CC2652R: Adapting Thread temp sensor example from MTD to SED

cabled

Intellectual 425 points

Part Number: CC2652R
Other Parts Discussed in Thread: ENERGYTRACE, LAUNCHXL-CC26X2R1

Hi gents,

With reference to this example

Am wondering what happens to the power state of the mcu during periods in between the reporting interval?

Based on my understanding, I think it remains powered, or at least the radio does since I am able to do coap polls.

Is there a way to put the mcu into some sort of low power mode between reporting cycles? It is ok for the functionality of the unit to change from MTD to SED if necessary (so the radio can be turned off).

over 7 years ago

0 Hector_r over 7 years ago

TI__Genius 12415 points

Your post has been assigned to an expert, expect a reply soon

0 cabled over 7 years ago in reply to Hector_r

Intellectual 425 points

anyone?

0 Hector_r over 7 years ago in reply to cabled

TI__Genius 12415 points

By default I believe the temp sensor is configured as an SED and it should be going to low power mode. Have you confirmed that this device doesn't go to sleep by measuring the power consumption?

0 cabled over 7 years ago in reply to Hector_r

Intellectual 425 points

I'm getting about 60mA on energytrace. Is this what I should be expecting?

0 cabled over 7 years ago in reply to Hector_r

Intellectual 425 points

anyone?

0 Severin Suveren over 7 years ago in reply to cabled

TI__Mastermind 23730 points

Hi cabled,

The EnergyTrace version on your LAUNCHXL-CC26x2R1 (pg1.1) is not suitable for current consumption characterization. For proper current consumption measurements you have to use a power analyzer.

However, you could potentially get better results by powering the chip with an external power supply instead of the XDS110.

0 cabled over 7 years ago in reply to Severin Suveren

Intellectual 425 points

Hi Edvard,

Noted on the energytrace. The chip is currently powered off an external battery and not the XDS110 yes. Would you be able to assist with my other question regarding the power mode/state based on the example programme?

0 Severin Suveren over 7 years ago in reply to cabled

TI__Mastermind 23730 points

Have you removed all jumpers from the XDS110 to the chip? Note that the UART RXD jumper should not be removed if it is configured in your application (it is by default).

0 cabled over 7 years ago in reply to Severin Suveren

Intellectual 425 points

I've removed the ground, 5v and 3.3v jumpers as well as moved the power jumper from xds to external. rxd is not removed.

0 cabled over 7 years ago in reply to Severin Suveren

Intellectual 425 points

anyone able to assist?

0 Severin Suveren over 7 years ago in reply to cabled

TI__Mastermind 23730 points

Hi cabled,

Sorry for the late reply. Can you take a picture of the setup?

0 cabled over 7 years ago in reply to Severin Suveren

Intellectual 425 points

Hi Edvard,

I assume you mean a picture of the jumper setup since that was what we were on previously. Please let me know if that's not the case.

At this point I'm getting about 10000 transmissions off a 2400mAh LiSOCl2. That's not great and based on figures I've been able to find seems to indicate taht the MCU stays on active mode between transmissions, hence my original question about whether code example.

0 YiKai Chen over 7 years ago in reply to cabled

Guru 735695 points

If you use power Analyzer to give 3.3V to measure power consumption, I suppose you have to remove all those jumpers between XDS110 and CC2652.

0 cabled over 7 years ago in reply to YiKai Chen

Intellectual 425 points

Hi YK,

Just to clarify, are we saying that the XDS110 (in addition to the cc2652) in my case is actually being powered by the external source, hence the seemingly higher consumption? And that removing all the jumpers should help with this?

i.e. I cannot monitor processes and do consumption gauging simultaneously?

0 YiKai Chen over 7 years ago in reply to cabled

Guru 735695 points

Those jumpers would impact power consumption measurements so please remove them when you do power measurements.

0 cabled over 7 years ago in reply to YiKai Chen

Intellectual 425 points

ok, thanks for the suggestion. I shall try that when i have time to run more tests.

Would you happen to have a rough idea on what sort of consumption overhead difference this might result in? Or is the connected XDS somehow preventing the mcu from switching power states?

0 YiKai Chen over 7 years ago in reply to cabled

Guru 735695 points

I don't have rough idea on what sort of consumption overhead difference this might result in and it supposes there are current leakages on those pins. I don't think the connected XDS somehow preventing the mcu from switching power states.

0 cabled over 7 years ago in reply to YiKai Chen

Intellectual 425 points

alright, thanks for your input. I'll give the suggestion a try and come back with an update either way.

0 cabled over 7 years ago in reply to cabled

Intellectual 425 points

Unfortunately it does not seem like removing all the jumpers makes a huge difference in the consumption.

which leads me back to my original questions...

1) Is the mcu actually supposed to go to standby between transmissions?
2) If it is, what's preventing this from happening and
3) if it hasn't been coded as such, is there an available reference on how to force a switch in power state?

The technical manual seems to imply that the mcu should move itself between power states but i'm guessing that being made to poll for button inputs and listen for coap transmissions might be preventing the switching.

0 Severin Suveren over 7 years ago in reply to cabled

TI__Mastermind 23730 points

If there is no required computation between transmissions, then yes the MCU will go to standby. Could you tell me which example your example is based on?

Have you tried following this guide? dev.ti.com/.../energy-trace.html

At least double check that you have configured it correctly.

Either way, if you want accurate power consumption characterization, use a power analyzer and not EnergyTrace. EnergyTrace will not give you accurate enough measurements to do power consumption characteristics.

0 cabled over 7 years ago in reply to Severin Suveren

Intellectual 425 points

Hi Edvard,

Yes I did read that energytrace guide. I'll double check it again.

I was working off this example.

I do understand the part about the analyzer. I'll arrange that next week. Just used energytrace as a ballpark because I'm consuming battery at a rate that seems to be more in-line with always-on mcu than standby mcu consumption.

0 Severin Suveren over 7 years ago in reply to cabled

TI__Mastermind 23730 points

The Temperature Sensor example is configured to be low-power by default. In that case, there must be something wrong in the setup of either EnergyTrace or the LaunchPad.

0 cabled over 7 years ago in reply to Severin Suveren

Intellectual 425 points

Hi Edvard,

with an analyzer and also back calculating from a consumed battery, i'm running around the range of 3.05 mA, which seems more consistent with always-on.

The only part of the program that is not stock is that the ipv6 address used by the sensor example has been changed to another address and the tx interval moved to 3 minutes.

Would you be able to offer any advice on what might be wrong with the launchpad setup? if necessary, I could send you the code...

0 Severin Suveren over 7 years ago in reply to cabled

TI__Mastermind 23730 points

Is the Thread device commissioned and joined to a Thread network?

0 cabled over 7 years ago in reply to Severin Suveren

Intellectual 425 points

yes, its joined to a thread network and sending its data to the thermostat example

0 Severin Suveren over 7 years ago in reply to cabled

TI__Mastermind 23730 points

Then the setup of the launchpad and the power analyzer must be setup correctly. Could you again show me the setup as well as some snippets of the power measurement results?

0 cabled over 7 years ago in reply to Severin Suveren

Intellectual 425 points

alright, i'll upload the images when I get back to my rig. am outstation for a couple of days. thanks for the assist

0 Severin Suveren over 7 years ago in reply to cabled

TI__Mastermind 23730 points

Ok, keep me updated.

Zigbee & Thread

Zigbee & Thread forum

CC2652R: Adapting Thread temp sensor example from MTD to SED