Your setup Looks awesome!
You can get rid of LT3042 for DVDD with same ultracapacitor rail. I didn't feel any difference. The ultarcapacitor works like a black hole
Ian
Sorry for the confusing.
For the 1.2V rail, you can use both internal reg or external 1.3V reg. But ultraCapacitor can use for all other 3.3V rails.
Ian
Another one 
I always favoured voltage mode from akm over one from sabre.
All is powered by ultracap@5V.
The are some low noise 3.3V LDOs. Separate for clock and digital part using ultracaps as pre reg.
AVDD for ak4490 is powered directely from ultracap @ 5V.
Fifo reclock -> ak4490 -> LL1684 (super v/i converter)
ak4490_ultracapacitor.jpg - Google Drive
This is just amazing!
Those ultracaps really rock! They simply bring ANY dac chip into new level! Say RIP to regs!
I always favoured voltage mode from akm over one from sabre.All is powered by ultracap@5V.
The are some low noise 3.3V LDOs. Separate for clock and digital part using ultracaps as pre reg.
AVDD for ak4490 is powered directely from ultracap @ 5V.
Fifo reclock -> ak4490 -> LL1684 (super v/i converter)
ak4490_ultracapacitor.jpg - Google Drive
This is just amazing!
Those ultracaps really rock! They simply bring ANY dac chip into new level! Say RIP to regs!
Another one
All is powered by ultracap@5V.
The are some low noise 3.3V LDOs. Separate for clock and digital part using ultracaps as pre reg.
AVDD for ak4490 is powered directely from ultracap @ 5V.
Fifo reclock -> ak4490 -> LL1684 (super v/i converter)
ak4490_ultracapacitor.jpg - Google Drive
This is just amazing!
Those ultracaps really rock! They simply bring ANY dac chip into new level! Say RIP to regs!
Interesting
Agree 're the regs their a bottleneck
*Flamesuit on* Lol
Please don’t hate on me for this post.
The whole current vs voltage mode discussion IMHO seems full of contradictions. Maybe I’m wrong here but here’s what I’ve seen over the years...
As soon as people bring out the measuring equipment, current out mode gets put in the spotlight and “proclaimed as the best”. But rarely (if ever) have I seen people claiming sonic superiority.
If anything I’ve seen people saying they like the voltage mode sound better, even if it doesn’t measure as well. A very rough extremely simplified summation seems to indicate that a lot of people describes the sound as more meaty, natural and musical.
I seem to remember there was a pretty serious study made on sound levels and “masking”. Supposedly the human hearing looses the ability to pick up sounds 75db lower than the peak level sound.
So, how important is a number saying -126db vs -136db?
Same thing goes for frequency range and the number of bits, at what point will diminishing returns make “better numbers” irrelevant?
A lot of people love their tube amps, but I don’t think I’ve ever seen anyone claiming them to measure great.
I guess my point is this, IMHO we need to be careful and not let measurements become more important than actual listening. Sometimes at least myself prefer a particular sound even if it isn’t the most pure.
When you go to a concert and listen to music live, do you sit there and say “the music shouldn’t sound this way”?
Even though concert speakers measure much much worse than hifi speakers, they are still the reference since this is what it sounds like to see the artist irl.
So, what is our goal? And how do we get there?
I can only speak for myself but for me the “air guitar factor” is one of the most important properties when I listen to music, how do you measure that?
I’m sorry if this is OT, I just think it’s always important to take a step back and look at where you are, where you want to go and where you are currently heading.
Hi markusA,
I totally agree with you.
OP amp I/V stage with deep feedback has the best testing numbers, but the sonic performance is not as good as I/V stage with no or low feedback. What you are saying does make sense.
Actually I don't think a transformer I/V works in voltage mode. It relies on the changing of input current to convert the audio energy. The only thing is that it's not an ideal load for a current source because the impedance is not 0 or very close to 0.
However it's absolutely not in voltage mode because the transformer I/V is not a high impedance load for the DAC.
In this case, all the good things associated with current mode apply to transformer I/Vs.
Ian
Although it would be more expensive, if there were two ultra cap banks, you could have the idle one on charge while the active one is isolated and then switch once the voltage drops.
Just started following this thread, really interesting stuff going on here.
A 1:1 transformer hung on the output of a DAC isn't an I/V, it's voltage in/voltage out. The DAC chip itself sees whatever load the transformer is driving, such as the input impedance of an amplifier. In other words, it sees a high impedance and therefore is on voltage mode.
A 1:1 transformer hung on the output of a DAC isn't an I/V, it's voltage in/voltage out. The DAC chip itself sees whatever load the transformer is driving, such as the input impedance of an amplifier. In other words, it sees a high impedance and therefore is on voltage mode.
Current source and voltage source can be converted in between.
Source transformation - Wikipedia
So, an ESS DAC can be looked upon as either of them.
In which mode is decided by the load connected to the DAC output.
It will be in current mode if it is a low impedance load;
And it will be in voltage mode if it is a high impedance load.
I don't think we can say a 1:4 transformer with input resistor in parallel works in voltage mode.
Regards,
Ian
This is what Russ White (buffalo III designer) says about current mode from sabre ess dacs:
"Despite what you have been told, *ANY* resistance will have the DAC working in voltage mode. The fact is for all intents and purposes it is a voltage source unless it is working into a virtual ground.
In order for the DAC to act like a current source it really has to operate into a virtual ground. (though this will usually not be 0V but closer to AVCC/2)
Anyone who tells you differently is completely mistaken.
The effect of adding a resistor to GND is the same as having a voltage divider of RZ/195R where RZ is your resistor to GND.
So Vout = Vin * RZ/195R.
The reason the DAC *appears* as a current source into a virtual ground is because the output impedance is many times higher(as close to infinity as possible) than the input impedance of a virtual ground.
20R/195R is no where near high enough a ratio. Now .1R/195R would be much better."
Source:
Questions on es9018 outputs
See post #2
"Despite what you have been told, *ANY* resistance will have the DAC working in voltage mode. The fact is for all intents and purposes it is a voltage source unless it is working into a virtual ground.
In order for the DAC to act like a current source it really has to operate into a virtual ground. (though this will usually not be 0V but closer to AVCC/2)
Anyone who tells you differently is completely mistaken.
The effect of adding a resistor to GND is the same as having a voltage divider of RZ/195R where RZ is your resistor to GND.
So Vout = Vin * RZ/195R.
The reason the DAC *appears* as a current source into a virtual ground is because the output impedance is many times higher(as close to infinity as possible) than the input impedance of a virtual ground.
20R/195R is no where near high enough a ratio. Now .1R/195R would be much better.
"Source:
Questions on es9018 outputs
See post #2
It's not black and white as Russ says.
There is no defined impedance that says the DAC is working one way or the other. Driving any finite impedance, the DAC will output BOTH current and voltage. Which mode it's said to operate in depends on which one dominates, so driving a small value resistor it's safe to say it's working in current mode.
There is no defined impedance that says the DAC is working one way or the other. Driving any finite impedance, the DAC will output BOTH current and voltage. Which mode it's said to operate in depends on which one dominates, so driving a small value resistor it's safe to say it's working in current mode.
Hi Spartacus,
Agree with you that a low value resistor makes DAC more to the current mode.
I could be wrong, but I think the transformer Vin/Vout is just an equivalent equation for calculation, it's not the principle.
Based on the principle of electromagnetic induction(Faraday's law ),
Electromagnetic induction - Wikipedia
a transformer is really doing a current to voltage converting job.
"A varying current in the transformer's primary winding creates a varying magnetic flux in the transformer core and a varying magnetic field impinging on the secondary winding. This varying magnetic field at the secondary winding induces a varying EMF or voltage in the secondary winding due to electromagnetic induction."
Transformer - Wikipedia
The equation Vin/Vout=Np/Ns was just calculated from above principle, but physically, a transformer converts the AC current into AC voltage.
Regards,
Ian
Last edited:
HI Ian,
Wikipedia is of course correct. With any finite impedance, both current and voltage will be present. But within the context of the circuit it's employed in, the transformer is converting a voltage on it's input into a voltage on it's output. It's not working the same way as an I/V converter which has to present a small impedance to whatever is driving it.
Wikipedia is of course correct. With any finite impedance, both current and voltage will be present. But within the context of the circuit it's employed in, the transformer is converting a voltage on it's input into a voltage on it's output. It's not working the same way as an I/V converter which has to present a small impedance to whatever is driving it.
With transformer IV, the impedance at the secondary of the transformer is reflected back to the primary (according to the square of the turns ratio). Since a high input Z amp is usually connected to the secondary, it is the low value resistor at the primary that is doing all the IV conversion.
OPA1632 IV stage
OPA1632 is a fully differential audio amplifier. It has very high speed up to 180MHz bandwidth, 50 V/μs slew rate, ultra Low distortion and 1.3 nV/√Hz low noise. Seems more suitable for ESS DAC I/V stage than the standard OP amplifiers.
Twisted Pear uses it for both IVY and Mercury. The sound quality is suppose to be good. I want to give it a try on my ESS DAC HATs so I designed this OPA1632 I/V stage. One thing I'm worry about is the deep feedback. Could it affect the real listening experience? I'll figure it out by getting it compared with my other I/V boards soon. I'll set up another listening test then.
OPA1632IV1 by Ian, on Flickr
Ian
OPA1632 is a fully differential audio amplifier. It has very high speed up to 180MHz bandwidth, 50 V/μs slew rate, ultra Low distortion and 1.3 nV/√Hz low noise. Seems more suitable for ESS DAC I/V stage than the standard OP amplifiers.
Twisted Pear uses it for both IVY and Mercury. The sound quality is suppose to be good. I want to give it a try on my ESS DAC HATs so I designed this OPA1632 I/V stage. One thing I'm worry about is the deep feedback. Could it affect the real listening experience? I'll figure it out by getting it compared with my other I/V boards soon. I'll set up another listening test then.
OPA1632IV1 by Ian, on Flickr
Ian
Hey Ian, when you’re working on the RPi power supply, you might want to incorporate a shutdown routine for the pi ending with a hard shut down.
This would be a great feature for the RPi specifically, since it doesn’t have a power off itself.
It’s actually quite power hungry in standby or even in the “soft” off mode.
This would be a great feature for the RPi specifically, since it doesn’t have a power off itself.
It’s actually quite power hungry in standby or even in the “soft” off mode.