But I don't think I2S is differential, is it? I know USB is but I'm talking about leaving off the ground in I2S. That I2S spec link above just gives the high & low voltage thresholds but here's something that was just posted on the slimdevices forum where I posted this info also:
Just to clarify my set-up - At the ESS DAC, I'm using the MCLK coming from I2S & not providing a local clock to the ESS DAC!
Last edited:
Per Lundahl recommended a Lundahl transformer? Wow!
Seriously, I2S or any other single polarity digital signals make no sense without a ground reference. Optocouplers or transformers can be used for galvanic isolation but both will cause some degree of signal distortion and introduce jitter. Whether this will be preferable to ground noise is far from obvious and will depend greatly upon the asrc.
Seriously, I2S or any other single polarity digital signals make no sense without a ground reference. Optocouplers or transformers can be used for galvanic isolation but both will cause some degree of signal distortion and introduce jitter. Whether this will be preferable to ground noise is far from obvious and will depend greatly upon the asrc.
What about using a big choke in the USB ground - has anybody successfully done this? I had a PCM2707 based board from some company, I can't remember the name at the moment. It was posted about in here & it had a problem when operated in I2S mode - it wouldn't work. The problem boiled down to a choke on both the VBus & USB ground lines. I seem to remember the choke on the USB ground caused the problem! Any chance of this working to reduce USB ground noise with suitable sized/types of chokes?
afaik they use fast optocouplers which are just fine at usb1.1 speed.. even if jitter is introduced, they always move data asynchronously to controller.
no way at 2.0 hi speed..
jkeny: you replied faster
Vale
Hi John,
those lifepo batteries are quite a bitch to solder..
I was running two alkaline AA cells before introducing lifepo cell and my hiface was already quite modded .. so, direct comparison between stock and full modded lifepo version is not available, yet.
anyway.. I confirm my initial impression.. battery supply on clocks makes a difference.. listening is really effortless.. super!
my system now is: TP wolfson 8742 + aleph p.1.7 + F4 + fostex 103 sigma in tqwt box .
I'll compare stock (I have one unmodded) to mine soon
regards
ciao
Vale
afaik they use fast optocouplers which are just fine at usb1.1 speed.. even if jitter is introduced, they always move data asynchronously to controller.
no way at 2.0 hi speed..
jkeny: you replied faster
Vale
So the trick is to isolate it during the asynchronous state (USB data stream). Seems like no way to do that with USB 2.0. Anything else like doing an ethernet asynchronous receiver is probably beyond the ability of anyone in these forums. At least anyone who help out. Or we'll have to wait for Light Peak
Told you about thatHi John,
those lifepo batteries are quite a bitch to solder..
- be careful to use a big high wattage iron with a large tip (especially on the positive end which pulls the heat away quickly) - you want to only have the iron on there for a couple of seconds - can damage the battery if heated for too long. Have you lightly sanded the ends & applied some flux first? so you had already removed the series inductor that the clock PS goes through & connected the new independent clock supply at that point?
Have you tried my other mod - supplying independent 3.3V to Vccio pin of Xilinx? This makes a difference too
Very nice kit!
Look forward to it
So the trick is to isolate it during the asynchronous state (USB data stream). Seems like no way to do that with USB 2.0. Anything else like doing an ethernet asynchronous receiver is probably beyond the ability of anyone in these forums. At least anyone who help out. Or we'll have to wait for Light Peak
To my understanding, yes.. since audio clock is not derived from asynch usb stream, as long as optocouplers don't corrupt (10x faster than stream I guess) data transfer everything should be just fine..
Vale
yes, otherwise near to impossible with 60w iron.. heat spread very quickly!
yes, I followed your instructions.. thanks for photos
I'll try soon.. waiting for extra batteries to arrive
thanks!
Look forward to it
soon
ciao
Vale
Empirical Audio is Supposedly doing galvanic isolation in their new 24/192 Pacecar product.
Async 192 USB upgrade to Pace-Car 2
Maybe they are still using USB 1.1?
Hi labjr,
yes, it's interesting product, based on m2tech drivers and firmware, so UBS2.0 hi speed, for sure.. USB1.1 is not fast enough
probably they are using fast enough optocouplers or they are using a different approach currently obscure to me. dunno
regards
Vale
Hi labjr.
from audiocircle thread Empirical says: "2) USB is galvanically isolated (this can only be done in a Pace-Car type reclocker to avoid adding jitter)"
since this option is only available in their reclocker it probably means that they are isolating USB interface at i2s .. added jitter introduced by optocouplers is then removed by reclocking i2s
of course, I'm guessing.. I don't know their products.
regards
Vale
Hey, guys,
I'm offering a service to mod the HiFace USB transport for PS improvements & to allow I2S output (Price 50euro) The PS mods really allows this transport's low jitter clocks to work properly (as evidenced here) & the I2S will allow it to plug directly into an ESS DAC board here http://www.diyaudio.com/forums/digital-line-level/147817-ackodac-based-es9018.html or my ESS Vout DAC based on the ES9022 DAC or any DAC that you can feed I2S to (& know how to do it).
Anyone interested? Email me rather than PM me if you are interested as I want to keep control of my correspondence!
I'm offering a service to mod the HiFace USB transport for PS improvements & to allow I2S output (Price 50euro) The PS mods really allows this transport's low jitter clocks to work properly (as evidenced here) & the I2S will allow it to plug directly into an ESS DAC board here http://www.diyaudio.com/forums/digital-line-level/147817-ackodac-based-es9018.html or my ESS Vout DAC based on the ES9022 DAC or any DAC that you can feed I2S to (& know how to do it).
Anyone interested? Email me rather than PM me if you are interested as I want to keep control of my correspondence!
There's a review of thw Hiface on enjoy the music. Claims they use galvanic isolation. I don't believe this is true is it?
M2Tech hiFace 24-bit/192kHz USB Digital Audio Interface A journey into the world of PC audio. Review By Mike Galusha
M2Tech hiFace 24-bit/192kHz USB Digital Audio Interface A journey into the world of PC audio. Review By Mike Galusha
It's a bit misleading. What he's saying is that the SPDIF pulse transformer offers galvanic isolation from the PC - the same as every other USB unit that uses a SPDIF transformer so there's nothing special about it & it's not a feature specific to the HiFace!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.