BlueSaab starts on bench, not in car

[homerisback]

Hi guys,
I'm experiencing a confusing behaviour from a v5.x module. It comes from a bunch of 3 where the two others work like a charm.
As written in title, the module starts properly on bench while either powered by 3.3V from ISP, 5V from USB or 12V via CD changer connector (like it would in car) : it does the boot sequence and puts RN52 into discovery mode, when paired it starts playing music. Everything normal.

When I plug it into the car (tested in both my 9.3s where other modules ever worked), power lights up, little blinkings at RX & TX leds, then ... nothing. I mesured voltages at different points, everything is ok and car battery is healthy enough.
It looks like firmware won't start (forgot to say that I tried v4.1, v5.1b and my v5.1c with character mapping).
This BlueSaab remains invisible to car and no RN52 start either.

Any ideas that pop up ?
Thanks guys for your help

[Seth]

Sounds like a possible canbus communication issue...? I would look at that area and make sure there aren’t any shorts or missing solder joints.

Could also be an issue with the mcp2515 resonator, if it’s not giving the mcp a good clock signal.

Seth

On Wed, Jun 24, 2020 at 14:17 homerisback [via BlueSaab Forum] <[hidden email]> wrote:

Hi guys,
I'm experiencing a confusing behaviour from a v5.x module. It comes from a bunch of 3 where the two others work like a charm.
As written in title, the module starts properly on bench while either powered by 3.3V from ISP, 5V from USB or 12V via CD changer connector (like it would in car) : it does the boot sequence and puts RN52 into discovery mode, when paired it starts playing music. Everything normal.

When I plug it into the car (tested in both my 9.3s where other modules ever worked), power lights up, little blinkings at RX & TX leds, then ... nothing. I mesured voltages at different points, everything is ok and car battery is healthy enough.
It looks like firmware won't start (forgot to say that I tried v4.1, v5.1b and my v5.1c with character mapping).
This BlueSaab remains invisible to car and no RN52 start either.

Any ideas that pop up ?
Thanks guys for your help

01' 9.3 SE Conv - AS3 - Nexus 5 Android M 6.0.1 - BlueSaab v4.2 + Amp - Latest CDC repo.

---

If you reply to this email, your message will be added to the discussion below:

http://bluesaab-forum.2349123.n4.nabble.com/BlueSaab-starts-on-bench-not-in-car-tp1657.html

To start a new topic under BlueSaab Forum, email [hidden email]
To unsubscribe from BlueSaab Forum, click here.
NAML

[Karlis]

I'd start with putting a debug print in, say, CDC.cpp line 121; maybe another one in setup or loop. CDC.cpp debug print will show you if CAN bus comms are actually working and BlueSaab sees any CAN frames on the bus. A print in setup or loop will tell you if the firmware is actually running.

Once prints are put in the code, plug the module in the car and observe serial console.

[homerisback]

Thanks to both of you, I double checked every solder joint with binocular magnifying glass, there is very little possibility of shorts.
I thought of this CAN loss of communication, let's go this way after swapping resonator.
Crossing fingers that serial monitor will give an answer ;)

[homerisback]

Mystery remains
I hadn't been able to get any serial monitor feedback except start up sequence with RN52 configuration, even supposedly activating debugmode in CPP ; maybe I'm doing wrong with this.

By the way, as it starts properly while USB powered, it works perfectly in car. It is recognised by car, plays music, change tracks with SEEK buttons... I assume CAN comms are ok.

Removing power from USB let it keep going on.
I gave a try by connecting module directly to car's 12V CDC connector without other pins and it starts !

Under proper CDC connection, it would start if I shorten 3.3V and PWREN pins from RN52 chip... it looks like PWREN signal isn't sent by 328P when connected to audio lines and CAN bus, code seems to stop but why ?
I use v4.1 code revision to ensure nothing related to text scrolling would interfere.
I keep following white stones in the code to understand starting steps...