nRF54L15-DK 0.9.1 pre-production board is unable to flash and debug custom nRF54L15 board

[abvio]

I am debugging my custom nrf54L15 board, and want to use the nrf54L15-DK to flash and debug the board. My SWD debug probe and cabling was used for a previous version of the board and worked with the nrf52840, but I've had to upgrade my DK to the nRF54L15-DK. I've double checked all the SWD traces and they seem correct.  

I've been unable to get it to work with my new board. I suspect it is because my DK is pre-production (PCA10156 0.9.1).  

Would you agree that this is why my pre-production DK is unable to flash an nrf54L15 custom board?

Here is information from nrfutil:

 >>> nrfutil device device-info --serial-number 1057708942
serial_number: 001057708942
        boardVersion: PCA10156
        deviceFamily: NRF54L_FAMILY
        deviceName: nRF54L15
        deviceVersion: NRF54L15_xxAA_REV1
        jlinkObFirmwareVersion: J-Link OB-nRF5340-NordicSemi compiled Apr  1 2025 10:13:35

One data point:  If I disconnect the debugging cable to my custom board, pin 1 (VDD), pin 2 (SWDIO) and pin 3 (CPU_RESET) are all 3V.  But when I connect the cable to the DK, pin 3 is dragged down to 0.19V permanently.  

I've ordered a 1.0.0 version of the DK, but was wondering if there is some way to upgrade the one I have.

Thanks,

Steve

[Asbjørn]

Hello Steve,

the version of the DK should be fine to program other boards with. I have the same version on my desk and have used that to program other boards.

The voltage on the DK would normally be 1.8V, but there are level shifters to handle the difference. However, we've seen that some customers have had som situations where the level shifters did not behave as expected when there was a +10cm long cable between DK and custom board. Assumption that there's not enough power in the level shifters in some situation to drive the IO with the extra capacitance from cable length. We've updated the circuitry for version 1.0.0 of the kits, but it should still work for the 0.9.1 you have. 

Q1: How long and what sort of cable are you using between DK and custom board?

Q2: pin3 (CPU_RESET) shouldn't be pulled down unless it's trying to reset, do you have a pull up on your custom board for the RESET pin?

Best regards

Asbjørn

[abvio]

Thanks for your reply, Asbjørn.  

Q1: How long and what sort of cable are you using between DK and custom board?

It's 30cm.  It uses a Tag-Connect probe plus an adapter and cable from 0.1" to 0.05".  I've ordered a TC2050-IDC-050-ALL to get closer to the 10cm recommendation.  

Q2: pin3 (CPU_RESET) shouldn't be pulled down unless it's trying to reset, do you have a pull up on your custom board for the RESET pin?

I misspoke in my original post.  CPU_RESET is pin 10 on the SWD, not pin3.  

I do not have a pull-up on it, but I believe it is internally being pulled up.  When I disconnect the cable but power the custom board with a battery, pin 10 reads 3V.  It's only when I connect it to the DK that the pin drops to 0.19V. 

I'm due to receive a new nrf54L15-DK (hopefully 1.0.0) from Mouser tomorrow, so I'll report back then with the results.  

Best,

Steve

 

[Asbjørn]

Hi Steve,

Q1: with 30cm long cable I can definitely see this happening. We've seen it on cable that are 10cm, but not with the short 5cm. Do you have an alternative programming device? Segger Base or similar to compare with? They are more robust in performance in general.

Q2:

So it is Pin10 that is pulled down from 3V to 0.19V when you connect the JTAG cable? Above is the connections on the nRF54L15-DK. Do the connector for the DK on your JTAG cable the little notch on it to prevent it placed 180degress wrong? We have some of them floating around here and whenever they're in use we'll see the same failure message as the cable fits in both directions.

Best regards

Asbjørn

[abvio]

Hi Asbjørn,

Thanks for your response.  

You have solved the issue.  I bought a TC2050-IDC-050-ALL which provides a 10cm cable directly to the DK, and now it works with my 0.9.1 nrf54L15-DK board. The 30cm cable with an adapter was apparently too long.  It was keyed and correctly hooked up, but the capacitance must have been too much.  

I also have a couple of 1.0.0 DK boards now that will come in handy for flashing production units. 

Thank you!

Steve