I'm not having any luck getting Hugo to look into the matter of whether there are I2S or similar format signals present somewhere in the Rio Receiver. I'd like to buy one, but only if I can get digital out enabled. Anyone care to open theirs to see if there is anything obviously labled as there is in the RioCar and Empeg?

Thanks,
Stu

More than happy to open up one of mine if you can tell me what to look for . Would love to have a digital out!

It's funny, I was about to reply with: "hey, there's this guy doing digital outs for the car player, maybe he would know..."

Gotta look at those usernames.

Ideally, there should be a +5 volts available somewhere too.

Stu

Hmm. I have one of my units opened up on my desk. There are ground and AVCC points (JP10 and JP12), but other than that, what should I be looking for?

Worst case (it will require some fine soldering) download the PCM1716 datasheet from www.ti.com and you'll find where the MCLK, LRCLK, DATA and CLOCK lines of the I2S stream can be found...

Hugo

OK, definitely no clear and easy-to access point for I2S. The stuff around the PCM 1716 is very tight, and very hard to track (lots of it on the other side of the board). Might have to take out the whole board. But there's an empty surface mount pad next to the 1716 labeled "U11". Does not seem be a circuit on the bottom side of the board on those pads.

A brief look at the doc. indicates pins 1,2,3,4 are the I2S ones needed. Anyone care to send us their player for a free digital out board? If we can get it to work you get the digital out installed, etc. If we can't, you get your player back (working) and a not very useful paperweight (the digital board).

Thanks Hugo and "julf."

Stu

More than happy to send one of mine, but where are you located? There is probably somebody closer, as I'm in Amsterdam, The Netherlands.

Yeah, thanks, but we are in the New Orleans area, USA. That would be mighty expensive to ship, I would think.

Stu

Not to mention time-consuming and risky. During the past year I've lost faith in the transatlantic mail system.

I'd forgotten about U11; this is an alternate pinout for a Cirrus CS4341 DAC. ISTR that the package is bigger and so it'll be easier to pick signals off it.

SDATA is pin 2, SCLK is pin 3, LRCLK is pin 4 and MCLK is pin 5. If you need it (at 3.3v), pin 6 is I2C SCL and pin 7 is I2C SDA.

There's no digital 5v on the board; any CS8405 board is going to need 5v, so you'll have to regulate it down from the 12v supply on the main board. There *is* a 5v linear regulator supplying the DACs, but this is supposed to be a clean DAC feed...

Hope this helps!

Hugo

Thanks Hugo,

One last question for anyone with access to a RioReceiver- What frequency crystal is used for MCLK? The current digital board is configured for 11.2896 MHz (256 x 44.1 KHz). If it's something other than 11.2896 MHz, it's going to require rework. Anyone see an 11.2896 MHz crystal near the DAC?

Stu

The only xtal seems to be 20 Mhz

Well, 20 Mhz isn't a multiple of any standard audio sampling frequency. There must be an explanation though.

Thanks,

Stu

The MCLK is 11.2896MHz. Everything on the board is generated from one xtal and the Cypress CY2292 chip which has programmable (at production, anyway) dividers/PLLs/etc, which means you don't have to cover the board in loads of xtals.

Hugo

OK, but the chrystal (Y2) next to the CY2292 is a 20 MHz one? The only label on it says Q20.000C0.

I figured it was something like that. Now all I need is to build a small +5v secondary power supply board and a Rio Receiver to play with. The ones on eBay are too steep for something that might not work anyway.

Thanks!

Stu

Yes, it is a 20MHz xtal. The 2292 generates buffered 20MHz (ethernet), 64MHz (PNA), 3.6864MHz (CPU) and 11.2896MHz (audio)

Hugo

Here's a progress report. The +5v power supply is made and seems top work. The connections have all been made on the Rio circuit board- no shorts. However, upon powering up the Rio with everything attached, the Rio's main power supply died. I never got a chance to test the digital out, because the failure was instantaneous. The fuse is still good. It bothers me that the Rio's power supply is so fault intolerant. Haven't had any luck finding what part went bad. I hate switching power supplies!

Anyway, I ordered two replacement supplies and will try to see what caused the other one to die. So far no real clues.

Stu

Very strange, it's a fairly meaty supply (40W, 12V) to supply the audio amps.

As it died when hardware was added, I would suspect it might be something to do with the changes you made (or maybe even something dropped into the PSU by mistake?). Blown PSU is not a cause a failure in any of the dead receivers I've seen (mostly, it's the LCD cable fracturing, followed by one of the figure-of-8 power pins getting pushed into the unit).

Hugo

Thanks,

Yes it's very strange. Power is taken from JP11 (used as ground) and JP7 (source of 12 volts). I didn't see anything that fell into the power supply. I'll keep investigating.

Stu

The power supply was replaced and the unit is up and running again. We also got to test the digital out board in the Rio. We got sound, but it was quite garbled and L/R was all out of whack. Are you sure the signals going to U11 are in I2S format and not something else like left or right justified? That would require some modification to the board to get the CS8405 running in those formats, but can be done of course. Later, I'll try our original prototype board with jumpers that lets us change formats and see what happens.

Stu

I tried all the supported formats on the prototype board, but it still isn't working any better. Left justified gives the opposite effect of the I2S setting. Right justified gives no sound. Not sure what the problem is. I can hear something resembling what is playing, but it is riddled with distortion. In I2S mode, there is more distortion in the right channel and in left justifed, the opposite is true. There must be something wrong with the timing of the signals, but I'm not sure what. It's odd that it works fine in the Empeg and not in the RioReceiver.

I'll keep investigating.

Stu

Hugo, could you give me a hint as to which format of I2S is going to the PCM1716? Right now all I am getting is a very distorted output regardless of format, but it would help if I knew what format to leave the board set on.

I recorded a sample of the output which I have attached.

Stu

Just to let everyone know, it looks like the I2S signal from the RioReceiver is incompatible with the CS8405 transmitter chip used in the RioCar S/PDIF board. Another chip will need to be used. Perhaps a CS8420. This will have the added benefit of jitter reduction if it works. Costs, will go up a bit though. This chip costs 3x the CS8405.

Stu

Ah, sorry; I've not been reading this for a while. The output is whatever you get from a Cirrus 7212's DAI port (try the cirrus website).

The audiotron, which also uses the 7212/7312, uses an 8405 for S/PDIF out, which means that it should work... see http://www.gweep.net/~prefect/eng/audiotron/audio.html

Hugo

Thanks Hugo. The Audiotron uses the 7312 chip rather than the 7212. I don't know how similar the two are, but I'm almost positive I have the CS8405 configured properly and it doesn't work right. I was unable to get a hold of any 7212 docs, but assuming they are basically the same, looking at the 7312 docs wasn't much help anyway. I couldn't find mention of the format the chip outputs through the DAI port, only the pinout on the signals.

I've had some people take a look at the signals heading to the DAC on the receiver and they seem to think that the channel frame size is 64 bits and that it is problematic for the CS8405, as it is at the extreme end of its capabilities.

Let's see what happens with a CS8420 when it comes in.

Stu

Okay, I tried the CS8420. Same deal. Intense noise in the right channel, and noise at higher digital levels in the left channel. No idea why it is doing this. The circuit works fine in the Empeg. I tried the clock out of the RioReceiver and an independent clock. Same result either way. It's gotta be something with the timing of the I2S, but what and why? I don't have a suitable scope to anaylze the timing, so I'm S.O.L. there. Anyone want to step in and finish this? Maybe we can work something out.

Stu