First I apologise as I don't fully understand the implications of this.
The spec sheet for the ES9018 and ES9018K2M (which I am considering buying) both state they are 32bit 2-channel which, I assume, equates to the 64bit word. The ES9023 (which I have built to work with the BBB/Cape) states its 24bit 2-channel, thus 48bit word by the same assumption. So will the ES9023 be supported by cape and drivers as they currently stand?
Last edited:
It will be easy, but I'd like to finish other things first.
Unfortunately I do not have functional BBB at the moment and development on my side is halted. The possibility to connect things directly to CPU pins is not only the advantage, but also the greatest weakness, because you are always one small step to "damage it yourself".
I'd like to share my story about three BBBs:
- the 1st original rev. B, accidentally damaged while disconnecting power adapter and then again damaged even more by crazy diagnosis of me ... later sent to Circuicto ... official diagnosis states: "the power rail VDD_3V3A has been found shorted to ground" ... possible to replace CPU for $25 + $20 shipping to EU
- 2nd clone from element14, damaged just yesterday by accidental pin connection to VD pin on B3 (5.25V !!!) -- ouch, they are so close, be careful! warranty void, of course; distributor logicsupply.eu does not repair; circuicto does not repair clones -- request sent to beagleboard.org RMA
- 3rd on original rev C from exp-tech.de (still available), on the way to me...
Who knows how will this story end, let's hope for the best...
- the 1st original rev. B, accidentally damaged while disconnecting power adapter and then again damaged even more by crazy diagnosis of me ... later sent to Circuicto ... official diagnosis states: "the power rail VDD_3V3A has been found shorted to ground" ... possible to replace CPU for $25 + $20 shipping to EU
- 2nd clone from element14, damaged just yesterday by accidental pin connection to VD pin on B3 (5.25V !!!) -- ouch, they are so close, be careful! warranty void, of course; distributor logicsupply.eu does not repair; circuicto does not repair clones -- request sent to beagleboard.org RMA
- 3rd on original rev C from exp-tech.de (still available), on the way to me...
Who knows how will this story end, let's hope for the best...
Last edited:
It does with (sampling rate) x (number of bits transferred)...
E.g. for ES9018 and 44100Hz I need to program BBB to divide 22579200Hz clock signal by ratio 8. This produce clock for 64bit data frames with 44100Hz frequency.
When DAC like PCM1794A requires 48 bit data frames (24bit stereo) input and 44100Hz then clock signal must be divided by 48. But with 22579200 clock this is non-integer number: 10.67
So for 24bit DACs that does not support 64bit data transfer the different set of clocks will be needed, e.g.: 16934400Hz and 18432000Hz.
E.g. for ES9018 and 44100Hz I need to program BBB to divide 22579200Hz clock signal by ratio 8. This produce clock for 64bit data frames with 44100Hz frequency.
When DAC like PCM1794A requires 48 bit data frames (24bit stereo) input and 44100Hz then clock signal must be divided by 48. But with 22579200 clock this is non-integer number: 10.67
So for 24bit DACs that does not support 64bit data transfer the different set of clocks will be needed, e.g.: 16934400Hz and 18432000Hz.
Ok, so not to argue but understand, that makes sense, so the question I have then is, how are others doing 24 bit? Is it a pll scenario that is generated in the cpu or the like? The same sorta thing the Pi is doing?
Is there a clock freq available that will do the 3 bit depths?
Or, is it we specify which we want bit depth wise and that defines the clock speeds? If for instance the dddac, could the BBB be setup to only output 24 bit data even if fed 16 bit, padding the remainder of bits as required to 24?
This then is where I query the 9018, aas unless it is fed the signal mck to run synchronous, it only has 100MHz clock to divide down and that matches no frequency at all.
Chuz,
Drew.
Is there a clock freq available that will do the 3 bit depths?
Or, is it we specify which we want bit depth wise and that defines the clock speeds? If for instance the dddac, could the BBB be setup to only output 24 bit data even if fed 16 bit, padding the remainder of bits as required to 24?
This then is where I query the 9018, aas unless it is fed the signal mck to run synchronous, it only has 100MHz clock to divide down and that matches no frequency at all.
Chuz,
Drew.
BBB is capable of perfect output, because it can just take clock and generate data EXACTLY on each X-th tick.
rPi is not, because it has fixed clock and it generates audio data with frequency AVERAGE to the requested one.
Theoretically one could use other frequencies that would be usable for both 24 and 32 bit DACs:
- 16934400Hz and 18432000Hz ... enough for 96kHz
- 33868800Hz and 36864000Hz ... enough for 192kHz
- 67737600Hz and 73728000Hz ... enough for 384kHz
But I'm not sure:
- if quality oscillators on such frequencies are commonly available
- if it would not be too fast for BBB
rPi is not, because it has fixed clock and it generates audio data with frequency AVERAGE to the requested one.
Theoretically one could use other frequencies that would be usable for both 24 and 32 bit DACs:
- 16934400Hz and 18432000Hz ... enough for 96kHz
- 33868800Hz and 36864000Hz ... enough for 192kHz
- 67737600Hz and 73728000Hz ... enough for 384kHz
But I'm not sure:
- if quality oscillators on such frequencies are commonly available
- if it would not be too fast for BBB
Last edited:
I am surprised to hear of problems with the BBB on power up/down. I have several BBBs that I use for ham radio APRS and EchoLink applications. I use these boards in unattended, remote locations, so while testing I randomly disconnected and connected the power many times to make sure that the systems would boot up and resume operations reliably. The BBBs have weather sensors and t/r relays and other digital and analog connections.
Bottom line is that I have not had a beagle bone failure due to any of this - I did brick one a few times with software glitches.
Ralph
Bottom line is that I have not had a beagle bone failure due to any of this - I did brick one a few times with software glitches.
Ralph
Bummer, I thought it would be a good solution to get a cleaner s/pdif signal to my dac with impeccable timing... I must have mis-read the design goal of this unit.
Ralph,
I don't think folks are pointing to issues restarting remotely through command line, but rather, pulling the power plug without issuing the reboot, or shutdown command. The way you are doing it is the proper way to shutdown.
OK, I thought early on in the thread Russ indicated it would have a s/pdif output included. Sorry to hear that changed as the project progressed. Maybe in rev2?
SP/DIF input rather than output, if I remember correctly. To enable other gear to go via the BBB and end up on the same i2s output.
I don't want to make work for anyone, but a short summary of the features (HW and SW) would be nice. It would be really nice if it could be added to the beginning of this thread but I don't know if that's possible.
IMHO, in the long run, this should cut down on questions.
BTW, I think this is a really cool project. I bought a BBB just for this.
Miero,
If there was a way, I would donate a few bucks to you towards getting another BBB for your development. I'm pretty sure others would too. So maybe if you setup a paypal donation page or something like that.
I don't think this would violate any diyaudio rules since it's not intended for profit, just to offset your expenses.
Randy
IMHO, in the long run, this should cut down on questions.
BTW, I think this is a really cool project. I bought a BBB just for this.
Miero,
If there was a way, I would donate a few bucks to you towards getting another BBB for your development. I'm pretty sure others would too. So maybe if you setup a paypal donation page or something like that.
I don't think this would violate any diyaudio rules since it's not intended for profit, just to offset your expenses.
Randy
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.