TP wires tend to have Z0 up around 100-ohms. Somewhere around there. If only coax is any good then why do we use CAT-6 now instead of coax for networking. According to what you appear to be claiming only coax can be a 'proper' solution, not TP.
Beyond that IIRC Chris719 agrees with me that so-called series-termination resistors at CMOS outputs are better termed as damping resistors, since the output impedance of a CMOS driver is a not constant. Therefore no perfect series-termination resistor exists. Also IIRC Syn08 accepted Chris' explanation of damping as satisfactory. From you however, I get lectured to the effect that your way is the final word on what is okay or not.
In addition, the treatment of uniform TL are covered in all sorts of textbooks. Non-uniform TL can also work and be well managed for selected cases. Distributed damping can work better than solely end-termination if a TL is non-uniform.
Last edited:
I honestly don't understand your ranting especially about topics not discussed in this thread. Damping resistor normally refers to something else than series termination resistor (try googling "damping resistor"). But anyhow that is just terminology. And I have never said that coax is the only solution for I2S but proper termination minimizes reflections and not just in theory.
Your comment above appears to be in response to a pic I posted before of the best way I found to interface a USB with an AKM designed AK4499 evaluation board. Since you re-posted the pic and made derisive comments multiple times, there is clearly a pattern to your behavior. As I explained to you before AKM put a pin header I2S connector on the top of their evaluation board which then allowed EMI/RFI coupling into the dac circuitry when I2S was used. The best solution I found that allowed the AKM board to work well was to put the USB board under the eval board ground plane and connect it with distributivity damped TP TL. Considerable time and effort was expended on finding the best solution under the circumstances. Since then you have continued to post that pic and complain about it. I tried explaining to you again that there was no better solution found for this particular problem, and that it actually works quite well in practice. So, it should not be surprising if I get tired of having to reexplain the whole thing over and over again every time you try to use it to discredit correct advice I have posted.
It has also been tiresome for you to have kept insisting as you have up until just now that damping is not a proper term for absorbing TL reflections, and the only correct term is termination. Thank you for acknowledging now that damping can be acceptable terminology.
Anyway, I am happy to put these things in the past if we can end it here.
Last edited:
I tried the u.fl coax cable verses pin header I2S option with Iancanada McFIFO, and also with Twisted Peak Cronus clock board. In both cases u.fl performance was unacceptable, and pin header connections worked better for short distances. Having said that, maybe using properly torqued SMA coax would have been best. Didn't try that since the two boards didn't accommodate that option for all the I2S lines.
BTW, I am not disagreeing with TL theory or practice. I would however note that recent research indicates that TP TL can work effectively up to 2GHz and with negligible stray coupling (with proper implementation, of course).
I will leave off here for now. Sorry to the readers for OT.
BTW, I am not disagreeing with TL theory or practice. I would however note that recent research indicates that TP TL can work effectively up to 2GHz and with negligible stray coupling (with proper implementation, of course).
I will leave off here for now. Sorry to the readers for OT.
You seem to be forgetting that TP in Ethernet cables is used for differential signals. I2S pin headers are not differential. There is no guarantee that the return current of e.g. BCK follows the same GND wire that is twisted with BCK wire. Actually it is better to use a flat ribbon cable with signal wires between gnd wires.
Right, good points.
I think part of the response to your first point, and what andora76 was asking about, can apply somewhat to parallel coaxes. Basically, I agree that forced balancing can be very important. There are some other issues specific to u.fl too. Short answer to andora76 is sound was affected like one might expect from bad jitter. Maybe more on that tomorrow morning.
Regarding ribbon cable, agree that can be work well if done right. Problem with the AKM board and mating to the I2SoverUSB board is that neither board was designed with alternating grounds and signal lines that would make it practical to try it. The signals were out of order, there were gaps header pin assignments between I2S signals, etc., so it would have required hand cross connecting the ribbon wires and soldering on. Unfortunately, ribbon wires tend to be delicate, brittle, and easily fatigue crack if soldered instead crimped. Therefore it was not a practical solution in that case.
here are URLs with images from the main board:
https://www.hi-files.com/forum/index.php?/topic/38487-smsl-m8a-dac/
https://mysku.club/blog/aliexpress/79127.html
https://www.audiophonics.fr/en/dac-...8q2m-32bit-768khz-dsd-xmos-xu208-p-12934.html
Only for use with outdoor power supply.
It seems not many have found that error. So links, pictures, overload of information and reviews are not always necessary or helping at all. Not even 1 reviewer noticed it. All reviews are positive while the device was not good at all due to the error. Used with USB is even stuttered now and then.
There is a thread describing the simple and cheap 1 part fix. I kept a M8A for myself as it is good enough for me. Of course it is used with a linear PSU as mine had a 9V SMPS delivered instead of the required 5V. The micro/nano display indicated too high supply voltage which was a first for me.
There is a thread describing the simple and cheap 1 part fix. I kept a M8A for myself as it is good enough for me. Of course it is used with a linear PSU as mine had a 9V SMPS delivered instead of the required 5V. The micro/nano display indicated too high supply voltage which was a first for me.
Last edited:
Last week I have heard this DAC model against an old version of PCM1702 DAC - used in Thule Audio's CD150B - go to second image in post #40 under
https://www.diyaudio.com/community/threads/thule-audio-components-overview.180929/page-2
for an image of this DAC PCB.
Several years ago I/U section of this DAC board was replaced by that one used in Pass' D1 version - go to
https://www.diyaudio.com/community/threads/pass-d1-dac-question.369812/
(not to find on the mentioned image)
Several differences were clearly audible.
The most noticeable was a slightly higher resolution, but there was also a flat image and a more sharp and crispy sound - definitely not really better at all than Thule Audio's CD150B in this condition (with the genuine I/U converter, which is OP-amp based, the M400 would probably have been preferred).
If there is time, I will check the sonic character against this DAC made in Germany at an other owner, who lives in the same area:
https://www.diyaudio.com/community/...te-dac-horch-steps-1-schematic-wanted.345211/
Maybe after done this it is better to estimate, how much influence there is by using an old fashion 50/60Hz power supply together with such large electrolytic caps.
Don't surprise us with what many already know
Retest with a recent DAC and then with linear PSUs.