Thanks for the advice.
1) The input voltage can be freely adjusted through the FB resistor of LT3471.
4) Yes. The distance between the oscillator and the DAC MCLK pin is only 16mm, so overkill will be correct.
1) The input voltage can be freely adjusted through the FB resistor of LT3471.
4) Yes. The distance between the oscillator and the DAC MCLK pin is only 16mm, so overkill will be correct.
Maybe 🙂
But this "overkill" is no so compatible with the use of a DC-DC switching mode power supply.
Alex.
@altor
sorry. I didn't quite understand.
Does this mean that the noise performance of Switcher+LC+REG will not be good enough?
Or the possibility of AC magnetic coupling by pcb trace?
Actually, I am a third-year college student. I only know the little contents of the textbook,
but I don't know the more detailed or practical technic...
sorry. I didn't quite understand.
Does this mean that the noise performance of Switcher+LC+REG will not be good enough?
Or the possibility of AC magnetic coupling by pcb trace?
Actually, I am a third-year college student. I only know the little contents of the textbook,
but I don't know the more detailed or practical technic...
Last edited:
Maybe yes but I still would put two resistors. Just in case, for debug purposes.
One really have to look at SMPS frequency and PSRR of the LDO. There is another thread about that created couple of days ago.
Have you ever used modern DC-DC together with LDO to reduce ripples? Do you think it is worth it?
For some reasons higher frequency is better, but LDO's PSRR at these frequencies is zero.
Used and worth.
But in "no compromise", "overkill style" I prefer not to use SMPS if possible.
(even though designing SMPS was one of my main jobs).
And yes, I've used this in some designs.
Alex.
Last edited:
Something like TPS65131 operates at 1.3-1.5MHz and considering higher harmonics of that - an LDO looks completely useless. I was trying to find some experimental data but no luck yet.
A similar example is a silent switcher. Since the circuit diagram of the silent switcher has not been made public, it is not known if it is completely the same. I think it will be almost similar to my design. (LDO products, resistance values, and capacitor values are applied temporarily.)
And Analog Devices also recommends this design for high-performance DAC applications. (Refer to ADP5070 document)
That's why I adding an C-L-C filter.
Last edited:
what is the part number?
Please pay attention that PSRR is dependent on input/output voltage and the headroom for a given LDO.
Also it is not so easy to filter out switching power supply spikes (due to higher harmonics content) with active device like LDO, CLC might be significantly more effective.
Just to clarify: I'm not trying to say that LDO is completely useless for that purpose. But I am not sure because haven't seen experimental data 🙂
TPS7A47, LT3042.
Those variables are almost entirely specified in the datasheet.
you're right. Since it is almost a sawtooth wave, there are many harmonic components.
Regarding the use of a clock buffer, unless AK4497 includes an internal ASRC, the local dac MCLK signal is probably going to need to be exported to whatever device drives the dac. In that case a buffer probably makes a lot of sense.
Regarding SMPS, they can radiate EMI/RFI. Also, the versions of silent switcher I tried caused noise problems due to using a wall wart power source. EMI/RFI leakage currents from the wall wart went down the USB cable and allowed CM noise to pass right through silent switcher ground. It got into a dac and caused audible problems. It only worked okay if I drove SS input from a linear 5v power supply. I even tried a floating USB power pack but it has an SMPS in it to regulate the 5v. That was also enough to produce noise that propagated down the USB power cable like it was an antenna. So that didn't work either for my purposes.
Regarding SMPS, they can radiate EMI/RFI. Also, the versions of silent switcher I tried caused noise problems due to using a wall wart power source. EMI/RFI leakage currents from the wall wart went down the USB cable and allowed CM noise to pass right through silent switcher ground. It got into a dac and caused audible problems. It only worked okay if I drove SS input from a linear 5v power supply. I even tried a floating USB power pack but it has an SMPS in it to regulate the 5v. That was also enough to produce noise that propagated down the USB power cable like it was an antenna. So that didn't work either for my purposes.
I needed +-15v and +5v. So silent switcher seemed like it could be convenient. In the end I just used linear supplies.
My feeling is that for development of, say, the best dac you can figure out how to make at this point in time, maybe its good to play it safe, use some overkill, and see how good you can do. Then if you want to find out how much you can simplify the design to reduce costs, at least you have an overkill dac to compare with. That way you know when certain design changes of the final low-cost design are causing problems, and or know when you found a better way than than in the overkill prototype.
The reason I suggest to have a prototype dac to compare with is because I believe not all things that are audible to humans are easy to measure using standard measurement instruments. Sound stage could be one example. So sometimes it may be necessary to use humans to evaluate things that require something more like is done in human factors engineering.
My feeling is that for development of, say, the best dac you can figure out how to make at this point in time, maybe its good to play it safe, use some overkill, and see how good you can do. Then if you want to find out how much you can simplify the design to reduce costs, at least you have an overkill dac to compare with. That way you know when certain design changes of the final low-cost design are causing problems, and or know when you found a better way than than in the overkill prototype.
The reason I suggest to have a prototype dac to compare with is because I believe not all things that are audible to humans are easy to measure using standard measurement instruments. Sound stage could be one example. So sometimes it may be necessary to use humans to evaluate things that require something more like is done in human factors engineering.
Last edited:
Yes, original SilentSwitcher is somewhat similar. But the schematic itself is only a start. Layout and component selection is what matters for good performance.
4191 includes, but 4497?!
Sure, but for THIS device, not for DC (if oscillator is close to DAC.
About SMPS:-there is a good solution, but not always possible -to synchronize SMPS with I2S (for example - to use divided MCLK or BCLK).
Alex.