12 posts / 0 new
Last post
boydy1989
Offline
Last seen: 1 month 1 day ago
Joined: 2017-10-18 16:11
High sleep current

SDK 6.0.8

Hi

I have developed a PCB based on your reference designs, including an I2C sensor device.

I have burned the SW to the OTP, and functionally the device appears to operate as expected, waking on a button press and performing a POR. However, a current measurement of the device reports a reading of 110uA, in both extended and deep sleep modes.

I have removed the sensor from the board, and the current usage is still high.

This current reading is taken with a CR2032 cell powering the device, with no debugger etc connected at all - just the bare device, using an accurate DMM.

I can send details of my design if there is a location I can send them direct to.

Any thoughts?

Thanks!

Device: 
PM_Dialog
Offline
Last seen: 2 days 5 hours ago
Staff
Joined: 2018-02-08 11:03
Hi boydy1989,

Hi boydy1989,

The 110uA is an unexpected current consumption in both extended and deep sleep mode configuration. I strongly suggest you to read the Tutorial 4 : configuring sleep mode on DA1458x from our support page in order to make sure that you are measuring the current consumption with the correct way. Also, be aware that something on you custom PCB probably increases the current consumption, so please review it first.

Thanks, PM_Dialog

boydy1989
Offline
Last seen: 1 month 1 day ago
Joined: 2017-10-18 16:11
Hello

Hello

I have followed this tutorial in some detail already, in fact I used it as the basis for establishing these measurements.

I have also removed everything else from the PCB that could draw current - the only thing that reduces the current usage is removing the power to the DA14585.

Please can you offer some more thoughts, again I am happy to send schematic or code over privately if this is possible.

Thanks!

PM_Dialog
Offline
Last seen: 2 days 5 hours ago
Staff
Joined: 2018-02-08 11:03
Hi boydy1989,

Hi boydy1989,

If you have followed correctly our reference design you should not have too much current consumption in the sleep mode configuration. Could you please make sure that the chip goes into sleep?

Thanks, PM_Dialog

boydy1989
Offline
Last seen: 1 month 1 day ago
Joined: 2017-10-18 16:11
Hello

Hello

I have just checked, and changing
static const sleep_state_t app_default_sleep_mode = ARCH_EXT_SLEEP_ON;

to

static const sleep_state_t app_default_sleep_mode = ARCH_SLEEP_OFF;

makes the device draw ~550uA, rather than the ~110uA with pulses for advertising packets. This would suggest to me that the device is entering extended sleep in at least some form.

Anything further I can check?

Thanks!

boydy1989
Offline
Last seen: 1 month 1 day ago
Joined: 2017-10-18 16:11
I have run some more tests.

I have run some more tests. Using the prox_reporter project, I get a higher sleep current draw of ~250uA.

The only difference between the packaged project and what I have flashed is that I am using the internal LF clock, not a 32k crystal.

I have attached a current trace showing an advertising package. (Red line is raw data, green line is a 10 sample rolling average)

Reverting to my code, I see ~110uA again. I have also attached a trace of an advertising package. (second page)

Next I will try the same with a DA14585 daughterboard, not using my own PCB to compare.

In the meantime, can you please offer some more ideas?

Thanks!

Attachment: 
PM_Dialog
Offline
Last seen: 2 days 5 hours ago
Staff
Joined: 2018-02-08 11:03
Hi boydy1989,

Hi boydy1989,

As I am able to understand from you post, the current consumption is 550 nano Ampere in active mode configuration? Did I understand correctly? Be aware that is a wrong measurement, because in active mode the current consumption is ~0,6mA. Did you use power profiler from SmartSnippets studio in order to check if the device advertises? If yes, the current consumption is 550nA? With our DA14585 DK, what current consumption did you get? If I understood correctly from your second post, you do not have an external XTAL32, so are you using the internal RCX20 clock as low power clock?

Thanks, PM_Dialog

boydy1989
Offline
Last seen: 1 month 1 day ago
Joined: 2017-10-18 16:11
Apologies - nA is an error,

Apologies - nA is an error, the correct value is 550uA (0.6mA) for active mode, 110 uA for sleep. I have corrected the post above.

I am using my own power profiler to take the advertising current, using a self designed tool. I have verified the accuracy, and can confirm that it is accurate enough for what I require.

I can confirm the device advertises correctly, and all other functions work as expected - I have verified this with a bluetooth scanner and our own in house app which is in development. The current spikes visible in my previously attached PDF align with the received advertising packets.

Your understanding is correct, I am using the internal RCX20 for the low power clock.

Thanks!

boydy1989
Offline
Last seen: 1 month 1 day ago
Joined: 2017-10-18 16:11
Hello

Hello

I think I have solved this. On my PCB I had connected pin 5 (VCC_FLASH/P3_0) to 3.3v. If I cut this track, the current during sleep drops to ~3uA as expected.

Can you please suggest why this may be?

Thanks!

PM_Dialog
Offline
Last seen: 2 days 5 hours ago
Staff
Joined: 2018-02-08 11:03
Hi boydy1989,

Hi boydy1989,

Could you please clarify which chip are you using in your custom board? The DA14585 or the DA14586? In the thread that you have posted, you have tagged the DA14585 SoC. Although, the DA14585 doesn’t have VCC_FLASH. Only DA14586 VCC_FLASH GPIO because it has embedded SPI flash. In case you are using the DA14586, according to the datasheet, the VCC_FLASH should be supplied with 1.8 V to 3.6 V. If you cut this track, you will not able to access the SPI flash, so I suggest you to keep it to 3.3V. The most possible reason is that you are not power the SPI flash down, so that’s why you get high current consumption. Please try to power it down and check again the current consumption.

Thanks, PM_Dialog

boydy1989
Offline
Last seen: 1 month 1 day ago
Joined: 2017-10-18 16:11
Hello

Hello

I am using the DA14585. My initial prototypes were using a 586, but I realised I didn't need the SPI flash so I switched to 585.

I have verified the fix on a second PCB now, and the current is also reduced to ~3uA.

Thanks!

PM_Dialog
Offline
Last seen: 2 days 5 hours ago
Staff
Joined: 2018-02-08 11:03
Hi boydy1989,

Hi boydy1989,

Glad that your problem has been fixed. If you found any of the above answers useful please mark one of them as accepted.

Thanks, PM_Dialog