LPC1857 not booting and getting very hot

[lpcware]

Content originally posted in LPCWare by paul.nichols@grabba.com on Sun Feb 21 19:25:33 MST 2016
Hi,

I am having issues with some of our custom PCB based on LPC1857. On power connection to the CPU and peripherals the processor will not boot and gets very hot. The processor pulls ~750mA at 1.1V (the 750mA draw is enough to hold our 3.3V switchmode down to 1.1V)

A quick removal of power followed by re-connecting power results in a successful boot each time. This error condition is not observed on all boards (approximately 50% of 20 board prototype run)

Notable Items:
[list]
  [*] If we have a Red Probe+ attached(through JTAG) when the battery is plugged the board boots everytime (the error condition is never observed)
  [*] With no power connected and the Red Probe+ JTAG interface connected the 3.3V rail pre-charges to ~1V,TDO & TDI are at ~1.5V and TCK has 50kHz non-continuous clock signal
  [*] If we have a LPC-Link2 attached (through JTAG) the problem condition is observed
  [*] Holding reset low during battery insertion/pwoer connection) has seemingly no effect on if the problem state is entered or not
  [*] The 3.3V rail has ~100uF of bulk capacitance on the output of the switch mode (3.3V)
  [*] The battery net has ~50uF of bulk capacitance at the input of the switch mode (3.3V)
  [*] There is no VBUS present at the time of test. The VBUS pin is 0V
  [*] DBGEN and TRST are tied to 3.3V through 10k pull-ups
  [*] TDO, TDI, TMS are tied to 3.3V through 10k pull-ups
  [*] TCK is tied to ground through a 10k pull-down
  [*] P2_7 (boot source) is tied to 3.3V through a 10k pull-up
[/list]

The problem seems to happen prior to the CPU booting (as holding reset has no effect on whether the problem condition is observed or not)

I have attached schematics of our power and JTAG connection

Any advice would be greatly appreciated.

Thanks
Paul

[Carlos_Mendoza]

Hi Gerrit,

Does the same problem was present when the MCUs were unprogrammed?

Could you please try flashing the boards presenting the issue with the periph_blinky example from the LPCOpen package and let me know if the problem is also present? This is to check if the software is in some way causing a short circuit.

http://www.nxp.com/products/developer-resources/software-development-tools/software-tools/lpcopen-li...

For flash-based parts, the LPC18xx boots from internal flash by default (boot pin P2_7 is HIGH). If the boot pin is sampled LOW on reset, the boot source is determined by the setting of the OTP or the states of pins P2_9, P2_8, P1_2, and P1_1. How are you configuring the part to boot?

Hope it helps!

Best Regards,
Carlos Mendoza
Technical Support Engineer

[lpcware]

Content originally posted in LPCWare by paul.nichols@grabba.com on Thu Mar 10 17:44:55 MST 2016
Unfortunately I haven't found a solution yet no.

I tried your suggestion of removing FB1 and FB2 but it didnt resolve the issue.

Any other ideas?


Thanks
Paul

[zjeriet]

Hello Paul

Were you able to resolve this problem?

We're experiencing the exact same thing on one of our custom PCBs.

Kind regards,

Gerrit

[Carlos_Mendoza]

Hi,

This is probably a problem with the logic levels between the computer and your target MCU, I would recommend you to review the following document about the design considerations for debug, especially the “Logic Levels and Ground” section. Let me know if this helps.

https://community.nxp.com/message/630601

Hope it helps!

Best Regards,
Carlos Mendoza
Technical Support Engineer

[zjeriet]

Hi Carlos,

Thanks for your reply.

We're using the Keil ULINKpro debugger and we're not facing the issue when the probe is connected.

The problem (on our side) appears when the board is running in stand-alone mode (no programmers connected)

There are not latch-up scenario's that could induce this problem.

We even removed all our peripheral HW IC's to check this.

We also tried different rise times on the power supply with no effect.

What we see is:

- Reset line of MCU is kept low at all time and the MCU is consuming lots of current. (MCU consumes >0.5A at 1.14V)

- Quickly power cycling the MCU will allow you to have a normal boot scenario

- Waiting for about 30s will put you back into the high current consumption scenario

- It's not isolated to one PCB. We have three PCB's and two are having this issue.

Kind regards,

Gerrit

[lpcware]

Content originally posted in LPCWare by bavarian on Wed Mar 09 07:13:42 MST 2016
Did you find a solution already?

I would take FB1 and FB2 out for a test. Inductors in the power supply path could heavily impact the startup behavior.

Regards,
NXP Support Team