Been tracing out this new es9039q2m DAC to give it cleaner supply to the relevant parts. Does anybody know how the voltage ist regulated in the following configuration?
I can't find where the pins 3 and 5 of the OPA1612 are going to. They certainly control the output voltage. Haven't seen this regulation scheme before.
I can't find where the pins 3 and 5 of the OPA1612 are going to. They certainly control the output voltage. Haven't seen this regulation scheme before.
Both OP are followers, output voltage is the same as the reference voltage at pin 3 (and pin 5).
Reference voltage can be well filtered voltage from additional or existing regulator.
P.S. Bur OPA1612 is not a good choice for this, much better are AD8397, OPA1656, etc.
Reference voltage can be well filtered voltage from additional or existing regulator.
P.S. Bur OPA1612 is not a good choice for this, much better are AD8397, OPA1656, etc.
Thanks Altor
Can this really be considered as two separate supplies?
I think I will use two LT3045 for avcc.
Can this really be considered as two separate supplies?
I think I will use two LT3045 for avcc.
The schematic has errors. With 12V input, and unity gain, the output can only be 3.3V if the +input of the opamps is 3.3V.
So that has to be present at pin 3 and pin 5.
If that 3.3V reference is low noise and stable, it will be an excellent voltage regulator.
You can call it a voltage regulator but you can also see it as a unity gain DC amplifer with a 3.3V input.
Jan
So that has to be present at pin 3 and pin 5.
If that 3.3V reference is low noise and stable, it will be an excellent voltage regulator.
You can call it a voltage regulator but you can also see it as a unity gain DC amplifer with a 3.3V input.
Jan
Yes pin 3+5 have a common 3.3v source.
The reason for this approach escapes me.
You need a very good regulated 3.3v to make this a very good 3.3v regulator. 🤔
The only reason I can think of would be perfect symmetry for both avcc supplies. It is indeed perfectly symmetrical to three digits behind the dot.
What about the+12v supply, does it also have to be high quality?
The reason for this approach escapes me.
You need a very good regulated 3.3v to make this a very good 3.3v regulator. 🤔
The only reason I can think of would be perfect symmetry for both avcc supplies. It is indeed perfectly symmetrical to three digits behind the dot.
What about the+12v supply, does it also have to be high quality?
You always need a good voltage reference when you want to make a good voltage regulator. A voltage regulator is essentially a voltage reference with a buffer or amplifier, with or without a low-pass filter between the reference and the amplifier.
Why not? The OPA1612 has far less voltage noise than your alternatives.
The opamp has a very high power supply rejection ratio, meaning that any effects from the power supply on the output is reduced by 100,000 or so. The output depends almost exclusively on the quality of the 3.3V reference.
As Marcel mentioned, an integreated voltage regulator is a reference and 'opamp' in one package, exactly the same concept.
The reason for doing it the way they did could be both quality and cost. Avcc symmetry almost certainly no factor.
It is very unlikely that sub-optimzing at this location has any useful impact on the overall final result.
Jan
Yes, sure.
I often use this for 2 channels PS/
And also place a big electrolytic capacitors at the OP's outputs (100-470uF).
Yes, my alternatives have a slightly higher noise level (which is usually enough), but also a higher output current.
Alex.
ESS datasheets and evaluation boards have promoted similar scheme for many years (see e.g. ES9028Q2M datasheet). The buffer opamp is preceded by a cheap regulator + LPF. This results in very low THD+N within audio frequencies so good for the "numbers game", but 1/f noise is poor. No wonder many cheap DACs use it.
Ok guys these are quite some contradicting opinions.
I know that avcc of ES Dacs need very clean power, so I will use my all time favorite LDO the LT3045 for that purpose.
Thanks anyway
Klaus
I know that avcc of ES Dacs need very clean power, so I will use my all time favorite LDO the LT3045 for that purpose.
Thanks anyway
Klaus
Did you ever do a SQ evaluation of this? Is this something to chase? Whats your opinion?
//
I haven't done any SQ evaluation of this. However with poor 1/f noise use of attosecond italian clocks seems futile.
Electric circuit 1/f noise is uncorrelated but the clock noise would be correlated - right? So perhaps its not a real SQ problem when coming from circuits? This is nt to say that clock noise if a problem - maybe it is - it seems not really investigated properly - my test/experiance show perhaps not..
//
//
But the impact of jitter is signal correlated distortion - no?
Noise in circuits, like from a transistor, is just riding along with the signal - no modulation, or?
//
Noise in circuits, like from a transistor, is just riding along with the signal - no modulation, or?
//
It's not additive noise. Low-frequency noise on a DAC or ADC voltage reference amplitude-modulates the signal, causing sidebands around the spectral peaks of the signal. Conventional wisdom has it that such sidebands are well-masked, but I doubt that everyone would agree to that. White noise on the reference could cause a noise floor that varies with the signal or just add constant noise, depending on the details of the DAC.
I would focus on the reference and LPF in front of the OPA1612, I wonder whether your proposed solution could actually equal or better the performance of the current AVCC.