Hi all,
I've just been thinking, in preparation for a USB DAC build, about, strangely enough, USB. Plan is to use the avaiable Twisted Pear modules to output I2S.
Anyway - I've been thinking about ground. It seems to me that for noise reasons, sharing a ground with the PC is extremely undesirable, but since USB requires a ground connector I can't see a way around this.
So, how can this be avoided? Conversion to SPDIF over TOSLINK seems to be one way to avoid connecting the grounds, but it's not an option I'm terribly attached to.
Two questions: Is this actually the problem I think it is?
Secondly, is it feasible to use transformers in order to couple the signal without a direct connection? If the output side of the transformer is referenced to the clean(er) ground of the DAC, will this reduce the ground noise, or would the ground noise simply be modulated into the signal anyway? But wait, since USB is a digital signal, wouldn't that noise be insignificant, but we have, in the process, gained ourselves a clean ground for audio output?
Thanks,
Chris
I've just been thinking, in preparation for a USB DAC build, about, strangely enough, USB. Plan is to use the avaiable Twisted Pear modules to output I2S.
Anyway - I've been thinking about ground. It seems to me that for noise reasons, sharing a ground with the PC is extremely undesirable, but since USB requires a ground connector I can't see a way around this.
So, how can this be avoided? Conversion to SPDIF over TOSLINK seems to be one way to avoid connecting the grounds, but it's not an option I'm terribly attached to.
Two questions: Is this actually the problem I think it is?
Secondly, is it feasible to use transformers in order to couple the signal without a direct connection? If the output side of the transformer is referenced to the clean(er) ground of the DAC, will this reduce the ground noise, or would the ground noise simply be modulated into the signal anyway? But wait, since USB is a digital signal, wouldn't that noise be insignificant, but we have, in the process, gained ourselves a clean ground for audio output?
Thanks,
Chris
I doubt that you could get USB signalling to work across a transformer: each end - client and host - need to be able to pull the data lines to DC levels.
But you COULD do it with opto-couplers. It would need a bit of driver circuitry to achieve - not really a project to try if you aren't up on logic design.
Such things can be bought ready made:
E.G.:
http://www.busytrade.com/products/info/740601/Usb_Isolator_Cable.html
There is talk that the timing issues with USB make it not great for top-quality sound; I have no personal experience - I've never put 'live' audio through USB. Whatever the issue, an isolater might contribute to this. Or might not.
But you COULD do it with opto-couplers. It would need a bit of driver circuitry to achieve - not really a project to try if you aren't up on logic design.
Such things can be bought ready made:
E.G.:
http://www.busytrade.com/products/info/740601/Usb_Isolator_Cable.html
There is talk that the timing issues with USB make it not great for top-quality sound; I have no personal experience - I've never put 'live' audio through USB. Whatever the issue, an isolater might contribute to this. Or might not.
Optically Isolated USB
http://3dotaudio.com/opticis.html
Those cables appear to be the sort of thing that would do the job.....rather expensive though, not the sort of thing that I'd throw money at just for fun.
So, back to the original question: is ground contamination from the PC source a significant issue for a USB DAC, considering that the same ground will be used all the way through the system including the analogue output stage?
What about opto-isolated I2S? Or would it be easier, in such a case, to just output SPDIF over TOSLINK back into a SPDIF reciever?
http://3dotaudio.com/opticis.html
Those cables appear to be the sort of thing that would do the job.....rather expensive though, not the sort of thing that I'd throw money at just for fun.
So, back to the original question: is ground contamination from the PC source a significant issue for a USB DAC, considering that the same ground will be used all the way through the system including the analogue output stage?
What about opto-isolated I2S? Or would it be easier, in such a case, to just output SPDIF over TOSLINK back into a SPDIF reciever?
I had a look round, isolating USB is much harder than I guessed, hence he high cost of the boxes to do it. If you are *really* into DIY, here is a possible project:
http://www.maxim-ic.com/appnotes.cfm/an_pk/3891
SPDIF, via transformer, or optical (oslink) seems way easier and cheap!
Ground loops will not be a problem for the USB or digital signal as such; the problem is that through that cabling, you are grounding the analogue circuitry - which CAN cause problems.
An acceptble solution there might be to power the DAC either from the USB line, or better from its own isoltated supply (its own mains transformer - or even better, batteries), and the use a good audio transformer to couple into other audio (analogue side) boxes. (maybe you described this situation, i couldn't quite follow this part in your original post).
http://www.maxim-ic.com/appnotes.cfm/an_pk/3891
SPDIF, via transformer, or optical (oslink) seems way easier and cheap!
Ground loops will not be a problem for the USB or digital signal as such; the problem is that through that cabling, you are grounding the analogue circuitry - which CAN cause problems.
An acceptble solution there might be to power the DAC either from the USB line, or better from its own isoltated supply (its own mains transformer - or even better, batteries), and the use a good audio transformer to couple into other audio (analogue side) boxes. (maybe you described this situation, i couldn't quite follow this part in your original post).
Hello all-
There is a product that might be fairly close to what you are looking for, made by B & B Electronics:
B&B UISOHUB4 Isolated Hub datasheet(PDF)
I have found that B&B makes very good quality, robust products, but they often charge a little more than some other Mfg. do for similar items. While not exactly a transformer type of isolation, perhaps it might give you a clue as to what items possibly are needed to accomplish this for your uses....
HTH-
There is a product that might be fairly close to what you are looking for, made by B & B Electronics:
B&B UISOHUB4 Isolated Hub datasheet(PDF)
I have found that B&B makes very good quality, robust products, but they often charge a little more than some other Mfg. do for similar items. While not exactly a transformer type of isolation, perhaps it might give you a clue as to what items possibly are needed to accomplish this for your uses....
HTH-
Hi all,
My concern is not so much about ground loops, as they're easily understood and avoided, but about ground noise, from the pc itself, rendering the ground a less than perfect reference - not so much a problem for the digital stages, but it seems like ground contamination could be a problem by the time it reaches the analogue stages. But then, maybe that's where I'm wrong....since the output signal IS referenced against ground, does it matter if ground is an imperfect reference, so long as the output level of the signal itself floats with it?
@Cliff45 - yes, opto-isolators can be used to do it, and that's what the product I linked to, as well as the product you linked to, utilise for the job. However, they're a) expensive, and b) said to introduce additional jitter - while I'm not sure jitter is the problem some say it is, it still seems worthwhile to avoid.
My concern is not so much about ground loops, as they're easily understood and avoided, but about ground noise, from the pc itself, rendering the ground a less than perfect reference - not so much a problem for the digital stages, but it seems like ground contamination could be a problem by the time it reaches the analogue stages. But then, maybe that's where I'm wrong....since the output signal IS referenced against ground, does it matter if ground is an imperfect reference, so long as the output level of the signal itself floats with it?
@Cliff45 - yes, opto-isolators can be used to do it, and that's what the product I linked to, as well as the product you linked to, utilise for the job. However, they're a) expensive, and b) said to introduce additional jitter - while I'm not sure jitter is the problem some say it is, it still seems worthwhile to avoid.
At work I've built a strain gage amplifier with USB intf and isolation - between the usb chip and my data aquisition and amplifier circuitry
I used new fast micro transformer isolator chips:
http://www.analog.com/en/interface/digital-isolators/products/index.html
The i2s isolator is a little odd, requiring care in pointing the "i2s compliant" and "i2s interoperable" sides in the right orientation
I used new fast micro transformer isolator chips:
http://www.analog.com/en/interface/digital-isolators/products/index.html
The i2s isolator is a little odd, requiring care in pointing the "i2s compliant" and "i2s interoperable" sides in the right orientation
Re: Optically Isolated USB
This is IMHO the easiest way. I used this for many years to avoid a terrible ground loop between the PC and the audio system.
While theoretically Toslink is inferior to coaxial, in reality this may be hardly audible. I'm not a bits-is-bits guy as I once heard a more than subtle difference between two CD-players that were connected to the same DAC.
But I also listened to one player that was connected to the DAC both ways at the same time, just a matter of switching between coaxial and optical inputs. I thought Toslink to be inferior, until I noticed that was only the case when I knew which source was selected. Repeated double-blind, I couldn't tell the difference. Now, this was hardly high-end equipment I used at the time for this comparison, but it IMHO goes to show that Toslink may not be so inferior as many say it to be.
TheSeekerr said:
This is IMHO the easiest way. I used this for many years to avoid a terrible ground loop between the PC and the audio system.
While theoretically Toslink is inferior to coaxial, in reality this may be hardly audible. I'm not a bits-is-bits guy as I once heard a more than subtle difference between two CD-players that were connected to the same DAC.
But I also listened to one player that was connected to the DAC both ways at the same time, just a matter of switching between coaxial and optical inputs. I thought Toslink to be inferior, until I noticed that was only the case when I knew which source was selected. Repeated double-blind, I couldn't tell the difference. Now, this was hardly high-end equipment I used at the time for this comparison, but it IMHO goes to show that Toslink may not be so inferior as many say it to be.
oops, i2c is provided byADuM125x isolators, that I used, too slow for digital audio
i2S serial digital Audio uses unidirectional signals, any one of several of the ADuM Standard Digital Isolator series with the correct number/direction channels should do
I've used the ADuM3400/1C 90MBps isolators for the parallel bus isolation in the same product
i2S serial digital Audio uses unidirectional signals, any one of several of the ADuM Standard Digital Isolator series with the correct number/direction channels should do
I've used the ADuM3400/1C 90MBps isolators for the parallel bus isolation in the same product
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.