Samb11 ULP Power Consumption Wildly Varies vs. VBAT/VDDIO Voltage

Go To Last Post
4 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi,

 

I have a custom PCB using the ATSAMB11-MR210 module. It's working fine and goes into ULP mode perfectly and wakes up with AO_GPIO_0 interrupts and AO_TIMER as expected.  However, there's something very odd going on with ULP sleep current readings.  I'm using a Fluke 45 bench DMM to take these readings, which are fairly accurate.  Here's what I see at various VBAT/VDDIO voltages (as powered by different CR2032/LI2032 coin cells):

 

On a CR2032 battery at 3 VDC or less, ULP draws just 2 to 3 microamps, as it should and is very stable.

On a CR2032 battery at 3.3 VDC, ULP draws about 65 to 80 microamps, similar to what I see on Xplained Pro board using Data Visualizer and is somewhat unstable.

On a LI2032 battery at 3.8 VDC, ULP draws about 300 to 400 microamps! It's fluctuating a lot.

 

The only thing I'm changing out here is the battery. Both VBAT and VDDIO are tied to VCC3, a single power bus on the PCB.  These voltages are well within the specified ranges for both VBAT and VDDIO for the AtSamb11 SoC and MR210 module.

 

I haven't been able to find a schematic of the ATSAMB11-MR210, so not sure what else may be going on here.  This has me stumped.  Here's a link to the MCU schematic.

 

Has anyone seen this behavior or know how to approach troubleshooting?

 

Rick

 

P.S.  I originally thought this was a ASF power management issue and posted this.

 

Last Edited: Mon. Sep 25, 2017 - 03:29 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Have you got some capacitors with too-low voltage rating ?

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The only cap I have on my PCB is a 10 uF 6V cap and we are well below that level at 3.8 V.  I also removed that cap and no change.  I don't know exactly what's on the MR210 board, but one would think it's capable of these low level supply voltages that are less than the maximum specs.

 

The current definitely scales up exponentially with an increase in voltage.  Using my bench supply, I varied the voltage from 2.5 volts (reads 2 microamps ULP current) up to 4 V, which reads nearly 450 microamps... by only varying the supply voltage. In this test configuration, all I have on my PCB is the SAMB11-MR210 board plus CR2032, a few pull-up resistors and a MOSFET for TTL leveling. I tried removing the MOSFET and it's not the cause.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I now suspect some undocumented behavior from the built-in DC/DC buck converter inside the Samb11, as there's no other logical explanations, given the simplicity of my test setup. Of course, if true, it would mean the Samb11 specs are false...

When a part is spec'd to support up to 3.6V (VDDIO) and 4.2 for MCU, with a ULP sleep current of a few microamps, it should deliver it. I would not have selected this Atmel (Microchip) component but for the low power operation, as many others I'm sure.

I'd sure love to find a mistake in my prototype or firmware that could cause this but given it's simplicity I don't see it.

Alas, I'll probably solve it by switching from LI2032 (3.6V) to ML2032 (3V) and move on.

Can't even register a new Atmel Support account that can be logged into to get help.

Disappointing.