Odd I2S observation

This doesn’t really fit anywhere. I just thought I’d mention it.

My Pi3 has a PiFi DAC+2.0 and my Pi Zero has a bare ??? PCM5122 based DAC too. Both have an active LED.
I’ve noticed that for some reason eventually the LED gets turned on with both in RISC OS. Not sure how, but it is.

I sure haven’t gotten around to working on that yet.

Does RISC OS probe I2S to see what devices are active on the bus?

What needs to be written to said device to activate it?

It could be a rogue write to I2S, or a clash with another device possibly?

What I know about I2S could be written on the back of a postage stamp with a blunt crayon, but I think it only has a master and slave. Hmm, perhaps the nonmenclature of “worker” like my car’s service manual uses sounds better.

My only thoughts are that it’s something to do with the VC hardware or firmware that I haven’t noticed, or as you said, some kind of errant write.
The DAC on my Pi Zero only has an I2S interface. What would be controlled by an I2C demux chip on proper DAC boards like on my Pi3 is wired to what I want with a header and a bunch of jumper wire.

I just find it interesting that something trips it when nothing should be.

On I2S, I did a little reading about the Pi’s relationship with it. It’s about as good as it’s relationship with USB, UARTs and interrupts.

It would appear that due to clocking issues (stretching and jitter I think) the Raspberry Pi series can’t satisfactorily drive an I2S device.
People get around this in two ways. The first is by using a DAC chip like the ones I have, which can operate in master mode, providing the clock source to the RPi, among other things. The other seems to be some kind of black magic using a special clocking board.

I have to say, I think RISC OS really should have some degree of I2S support in the HAL. I think the standard was introduced in 1989? and it isn’t showing any sign of going away. It’d probably save people reinventing the wheel repeatedly too.

Just a guess, the activity light is possibly triggered when the clock is activated or a channel is selected. Its possible the blob is doing something with I2S, running Linux might test that as it probably doesn’t include I2S support unless you specifically install it.

If it were possible, I could be a highly paid consultant in losing things. But it’s not. So here’s a little info I don’t want to lose.

ALT 0

=

GPIO 18 – PCM_CLK

GPIO 19 – PCM_FS

GPIO 20 – PCM_DIN

GPIO_21 – PCM_DOUT

Armed with this I used *GPIODevices to check the GPIO states.
It doesn’t say alt functions, however all the pins are set to inputs and have a 0 state.
So it possibly is happening during the VC boot stage.

A little correction. My Pi Zero is using a PCM5102. The Pi 3 has the proper 5122 DAC board, which is really nice by the way. No complaints.