May I ask, are you looking to simply enable a balanced, box-to-box, signal interconnection, and if not that, what you hoping to achieve via a differential DAC configuration?
A general answer why people would design / manufacturer / sell / buy balanced (fully symmetrical differential) gear :
1) lower distortion by even harmonic cancellation
2) signal immunity to noise pick-up
3) higher output amplitude
4) reduction of random noise
Whether each individual wants any of the above is a personal choice.
There are hi-end gear that are fully balanced, there are those that are single ended.
Here an example of a Topping D10 SE vs Bal version :
https://www.audiosciencereview.com/...topping-d10s-usb-dac-and-bridge-review.14859/
https://www.audiosciencereview.com/...ds/topping-d10-balanced-review-usb-dac.25094/
Cheers,
Patrick
I concur with your #1 on your list, and with #2 under certain conditions. I view #3 as not easily lending itself to practical use. Benefit #4 is an interesting case, because for a differential configuration, the DAC outputs are functionally in series. As a result, random noise is actually increased by +3dB. However, the signal is increased by +6dB. The net result is that SNR is improved by (+6dB signal - 3dB noise) = +3dB SNR. If the two DAC output were instead connected in functional parallel, that's when random noise would decrease by -3dB, while the relative signal level remains the same. Giving, (0dB signal + 3dB noise) = +3dB SNR
Weighing against the potential benefits, however, is the not insignificant greater cost of a second pair of PCM63 DAC chips, plus an increased implementation complexity. Just my 2 cents.
A pair of whatever DAC ICs is a fraction of the total cost of a complete build.
Any modern DAC, e.g. AK4493 or ES90xx, have inherent differential outputs.
People just throw it away with a differential opamp Sallen Key.
I have absolutely no desire to convince anyone one way or another.
I merely pointed out some technical facts.
And I always build balanced for myself, with very few exceptions (for headphones).
Cheers,
Patrick
Any modern DAC, e.g. AK4493 or ES90xx, have inherent differential outputs.
People just throw it away with a differential opamp Sallen Key.
I have absolutely no desire to convince anyone one way or another.
I merely pointed out some technical facts.
And I always build balanced for myself, with very few exceptions (for headphones).
Cheers,
Patrick
Sorry, Patrick, I wasn’t intending to be confrontational. I was only attempting to point out some additional technical facts.
I agree, that DAC cost is normally a small fraction of the BOM cost. The cost problem comes with these vintage DACs, such as the PCM63. I think I recall seeing some (I don’t recall the grade) offered on eBay for around $100 each, and you would need to buy an additional pair of those for differential, so four in total!
Those modern DAC chips featuring differential output largely began doing so in order for the chip to dispense with requiring a negative supply. In other words, it was primarily for power supply implementation cost savings.
Best regards,
Ken
Last edited:
For what it is worth, here our own distortion measurement of a balanced PCM56-K DAC.
Typical datasheet spec is -94dB THD.
We measured, in balanced mode without trimming, H2 at -122dB and H3 -112dB.
https://www.diyaudio.com/community/...st-tht-i2s-input-nos-r-2r.354078/post-6925148
Patrick
Typical datasheet spec is -94dB THD.
We measured, in balanced mode without trimming, H2 at -122dB and H3 -112dB.
https://www.diyaudio.com/community/...st-tht-i2s-input-nos-r-2r.354078/post-6925148
Patrick
Interesting. I’ve been intending to someday build a PCM56 based DAC, largely because of the persistent and compelling advocacy of member, bernhard. Is there a schematic of your differential PCM56 DAC in that thread which shows how the differential currents are subsequently summed? For example, by dual op-amp I/Vs, which then feed a single differential amplifier? Perhaps, something else?
Also, how the inverted output current half of each PCM56 differential DAC pair is produced. For example, by simple digital inversion of the sample data bits feed to the signal inverting DACs? Thanks.
Last edited:
The measurement was made with the internal opamp of the PCM56, so that we do not end up with measuring the distortion of the IV converter itself.
The negative phase has its data inverted using a 74xx86. Nothing special.
See Pass Labs D1 DAC schematics.
The measurement was made using the differential input of the distortion analyser directly.
SDATA are 8x over-sampled.
Patrick
The negative phase has its data inverted using a 74xx86. Nothing special.
See Pass Labs D1 DAC schematics.
The measurement was made using the differential input of the distortion analyser directly.
SDATA are 8x over-sampled.
Patrick
The cost of the "used" PCM63 went crazy high
... if you want the PCM63 as balanced, just stack 2 of my stereo PCM63 DACs, and invert DATA for the second DAC (inverter can be CD74HCT04 decoupled with ceramic capacitor and powered directly from DACs pcb) ... tricky part will be the I/V, you can use a differential opamp or a custom discrete build (ideally based on some functional diagram from some commercial DAC service manual).
I can build something in the future, but probably only with the AD1865N (the main reason is the simplicity (it is dual, +-5V PSU) and price availability of the AD1865, which will significantly reduce costs, moreover AD1865 is great DAC)
Thanks for the suggestion and excellent link, took me a while to digest it but in the end it was worth reading. It is not a matter of bitrate rather the clock. I am too lazy to design a new circuit for that. I guess Amanero/XMOS would still be ok.Dont bother with raspberry i2s it is verry bad
On a side note a gentleman by the name of "diyaudiophiler" has made PiHAT DACs around AD1865 and AD1868. Although he has not shared schematics but I could guess that he migh have designed a clock signal for the DAC.
Congrats on your new DAC rehanabid! I would suggest to remove the MSB trimming components if your not going to use a proper trimming procedure.
Also, there is no fruit in my audio system, no RPi used.
Sorry my bad, I thought you used RPi as audio source for your DAC. You are right, after listening to the DAC for a while I agree that there is no need for the trim pot circuit, will remove it soon.
New, K-grade, 40 USD each, THD+N -100dB typical :
https://www.diyaudio.com/community/threads/selling-pcm63-k-pcm1704u-k-sm5842-cs8412-etc.237988/
I will never use ICs that are "used" or "removed from equipment".
Disclaimer : I do not know the seller personally, and have no connections with him whatsoever.
Patrick
Actually your measurements shows -97dB THD. So only 3dB improvement over datasheet with balanced.
