Chipworkx not booting reliably. Power supply rise times?

NETMF & Gadgeteer
1

I’m working on a custom board using the Chipworkx modules and a 7in LCD.
The unit has an internal AC-DC power supply with the AC line switched.
The power supply outputs 5vdc which is them regulated down to 3.3v for the Chipworkx.

What I’m seeing is that about 90% of the time when power is applied the Chipworkx will not boot and the LCD has a white or white/black striped image on the display.
Pressing the reset button does not reset the Chipworkx. The only solution was to unplug the Chipworkx with the power on and the plug it back in. Then it would boot normally.

I suspect that there is something relating to the rise time of the power supply to get to a steady state 5v/3.3v. I measured the rise time and it is 3.8ms from 0v to steady 3.3v, and 6.4ms from 0v to 5v

I added a switch inline on my board on the 5V rail between the power supply and the 3.3v regulator. If I give the power supply a few hundred milliseconds to stabilize then switch on the 5v rail, the Chipworkx boots normally every time.

Is this the expected behavior of the Chipworkx?
Is the rise time of the 3.3v rail too long (3.8ms)?

Any insights or assistance is appreciated.

Thanks
Brian

2

Hi Brian,

I can offer no real value here, but I would expect the people who could will want to know a few things that I might be able to help ask up front.

What is the current capacity of your power supply circuit? Are you sure you can reliably provide that without the voltage sagging? How much is the CwX drawing at startup, and does that change in the good and bad startups?

Do you have a voltage brown-out chip? If so, what one and if configurable what parameters are you checking for?

Does the app reliably start without the LCD connected?

Are there any diagnostics coming from the CwX when it fails to boot ? What is “visible” from MFDeploy and device manager on a connected PC?

3

Do you use LCD EN signal. I think it is PB3.

4

Brett:
PS is rated at 5v 2A. No voltage sag at startup.
No brownout chip.
Chipworkx doesn’t boot with or without LCD connected.
When it fails the module is completely unresponsive.

Gus:
Yes, I’m using LCD_EN (PB3)

5

IIRC, This pin also also BMS (boot mode select). If you disconnect the pin then does the board boot every time?

6

Disconnecting PB3 (BMS) has no effect.

I’m going to take a shot in the dark here…
According to the CPU (AT91SAM9261S) datasheet it requires several voltages below 3.3v. I assume the Chipworkx module has regulators to supply these voltages.
As I don’t know the exact configuration of the regulator circuits on the Chipworkx, is it possible that the rise time of the 3.3v power to the Chipworkx is allowing those other voltages to initialize in some random order that causes the CPU to hang?

from the datasheet:
VDDIOM EBI I/O Lines Power Supply Power 1.65 V to 1.95V and 3.0V to 3.6V
VDDIOP Peripherals I/O Lines Power Supply Power 2.7V to 3.6V
VDDBU Backup I/O Lines Power Supply Power 1.08V to 1.32V
VDDPLL PLL Power Supply Power 3.0V to 3.6V
VDDOSC Oscillator Power Supply Power 3.0V to 3.6V
VDDCORE Core Chip Power Supply Power 1.08V to 1.32V

7

Not sure why you would have problems as we never used anything extra special to power up the module and never had similar issues.

Would wiring 3.3 from the dev sys board to your board help in isolating the problem?

You may also use this in your testing if you like http://www.ghielectronics.com/catalog/product/280