WaveIO ground
Check this post from Lorien: http://www.diyaudio.com/forums/digi...ronous-usb-i2s-interface-189.html#post3708036 .
Also, it is documented on the board itself:
Check this post from Lorien: http://www.diyaudio.com/forums/digi...ronous-usb-i2s-interface-189.html#post3708036 .
Also, it is documented on the board itself:
Attachments
I found both reclocking and isolation improved the sound of the Doede's dac with the rpi, if they don't improve the sound of the wandboard I will have to get one and try it - less components in the chain = less jitter! Thanks for pointing that out.
Does the wandboard's pll generate the exact frequencies, or does it fudge a little like the rpi?
This is a note from WaveIO;
Other notes:
Please do NOT use non-isolated SPDIF output since is was built for testing purposes only, use instead the isolated one instead. In addition, H1, H2 and H3 mounting holes are NOT connected to the ground planes while H4 is... so, be careful!
See also PDF
Hi,
I just bought some NVE isolators as Palmito suggested.
I will try this next week.
@Palmito: I assume you also need to isolate the ground connection to the dac? Otherwise you still get the ground pollution from the Pi.
I also ordered new power supply components: EI core 100VA transformer, shottky diodes, Lundahl chokes. I have some Back gate on stock so I think this will make a very nice raw power supply for the dac.
Regards,
Look at the datasheet for the isolator, it isolates the signal and the ground. The input side is galvanically isolated from the output side. Each side has to be powered by a separate 3.3vdc power supply. For the input side I tried both the rpi 3.3v from the P5 header and a separate power supply from the one powering the rpi, I can't say I could tell the difference. The output side you can power from Doede's 3.3vdc output. You need to power both sides and from different sources or, as you state above, the ground noise will be transmitted. Did I understand you question correctly?
I don't think so, as long as it's clean (the current draw is very small). The important one would be the output side, and we know Doede's 3.3vdc is clean. The isolator does add jitter, that's why many people reclock after isolation. Still, it seems to me like even with added jitter, the benefit of isolation improved the sound some (with the NVE one, I didn't like the sound with the TI one, but that one had ferrite beads in the power supply line). It will be interesting to hear your results! With the S03, Acko suggested using a 100mhz clock, since Doede's dac generates the mck from the bck. We tried the highest clock I had in his board. It did improve the sound. The lastest iteration of my S03 connected to Doede's dac has a 100mhz Crystek CCHD-950.
Last edited:
I had some fun last night updating the DDDAC and found out a few things along the way that might be of interest.
Two big changes I made where:
1- Upgraded the standard 30Vac to an 80Vac transformer to the DDDAC PSU & jumpered/bypassed R1&R2 (I already upgraded this PSU with Mundorf’s for C1&C3 and by leaving the fuse out, which I’ve put at the power inlet)
2- I put in the DIYINHK PSU for the WAVEIO board.
Ad1. So first big eye opener was the fact that the sound became fuller and significantly more dynamic also the gain for the four board stack is higher, possibly by a factor of 2 to 3…!. Secondly the heat sink of the voltage regulator is becoming much warmer. Previously is this was hardly warm to the touch, now it’s 51Celcius according to my heat gun. This leads me to suspect the Standard 30Vac transformer is no were near big enough to power a four board stack. Giving the Dac board plenty of power is an easy, obvious and big way to improve SQ.
Ad2. I was having quite a bit of high frequency noise (@ no input signal, full gain), which made me wonder if I was going to play around with a LP filter. However now that I’ve changed the PSU for the WaveIO board this is but all gone. I believe the ground plain of the WaveIO and DDDAC are connected so PSU ground noise here is sensitive. (As Palmito and Supersurfer have also suggested)
Two big changes I made where:
1- Upgraded the standard 30Vac to an 80Vac transformer to the DDDAC PSU & jumpered/bypassed R1&R2 (I already upgraded this PSU with Mundorf’s for C1&C3 and by leaving the fuse out, which I’ve put at the power inlet)
2- I put in the DIYINHK PSU for the WAVEIO board.
Ad1. So first big eye opener was the fact that the sound became fuller and significantly more dynamic also the gain for the four board stack is higher, possibly by a factor of 2 to 3…!. Secondly the heat sink of the voltage regulator is becoming much warmer. Previously is this was hardly warm to the touch, now it’s 51Celcius according to my heat gun. This leads me to suspect the Standard 30Vac transformer is no were near big enough to power a four board stack. Giving the Dac board plenty of power is an easy, obvious and big way to improve SQ.
Ad2. I was having quite a bit of high frequency noise (@ no input signal, full gain), which made me wonder if I was going to play around with a LP filter. However now that I’ve changed the PSU for the WaveIO board this is but all gone. I believe the ground plain of the WaveIO and DDDAC are connected so PSU ground noise here is sensitive. (As Palmito and Supersurfer have also suggested)
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Hi Stijn,
Thanks for posting your results. It also seems you are a bit quick with conclusions, so let me react:
ad1 Using a large transformer does make a difference, as in Audio SQ, not always easy to understand why, but that does not matter. Claiming the smaller transformer cannot power 4 boards is totally wrong. It easily can. The higher heat has to do with the higher Voltage after the diodes with your new transformer. Just measure it, you will see. The Standard transformer also does 8 stacks, without any Problem. On Improvement of SQ that is anpther story and has to do with our beloved Audio Hobby tweaks
ad2 the groundplanes are NOT connected, so this cannot be the cause. Have you been using another 5 Volt supply at the time you had noise Problems? or did you powered the USB Board throug your PC USB bus?
The Motherboard spacer holes are not connected to GND, so that is not needed ...
Hi Stijn,
Thanks for posting your results. It also seems you are a bit quick with conclusions, so let me react:
ad1 Using a large transformer does make a difference, as in Audio SQ, not always easy to understand why, but that does not matter. Claiming the smaller transformer cannot power 4 boards is totally wrong. It easily can. The higher heat has to do with the higher Voltage after the diodes with your new transformer. Just measure it, you will see. The Standard transformer also does 8 stacks, without any Problem. On Improvement of SQ that is anpther story and has to do with our beloved Audio Hobby tweaks
ad2 the groundplanes are NOT connected, so this cannot be the cause. Have you been using another 5 Volt supply at the time you had noise Problems? or did you powered the USB Board throug your PC USB bus?
Hi Doede, Thanks, Second point first. Yes you are right my old PSU for WaveOI was of inferior quality. Thanks for clearing the ground plain question. What I don’t understand then is how the high frequency noise could have reach the analogue domain or do you think this could be attributed to jitter/bit recovery errors?
As for the DAC PSU. I , take your word for it that the voltage is higher after the rectifier compare with the old transformer, I can no longer measure before and after without swapping them out again. Logically however, I don’t understand how this could be, as both the old and my new transformer have 2x15V secondary winding (which I hadn’t stated yet).Therefore, in order to explain the increase in heat put out by the regulator would have been an increase in current, hence demand, hence my "assumption" that the old transformer wasn’t delivering enough. Also I made the other PSU “mods” before I swapped the transformer and the regulator heat sink stayed cooler at that time, so I can’t explain the increase in SQ that way.. I hear what you are saying about, it being the combination of.., but I’m still a bit puzzled, and eager to learn obviously, so please don’t take my questioning the wrong way.
Last edited:
I got better results with this (compared to the TI put together per diyinhk bom):
www.digikey.com/product-detail/en/IL715E/390-1062-5-ND/964108
BTW, if you are using WaveIO, this is the same isolator as in the WaveIO board, different chip sizes, but the same isolator.
I jumpered L1 & L2.
I used these in C1/C3 (I used them not because I know what I'm doing, but because this is what Acko uses on his S03 board) :
www.digikey.com/product-detail/en/C1608X7R1H473K080AA/445-1313-1-ND/567686
Except for the pin headers, I did not populate anything else on the diyinhk board.
This is an actual photo of my board: http://www.diyaudio.com/forums/digital-line-level/224108-nos-192-24-dac-pcm1794-waveio-usb-input-148.html#post3882762 . When connecting to the input side of the board, remember this was designed to fit the Amanero usb->i2s board, so you need to look at the amanero pin diagram (or follow my photo above).
Last edited:
S03 with Crystek
So far this is the best the rpi has sounded on my system, your mileage may vary...
I'm looking forward to clivem's wandboard implementation or Russ' cape for the bbb, as they do promise better sound than the rpi. But right now with the rpi, I'm happy!
So far this is the best the rpi has sounded on my system, your mileage may vary...
I'm looking forward to clivem's wandboard implementation or Russ' cape for the bbb, as they do promise better sound than the rpi. But right now with the rpi, I'm happy!
