Hello,
I was testing the extended sleep mode with the DA14580 and it works, but the current consumption was 350 uA, and not 1.4 uA, as expected
For this test I followed the instructions of the Training_04_sleep_modes_current_measurement.v10.pdf. I am using the pro kit with a PAN1740 doughter board and the SDK 5.0.2.1.
First I used the power profiler of SmartSnippets tool, but the average current was 350 uA. I checked the current offset before doing this test.
Then, I used a DMM and the same results appear.
Finally, I used other projects like sleep_mode of the SDK guide, but the current level was again the same.
Maybe someone had this problem before and know how to fix it?. Maybe I am making a mistake during the measurement?. I really appreciate your help.
Best,
Matías
Hi mjara,
I am measuring near 350 uA when the sleep is not enabled (app_default_sleep_mode = ARCH_SLEEP_OFF;) and the watchdog is enabled, so an NMI interrupt occurs, perhaps your system hits the NMI handler and this is what you measure. Other than that, check you pins if there is any kind of leakage, remove all the jumpers from the dev kit.
Thanks MT_dialog
Hello MT_Dialog,
I am using,
sleep_state_t app_default_sleep_mode = ARCH_EXT_SLEEP_ON
So I had no problems with the WD timer. I checked the pins for any leackage, and there is not any significant difference. I am using the last SDK version, wich example do you recomend me to try the extended sleep mode?.
By the way, I am using the Debug to program the DA14580 (start debug, and then stop), not Smart Snippets.
Best,
Matías
Hi mjara,
The program that you are using is fine (as far i could understand when you say sleep_mode of the SDK guide, you mean ble_app_sleepmode in the 5.0.4 SDK) since it fallls in permanent sleep and you dont have to place a long adv interval in order to measure the lowest point of the power consumtion, so after the few seconds of advertising the device will fall in sleep so at that point you should be able to measure tha lowest power consumtion. Using a PAN on a dialog's pro kit i wasn't able to replicate the 350uA that you see, the only value close to the was when the sleep was disabled and no radio activity (about 442uA), are you sure that you dont disable sleep anywhere in the testing code (arch_disable_sleep()) ?
Thanks MT_dialog
Hello MT_Dialog,
I double checked and I never disable the sleep mode. Yes I am using the ble_app_sleepmode project in the 5.0.4 SDK .
The projects works fine, after 10 seconds, the ble stops advertising, but the current that I measured in "sleep mode" is near 350 uA.
I read inhttps://support.dialog-semiconductor.com/current-bleappsleepmodethat in ble_app_sleemode in the major current consumption comes from the flash.
Maybe that is the extra current that I am measuring. Do you know anything about this issue?
Best,
Matías
Hi mjara,
Do you have a 583 with an embedded flash or a 580 ? The reference design that you are using hasn't got any flash interaction (your program downloads straight to the sysram via keil) also in order to use the flash you have to place some jumpers on the specified pins (you can check the pins for using the flash at the schematic on the pro motherboard right beside to the J6 header), check if you mistakenly have placed the power supply jumper of the SPI (you can find more info regarding the pins and the pro motherboard at the UM-B-034 document).
Also you can perform the following test, increase the duration that the device stays awake and advertise (change the APP_ADV_DATA_UPDATE_TO from 1000 -> 10 seconds to 3000 -> 30 seconds or even more) and after you ve changed that, increase the advertising intervals (at the user_config.h file change the members .intv_min and .intv_max of the user_adv_conf structure) for over a 2 or even 3 seconds or even for 10.24 seconds which is maximum value. This will force the device to go into sleep mode between the advertising intervals and give the measuring instrument enough time to measure the power consumption in sleep mode between the advertising intervals (if the device is working properly before going to permanent sleep you should see the minimum power consumption between the advertising events).
Other than that i would suggest to try to use a different daughter board and perform the measurements.
Thanks MT_dialog
Hello MT_Dialog,
Yes, I am using a 580, so there is not an embedded flash. Also, I checked and the power supply jumber of the SPI is no connected.
I performed the test that you suggested. But I am still measuring the same current level when the BLE stops advertising. I tried with another dughter board, and the same happened.
Any other suggestion?. Thanks for your time,
Matías
Hi mjara,
I ve aksed you to do this in order to give you enough time to measure the power consumption between the advertising interval, not when the device falls asleep permanently, so what was the power consumption between the advertising intervals ?
You can perform the same with other SDK examples (increase the advertising interval and try to measure between the intervals, or kill the advertising at some point and dont start it again), perhaps also try on a fresh SDK, perhaps you ve made any changed on the current one.
Thanks MT_dialog
Hello MT_DIalog,
The current between the advetising intervals is 440 uA (measured with SmartSnippets).
I tried a fresh SDK, and the current level was never under 350 uA.
Best,
Matías
Hi mjara,
As the tutorial indicates you wont be able to measure the lowest power consumption of the device via Smart Snippets and its better to use a DMM. If this value is the same with what you are measuring with a DMM then i cant think of any other reason why your device doesn't go in sleep mode than a deflective motherboard, if you have checked all the above and still not measuring about 1.5uA when sleep then i am not able to think anything else that could be responsible to this kind of behaviour.
Thanks MT_dialog
Hello MT_dialog
After some test, I figured it out that if I use the Booter tool of Smar Snippets with Close Debug Sess to write in the SysRAM, then I can measure a low current of 2.2uA in sleep mode. But if I use Keil, although I detached the debug, the current level always increment in 300 uA.
Thank you for your time.
Matías
Hi mjara,
Are you using a dev kit or a custom board ? The expert kit when the debugger is attached on the board it consumes some extra power.
当使用凯尔和智能Sni下载代码ppets for using the power profiler, can you see if the 580 goes to sleep and wakes up ? Also which version of keil you are using ?
Thanks MT_dialog
Hello MT_dialog
I am using the development kit - pro.
If I used Keil, I cannot see that the 580 goes to sleep in the Power Profiler of Smart Snippets. I am using Keil V5.21.1.0
Best,
Matías
Hi mjara,
Thats probably due a known compability issue between the jlink driver you have, and the keil version that you are using, please check the FAQ section in the known issues section and in the corresponding FAQ "凯尔MDK521a compability问题older JLink drivers.". The FAQ doens't report this but we have observed that sometimes this issue can also prevent the 580 from going to sleep when downloaded via keil. You can either follow this or install an older version of keil.
Thanks MT_dialog