CCS/CC2642R: CC26X2 Leakage on MOSI pin

Frederic Griffe1

Part Number: CC2642R
Other Parts Discussed in Thread: SYSCONFIG

Tool/software: Code Composer Studio

Hello,

I am using the CC2642 (with simplelink_cc13x2_26x2_sdk_3_40_00_02) on a custom design where the CC2642 SPI port is connected to 6 slave devices.

I observe a leakage current during about 1s after any spi transaction.

I managed to get rid of it by introducing a pulldown on MOSI line (not MISO) either with a 100k external resistor or by activating the GPIO pulldown through direct access to the IOC register.

This is surprising because this kind of issue is usually observed on the MISO pin which is an input of the CC2642 when all the slaves are in standby (high Z).

If the same problem happens on the MOSI it means that the CC2642 does not drive properly the MOSI line when the CC2642 is in standby.

In the CC26X2 Technical Manual it is explained that in idle "The transmit data line SSIn_TX (ie MOSI) is arbitrarily forced low". And IOS are maintained during standby.

Any idea why this is not the case ?

Kind regards.

Frederic

over 5 years ago

0 IUsedToBeARealPerson (M-W) over 5 years ago

TI__Guru 58425 points

Hi Frederic,

I will look closer at it and get back to you, the "arbitrarily" part throw me of a bit. I will get back when I have had the chance to test this myself. Did you try this out on the default examples to see if you saw the same there? Also, which mode are you operating in?

0 Frederic Griffe1 over 5 years ago in reply to IUsedToBeARealPerson (M-W)

Prodigy 160 points

Hello M-W,

I did not try on the examples because I am using a custom design. However the code is coming fro the simple_peripheral example.

What do you mean by "which mode" ?

Kind regards.

Frederic

0 IUsedToBeARealPerson (M-W) over 5 years ago in reply to Frederic Griffe1

TI__Guru 58425 points

With mode I basically mean Frame Format :)

0 Frederic Griffe over 5 years ago in reply to IUsedToBeARealPerson (M-W)

Expert 1745 points

I am using the 8 bit frame format with CPOL=0, CPHA = 0.

I am trying to connect the pulldown by software. I managed to do it in CCS debug by changing the value of the IOC register by hand (it worked only once. I do not know why).

Now I am trying to do it by software using the Ioc.h driverlib IOCIOPortPullSet function but it does not work.

Any idea why ?

Kind regards.

Frederic

0 IUsedToBeARealPerson (M-W) over 5 years ago in reply to Frederic Griffe

TI__Guru 58425 points

Hi Frederic,

Where in the code (in relation to the SPI driver) do you try to do this call?

0 Frederic Griffe over 5 years ago in reply to IUsedToBeARealPerson (M-W)

Expert 1745 points

I tried just before and right after the spi_open() command but none of these option woks.

Regards.

0 IUsedToBeARealPerson (M-W) over 5 years ago in reply to Frederic Griffe

TI__Guru 58425 points

Hi Frederic,

Doing the call after SPI_open() should work and I have confirmed this by setting up my own example. If you observe the registers, you see nothing happening?

0 Frederic Griffe over 5 years ago in reply to IUsedToBeARealPerson (M-W)

Expert 1745 points

Hi M-W,

I checked again and it works indeed. Sorry I do not know what happened when I checked first.

Now, is it logical that we have to do that to avoid current leakage on the MOSI input pins of the SPI slave devices ?.

The CC26X2 should keep its MOSI output (and SCLK) in a defined state when in standby mode.

Kind regards.

Frederic

0 IUsedToBeARealPerson (M-W) over 5 years ago in reply to Frederic Griffe

TI__Guru 58425 points

Hi Frederic,

I have looked closer at the implementation and "floating" is the intended state. When in motorola mode and the module is "idle", the output enable signal will be forced high and the pad will be in HIGH-Z.

0 Frederic Griffe over 5 years ago in reply to IUsedToBeARealPerson (M-W)

Expert 1745 points

Hi M-W,

This is interesting.

Where did you look at to get this information because In the CC26X2 Technical Manual it is explained that in idle "The transmit data line SSIn_TX (ie MOSI) is arbitrarily forced low". ?

Is it also the case to the SCLK signal ?

Regards.

Frederic

0 IUsedToBeARealPerson (M-W) over 5 years ago in reply to Frederic Griffe

TI__Guru 58425 points

Hi Fredrick,

I had to go into the RTL module design documentation to find this out, it is not reflected in the TRM. This could be an oversight from the one writing the chapter.

The TX line as such as forced low, but this is from the internal point of view. the pad in question is disabled. This should not be the case for the clock signal.

0 Frederic Griffe over 5 years ago in reply to IUsedToBeARealPerson (M-W)

Expert 1745 points

Hi M-W,

Thanks for the answer

What is the "RTL" module and where can I find the this documentation ?

I am quite surprised because this "pad disabled thing" exposes all CC26X2 developpers to random over consumption soemtime very difficult to identify because it is documented nowhere. Everybody knows about the MISO pulldown thing but nothing on the MOSI line.

Regarding the clock signal "Should" or "is' there not problem ? Where can I check.

Best regards.

Frederic

0 IUsedToBeARealPerson (M-W) over 5 years ago in reply to Frederic Griffe

TI__Guru 58425 points

Hi Fredrick,

Simply put, it is the digital implementation of the hardware. This documentation is internal only and nothing that is publicly shared.

The design is most likely by purpose as there is no true standard on how to handle this. As there could be variations in implementations on different devices, leaving it up to the designer to choose if to pull it low, high, or leave it floating means that you are also open to all type of variations.

As for clock, this is a "is" as one thing is clear, there is no assumption on the state of the clock while in idle, it must not be floating. The clock pad would only be disabled when configured as an slave, never as master.

0 Frederic Griffe over 5 years ago in reply to IUsedToBeARealPerson (M-W)

Expert 1745 points

Hi M-W,

Thanks for the answer which improves my understanding.

a. regarding the MOSI pin the Pad becomes High Z when the SPI modules is in idle. Does Idle means when the spi module is not doing a transaction or outside of an spi-open() spi_close() section ?

b. Regarding the SCLK signal I understand the Pad is never left high Z in Master mode so the pad is always driven high or low. Correct ?

c. Regarding any other IO where in the documentation can I check if there is similar Pad disconnection cases ?

Kind regards.

Frederic

0 IUsedToBeARealPerson (M-W) over 5 years ago in reply to Frederic Griffe

TI__Guru 58425 points

Hi Frederic,

a) It means Idle as in not doing transaction. When outside open/close, the IO will be in the state configured as "not in use" in SysConfig.

b) Yes

c) There is o such dedicated section. If there is such information in the TRM, it would be on a per module basis.

Bluetooth®︎

Bluetooth forum

CCS/CC2642R: CC26X2 Leakage on MOSI pin