In general I think the length of i2s traces from a source to two different DAC chips should be insignificant. Where i2s cable length is more significant is for differences between wire length to the same dac chip for bck, data, lrclk, mclk signals ...
Another thing to remember is that just because it works, doesn't mean its been optimised. So just because others have done something before, doesn't mean it is correct!
Another thing to remember is that just because it works, doesn't mean its been optimised. So just because others have done something before, doesn't mean it is correct!
Hahaha...
It is audio, there is more than one "truth", tubes vs, solid state anyone? See Demian's response above for some truth, but, really, I have heard outstanding sounding DAC's which used a single chip, and outstanding sounding DACs which were dual mono... The DAC chips today are so good from a distortion, and signal to noise perspective, that the theoretical advantages of using more than one dual differential SDM chip seem pretty moot to me. Details like this just seem a lot less important than the I/V- analog stage design, the overall layout, and the power supply design, and the digital input implementation.
My DAC uses the ESS 9018, which running in stereo uses 4 DACs per channel on a single chip, two of these seem like serious overkill to me. I am not sure anyone could build an I/V stage which could take advantage of any more resolution than that offered by a single ESS 9018... I suppose there are those who would suggest that cross talk from the R to L channels in a single chip justifies the dual mono approach, even with something like the ESS 9018, perhaps this is true, but the measurements I have seen of commercially available DACs do not seem to demonstrate this.
I remain unconviced that there are real world (audible) advantages to using more than one DAC chip if one is using contemporary SDM chips. Now, perhaps 8 1704s might make sense...
Yet there are quiet a few examples of commercial or diy dacs that work with a dual mono implementation of two dac chips which share the same i2s signals without any buffers. And they seem to operate properly.......
Then where is the truth?
It is audio, there is more than one "truth", tubes vs, solid state anyone? See Demian's response above for some truth, but, really, I have heard outstanding sounding DAC's which used a single chip, and outstanding sounding DACs which were dual mono... The DAC chips today are so good from a distortion, and signal to noise perspective, that the theoretical advantages of using more than one dual differential SDM chip seem pretty moot to me. Details like this just seem a lot less important than the I/V- analog stage design, the overall layout, and the power supply design, and the digital input implementation.
My DAC uses the ESS 9018, which running in stereo uses 4 DACs per channel on a single chip, two of these seem like serious overkill to me. I am not sure anyone could build an I/V stage which could take advantage of any more resolution than that offered by a single ESS 9018... I suppose there are those who would suggest that cross talk from the R to L channels in a single chip justifies the dual mono approach, even with something like the ESS 9018, perhaps this is true, but the measurements I have seen of commercially available DACs do not seem to demonstrate this.
I remain unconviced that there are real world (audible) advantages to using more than one DAC chip if one is using contemporary SDM chips. Now, perhaps 8 1704s might make sense...
it would be interesting to know what delay a difference of 1 or 2 mm in cable length causes, depending on the speed of signal propagation in copper......
Figure 1-2 nS/ft for most cabling. 2mm is probably about 6 pS. If you were talking about HDMI 2.0 and 18 Gbps this becomes actually a consideration, but still not enough to matter. For audio applications the rise time will occupy more than 2 feet of wire. Finding that 6 pS difference will be really hard. And don't worry about getting the WC, data, bit cclock and master clock to arrive at the same instant. That may be the worst possible timing. Most chips buffer the incoming data so it should be stable at the clock edges, not transitioning. Having them all change at the same instant is a good prescription for ground bounce.
Now it's my turn to laugh (friendly) !
You just take the worst example one can find to argue ! Of course a dac chip with 4 dacs per channel is largely enough, no need to have a dual mono configuration. And it is the best dac chip one can find nowadays in terms of S/N ratio, dynamics, and so on.......
But with some older chips I think the dual mono configuration makes sense, if properly implemented.........
Anyway, as you say it is audio, and in this domain there are as many different opinions as there are people !
Kind regards, staki
Last edited:
Right...
Like I said, if one is choosing to use a 1704 then mutilple chips makes more sense. But say with the Wolfson 8742 even, not so much. My comments did reference contemporary SDM chips.
Now it's my turn to laugh (friendly) !
You just take the worst example one can find to argue ! Of course a dac chip with 4 dacs per channel is largely enough, no need to have a dual mono configuration. And it is the best dac chip one can find nowadays in terms of S/N ratio, dynamics, and so on.......
But with some older chips I think the dual mono configuration makes sense, if properly implemented.........
Anyway, as you say it is audio, and in this domain there are as many different opinions as there are people !
Kind regards, staki
Like I said, if one is choosing to use a 1704 then mutilple chips makes more sense. But say with the Wolfson 8742 even, not so much. My comments did reference contemporary SDM chips.
Moreover, maybe I'm completely wrong, but it seems to me that only certain DAC chips allow to be configured in mono mode by software (AD1955) or hardware (PCM1794).
There is no reason to put the other DAC chips in parallel dual mode if both chips still work in stereo mode.......
btw I'm really sorry, but with my poor english language, I'm not sure to speak properly!
I hope people understand what I mean.......
There is no reason to put the other DAC chips in parallel dual mode if both chips still work in stereo mode.......
btw I'm really sorry, but with my poor english language, I'm not sure to speak properly!
I hope people understand what I mean.......
Last edited:
I had the idea first to use two COD modules from Twisted Pear, with PCM1794A in dual mono mode, but suddenly I have seen this kit: http://www.dddac.com/dddac1794.html
I didn't know that Lorien was building WaveIO kits for Doede Douma !
The DAC kit with dual mono PCM 1794A in non-oversampling mode and passive output seems really interesting !
But I wonder about jitter with the used schematics.......
I didn't know that Lorien was building WaveIO kits for Doede Douma !
The DAC kit with dual mono PCM 1794A in non-oversampling mode and passive output seems really interesting !
But I wonder about jitter with the used schematics.......
WaveIO - 5V Power Supplies
If anyone is interested, I have now tried two alternative 5V power supplies for my WaveIO. Both might be considered to be strange and advisable!
1. Recommended to me by a commercial DAC manufacturer - a Chinese open-frame SMPS made by Cincon -
http://www.cincon.com.tw/data/products/cfm1_2/CFM10_15.pdf
Yes - Switched Mode Power Supply!! I'm not proud. Actually, it works really well and is noticeably dynamic, which the manufacturer said was probably caused by its very fast switching frequency. Being so high frequency, he also thinks that most switching noise is not audible being well out of the audible range.
2. LiPo Battery
Well actually an iPhone charging unit - EasyAcc 12,000 mAh LiPo
EasyAcc 12000mAh 4 x USB Portable External Battery Pack: Amazon.co.uk: Electronics
It has 4 USB Type A socket outlets output at 5V. I made up a lead using a Type A USB plug (using Cat 5 Ethernet cable), the other end with bare wires screwed into the WaveIO terminals. It has a switch on it to turn it on and off (not all do). 12,000 mAh should last about 24 hours with a 500 mAh draw. Charges with any Micro USB charger or PC.
SQ excellent. Clean and lovely warmth. Slightly more relaxed than the SMPS but not so dynamic.
Both options were first class and required little DIY skills and were cheap.
Lastly, as always Lorien is right! Using any reasonably good external power supply to the WaveIO is significantly better for SQ than running it off the 5V USB cable lines from the attached PC.
Jonathan
If anyone is interested, I have now tried two alternative 5V power supplies for my WaveIO. Both might be considered to be strange and advisable!
1. Recommended to me by a commercial DAC manufacturer - a Chinese open-frame SMPS made by Cincon -
http://www.cincon.com.tw/data/products/cfm1_2/CFM10_15.pdf
Yes - Switched Mode Power Supply!! I'm not proud. Actually, it works really well and is noticeably dynamic, which the manufacturer said was probably caused by its very fast switching frequency. Being so high frequency, he also thinks that most switching noise is not audible being well out of the audible range.
2. LiPo Battery
Well actually an iPhone charging unit - EasyAcc 12,000 mAh LiPo
EasyAcc 12000mAh 4 x USB Portable External Battery Pack: Amazon.co.uk: Electronics
It has 4 USB Type A socket outlets output at 5V. I made up a lead using a Type A USB plug (using Cat 5 Ethernet cable), the other end with bare wires screwed into the WaveIO terminals. It has a switch on it to turn it on and off (not all do). 12,000 mAh should last about 24 hours with a 500 mAh draw. Charges with any Micro USB charger or PC.
SQ excellent. Clean and lovely warmth. Slightly more relaxed than the SMPS but not so dynamic.
Both options were first class and required little DIY skills and were cheap.
Lastly, as always Lorien is right! Using any reasonably good external power supply to the WaveIO is significantly better for SQ than running it off the 5V USB cable lines from the attached PC.
Jonathan
you can bet it includes a switching regulator/limiter (the 12000mAH 'battery pack'), not an excellent option to power a DAC IMO. switchers can be good, if designed with noise in mind they are now very viable for even hi-END IMO, but its extremely unlikely this battery charger was designed with any of that in mind
Last edited:
no, what I mean is, the battery pack you linked already includes a switching regulator/DC buck-boost upconvertor internally. it is not just a battery and there is little/no chance its designed with keeping the switching frequency and its harmonics out of the audio band.
there is every chance it is noisier than even the cheapest AC powered linear regulator. I see people use this type of thing in audio thinking 'its a battery, it must be good' without relizing this type of thing by law must be both current limited and in this case also regulated for battery charging and that regulation will be switchmode
in short, it contains exactly the same type of noise you are trying to avoid by not using USB
Last edited:
thanks for the details
is there any chance to see some example battery psu that can be use for the Wave IO similar to Red Wine Battery PSU
Well yes, are "genuine" WaveIOs but without SPDIF outputs and related logic as Doede doesn't need it in his DAC (and frankly I understandI had the idea first to use two COD modules from Twisted Pear, with PCM1794A in dual mono mode, but suddenly I have seen this kit: http://www.dddac.com/dddac1794.html
I didn't know that Lorien was building WaveIO kits for Doede Douma !
The DAC kit with dual mono PCM 1794A in non-oversampling mode and passive output seems really interesting !
). As for measurements, from what I can recall he took his hobby quite serious as his latest addition is AN system two gear... per his statement found here.Cheers,
L
