Why do you think you need to understand everything?
99.9% of the people who enjoy the benefits of Aoip have absolutely no clue of the workings.
Design us a USB Interface we can purchase for $260, has 6 or more channels and sounds as good as the 96khz Monism Dante modules and I will be the first to switch.
Guys with all your expertize there certainly should be more in it for us diyers that mere critique of a certain technology.
You make it sound so simple in all your arguments, but year after year nothing productiv happens.
On the contrary those who make the best of there possibilites you try to discorage.
You make it sound so simple in all your arguments, but year after year nothing productiv happens.
On the contrary those who make the best of there possibilites you try to discorage.
You haven't heard DSD256 then?
Maybe so in your case, which would be perfectly fine. Probably unwise to assume everyone else hears the same way you do, is all.
Yeah, I am aware of what tends to get used in studios. Also aware of what gets used in some of the best mastering facilities. There is a range of equipment that gets used professionally. Not all of it can be the best of the best.
Not fooled here. Just know that some DACs sound better when fed DSD256. Its a quirk of some DACs. Has nothing to do with native per se.
People will be happy when the SQ is there. Maybe some day.
Some DACs are sounds better in all DSD than in PCM.
As for me - Rohn BD34301 one of them.
@samadi81 Thank you for sharing the information and schematics, I will definitely keep this in mind as I learn. I landed on this thread due to looking at the ISOUSB211 platform as an alternative ahead of the Amanero and here, the AN ADUM devices are used and talked about. Since I am not an engineer, I can only observe the debates, which are above my paygrade, but I can still engage a bit 
Are there any more or newer SQ info available regarding improvements with your isolator in place, have the device reached more users ?
Are there any more or newer SQ info available regarding improvements with your isolator in place, have the device reached more users ?
@samadi81 Thank you for sharing the information and schematics, I will definitely keep this in mind as I learn. I landed on this thread due to looking at the ISOUSB211 platform as an alternative ahead of the Amanero and here, the AN ADUM devices are used and talked about. Since I am not an engineer, I can only observe the debates, which are above my paygrade, but I can still engage a bit
Are there any more or newer SQ info available regarding improvements with your isolator in place, have the device reached more users ?
hi . Many aspects of the Adum4166 design and testing process seem to affect sound quality. The most influential factor is the quality of the external input power. The board I made uses the LT3045 regulator, and as shown in the LT3045 datasheet, the SET pin, Input, and Output capacitors also had an effect.
There are also variables depending on the board's ground and shield connections. Since I don't use a metal shield, I'm basically using an RC connection, but the results vary depending on whether I short it or not connect it.
Of the five boards I ordered with the support of JLCPCB, I gave one to someone else and am currently testing the remaining four by replacing and tuning the elements.
I am currently testing a separate external power supply, but there are some improvements.
Attachments
Thank you for sharing your experience with ISOUSB211. When I made and used the ADUM4166, it was not affected by the order of signal input, external power input, and DAC connection as described in the data sheet.
It depends of a device and it's driver.
I make ~100 pcs with TI and ~50 with ADUM. Most of them customers used with Cosmos ADC, but only few user reports about this issue.
So, I think that it maybe also depends of a PC hardware.
P.S. I don't understand why you use 4166.
Difference between 316x and 416x - only the maximum input-output voltage, which is not an issue in audio at all.
Difference between x165 and x6166 is more complicated, I believe the right way is 65 and not 66.
Alex.
I was mistaken for the previous model. I just looked up adum316x. Thank you for letting me know
That's right, it's as you said^^. I guess 316x is a model after 416x. I wonder if there is any improvement in 316x compared to 416x other than the maximum insulation voltage. The pin arrangement is the same, so I'm going to order it, remove the 416x, replace it, and test it.You are welcome
I think you confused 3165/66 with 3160, which is just Full Speed, not High.
Alex.
I don't know, but both datasheet marked (c)2022.
It is usually just a distance between pins and a package. Internally maybe the same.
I made my own design.
Alex.
Alex,
The ADUM problem you see with sequence has to do with the boot time. This is in the data sheet for them. If the HOST side has a slow ramp up voltage it will in cases lock out the DEVICE side, if the DEVICE side is already powered up. I had this problem when I regulated the HOST side.
I have the TEK USB Compliance Test Set here because of all my work with USB. Comparing the TI to the AD gave the edge to the AD part. But both of these parts require some extra work as the differences in jitter can be really evident. You can really tell that from the eval boards of both of them.
Jitter in these types of devices especially the 66 type will have to do with the 3V3 supply noise and the oscillator or crystal setup used for re-transmission. Transmission is really the key to audio that is why I choose the 4166 (plus the 3xxx versions were not out yet). Since any USB UAC/UAC2 device will basically be a receive only device other than enumeration and the feedback pipe. It was best to use the 4166, feed it with a ultra low noise 3V3 regulator and a quality crystal. I could have used a Crystek or NDK oscillator for this but choose not to.
Most of the jitter measurements in the data sheet are based on the eval board. Since there is no re-transmission on Full Speed the numbers are a lot higher.
You can get better results in High Speed with the re-transmission.
One note on all this stuff is they had to make room in the protocol to buffer and re-transmit. Therefore cable length can make or break a device like this. Just remember really 2M total length is just about the max using these devices.
Thanks,
Gordon