TPS65930 and DM3730 Not booting

Hi,

NRESWARM is not a reset signal that needs to be asserted at power-up. This is a signal controlled by DM part to put TPS65930 in a warm reset state (same as a soft reset). So, I dont think you need to worry about this signal when you power up. This is an input to the PMIC.

Hope this helps.

I also do not see any activity on any of the peripherals indicating an attempt at boot. It seems possible that nRESWARM won't be released by DM until later on in boot process. I just assumed this would happen almost immediately after the global reset was released.

Are the power supply rails for the DM3730 coming up in a consistent manner to the power-up sequencing requirements of the DM3730 described in the datasheet in Figure 3-2?  I would also suggest looking at both the 32KHz clock input and the high-frequency clock input (sys_xtalin, etc.) to ensure those are stable prior to the release of the sys_nrespwron input.

Is there any activity on the DM3730 I2C4 interface between it and the TPS65930?  This would give you some clues if the DM3730 is actually alive.

As best I can tell the sequencing is all fine. I do have the TPS65930 connected to the I2C1 interface, not I2C4. There is no activity on I2C1, and  I have no way of probing I2C4.

Hi Jon,

When you say "As best I can tell the sequencing is all fine".

Do you mean that all expected power (power up sequence on PMIC) rails on TPS65930 are powering up fine but you do not see DM part powering up?

If NRESPWRON is asserting (transitioning to high) then I would expect DM part to power ON. If you see no activity on I2C1 then you are not sending any commands to PMIC. So, NRESWARM shouldnt be causing any problem.

How are you booting the platform, any peripheral boot - MMC, USB, etc. Is DM part getting the correct boot file. Can you check the sysboot option on DM. I dont know details about this. If PMIC is providing power then I would suspect some issue with DM boot logic, something similar to the sysboot configuration.

When you say "As best I can tell the sequencing is all fine".

Do you mean that all expected power (power up sequence on PMIC) rails on TPS65930 are powering up fine but you do not see DM part powering up?

I can see all of the rails coming up fine. I have taken some measurements to determine there timing as compared to nRESPWRON and everything seems ok. I do not have the equipment necessary to quantify  the stability of the clock sources, but 32k seems to become stable within just a few cycles. I am using a 26mhz clipped sin oscillator which is potentially picking up system noise and introducing jitter in the clock. 

nRESPWRON is going high, but there are no signs of life from DM except for the internal LDO's powering up. I have checked SYSBOOT several times and it is configured for MMC1, USB, UART3. I see no activity on any of those interfaces. MMC1 clock does not toggle, ULPI has no clock, and there is no data on UART3.

If PMIC is powered on correctly then you are having some problems with DM for sure.

How does DM receive 26MHz clock? You said you have a clipped sine wave. Normally the clock is as follows:

26M oscillator -> PMIC HFCLKIN -> PMIC HFCLKOUT -> DM37xx

Do you see 26M on HFCLKOUT pin from PMIC?

One more comment - HFCLKOUT (input to DM3730) should be present before NRESPWRON is asserted. If this is not the case then DM part will not function as expected. In your case not seeing any activity on DM is a bit strange.

Yes it is wired Osc->HFCLKIN HFCLKOUT->DM. The 26mhz square wave is present on this line.

I found a problem with the HF oscillator. I am using a Pericom UJWIFI026 which has an enable to clock out of 10ms. Obviously larger than the 5.3ms max required. On the scope I do see HFCLK starting at least a few cycles before nRESPWRON going high, so not sure if this is actually a problem on my particular board.

I think this is your problem. Try another oscillator if you can or try to enable the oscillator pin with an external source that will provide output much before or just after power up.

What is the enable on the oscillator at this point?

I tied the oscillators enable to VIO in an effort to get faster startup. Not sure if it is fast enough. I will try externally powering it so it is running long before powering the board.

Hi Jon,

I am also facing the same problem of not seeing any activity on MMC1 USB & UART, have u able to solve that problem? i have also connected 26MHz oscillator OE to VIO1.8 from TPS65950.

bhupesh

An unmodified design later worked on another build of the PCB. The new PCB's were made somewhere else, assembly process was different so its difficult to say exactly what the problem was. I suspect that the particular build I was mentioning actually did work because we were detecting 3.3v on VDD_MCC. I now know that this only can happen from successful boot rom boot and i2c communication with the PMIC. Its probably just that our scope wasn't triggering and we later butchered a perfectly good assembly.

Thanks Jon,

Is there any way u can share the chematic of DM3730 & TPS65950 or i can send you our schematic to see whether we are missing some connections.

bhupesh

Hi All,

I am also facing the similar kind of problem.

VIO, VDD1, VDD2, HFCLKOUT and 32K CLKOUT are properly coming from TPS65950 to DM3730. RESPWRON is also released.

Still i am not seeing any activity on the processor.

I am using DM3730 CBP package with PoP memory and trying to flash through UART. I am using Mode2 function UART3 which is muxed with HSUSB0_D0 and D1 as a boot uart as per the EVM.

Please let me know if there is anything else that i have to take care.

Thanks,

Bharathi

I dont see anything missing from the power perspective. Please check connections on the board.