i actually could place a CC pad on the backside, so if one desires can remove R2, tho the current ferrite bead is only rated for 1,9A, which makes this a bit problematic
Edit: well i might just include it like this
Edit: well i might just include it like this
Last edited:
@Ghoostknight : The datasheet for a USB-C male connector https://www.tme.eu/Document/33f47f021aa23f1431a5aebac3049b3b/USB4151.pdf says:
IIUC it means the connector itself has the shield connected to A1/A12/B1/B12 internally, which is what @bohrok2610 is saying, IIUC.
IIUC it means the connector itself has the shield connected to A1/A12/B1/B12 internally, which is what @bohrok2610 is saying, IIUC.
CLC filter on Vbus needs serious verification that it does not ring and does not do more harm than good. When looking at the whole thing in use, there is actually a CLCLCLC filter on Vbus, the outer Cs are on the host and device and the outer Ls are the cable (or extensions etc), and the inner CLC is what you add. If you short the central ferrite, it is still CLCLC.
well i was hoping i could spare the one shield latch because it ruins the layout quite a bit but will see what i can do, tho im 100% sure it would work, just not "on spec"
im using this one: https://www.molex.com/content/dam/m.../salesdrawingpdf/105/105444/1054440011_sd.pdf
would there be an alternative? i mean, if oscillation could happen doesnt it come down to both devices and the cable combination, so swapping any of the 3 options the resonancefrequency would change?
------
i actually also have to also add a 10k pullup resistor on CC on the device end as i just found out, so either make room for solderbridges to choose or make 2 pcbs
this CC stuff seems pretty dumb tho... if i understand correctly and you have a "dumb" usb port (or some adapter also utilizing pullup resistors without further circuitry) you can actually short out two power supplies.... also the cable becomes directional, i can place some "audiophile markers" then 😀
but its the only option if i dont wanna run a CC line....
The point is to make sure the ferrite is already dominantly resistive at the frequencies where the capacitors short so that the RF energy is converted to heat rather than reflected.
Sometimes it makes sense to parallel a (low-capacitance) resistor to the ferrite to make it resistive at lower frequencies.
Simulation of the scenario helps, looking into the source impedance the Vbus end consumer sees.
looks like the resonant frequency of the cap is around 4 Mhz https://www.lcsc.com/datasheet/lcsc_datasheet_2409272232_YAGEO-CC0603KRX7R9BB105_C559769.pdf (can i take the resonant frequency from the ESR vs frequency graph like that?)
https://www.murata.com/en-eu/products/productdetail?partno=BLM18SP101SH1# where the new ferrite bead i chose (which goes up to a few ghz) is at around 20 ohm
https://www.murata.com/en-eu/products/productdetail?partno=BLM18SP101SH1# where the new ferrite bead i chose (which goes up to a few ghz) is at around 20 ohm
- removed the rc on shield (only one solderbridge for shield -> GND connection)
- added seperate pullup/pulldown resistors (3A) choosable via solderpads for "source" and "device" plug, it makes it a directional cable but really the only option if you dont wanna put a CC line
- added solderpad for CC line if your cable supports it, tho i imagine there are basicly non 2.0 data cables with cc line
- connected both shield pads
what i might experiment with is not putting shield on the board at all on the device end (or dont connect GND)
Any more thoughts on this?
i checked cost.... 30 pieces would cost 130 at jlcpcb
which comes out around 4-5 euro per piece
probably cheaper if you take 50+ but i dont wanna commit that much without having heared it in action, just fairly certain that the last revision should work out just fine
i checked cost.... 30 pieces would cost 130 at jlcpcb
which comes out around 4-5 euro per piece
probably cheaper if you take 50+ but i dont wanna commit that much without having heared it in action, just fairly certain that the last revision should work out just fine
cant really argue about the effectiveness till i heared it...
well, i have heared CMC's on the data line already and it was an improvement so im fairly optimistic here
very few dacs under 1-2k use usb isolation inside the dac and even then, theoretically this should also reduce radiation from the cable, not just what is going into the dac (how much you weight this as a pro is up to you tho)
pretty much any dac i heared with standard xmos/amanero implementation suffers from audible degredation because of USB and its tweakable with isolator/filters/cables
Well if you run a true PC setup there are very few alternatives and even fewer above 192khz (not that you need it but its argueably the most future proof)
Then avoid the true pc setup and USB. You can easily do without either of those and still have high res audio (without RF).
If one jumps in the water one should not complain the water is wet.
If one jumps in the water one should not complain the water is wet.
Last edited:
Have you made any measurement that proved what you say , it will be interresting to see them and maybe talk about
.
might work for music but not for movies/games/production, atleast non hasslefree solutions im aware of (DLNA too much latency for games for example)
the thing is i had an ian canada transport and ultimatively a tweaked usb source beat it while an untweaked straight wire usb source sounded worse
no, im pretty sure its measureable if you look "beyond" the 20-22.05khz audible output of a dac, but noone of the objective camp does for obvious reasons
beside this i might link to some ASR reviews about usb cables: https://www.audiosciencereview.com/...s/do-usb-audio-cables-make-a-difference.1887/
the picture is definitely not a clear cut, people just keep arguing whether its audible..... its exhausting to say the least
Also since i dont have the equipment.... if anyone wants to measure this in action, hit me up, some samples for free of course 🙂
the picture is definitely not a clear cut, people just keep arguing whether its audible..... its exhausting to say the least
Also since i dont have the equipment.... if anyone wants to measure this in action, hit me up, some samples for free of course 🙂
How would you know what a measurement proved?
From what I have seen so far most people don't seem to know all that much about measuring noise in all its guises, and or know whether or not particular noise effects have been studied for audibility in particular dac topologies.
Last edited:
imo there are many off the shelf dacs (pretty much any sub 1k interface/dac) that could benefit from additional filtering between isolator and dac, listening tests have to confirm but they are alot of usb tweaks that seem to be beneficial, atleast in my expierence
if we talk higher end dacs with isolation build-in the picture might change a bit tho, i would still give a shot, specially if the prices for these cables are reasonable
and like you say, if data arrives all should be fine, so i rather take a bit additional filtering 🙂
That may very well be true, but dacs like that I have opened up and looked at appear to have multiple problems affecting SQ. USB noise incursion is only one of them. If the dac was designed right in the first place it wouldn't have any of those problems. But then it wouldn't sell because it would cost too much for the same SINAD numbers when measured by an AP. However the AP isn't measuring everything, such as susceptibility to USB noise incursion. So everybody will say its overpriced in terms of SINAD/$. That's how people end up with cheap dacs like that. Then they add on a few hundred more $$$ trying to fix it so it sounds good.