Hi Miero,
I measured the voltage on the pins before playing:
3.3 Volts are 3.3 V , MCLK = 0.8V, BCLK = 0V, LRCK = 0V
During playing and after the error message:
3.3 Volts are 3.3 V , MCLK = 0.8V, BCLK = 0V, LRCK = 3.3V
Thanks and greetings,
Sjoerd
How long does it take until "play"fails with the error "Device or resource busy"?
Is it immediately or it takes more than that?
If it is instant fail, then another program is blocking the sound output (e.g. MPD or anything other you have installed).
If it take a while until it fails, then the clock signal is not correctly delivered to the BBB. In this case check 2 things:
1) exec the following command and check that it prints "3":
cat /sys/module/snd_soc_botic/parameters/ext_masterclk
2) Measure a voltage on Hermes-BBB (with everything powered on) between P9_25 and DGND pins of the BBB. There should be the MCLK signal after it passes through isolator on the Hermes_BBB. It should be non-zero value. Check the following image for locating pins: http://beagleboard.org/static/images/cape-headers.png
Is it immediately or it takes more than that?
If it is instant fail, then another program is blocking the sound output (e.g. MPD or anything other you have installed).
If it take a while until it fails, then the clock signal is not correctly delivered to the BBB. In this case check 2 things:
1) exec the following command and check that it prints "3":
cat /sys/module/snd_soc_botic/parameters/ext_masterclk
2) Measure a voltage on Hermes-BBB (with everything powered on) between P9_25 and DGND pins of the BBB. There should be the MCLK signal after it passes through isolator on the Hermes_BBB. It should be non-zero value. Check the following image for locating pins: http://beagleboard.org/static/images/cape-headers.png
How long does it take until "play"fails with the error "Device or resource busy"?
Is it immediately or it takes more than that?
If it is instant fail, then another program is blocking the sound output (e.g. MPD or anything other you have installed).
If it take a while until it fails, then the clock signal is not correctly delivered to the BBB. In this case check 2 things:
1) exec the following command and check that it prints "3":
cat /sys/module/snd_soc_botic/parameters/ext_masterclk
2) Measure a voltage on Hermes-BBB (with everything powered on) between P9_25 and DGND pins of the BBB. There should be the MCLK signal after it passes through isolator on the Hermes_BBB. It should be non-zero value. Check the following image for locating pins: http://beagleboard.org/static/images/cape-headers.png
Hi Miero,
It takes 10 sec when the error occured. I think this is not immediatly in computertime
Ext_masterclk = 3
But I measured 0V on P9_25 (doublechecked if I had the right pin).
Ok, another attempt. Check a voltage on components of the MCLK signal on the isolated part of the Hermes-BBB.
Notice, you have to use other GND as reference (e.g. on the battery connector) for measuring voltages on the BBB and part of Hermes.
Hi Miero,
I checked the several points in the picture:
Used "red GND": All the red marked points stays 0V. The pink marked points are 3.3V (against the red GND)
Used the Cronus GND: The yellow marked points are 3.3V, the blue marked points are 0.8V
regards,
Sjoerd
Ok, so at least 3 possible sources of error:
- there is a short to GND somewhere along the red line,
- the chip that translates "blue" to "red" is damaged
- software configuration error? The dai_format has to be 16385-16387.
I suggest you to write an email to Brian, i.e. brian@twi... and ask for the instructions how to proceed further.
- there is a short to GND somewhere along the red line,
- the chip that translates "blue" to "red" is damaged
- software configuration error? The dai_format has to be 16385-16387.
I suggest you to write an email to Brian, i.e. brian@twi... and ask for the instructions how to proceed further.
Hi Miero,
Tried several dai_formats, but default is already 16385. It doesn't help. If it's a short I could detect this with the multimeter? So I guess it's the chip.
I will contact with TPA.
Again, thank you for your help. It's very interesting how to trouble-shoot this and also how you handle/explain this in a easy way for me.
Best,
Sjoerd
...then probably also worth heating the isolation chip pins in question with a "wet-tip" soldering iron, in case one is dry or poorly connected...
...then probably also worth heating the isolation chip pins in question with a "wet-tip" soldering iron, in case one is dry or poorly connected...
Good idea, I did it, also the resistor R4. I checked conductivity from chip to P9- 25. I checked the voltage-values: voltage from the Cronus is still 0.78V and the voltage on the Hermes is still 0.0 V
(Also the error remain).Best,
Sjoerd
Yes! , problem solved.
It was the isolator-chip. Comparing my board with this one:
http://www.diyaudio.com/forums/twis...otic-cape-beaglebone-black-9.html#post4300592
isolatorchip1 and isolatorchip2 seems to changed
I contacted TPA and they send me immediately a new board, I installed it and it works flawless. great service,
so now I can play around with the software.
Thank you all again for helping and thinking,
best,
Sjoerd
It was the isolator-chip. Comparing my board with this one:
http://www.diyaudio.com/forums/twis...otic-cape-beaglebone-black-9.html#post4300592
isolatorchip1 and isolatorchip2 seems to changed
I contacted TPA and they send me immediately a new board, I installed it and it works flawless. great service,
so now I can play around with the software.
Thank you all again for helping and thinking,
best,
Sjoerd
Getting back to this great project, I have been able to troubleshoot a bit:
Following the clock signals from Cronus and clock select pin from the BBB it seems the isolators are doing far to good a job by blocking all signals.
The voltage across their decoupling caps is ~3,4 V on the BBB side but only about 1,3 V on the Cronus side.
Same 1,3 V across decoupling caps on Cronus board.
Lab-PSU is steady at 5 V. I disconnected the BBB and removed the clocks, still 1,3 V.
So, I suspect the regulator on the Cronus and would try to replace it.
Is this IC5 which is marked 715102?
Would this be the one to order:
http://www.mouser.fi/ProductDetail/...=sGAEpiMZZMsGz1a6aV8DcEg4Tjc2AQrmMjy9blLOYc4=
Best regards,
Tor-björn
Following the clock signals from Cronus and clock select pin from the BBB it seems the isolators are doing far to good a job by blocking all signals.
The voltage across their decoupling caps is ~3,4 V on the BBB side but only about 1,3 V on the Cronus side.
Same 1,3 V across decoupling caps on Cronus board.
Lab-PSU is steady at 5 V. I disconnected the BBB and removed the clocks, still 1,3 V.
So, I suspect the regulator on the Cronus and would try to replace it.
Is this IC5 which is marked 715102?
Would this be the one to order:
http://www.mouser.fi/ProductDetail/...=sGAEpiMZZMsGz1a6aV8DcEg4Tjc2AQrmMjy9blLOYc4=
Best regards,
Tor-björn
Back looking for help
Thanks francolargo for your reply on this. I gave up on it for a while, and recently started over. Currently using botic4 image, which I saw you recommend to someone else who is struggling.
Trying to get BBB to play I2S on B3 in stereo mode. The B3 plays I2S from an RPI but not with BBB-Hermes-Cronos. I didn't change the B3 but wonder if i need to.
Here is what I have, not sure what would be helpful for you or anyone else to advise me.
Speaker-test -c2 produces no sound on my speakers
aplay -c2 -D hw:0 pathto/somefile.wav produces a signal on bbb pin 30, which I think is correct for I--- (please correct me if i'm wrong)
amixer sset Master unmute gets an error (amixer: Unable to find simple control 'Master',0) - haven't tried researching this yet
SoX failed because SoX command was not found
My parameters:
root@botic:~# for i in /sys/module/snd_soc_botic/parameters/*; do echo "* $i"; cat $i; done
* /sys/module/snd_soc_botic/parameters/blr_ratio
64
* /sys/module/snd_soc_botic/parameters/clk_44k1
45158400
* /sys/module/snd_soc_botic/parameters/clk_48k
49152000
* /sys/module/snd_soc_botic/parameters/dai_format
16385
* /sys/module/snd_soc_botic/parameters/dsd_format_switch
1
* /sys/module/snd_soc_botic/parameters/ext_masterclk
3
* /sys/module/snd_soc_botic/parameters/pinconfig
default
* /sys/module/snd_soc_botic/parameters/serconfig
I---
I have not touched mpd.conf
I bought a cheap logic analyzer and like I said, when I use aplay or ympd to play, pin 30 on bbb shows a signal (and doesn't when the song ends). At the other end of hermes (cronos side), I don't see anything on the corresponding data pin (pin 5 i believe). Probably wrong about this, but it feels like this signal is not getting thru hermes. Can someone confirm or correct if this assumption sounds right and if so, suggest what should I look at next?
I used the signal analyzer to check the clock signals on the hermes to chronos header, and all are show signals. Also checked voltage on them and they all have about 1.6v. The two 3.3v pins have 3.3 volts as they should (I saw this in another post where Miero was helping someone else).
If anyone has time to read this and give me some pointers, I would appreciate it. I'm in awe of those of you who have gotten this thing to work and still have hope that I can as well.
Thanks,
Doug
Thanks francolargo for your reply on this. I gave up on it for a while, and recently started over. Currently using botic4 image, which I saw you recommend to someone else who is struggling.
Trying to get BBB to play I2S on B3 in stereo mode. The B3 plays I2S from an RPI but not with BBB-Hermes-Cronos. I didn't change the B3 but wonder if i need to.
Here is what I have, not sure what would be helpful for you or anyone else to advise me.
Speaker-test -c2 produces no sound on my speakers
aplay -c2 -D hw:0 pathto/somefile.wav produces a signal on bbb pin 30, which I think is correct for I--- (please correct me if i'm wrong)
amixer sset Master unmute gets an error (amixer: Unable to find simple control 'Master',0) - haven't tried researching this yet
SoX failed because SoX command was not found
My parameters:
root@botic:~# for i in /sys/module/snd_soc_botic/parameters/*; do echo "* $i"; cat $i; done
* /sys/module/snd_soc_botic/parameters/blr_ratio
64
* /sys/module/snd_soc_botic/parameters/clk_44k1
45158400
* /sys/module/snd_soc_botic/parameters/clk_48k
49152000
* /sys/module/snd_soc_botic/parameters/dai_format
16385
* /sys/module/snd_soc_botic/parameters/dsd_format_switch
1
* /sys/module/snd_soc_botic/parameters/ext_masterclk
3
* /sys/module/snd_soc_botic/parameters/pinconfig
default
* /sys/module/snd_soc_botic/parameters/serconfig
I---
I have not touched mpd.conf
I bought a cheap logic analyzer and like I said, when I use aplay or ympd to play, pin 30 on bbb shows a signal (and doesn't when the song ends). At the other end of hermes (cronos side), I don't see anything on the corresponding data pin (pin 5 i believe). Probably wrong about this, but it feels like this signal is not getting thru hermes. Can someone confirm or correct if this assumption sounds right and if so, suggest what should I look at next?
I used the signal analyzer to check the clock signals on the hermes to chronos header, and all are show signals. Also checked voltage on them and they all have about 1.6v. The two 3.3v pins have 3.3 volts as they should (I saw this in another post where Miero was helping someone else).
If anyone has time to read this and give me some pointers, I would appreciate it. I'm in awe of those of you who have gotten this thing to work and still have hope that I can as well.
Thanks,
Doug
First - sorry for your frustration - I will see if I can help.
Does it appear that the master clock is making it to the BBB from the Cronus?
A couple of general pointers:
- Double check SMT solder joints on the Hermes-BBB and the Cronus - rarely one might be dry - a little careful touch up can resolve that issue.
- If you have a scope (actually you can do this with a dmm too) check each cronus signal individually - especially the 3 used for 2 channel I2S. You can do this by powering the cronus with clocks completely independantly (not connected to any other module) and using a wire jumper to bring the signal high on the "in" side and checking that it indeed goes high on the "out" side - do the same to GND for each channel and observe that out goes low. It even better to do this with scope and a signal generator.
You can do the same kind of dignostics on the hermes. - just carefully observe the direction of the signals. Remember that master clock on the Hermes header is an output!!! So please don't short it high or low.
I hope that helps!
Cheers!
Russ
Does it appear that the master clock is making it to the BBB from the Cronus?
A couple of general pointers:
- Double check SMT solder joints on the Hermes-BBB and the Cronus - rarely one might be dry - a little careful touch up can resolve that issue.
- If you have a scope (actually you can do this with a dmm too) check each cronus signal individually - especially the 3 used for 2 channel I2S. You can do this by powering the cronus with clocks completely independantly (not connected to any other module) and using a wire jumper to bring the signal high on the "in" side and checking that it indeed goes high on the "out" side - do the same to GND for each channel and observe that out goes low. It even better to do this with scope and a signal generator.
You can do the same kind of dignostics on the hermes. - just carefully observe the direction of the signals. Remember that master clock on the Hermes header is an output!!! So please don't short it high or low.
I hope that helps!
Cheers!
Russ
Did you check the voltage across the local decoupling caps for the hermes-cronus-isolator IC:s?
Thanks Russ. I've checked the solder. Will do again. Have to think about your other suggestions. Not sure I know how to do it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.