CC8520: Seems to immediately "wake" from power off

Arthur Thompson27

Part Number: CC8520
Other Parts Discussed in Thread: TLV320DAC32

Hi,

We're using the CC8520 as a SPI slave and we need to be able to configure it to an off state via the external host interface to minimize power draw. We have been seeing an issue where after sending the pm_set_state OFF command the current consumption goes down to near zero, and then comes back up again to a lower (~20mA but not off) current draw after a few seconds.

Some observations

the OFF command causes the crystal to turn off and stay off
the OFF command causes the DCPL pin (digital LDO) to turn off and stay off
The RSTN line remains high consistently after the OFF command (no resets are generated)
The CSN line also remains high consistently after the OFF command
I don't see any activity on the GIO that's configured as the EHIF interrupt

I'm attaching an oscilloscope capture showing the current draw (green, measured via current probe) decreasing and then going back up, while the SPI CSN line remains inactive.

Any debugging advice to figure out what is going on and why the power consumption is only staying low for a few seconds?

Thanks,

Arthur

over 5 years ago

0 Fredrik K over 5 years ago

TI__Guru** 106705 points

Hi Arthur,

Hi have notified a PPW expert about your post.

Cheers,
Fredrik

0 Fredrik K over 5 years ago in reply to Fredrik K

TI__Guru** 106705 points

Hi Arthur,

I apologise for the late answer on this.

Can you share a scope capture of all 4 SPI signals when sending the PM_SET_STATE command?

Cheers,
Fredrik

0 Arthur Thompson27 over 5 years ago in reply to Fredrik K

Expert 1725 points

Here you go Fredrik,

Thanks,

Arthur

0 Fredrik K over 5 years ago in reply to Arthur Thompson27

TI__Guru** 106705 points

Thanks Artur,

This is only one byte though. Can you capture the whole command? Also, the first 16 bits (status word).

Regards,
Fredrik

0 Fredrik K over 5 years ago in reply to Fredrik K

TI__Guru** 106705 points

Looking at the CMD_REQ figure on page 90 in the CC85xx User´s Guide, can you verify that MISO is infact high before the command is initiated.

0 Fredrik K over 5 years ago in reply to Fredrik K

TI__Guru** 106705 points

One more thing: Could you try lowering the clock rate . What are you currently using btw?

0 Arthur Thompson27 over 5 years ago in reply to Fredrik K

Expert 1725 points

Hi Fredrik.

SPI clk is 10MHz and MISO is high before we begin any SPI writes. (Assert chip select, wait for MISO, then begin write).

Here is a picture with the whole last command.

Thanks,

Arthur

0 Fredrik K over 5 years ago in reply to Arthur Thompson27

TI__Guru** 106705 points

Hi Arthur,

As far as we can see there are no issues with the SPI communication. Are you sure the current is actually consumed by the CC85xx? Have you checked for floating IOs and or incorrectly configured pull-ups or -downs?

Regards,
Fredrik

0 Arthur Thompson27 over 5 years ago in reply to Fredrik K

Expert 1725 points

Hi Fredrik,

Since I have internal power vias, a direct current measurement of the CC8520 is not possible. However, I can let the system sit in this failed-power-down state and see a slight ~1 deg C temperature rise on the CC8520 chipset. The CC8520 is the warmest chip on the board in that state.

Thanks,
Arthur

0 Fredrik K over 5 years ago in reply to Arthur Thompson27

TI__Guru** 106705 points

Hi Arthur,

Do you have a CC85xx EVM which you can replicate this on? How many of your custom boards are you seeing this on? There is really no natural way for the CC85xx to behave as you describe, so it sounds like either a broken device or that the current is flowing somewhere else.

Regards,
Fredrik

0 Arthur Thompson27 over 5 years ago in reply to Fredrik K

Expert 1725 points

Fredrik,

We are working to replicate on an EVM.

Can you tell me how long the power down sequence is supposed to take to reach the minimum current consumption after the EHIF power down command is received? Is it on the order of a couple of seconds?

Thanks,
Arthur

0 Arthur Thompson27 over 5 years ago in reply to Fredrik K

Expert 1725 points

Hi Fredrik,

Also, we just found TI's example EHIF code here www.ti.com/.../litabsmultiplefilelist.tsp
We learned that the example power down code resets the CC85xx right before sending the power down command. When we tried this it seemed to help. Why would it be a good idea to reset the CC85xx before powering it down?

Thanks,
Arthur

0 Fredrik K over 5 years ago in reply to Arthur Thompson27

TI__Guru** 106705 points

Hi Arthur,

What do you mean by "it seemed to help"?

The OFF command should work regardless of doing a reset first or not.

Regards,
Fredrik

0 Fredrik K over 5 years ago in reply to Fredrik K

TI__Guru** 106705 points

Hi Arthur,

I have checked the power down timing, it should be in the order of a couple of milli seconds.

I have asked a few questions in this thread that are still unanswered. Can you please come back to me on:
- Do you see this behaviour on multiple boards, or just one?
- Have you verified that the current is not flowing through incorrectly configured pull-ups or downs (or similar)?
- Can you clarify what you meant by performing a reset before sending the off command "seemed to help"?

Also, not that it should matter, but is the behaviour dependent on which mode the device is in when sending the off command?

Regards,
Fredrik

0 Arthur Thompson27 over 5 years ago in reply to Fredrik K

Expert 1725 points

Hi Fredrik,

I believe I have some small amount of resolution. It seems to be related to the DAC slave that's connected to the CC8520. We were turning off the analog LDO that feeds the TLV320DAC32, which also feeds the core digital power LDO inside the DAC32. Meanwhile the I/O rail was staying high. Leaving the LDO powered on when entering the power down state has eliminated this behavior.

The explanation may be that when shut down, the CC8520 I2C pins were sinking current from the I2C pins on the TLV320DAC32 slave.

Thanks,

Arthur

0 Fredrik K over 5 years ago in reply to Arthur Thompson27

TI__Guru** 106705 points

Hi Arthur,

Thanks for the update. That sounds like something I would expect to be causing your problem. Let me know if you conclude on this being the root cause.

Regards,
Fredrik

Other wireless

Other wireless technologies forum

CC8520: Seems to immediately "wake" from power off