BLE connection issue (error 8) with nRF54H20 + Zephyr

[Mykhaylo]

Hello,

We have the following environment:

1) nRF54H20-DK v0.9.1 and our custom board

2) nRF Connect SDK v3.2.1

3) Our simple project for testing BLE

4) nRF Connect for Mobile app

5) Our custom board has an external 32.768 kHz crystal

Before testing we have connected and bonded our device with nRF Connect for Mobile app.

Then we turn device off and on (power cycle) and try to connect to our device from app.

If we use nRF54H20-DK the connection is established successfully every time.

If we use our custom board sometimes we get the connection error 8. Both device and app show this error.

Also we have other samples of our custom boards. Some of them work well, but some ones have the same connection issue.

How can we debug and fix the issue?

Thanks

[Simonr]

Hi

I was asked to take a look at this ticket, and everything here points to the issue here being with either the HFXO or LFXO crystal on your custom boards either being mounted poorly or loaded incorrectly.

These error messages in the past and the fact that some boards work correctly points very much to this being the case. 

Can you:

• Upload the specifications for the crystals you're using in your design so we can verify they are suitable for your design. (Both 32.768kHz and 32MHz)
• Share details on how they are loaded (I assume you're using internal capacitors and not external ones)?
• Verify with a scope or similar that the external crystals are outputting the expected frequencies and that they are mounted physically correct on the boards. A common mistake here is to mount either crystal 90 degrees the wrong way if it has 4 mounting pads, and that would result in these kinds of errors as well.

Best regards,

Simon

[Mykhaylo]

Hi,

1) Could you try to change the LFXO accuracy config in BICR to 500 ppm just to see if that improved the performance?

2) Could you try to double the HFXO startup time config in BICR and see if that changes anything?

Fullscreen bicr_orig.json Download

{
  "power": {
    "scheme": "VDDH_2V1_5V5"
  },
  "ioPortPower": {
    "p1Supply": "EXTERNAL_1V8",
    "p2Supply": "EXTERNAL_1V8",
    "p6Supply": "EXTERNAL_1V8",
    "p7Supply": "EXTERNAL_1V8",
    "p9Supply": "EXTERNAL_1V8"
  },
  "ioPortImpedance": {
    "p6ImpedanceOhms": 50,
    "p7ImpedanceOhms": 50
  },
  "lfosc": {
    "source": "LFXO",
    "lfxo": {
      "mode": "CRYSTAL",
      "accuracyPPM": 20,
      "startupTimeMs": 600,
      "builtInLoadCapacitancePf": 15,
      "builtInLoadCapacitors": true
    }
  },
  "hfxo": {
    "mode": "CRYSTAL",
    "startupTimeUs": 850,
    "builtInLoadCapacitors": true,
    "builtInLoadCapacitancePf": 14
  }
}

Fullscreen bicr.json Download

{
  "power": {
    "scheme": "VDDH_2V1_5V5"
  },
  "ioPortPower": {
    "p1Supply": "EXTERNAL_1V8",
    "p2Supply": "EXTERNAL_1V8",
    "p6Supply": "EXTERNAL_1V8",
    "p7Supply": "EXTERNAL_1V8",
    "p9Supply": "EXTERNAL_1V8"
  },
  "ioPortImpedance": {
    "p6ImpedanceOhms": 50,
    "p7ImpedanceOhms": 50
  },
  "lfosc": {
    "source": "LFXO",
    "lfxo": {
      "mode": "CRYSTAL",
      "accuracyPPM": 500,
      "startupTimeMs": 600,
      "builtInLoadCapacitancePf": 15,
      "builtInLoadCapacitors": true
    }
  },
  "hfxo": {
    "mode": "CRYSTAL",
    "startupTimeUs": 1700,
    "builtInLoadCapacitors": true,
    "builtInLoadCapacitancePf": 14
  }
}

We have tried the updated BICR with changes above but the issue is still reproduced.

Mykhailo

[Susheel Nuguru]

Hi Mykhailo, 

I am Susheel Nuguru, and I am technical contact engineer responsible for Corsair. Thanks to Simon and Amanda for assisting here while I was on Easter holidays.

Mykhaylo said:

1) Could you try to change the LFXO accuracy config in BICR to 500 ppm just to see if that improved the performance?

2) Could you try to double the HFXO startup time config in BICR and see if that changes anything?

I am not sure if I understand the above? Are you asking me to test these settings on our EVK or your hardware? 

BICR settings are board based configs and I do not see changing these values as you suggested and testing on the EVK. Even if the failure happens on EVK with these new settings, I think it does not help us diagnose the issue on the custom board you have.

[Mykhaylo]

Hi Susheel,

The questions above were from Nordic. I just left our test results here.

I also have done some tests with different values of internal capacitor (1-25pF). Looks like two my samples work well with capacitor value about 25pF.

I am not sure if the value 25pF will be good for all Invasion samples. How can we get the right value for our Invasion boards?

Mykhailo

[Susheel Nuguru]

Mykhailo, 

A lot is happening through the emails and it seems that Brian and you guys are getting closer to finding the root cause of the issue. Let us keep that momentum in the emails for now.