Reducing SDA MCU Current When Idle

[alexdean]

I've been looking into the power consumption of the Freedom KL25Z and the SDA portions are still drawing a lot of power even when the USB connection is unplugged.

I measured the current for P_SDA through J3 (cutting the trace on the back on the Rev C board). I get similar readings on the Rev E board.

• When I power the board through the SDA USB connection, the P_SDA current is 21 mA.
• When I power the board through the P5-9V_VIN connection the P_SDA current is 8.9 mA.

Is this second current normal? Is there a way to reduce it? It is still very large compared with the current of the KL25Z target MCU and LEDs.

I checked the /reset signal for the Debug MCU (SDA_RST) and it is asserted (0.5 V).

Thanks for any help!

[Paul_Tian]

Hi, Alex

For your first question, power from 9V_VIN and USB port are different. As you can find from schematic, P5V_SDA(from USB), is not only provide power to MCU but also to K20 USB regulator and generate current on K20 reset pin. So it will consume higher current than 9V_VIN.

For your second question. Sorry, I do not know why you would like to test and reduce P_SDA current. It is used for K20 of OpenSDA function. As normal, it is not used on real product. I think you would like to measure KL25 current. So you should measure J4.

Hope my reply can help you.

Best Regards

Paul

[alexdean]

Hi Paul,

Thanks for your quick reply. I'm developing course material centered on this board, so the Freedom board is the "real" product in my application. I use an ultracapacitor to power the board and measure run-times to calculate energy use and evaluate the impact of power and energy optimizations. The problem with the SDA power is that it dominates the total power use and limits the impact of KL25Z optimizations.

Is there a safe way to disconnect the SDA circuitry and power just the application (e.g. P3V3_KL25Z) portion of the circuit?  For example, can I remove R74 and leave J3 open (disconnecting P3V3_SDA), powering the circuit through P5-9V_VIN?

Or is there a different way to accomplish this goal?

Thanks,

Alex

[Paul_Tian]

Hi, Alex

If you after debug, I think you can do as remove J3. I think it is the completely disable solution.

Hope my reply can help you.

Best Regards

Paul

[alexdean · Accepted solution]

Hi Paul,

Thank you for that information. I found out that I also need to disconnect J14 (and cut the shorting trace underneath) in order to release the reset input of the KL25Z and then the target MCU will run fine.

I tracked down the currents and realize that the linear voltage regulator has a large quiescent current. I measured a difference of 4.12 mA between the input current and output current. This is similar to what the data sheet says (6 mA with a 15 V input and 3.3 V output).

So then I tried powering the P3V3 rail directly with 3 V (after cutting the shorting trace under jumper J20). This gives me currents which I expect, with difference of only a 100 to 300 microamps.

Now I can proceed with my low-power investigations! :smileygrin:

Thanks again,

Alex

[lupogrigio]

Hi Dean,

are you plannig to make public your investigations?

[alexdean]

Yes, they will be available later this summer.  If you have any specific questions please ask now and I'll answer what I have figured out so far.

[lupogrigio]

Thank you Dean for your willingness.

Let me explain the reason of my interest:

I'm designing a KL25 Board in which power saving is critical. To accomplish power requiremnts

I' m using MIC20xx IC as Power Distributor Switches to enable different power rails only when

necessary. So when I need MAx3232 transceiver I enable a dedicated power supply line, on which

theese power greedy components resides, similar for Bluetooth and Lcd, and so on. My design is

at an intermediate stage. The most difficult part is the power section, as ever, because I want

to use only one switching regulator in order to keep low EMC, but I need both 5V and 3V3 power

rails, regulated from a main Battery of 9V. This is the kind of work I'm following. The final

stage of this project will be an intelligent power saving oriented board that will be able to

rule power consuption, in order to 'survive' working in a degradated fashion when the battery is

sensed to be low. The final stage will integrate also a little (and cheap) solar cell, to feed

the battery when necessary.

Why I'm writing to you and more generally on this in this forum?

I think the effort I'm facing in this work could be of general interest, and we, as community of

designers, could contribute to develope a solid and reusable common platform to develop our

own embedded projects, with the right mixture of hardware and  software, shared and private

stuff.

For now it is an idea!

If (you) Dean, but more generally, anybody user of this forum agree with me and is interested

in this Idea/Common Platform, just write here!

Best Regards.

[alexdean]

Interesting requirements! They sound similar to the constraints I have in a different project, but which I hope to apply to the KL25.  You have many alternatives, but what makes the most sense depends on your own specific circumstances. For example, choosing between switching and linear regulators depends on their relative efficiencies, which in turn depends on load current and voltage drop. That's just the beginning.

Good luck!