No problems. Happy to help
Actually, while assisting you I realized the there can be some simplifications to the build when used this way. ie. S03 to pcm1794, TDA1541 or similar DACs. With reference to the original post by Palmito:
The following components need not be loaded - some cost savings and the heaps of efforts soldering these tiny parts! If already loaded you may remove them to reduce switching noise and power consumption:
U5, U8, C13, FB4, R4, R13, R2, C16, C17, FB5, C8, R18, X2, R7, R5, Q2
Actually, while assisting you I realized the there can be some simplifications to the build when used this way. ie. S03 to pcm1794, TDA1541 or similar DACs. With reference to the original post by Palmito:
The following components need not be loaded - some cost savings and the heaps of efforts soldering these tiny parts! If already loaded you may remove them to reduce switching noise and power consumption:
U5, U8, C13, FB4, R4, R13, R2, C16, C17, FB5, C8, R18, X2, R7, R5, Q2
Last edited:
Hi acko,
You mention to remove R2... this should be the input of the clock for the flip flop... are you sure is to be removed?
Oops! You are right. Should have been R3 instead
Thanks for pointing this out
Updated:
U5, U8, C13, FB4, R4, R13, R3, C16, C17, FB5, C8, R18, X2, R7, R5, Q2
Last edited:
I am afraid you may have to get a new XO. This time test it separately first (with wires attached like earlier) to see if the factory build is alright before mounting on board. In the meantime you can test the rest of the circuit like the buffer and FFs by applying appropriate logic inputs and see if the outputs switch correctly.
tested some CPLD and CPU combinations on full turbo sync with Buffalo III
I wasn't sure where to post this but because it involves Acko recklocker here it is.
After Win7 new, clean installation I wanted to test some behavior of different CPLD and CPU firmwares. In ConfigTool, pin 11 is set for 24.576 MHz selector.
First, I've flashed slave_for_1080 CPLD and 1080 CPU firmware. Only 384 KHz sample rate doesn't produce any sound, but 352 KHz gave me some really bad crackling and noise. I didn't expected it because it appeared after the played file and it quite scared me so I had to mute the amp and then only Arta was used as spectrum analyzer. I didn't try DSD files.
After this, slave2244 CPLD was flashed. In this case 352 KHz doesn't produce any signal on output and after 384KHz sample noise appears again. I must mention noise starts sometimes several minutes after file played. I used foobar in ASIO and kernel streaming mode, it doesn't matter. If I play file again its normally reproduced but noise is here again after few minutes. If any other sample rate is used (44.1, 48,..,192) there's no such problems.
With CPLD_for_1080 and cpu firmware 1080 plays all sample rates but with 352 and 384 we have noise after file is played.
Bitclk CPLD and CPU firmwares doesn't play 352 and 384 sample rates.
One more thing, probably irrelevant, that I have come across when I tested this firmwares - I was using 2 channel 32 bit samples and Amanero doesn't send full 32bit samples to the DAC because it doesn't get full 32 bit data through USB. Tried WaveIO and JLSound which is both XMOS and data stream was the same - no full 32bit.
I wasn't sure where to post this but because it involves Acko recklocker here it is.
After Win7 new, clean installation I wanted to test some behavior of different CPLD and CPU firmwares. In ConfigTool, pin 11 is set for 24.576 MHz selector.
First, I've flashed slave_for_1080 CPLD and 1080 CPU firmware. Only 384 KHz sample rate doesn't produce any sound, but 352 KHz gave me some really bad crackling and noise. I didn't expected it because it appeared after the played file and it quite scared me so I had to mute the amp and then only Arta was used as spectrum analyzer. I didn't try DSD files.
After this, slave2244 CPLD was flashed. In this case 352 KHz doesn't produce any signal on output and after 384KHz sample noise appears again. I must mention noise starts sometimes several minutes after file played. I used foobar in ASIO and kernel streaming mode, it doesn't matter. If I play file again its normally reproduced but noise is here again after few minutes. If any other sample rate is used (44.1, 48,..,192) there's no such problems.
With CPLD_for_1080 and cpu firmware 1080 plays all sample rates but with 352 and 384 we have noise after file is played.
Bitclk CPLD and CPU firmwares doesn't play 352 and 384 sample rates.
One more thing, probably irrelevant, that I have come across when I tested this firmwares - I was using 2 channel 32 bit samples and Amanero doesn't send full 32bit samples to the DAC because it doesn't get full 32 bit data through USB. Tried WaveIO and JLSound which is both XMOS and data stream was the same - no full 32bit.
This is signal captured by USB Transit 96/24 when 384KHz file was played through Foobar (foo_dsd_asio) on Amanero used in full turbo sync with Acko's reclocking boards (90/98 clocks) and CPLD_for_1080 and CPU firmware_1080 and Buffalo III DAC.
Warning!!!
Do not left your amp unattended when you play 352 and 384 KHz files or your speakers can/may be damaged!! Or better do not use files with these sample rates at all. Forget them until the problem is resolved.
No problems with 192KHz or lower.
https://www.dropbox.com/s/8qt8ydj4r9jj7td/384KHz and noise.7z
Warning!!!
Do not left your amp unattended when you play 352 and 384 KHz files or your speakers can/may be damaged!! Or better do not use files with these sample rates at all. Forget them until the problem is resolved.
No problems with 192KHz or lower.
https://www.dropbox.com/s/8qt8ydj4r9jj7td/384KHz and noise.7z
Hi Miksi,
Thank you very much for your efforts -all noted and relevant. I haven't check your noise measurements yet but do you have a 100MHz scope that could be use to monitor the actual outputs (BCK & Turbo MCK) from the S03 going into the DAC. These are the controlling signals to the DAC and ideally should be picture perfect as shown in the timing diagram before:
I am assuming that you have set the S03 divider to /4 and Amanero prescaler to MCLK/1. Also, the CKSEL is switching correctly for different Fs right up to 384KHz
Thank you very much for your efforts -all noted and relevant. I haven't check your noise measurements yet but do you have a 100MHz scope that could be use to monitor the actual outputs (BCK & Turbo MCK) from the S03 going into the DAC. These are the controlling signals to the DAC and ideally should be picture perfect as shown in the timing diagram before:
I am assuming that you have set the S03 divider to /4 and Amanero prescaler to MCLK/1. Also, the CKSEL is switching correctly for different Fs right up to 384KHz
Yes, I have oscilloscope and already checked data and clocks stream. There's nothing wrong with them while music plays. Behavior with 352 and 384 is probably due to a firmware. What bothers me is that I have amanero more than a year (first generation) and I didn't notice this before, because I didn't use these sample rates. But I thought it's fair to warn people about this. Noise spectrum only resembles white noise.
The divider is set to /4 and Amanero prescaler to MCLK/1 and the CKSEL is just fine with both 44/48 groups.
Only have to check what's going on with data and clocks when noise appears.
The divider is set to /4 and Amanero prescaler to MCLK/1 and the CKSEL is just fine with both 44/48 groups.
Only have to check what's going on with data and clocks when noise appears.
Last edited:
Cchd957 45.xxxx & 49.xxxx MHz are the ones you're looking for cox77
The 22.xx & 24.xx MHz variants will work for up to 192k with the i2s to PCM board by Ian. See his posts on the thread about that board, you will need to use the half speed option for 352/384k to work with the tda1541a
I'll probably experiment with this exact config into one of the new aya ii boards at some point too
The 22.xx & 24.xx MHz variants will work for up to 192k with the i2s to PCM board by Ian. See his posts on the thread about that board, you will need to use the half speed option for 352/384k to work with the tda1541a
I'll probably experiment with this exact config into one of the new aya ii boards at some point too
Last edited:
Hi Miksi,
This might not apply to your Buffalo III implementation, but I found that trying to play 352/384 via amanero/acko isolator with P-sync, I had lots of noise/glitching. I found that the VDD(1.2v) Trident couldn't supply enough current to prevent VDD voltage droop causing the noise. After modifying the trident to supply more current, I've been able to play 352/384 files without any issues.
Hope this helps.
Cornelius.
This might not apply to your Buffalo III implementation, but I found that trying to play 352/384 via amanero/acko isolator with P-sync, I had lots of noise/glitching. I found that the VDD(1.2v) Trident couldn't supply enough current to prevent VDD voltage droop causing the noise. After modifying the trident to supply more current, I've been able to play 352/384 files without any issues.
Hope this helps.
Cornelius.
Hi Cornelius -
Would you explain the modification to the Trident that you did? I've been unable to get the Acko to function at the highest sampling frequencies, and would like to do so in order to bypass the internal filtering of the Sabre chip and to substitute software upsampling and filtering via Daphile.Thanks - Pat
No "modification" is required - just a resistor, the Trident is designed so that the current is adjustable by adding a through hole resistor at the CCS, This is because the load on the chip can be very dynamic, and you don't want the shunt Q shunting more current than it really needs to. In this case I would say adding 20-27R at the indicated spot on the trident will get you the current boost you want.
No "modification" is required - just a resistor, the Trident is designed so that the current is adjustable by adding a through hole resistor at the CCS, This is because the load on the chip can be very dynamic, and you don't want the shunt Q shunting more current than it really needs to. In this case I would say adding 20-27R at the indicated spot on the trident will get you the current boost you want.
Thanks, Russ, for the "non-stock configuration" info.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.