Can we push the concept further?
instead of going out in line level
can we drive a mosfet with the DSD data to drive the speaker.
this would be a full digital amp
IIRC DSD data are not exactly the same as the PWM signal at the output of a class D input stage comparator.
What did you think to do?
Can we push the concept further?
instead of going out in line level
can we drive a mosfet with the DSD data to drive the speaker.
this would be a full digital amp
In theory you can, but it will be extremely difficult to keep the reference (supply) clean enough.
But if you invert the bitsteram and have a balanced output you would surpress the power supply ripple... 🙂
Doable i guess..
Doable i guess..
That would certainly help, but only partly.
Even with a balanced output, the output signal is proportional to the supply, so any variation in the supply amplitude-modulates the output signal and causes sidebands. Low-frequency variations can be suppressed very well with some sort of regulated supply (series regulator or switchmode regulator).
Higher frequency disturbances can be suppressed with good decoupling, but the allowable level is really low. The problem is that a sigma-delta modulate has lots of out-of-band quantization noise and you don't want that to be converted back into the audio band by multiplying it with fast variations on the supply. At the same time the switching output stage draws a current out of the supply with lots of high-frequency switching transients. All in all, you need a very low supply impedance, especially at high frequencies.
Even with a balanced output, the output signal is proportional to the supply, so any variation in the supply amplitude-modulates the output signal and causes sidebands. Low-frequency variations can be suppressed very well with some sort of regulated supply (series regulator or switchmode regulator).
Higher frequency disturbances can be suppressed with good decoupling, but the allowable level is really low. The problem is that a sigma-delta modulate has lots of out-of-band quantization noise and you don't want that to be converted back into the audio band by multiplying it with fast variations on the supply. At the same time the switching output stage draws a current out of the supply with lots of high-frequency switching transients. All in all, you need a very low supply impedance, especially at high frequencies.
That would certainly help, but only partly.
Even with a balanced output, the output signal is proportional to the supply, so any variation in the supply amplitude-modulates the output signal and causes sidebands. Low-frequency variations can be suppressed very well with some sort of regulated supply (series regulator or switchmode regulator).
Higher frequency disturbances can be suppressed with good decoupling, but the allowable level is really low. The problem is that a sigma-delta modulate has lots of out-of-band quantization noise and you don't want that to be converted back into the audio band by multiplying it with fast variations on the supply. At the same time the switching output stage draws a current out of the supply with lots of high-frequency switching transients. All in all, you need a very low supply impedance, especially at high frequencies.
What about the DSD frequency?
Is it low enough?
An high frequency (video) buffer?
The switching frequency of DSD is indeed a bit high when you have to drive large power MOSFETs, if that's what you mean. I don't understand the remark about a video buffer.
Can we push the concept further?
instead of going out in line level
can we drive a mosfet with the DSD data to drive the speaker.
this would be a full digital amp
or how about feed the DSD stream into a class D amp. 😀
still trying to wrap my head on how the doubly modulated signal will become. but the audio frequency should remain.
The switching frequency of DSD is indeed a bit high when you have to drive large power MOSFETs, if that's what you mean..
how about a small MOSFET? if you want to use headphones or very sensitive speakers you only need a few watts for general listening. some people use 3watt tube amps with saba greencones while the volume knob is at 1/4
or how about a darlington pair?
There are small switched mode regulators with switching frequencies in the MHz range, so it should be feasible. You still have the issue that you need to keep reference (= supply) very clean, though.
There are small switched mode regulators with switching frequencies in the MHz range, so it should be feasible. You still have the issue that you need to keep reference (= supply) very clean, though.
can you post a link of these regulators?
supply could be batteries?
hi, and the reference voltage can be kept very very clean by a capacitance multiplier! 🙂
would a supercap come in handy here?
can you post a link of these regulators?
supply could be batteries?
I'm not sure if this helps, but see for example:
LTC3646/LTC3646-1 - 40V, 1A Synchronous Step-Down Converter - Linear Technology
LTC3861 - Dual, Multiphase Step-Down Voltage Mode DC/DC Controller with Accurate Current Sharing - Linear Technology
LTC3839 - Fast, Accurate, 2-Phase, Single-Output Step-Down DC/DC Controller with Differential Output Sensing - Linear Technology
LTC3838 - Dual, Fast, Accurate Step- Down DC/DC Controller with Differential Output Sensing - Linear Technology
If I were to try it, I would use a very well regulated supply, with the feedback sense point of the regulator right at the power stage, using a board with power and ground planes and with large ceramic decoupling caps right next to the power stage. No idea how good or how bad the performance would become, though. I only know that keeping the supply clean despite of the large switching currents drawn from it is quite critical.
I'm not sure if this helps, but see for example:
LTC3646/LTC3646-1 - 40V, 1A Synchronous Step-Down Converter - Linear Technology
LTC3861 - Dual, Multiphase Step-Down Voltage Mode DC/DC Controller with Accurate Current Sharing - Linear Technology
LTC3839 - Fast, Accurate, 2-Phase, Single-Output Step-Down DC/DC Controller with Differential Output Sensing - Linear Technology
LTC3838 - Dual, Fast, Accurate Step- Down DC/DC Controller with Differential Output Sensing - Linear Technology
If I were to try it, I would use a very well regulated supply, with the feedback sense point of the regulator right at the power stage, using a board with power and ground planes and with large ceramic decoupling caps right next to the power stage. No idea how good or how bad the performance would become, though. I only know that keeping the supply clean despite of the large switching currents drawn from it is quite critical.
thank you for the links and insights. if i read the data correctly it would easily make a few watts amp and it could be even powered by a 12v agm car battery
I just noticed this thread:
DSD Playback system, DSF Player + USB DDC + DSD Amplifier
It looks like they are doing what you are considering to do.
DSD Playback system, DSF Player + USB DDC + DSD Amplifier
It looks like they are doing what you are considering to do.
I just noticed this thread:
DSD Playback system, DSF Player + USB DDC + DSD Amplifier
It looks like they are doing what you are considering to do.
interesting! thanks for the hint. it looks like they want to use two mosfets for performance (high power)
i wonder if those LTC regulators posted above are capable of audio quality. it would be simpler to use those and good enough for sensitive drivers
Ive have really enjoyed this thread especially the positive approach of the main contributors - congratulations!!
I have just made it thru the 185 pages - WHAT A MARATHON
I left the thread after half a dozen or so pages to give it a go.
Just grabbed a basic DIYINHK XMOS usb board and connected it to an el cheapo Di transformer set I had made to couple phones and tablets to my HiFi system (following the advice from my son- low impedance drives to transformers really make them sing)
Any how I was blown away by the audio sound quality achieved by driving DSD directly into a transformer
No evidence of core saturating with 1.65V DC across 10kohm primary
For those interested it was a $10.00 transformer purchased from Altronics in Oz
Cya Peter
I have just made it thru the 185 pages - WHAT A MARATHON
I left the thread after half a dozen or so pages to give it a go.
Just grabbed a basic DIYINHK XMOS usb board and connected it to an el cheapo Di transformer set I had made to couple phones and tablets to my HiFi system (following the advice from my son- low impedance drives to transformers really make them sing)
Any how I was blown away by the audio sound quality achieved by driving DSD directly into a transformer
No evidence of core saturating with 1.65V DC across 10kohm primary
For those interested it was a $10.00 transformer purchased from Altronics in Oz
Cya Peter
Ive have really enjoyed this thread especially the positive approach of the main contributors - congratulations!!
I have just made it thru the 185 pages - WHAT A MARATHON
I left the thread after half a dozen or so pages to give it a go.
Just grabbed a basic DIYINHK XMOS usb board and connected it to an el cheapo Di transformer set I had made to couple phones and tablets to my HiFi system (following the advice from my son- low impedance drives to transformers really make them sing)
Any how I was blown away by the audio sound quality achieved by driving DSD directly into a transformer
No evidence of core saturating with 1.65V DC across 10kohm primary
For those interested it was a $10.00 transformer purchased from Altronics in Oz
Cya Peter
Hi Peter,
Glad to hear another positive feedback.
Wich one from Altronics bellow:
http://www.altronics.com.au/transformers/audio-transformers/&gws_rd=cr&dcr=0&ei=rnEUWurKLoSAafyVsbAH]Transformers | Audio - Altronics[/url]
I need to find the original trafo used here (wich is very hard to find) or a cheaper obtainable one.
Thanks in advance.
Ps: tank you all for you advices, i buyed instead of an intel nuc, an micro itx mobo full packed for multimedia + an amd 4890k+ 16gb ram at 2400mhz for this project. Still need to buy next year a pico itx adapter + power brick.
- Home
- Source & Line
- Digital Line Level
- The Best DAC is no DAC