HI All,
I got a question.
There are many DAC is balance output + and -.
In many application they need to use opamp to mix up to single end out to have better S/N.
I question about If I only take + signal only and all is single end design.
My idea is SE is better than PP.
BTW, there are a DC on DAC output.
ANy reference design can share??
I got a question.
There are many DAC is balance output + and -.
In many application they need to use opamp to mix up to single end out to have better S/N.
I question about If I only take + signal only and all is single end design.
My idea is SE is better than PP.
BTW, there are a DC on DAC output.
ANy reference design can share??
I was interested in what answer you may receive. As there are nobody to help, I will try. If you are looking for new experience, you may try to load (-) output with resistor. (+) output with i/u converter with same input resistance. As I understand, you need to know how output stage looks like for answer yours question. Without schematic you can only make intuitive suggestions. Beware, output performance of DAC will be degraded if passive i/u stage used. Just have fun :-D
differential output reduces noise from common gnd, supply coupling of the digital circuits in the DAC
it can also give lower distortion, and improves S/N
I don't understand your SE vs PP concern - that juxtaposition would usually refer to circuit bias for Class A operation
the conversion of a differential to single ended signal has nothing to do with push-pull operation or "single ended" bias
you can use all Class A circuitry, single ended bias too if you want in the I/V and differential to single ended Signal conversion
even with op amp I/V - the 6.2 mA center bias of the PCM1794 causes op amp I/V to operate Class A
you can add a ccs load to the summing op amp output too if you think it helps
it can also give lower distortion, and improves S/N
I don't understand your SE vs PP concern - that juxtaposition would usually refer to circuit bias for Class A operation
the conversion of a differential to single ended signal has nothing to do with push-pull operation or "single ended" bias
you can use all Class A circuitry, single ended bias too if you want in the I/V and differential to single ended Signal conversion
even with op amp I/V - the 6.2 mA center bias of the PCM1794 causes op amp I/V to operate Class A
you can add a ccs load to the summing op amp output too if you think it helps
Last edited:
Well...You could do a passive I/V conversion of both the + and - halves using a resistor to ground from each phase. Then just take a signal from the + side and ground through a buffer of your choice. BUT after looking at the ground contamination around the DAC chips from all the digital noise...I sure wouldn't think it would sound very good!
You would have to have a VERY carefully designed PCB to have the output be quiet enough for single ended operation.
Zc
P.s. (by PP I think he means differential)
You would have to have a VERY carefully designed PCB to have the output be quiet enough for single ended operation.
Zc
P.s. (by PP I think he means differential)
I suspect you have spent too much time reading Audiophile nonsense and don't actually understand the circuit theory well enough to make that judgment
I wouldn't use the exact circuit in the data sheet but that is how the performance of the DAC is measured, to get the S/N and distortion numbers - if you understand the numbers you should see there isn't much the DAC is supposed to be doing that is "missing"
just using one side will have more noise, distortion than combining both in a decent active circuit
if you insist on a "audiophoolish" method then you could use a transformer to convert the differential signal to gnd referenced - but any transformer will add low frequency distortion that is easily measurable
I wouldn't use the exact circuit in the data sheet but that is how the performance of the DAC is measured, to get the S/N and distortion numbers - if you understand the numbers you should see there isn't much the DAC is supposed to be doing that is "missing"
just using one side will have more noise, distortion than combining both in a decent active circuit
if you insist on a "audiophoolish" method then you could use a transformer to convert the differential signal to gnd referenced - but any transformer will add low frequency distortion that is easily measurable
Hi,
The PCM179x series of BB features such low distortions and noise, that its no trivial task to build a following stage that preserves that technical quality.
It may be the topic of endless discussions if a couple of dB plus or minus around -100dB is any more than just academic porn.
It for sure doesn´t say anything about sonic merits of a circuit.
Of course is a SE stage more prone to inferior layout design, ground issues etc. but then this doesn´t tell a thing about sonics either.
In any case it´s no issue to use just one of the PCM179x outputs in SE style.
You need to make sure that the second output sees an equal and low impedance.
You could either use two identical circuits like the ones in the attachment and take just one of two outputs, or connect a smallish resistor of 2-20Ohms from the second output to gnd.
Since the output of the attached circuits is free from dc it may be directly coupled to a buffer stage or a following device.
The left circuit requires well regulated clean power supplies, especially the positive supply.
The right circuit lessens the demand on the positive power supply at the expense of slightly more touchy setup, a bit more noise and slightly increased THD.
The left circuit is capable of THD better than -90dB (at fullscale) and better than -110dB in noise. But best is...it blows away OPamp-IVs within seconds.
See also the thread "ZEN -> CEN -> SEN, evolution of a minimalistic IV converter"
jauu
Calvin
The PCM179x series of BB features such low distortions and noise, that its no trivial task to build a following stage that preserves that technical quality.
It may be the topic of endless discussions if a couple of dB plus or minus around -100dB is any more than just academic porn.
It for sure doesn´t say anything about sonic merits of a circuit.
Of course is a SE stage more prone to inferior layout design, ground issues etc. but then this doesn´t tell a thing about sonics either.
In any case it´s no issue to use just one of the PCM179x outputs in SE style.
You need to make sure that the second output sees an equal and low impedance.
You could either use two identical circuits like the ones in the attachment and take just one of two outputs, or connect a smallish resistor of 2-20Ohms from the second output to gnd.
Since the output of the attached circuits is free from dc it may be directly coupled to a buffer stage or a following device.
The left circuit requires well regulated clean power supplies, especially the positive supply.
The right circuit lessens the demand on the positive power supply at the expense of slightly more touchy setup, a bit more noise and slightly increased THD.
The left circuit is capable of THD better than -90dB (at fullscale) and better than -110dB in noise. But best is...it blows away OPamp-IVs within seconds.
See also the thread "ZEN -> CEN -> SEN, evolution of a minimalistic IV converter"
jauu
Calvin
Attachments
That is just wishfull thinking at best. SE from 179x it just blows, that is it... Increased distortions, reduced dynamics, increased noise - all due to SE, resistor aproach.
Also those circuits above are very picky about adjustments, picky about components.
Last edited:
Differential output DAC chips aren't designed to be used that way, similarly, a crescent wrench isn't designed to be a hammer, but it can pound in a nail, it's just not the best tool for that job.
At the very best, your SNR will be 3 dB higher than a proper differential connection, and in practice with all the common mode noise, probably an order of magnitude higher than that. For single rail DACs (Crystal CS4398 for example), the zero signal output of each output is sitting at half the rail voltage. So you will need to block that DC as well, or factor that into your analog stage design.
Having just built a new DAC myself, I tried what you are proposing, and I found that transformer coupling to a single ended output was significantly better than just using one leg of the DAC chip. But that was just with the CS4398 and a two layer PCB, YMMV.
Sheldon
At the very best, your SNR will be 3 dB higher than a proper differential connection, and in practice with all the common mode noise, probably an order of magnitude higher than that. For single rail DACs (Crystal CS4398 for example), the zero signal output of each output is sitting at half the rail voltage. So you will need to block that DC as well, or factor that into your analog stage design.
Having just built a new DAC myself, I tried what you are proposing, and I found that transformer coupling to a single ended output was significantly better than just using one leg of the DAC chip. But that was just with the CS4398 and a two layer PCB, YMMV.
Sheldon
Hi,
SoNic You must be a golden ear and a wooden ear at the same ;-)
A golden ear if You´d be able to differentiate between -90dB THD at fullscale signal against a 10-20dB better Value.
A wooden ear if You can´t hear the sonic deficiencies of OPs (no pun intended)
Seriously, the THD-differences are just good as food for academic, off of praxis discussions.
Adding a Transformer is for sure a smart way to derive a galvanic ally isolated SE output signal.
But it is costly and adds components that may not be required in first place.
I also really doubt, that the transformer preserves the low THD-figures of the active stages.
jauu
Calvin
SoNic You must be a golden ear and a wooden ear at the same ;-)
A golden ear if You´d be able to differentiate between -90dB THD at fullscale signal against a 10-20dB better Value.
A wooden ear if You can´t hear the sonic deficiencies of OPs (no pun intended)
Seriously, the THD-differences are just good as food for academic, off of praxis discussions.
Adding a Transformer is for sure a smart way to derive a galvanic ally isolated SE output signal.
But it is costly and adds components that may not be required in first place.
I also really doubt, that the transformer preserves the low THD-figures of the active stages.
jauu
Calvin
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.