CC2650: Unusual high current consumption

Mathieu De Zutter

Part Number: CC2650
Other Parts Discussed in Thread: CC2640, CC2640R2F

Recently we discovered an increased current consumption in the latest production batch of our application. Between BLE advertisement intervals the MCU goes to sleep and normal current consumption is around 3-8µA. However the new production shows the idle current consumption is increased to 35-80µA.

Some background information:

The application is advertising every 1800ms a BLE packet
2 layer PCB design, 1.6mm thickness, FR4
Both layers do have ground plain
Only a button connected and a boost converter for driving blue LED
Powered from a CR2032 battery
Exact part number: CC2650F128RHB

We tested following:

Multiple 32kHz crystals -> no solution
Different load capacitor values for 32kHz crystal -> no solution
Resoldering all CC2650 connections with only flux -> temporarily resolves the issue. Current drop to 3-8µA but after a couple of days this effect is gone
Resoldering the X32K Q1 and Q2 pads with flux -> temporarily resolves the issue. Current drop to 3-8µA but after a couple of days this effect is gone
Different solder paste in production (ON338CPP and WS820) -> no difference
Washing the PCBs -> no solution
Removing the buck converter -> no solution
Matching the 2.4GHz antenna -> no solution
Adding some coating on 32kHz crystal pins of MCU (for example SL 1307 FLZ-HT and silicon coating) -> temporarily resolves the issue

We are out of possible solutions and have 4000 problematic PCB at our customers.

Are there any thoughts or solutions?

over 4 years ago

0 FI over 4 years ago

TI__Mastermind 21630 points

Hi,

I assume the device revision is the same in this production run?

It could be a soldering issue on the 32Khz xtal portion.

See if guidance here helps https://e2e.ti.com/support/wireless-connectivity/bluetooth-group/bluetooth/f/bluetooth-forum/554607/cc2650-current-consumption

Regards

0 Mathieu De Zutter over 4 years ago in reply to FI

Prodigy 20 points

Hello,

Thanks for your reply. We checked the PCB under X-ray and all solderings are good.

Even re-soldered 32k crystals with series resistor and 1M Ohm parallel resistor.

0 David Rubie over 4 years ago in reply to Mathieu De Zutter

Expert 1690 points

Are you sure the firmware logic is OK? If the part is consuming that kind of energy, but the circuit boards are not shorted/wet/cross linked it implies the CPU is doing something. Can you tell us which version of the libraries you are using, which compiler etc. I vaguely recall having a lot of issues with compilers, bluetooth stacks and energy consumption until we settled on 16.9.10 LTS which is still supported under Code Composer Studio 10.4. The bluetooth library I ended up using was 2_02_01_18 although I think you can move up to 2_02_06_07 (which might be the last one that supports the CC2650). We since moved on to the CC2640R2F but I still have a few thousand CC2640 boards floating around. Edit: I also ended up using a spray lacquer on our boards which helps a lot in humid/wet conditions.

0 Mathieu De Zutter over 4 years ago in reply to David Rubie

Prodigy 20 points

We found the solution. In previous batch there was another IC connected to the CC2650 but because of component shortage this one was not present on latest batch, which wasn't a big problem as it was only for future use. Because it wasn't assembled, there was a pin of the CC2650 in, what it looks like, an undefined state. We expected it to be input high impedance.

We removed the line of code where this unconnected pin was defined as input with no pull-up/down. After, current consumption dropped to normal values.

Thanks for the help.

0 David Rubie over 4 years ago in reply to Mathieu De Zutter

Expert 1690 points

Not sure I helped much but good to hear its sorted!

Bluetooth®︎

Bluetooth forum

CC2650: Unusual high current consumption