OK...I did an easy test: I plugged the USB-connector from the SoTM-Card and put it into the motherboard. With Linux this is possible without restart.
The sound is not bad. If you do not have the comparison, you will be happy for sure. USB back into the SotM-Card (all with FIFO in place): Like Turbo-Chargin it. Tones are now really vibrating. Wood sounds like wood and a piano's trebles are made of metal. A clear step towards "Live".
Interesting. So, the Fifo helps, but it needs still a good front-end. The old rule from Ivor Tiefenbrun (founder of Linn) holds true: A chain starts at the beginning and what you loose in the beginning, can't be recovered later.
This experiment will have some consequences for me...Focus will be to improve the Player and experiment with WavIO vs other solutions and their reaction on better PSUs (Currently the Sotm and the WaveIO gets its power from a LT1085, the Fifo from a Shunt Placid HD).
The sound is not bad. If you do not have the comparison, you will be happy for sure. USB back into the SotM-Card (all with FIFO in place): Like Turbo-Chargin it. Tones are now really vibrating. Wood sounds like wood and a piano's trebles are made of metal. A clear step towards "Live".
Interesting. So, the Fifo helps, but it needs still a good front-end. The old rule from Ivor Tiefenbrun (founder of Linn) holds true: A chain starts at the beginning and what you loose in the beginning, can't be recovered later.
This experiment will have some consequences for me...Focus will be to improve the Player and experiment with WavIO vs other solutions and their reaction on better PSUs (Currently the Sotm and the WaveIO gets its power from a LT1085, the Fifo from a Shunt Placid HD).
Last edited:
From what i understand Blitz is galvanically separated through the WaveIO but does not use Ian's isolators and subsequent reclocking. Still, in theory, at least, unless there is something i am overlooking, there should be no reason for the cabling/ports to be audible at all.
At this stage i have developed a healthy scepticism towards theoretically perfect solutions, but the audibility of the port is certainly disappointing.
At this stage i have developed a healthy scepticism towards theoretically perfect solutions, but the audibility of the port is certainly disappointing.
How do you power the isolated side of the WaveIO? Any potenial PS coupling to the USB side may be a reason for what you hear.
Well, I try to do the best to make it understandable:
I use the isolator of the WaveIO-Card, which is connected with the FIFO.
Each of the following has his own transformer and Regulator:
- Sotm-Card siolated: powered NOT by the PC, but from its own LT1085
- Wave-IO: Powered by its own X-Former and LM1085
- ISO-Chip on WaveIO: Powered by it own X-former and LT1085
- Alix: Normal 12V Switching-Supply
I simply plug now the USB-cable from one USB-Port from the Sotm-Card and connected it with the On-Board USB-connector of the Alix. Yes, I did know to what I am listening, but simply try it yourself before coming to the conclusion that stuff cant be because you dont want it to be...purpose of my comment is not to talk FIFO small, it does a very good job. I simply want to make you aware, that you have an additional option to improve the sound of your system (wit the Sotm-Card). Your choice if you want to try it. The Sotm Card has send my Lynx Aes 16e card into retirement for half of the price.
So, I believe it is A/B and I believe the Fifo is galvanic isolated as the rest of the DAC is.
I am not at all surprised. If you loose information earlier in the chain, you cant recover later.
This is not a computer-program. This is a continous datastream. We hear jitter, we hear cables and PSUs. We hear all kind of stuff like Windows vs Linux. This is still the case with FIFO. It is helpful, but not neutralizing the other effects completely. Too many crazy stuff happens in the components before the signal gets into the DAC...USB-BUffer-Sizes and latencies, 16 to 32 bit conversions etcetc...stuff these chips do without you even knowing it. Therefore be aware that you know only a fraction of what is really going on...so, I stopped to try to explain, I simply do my experiemnts and find what works better for me...as many of us I guess.
I use the isolator of the WaveIO-Card, which is connected with the FIFO.
Each of the following has his own transformer and Regulator:
- Sotm-Card siolated: powered NOT by the PC, but from its own LT1085
- Wave-IO: Powered by its own X-Former and LM1085
- ISO-Chip on WaveIO: Powered by it own X-former and LT1085
- Alix: Normal 12V Switching-Supply
I simply plug now the USB-cable from one USB-Port from the Sotm-Card and connected it with the On-Board USB-connector of the Alix. Yes, I did know to what I am listening, but simply try it yourself before coming to the conclusion that stuff cant be because you dont want it to be...purpose of my comment is not to talk FIFO small, it does a very good job. I simply want to make you aware, that you have an additional option to improve the sound of your system (wit the Sotm-Card). Your choice if you want to try it. The Sotm Card has send my Lynx Aes 16e card into retirement for half of the price.
So, I believe it is A/B and I believe the Fifo is galvanic isolated as the rest of the DAC is.
I am not at all surprised. If you loose information earlier in the chain, you cant recover later.
This is not a computer-program. This is a continous datastream. We hear jitter, we hear cables and PSUs. We hear all kind of stuff like Windows vs Linux. This is still the case with FIFO. It is helpful, but not neutralizing the other effects completely. Too many crazy stuff happens in the components before the signal gets into the DAC...USB-BUffer-Sizes and latencies, 16 to 32 bit conversions etcetc...stuff these chips do without you even knowing it. Therefore be aware that you know only a fraction of what is really going on...so, I stopped to try to explain, I simply do my experiemnts and find what works better for me...as many of us I guess.
Last edited:
OK, with that setup I'm also doubting your objectivity or rather, the ability to not being influenced by your ideas of what could/can/should happen...
Or you actually have data loss - thats the only acceptable explanation. But that usually don't have the audible impact you describe. So who knows
/
Or you actually have data loss - thats the only acceptable explanation. But that usually don't have the audible impact you describe. So who knows
/
The whole point is that you do not "lose information" anywhere. I am quite certain the WaveIO is bit-transparent and Ian's board should have no issues with the jitter coming off it. So, surprising or not, it is still a mystery.
Will repeat your experiments at first opportunity. It is rather annoying that the Exa uses a silly micro USB connector, which makes cable swapping awkward and my WaveIO board is currently hooked to another dac, but i'll think of something. My entire interest in this project was to eliminate the PC, cabling and even USB receiving interface from the list of items directly influencing sound quality. Never thought it would be easy but with Ian's help it seems we are at least getting a bit closer.
Different question: What are the best clock to try while waiting for the S571 ? I have already the crysteks 957 and the tentclock. I read somewhere ndk el, not sure which one exactly..., joe rasmussen voting for saw types, but for the fifo their frequencies are too high, anything else ?
Decoupling: what is best in terms of setup and parts ? Tent votes for a coil/ choke and a smal 0,1uf cap...any other advises ?
Best Regards
Decoupling: what is best in terms of setup and parts ? Tent votes for a coil/ choke and a smal 0,1uf cap...any other advises ?
Best Regards
what do you think is there inside a Crystek or a Tentlab oscillator?
I tell you that, a simple picogate oscillator and cheap crystal, phase noise specs talks itself.
Rakon has already design for you, just open the document, not more than six components (8 in case of overtone crystals). 74HCU04 have gain beyond 100MHz, so no speed problems.
Rf=2M2, Rlim=180R, C1=10p, C2=470p, are a good starting point.
Since probably you don't own an Agilent phase noise measurement system, just look at the waveform with a scope and listen how it sounds. Buy a bag of Citizens cylindrical crystal from farnell (0.3 USD each) and find the best. Better, if you have a source, buy a good crystal (that do not costs much more than a Crystek osc).
Finally you'll reach at least -115dBc@10Hz in phase noise.
All the above oscillators are already trimmed and tested for you, and perform worst than what you can build yourself (take a look at their specs).
I tell you that, a simple picogate oscillator and cheap crystal, phase noise specs talks itself.
Rakon has already design for you, just open the document, not more than six components (8 in case of overtone crystals). 74HCU04 have gain beyond 100MHz, so no speed problems.
Rf=2M2, Rlim=180R, C1=10p, C2=470p, are a good starting point.
Since probably you don't own an Agilent phase noise measurement system, just look at the waveform with a scope and listen how it sounds. Buy a bag of Citizens cylindrical crystal from farnell (0.3 USD each) and find the best. Better, if you have a source, buy a good crystal (that do not costs much more than a Crystek osc).
Finally you'll reach at least -115dBc@10Hz in phase noise.
All the above oscillators are already trimmed and tested for you, and perform worst than what you can build yourself (take a look at their specs).
I don´t know if anyone ask this before, but wouldn´t it be cool to have such an option also on the single/dual clock board?
