

1. The first thing I would do is verify that all of the DIP switches are set to their correct default values for NAND booting (SW4 all off and SW5 all off except for switch 5 for CPU voltage), if you somehow switched some switches and are in some other boot mode that could be a problem.

2. If the above does not work than try setting SW4 switches 2 and 3 to on, this should put you into the UART boot mode on the DM365, and if it is working you will see some repeating output on your serial terminal (something like '...Pôtig)Pôtig)...'). Though this does not actually fix the problem, this step will prove that the DM365 itself is alive, and that your serial port is actually functioning. If nothing comes out at this point you probably want to verify your power supply and UART cabling as well as that your serial terminal program is in a good state, beyond these you may be looking at something really wrong with the board (ESD, damaged power supply, pcb scratch/crack, etc...), in which case you may need to look into buying a new one or for future readers if this is happening out of the box new you should obtain an RMA from Spectrum Digital.

3. If the UART boot mode does respond, but you get no boot when in NAND boot mode from the first step than your flash is probably corrupted or otherwise empty. Reflashing the DM365 is a bit trickier than other boards as there is no serial boot recovery utility available due to a bug in the ROM boot loader (if you are curious that Pôtig)Pôtig) should be BOOTME BOOTME, but is not due to an incorrect UART clock). There is still hope however, as there are a couple of options I have seen people use, primarily the use of CCS/JTAG to recover the flash, or the SD card boot utility put together by Constantine Shulyupin.

Bernie Thompson said:

verify that all of the DIP switches are set to their correct default values for NAND booting (SW4 all off and SW5 all off except for switch 5 for CPU voltage),

Done: All settings were good.

Bernie Thompson said:

try setting SW4 switches 2 and 3 to on, this should put you into the UART boot mode on the DM365, and if it is working you will see some repeating output on your serial terminal (something like '...Pôtig)Pôtig)...'). Though this does not actually fix the problem, this step will prove that the DM365 itself is alive, and that your serial port is actually functioning. If nothing comes out at this point you probably want to verify your power supply and UART cabling as well as that your serial terminal program is in a good state, beyond these you may be looking at something really wrong with the board (ESD, damaged power supply, pcb scratch/crack, etc...)

I set bits 3 &4 on SW4 as above. I unfortunately still have nothing.
Power Supply: 5.2V 3A
UART Cabling: An FTDI cable but using a scope, there is just plain nothing on the DB9 connector. I can't fault the cable on this unfort.

What should the LEDS labelled DS2 through DS9 be doing on normal powerup? I have all 8 solid ON. I cannot recall if any of these were possibly previously flashing.

Bernie Thompson said:

If the UART boot mode does respond, but you get no boot when in NAND boot mode from the first step than your flash is probably corrupted or otherwise empty. Reflashing the DM365 is a bit trickier than other boards as there is no serial boot recovery utility available due to a bug in the ROM boot loader (if you are curious that Pôtig)Pôtig) should be BOOTME BOOTME, but is not due to an incorrect UART clock). There is still hope however, as there are a couple of options I have seen people use, primarily the use of CCS/JTAG to recover the flash, or the SD card boot utility put together by Constantine Shulyupin.

I have taken a look at the hyperlink. Ta. I will give it a proper read in the morning. Can I still attempt to implement this even if I received no comms in your previous UART suggestion?I have a 8GB SDHC card though. Is SDHC supported by the DM365?

While I am on this line of queries: where is the core bootloader actually stored? The ROM in the DM365 (reading from your comment)? On the NAND (I'd assume not)? Please feel free to treat me like a complete noob! What is the one-NAND on this board used for?

And in anticipation of the worst: Is Rev D the latest rev of DVEVM? Is the serial boot mode bug fixed? And are all DM365's shipped with the ROM boot preloaded?

Thanking you in advance for your time! I really do appreciate your input.
Rgds Nick

NickAllen said:

I set bits 3 &4 on SW4 as above. I unfortunately still have nothing.
Power Supply: 5.2V 3A
UART Cabling: An FTDI cable but using a scope, there is just plain nothing on the DB9 connector. I can't fault the cable on this unfort.

What should the LEDS labelled DS2 through DS9 be doing on normal powerup? I have all 8 solid ON. I cannot recall if any of these were possibly previously flashing.

Just to be clear, the switches should look something like the image below, if your board is powered and the switches are in this position you should see the repeating serial output, if not I am not sure there is much else that can be done.

As to the LEDs they should be on solid, this is normal.

NickAllen said:

Can I still attempt to implement this even if I received no comms in your previous UART suggestion?I have a 8GB SDHC card though. Is SDHC supported by the DM365?

You can certainly attempt it, however the probability of success is not so high, as if you cannot get into UART boot mode or see UART output the SD boot may not work either. SDHC support is trickier, the RBL will not support this so you could not use a SDHC card to boot, however the software stack in Linux can be made to handle SDHC cards.

NickAllen said:

While I am on this line of queries: where is the core bootloader actually stored? The ROM in the DM365 (reading from your comment)? On the NAND (I'd assume not)? Please feel free to treat me like a complete noob! What is the one-NAND on this board used for?

The first stage boot loader is stored in the ROM of the DM365, you may hear it called as RBL or ROM boot loader, it is masked into the chip during manufacturing and is not something that can be changed. The NAND will typically store the secondary boot loader(s) and potentially the OS image and application files, in a fully stand alone system the boot process would go RBL -> UBL -> U-Boot -> uImage / Kernel -> file system / application, where everything past the RBL is in the NAND. By default the board is booted out of NAND, I don't believe the OneNAND is being used in a typical boot case but is rather just there to be available to developers if needed.

NickAllen said:

And in anticipation of the worst: Is Rev D the latest rev of DVEVM? Is the serial boot mode bug fixed? And are all DM365's shipped with the ROM boot preloaded?

It looks like D is currently the latest revision. The serial boot mode bug is fixed with rev 1.2 silicon, however I am not sure if Spectrum Digital is using these parts for the shipping EVM boards, you may want to ask them to verify if this is a concern. All DM365 devices come with the RBL in place and set in silicon.

Hi Bernie, Thank you for your feedback & for your support / input. I have still not had any luck with this board and have requested a new board through our local TI channels. There appears to be a Revision E in stock in the eStore. We are battling to determine what Rev Silicon it contains. I will start a new thread though to query the collective brainpower of the TI community!

Hi Nick, I am searching the forum regarding usage of OneNAND and am curious about your board not working.  I have two revD boards and found the one doesn't start up beacuse the stupid socket the NAND sits in doesn't connect properly.  Opening and closing it after shuddering seeing the state that the IC's pins are in, it boots again but be careful as I broke the one's socket and now its just sitting there by grace.

Cheers, Jinh T.