• State Resolved
  • Locked Locked
  • Replies 7 replies
  • Subscribers 53 subscribers
  • Views 648 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

CC1310: EnergyTrace+ using XDS110 External debugger

Lasse Lundholm
Intellectual 280 points
Part Number: CC1310
Other Parts Discussed in Thread: MSP-FET, MSP430FR6047, ENERGYTRACE

Hi

Based on this post: https://e2e.ti.com/support/tools/ccs/f/code-composer-studio-forum/918854/ccs-energytrace-energytrace-not-working-with-cc1310-launchpad

It's not possible to use ET+ with the CC1310, however someone manage to do it here: https://embeddedcomputing.weebly.com/xds110-debug-probe-with-energytrace8203.html#

Is there any update on how to use the ET+ with the XDS110 external debug probe and the CC1310? Or any other way to get more detailed information of where, how and when the power is consumed.

The issue at hand is my application has an obscene consumption of 8-10 mA. This is a custom PCB with the CC1310 AND an MSP430FR6047 (USS_WATER_DEMO) on. So I have several issues to consider. Is it the MSP that consumes power or the CC1310? Using the MSP-FET's ET I can see the MSP spend 95% of it's time in LPM and 5% in active mode and if i just flash the CC1310 application (no MSP) to a Launchpad the consumption is around the 10mA. So the suspected culprit is the CC1310.

Now if I comment out my code and just run the main loop:

The consumption is lower but still rather high

Any suggestions on how to more accurately measure where the power is consumed?

Regards Lasse

  • 0
    TI__Guru**** 317180 points

    I don't have access to a ET+ so I'm not able to test myself. But from the documentation (https://www.ti.com/lit/pdf/sprui94) I get the impression that it's supported. EnergyTrace also only measure current and should not be directly dependent of the type of target. How have you tested ET+ with the board? 

    10 mA is a lot and indicate a HW error. If you use the empty example from the SDK and you don't have any LEDs connected to the DIOs you are toggling the current should be around 1 uA. 

    You can also try to set the MSP and/ or the CC1310 in active reset and check the current consumption. 

    We can do a quick schematic review to see if something pops up. 

  • 0 in reply to TER
    Intellectual 280 points

    Hi TER

    Thank you for your reply. Will try and answer most of your questions.

    I'm using three setups (explained below).

    The first one is the XDS110 external debug probe connected to a CC1310 LaunchPad (so not our custom board). The connections are basically what is showed in one of the links. The power is supplied via the supplyout from the AUX board and the configuration file changed to support that. The debug interface is connected via the CTI20-ARM10 adapter.

    On this setup I ran the empty example with all GPIO setup commented out, so just the while(1) loop with the sleep function. For this example I just press the EnergyTrace button from the Edit view. The current was measured to 3.5 mA which suggested that something was wrong with the setup. I changed from the cJTAG to the JTAG 1149.1 connection and removed the power jumper (read that in another thread). Now the current was 24 microA. So a lot better but still not 1 microA as expected.

    Then I flashed out program to the Launchpad and got 0.15 mA. This could very well be accurate (but still have the 24 to 1 microA difference in the back of my head). But for this example the CC1310 wasn't connected to the MSP.

    Now the second setup is our custom board with the CC1310 and the MSP. I have been struggling to connect the XDS110 external debugger to the CC1310 on this board. Using the TI14 JTAG adapter I get a connection error if I select the JTAG 1149.1. I can only use the cJTAG 2-pin choice.I suspect this might be because the TI14 adapter doesn't have the TRST ?!

    This gives a 4.9 mA current consumption. From before I know I cannot trust this measurement.

    A third setup was trying to bypass the connection issues from the external debug probe and using the XDS110 on the CC1312 and connect to the CC1310 via that:

    Still just able to choose the cJTAG 2-pin option for a successful connection. But current consumption is now 10 mA (which was the reason for the original post).

    Last but not least - and this might be the best indicator as to where the problem is. I took the LaunchPad and the external probe (the one with 0.15 mA consumption and connected it to the MSP on our custom board. The consumption jumped to 10 mA. Disconnecting the MSP made it return to 0.15 mA. So I think your absolutely right that it's a HW problem.

    For a review where can I send the schematics?

  • +1 in reply to TER
    TI__Mastermind 33080 points

    Hi Lasse,

    I have sent you an email reply from the design review. 

  • 0 in reply to RGW
    Intellectual 280 points

    Thank you very much. Just returned from an easter holiday. Will have a look at it today with our HW designer.

    I noticed that the reply was based around the CC1310, which makes good sense, since that's the review form I used. Did you take a look at the whole sensor or will it be an idea to send a review request about the MSP as well? Or did you have a look at that too?

  • +1 in reply to Lasse Lundholm
    TI__Guru**** 317180 points

    We only looked at the CC1310 part. It would make sense to test with a board with correct power supply first since the way you have used VDDR will heavily impact the current consumption. 

  • 0 in reply to TER
    Intellectual 280 points

    I promised a reply as a follow-up - and for other who might have the same problem. The problem that Richard and TER found was this connection

    The VDDR was connected to the DVCC. This was obviously a mistake. The connection wasn't there in some earlier versions of the design. So I was able to test on those. This showed improved power consumption but still around 4-5 mA.

    The search for any unwanted consumption moved to SW. Flashing a LPM4 program to the MSP and the PinStandby program to the CC1310. Making sure that the pins on both sides connecting these

    matched what we would expect in the Firmware code. This showed powerconsumption down to 3-4 uA. Back to the Firmware code and the pins was set based on the LPM/Standby and the consumption is now down to 0.3 mA. Still too high, but a few pin configurations still needs to be adressed. Confident that we will reach our power budget.

    Thank you for all your help on the issue.