Well, it probably already does perform better, but it is let down by flatcable connections etc. I'm working on a ES9016 design, that will replace the CS4397 dac permanently. This ES9023 design is just to get some experience.
Amp is a Kenwood L01a, according to specs, input impedance is 50k, which is not that high, but quite "normal". The difference between buffered and unbuffered was to be expected, because the opamp inside the ES9023 cannot do miracles. But I cannot put a finger on the reason why the dual supply version is an improvement without the buffer, but performs equal compared to single supply with buffer. Perhaps the charge pump inside the ES9023 doesn't deliver enough current?
Good luck with the ES9016 design, it should be a better chip so I am curious how things will turn out sound wise. It will be harder to get things right ( I just read some threads over here) but at least you won't have the charge pump 
Keep us informed please.
You will need a larger case, more transformers, more power supplies and more PCB's with the ES9016....and more money
Keep us informed please.You will need a larger case, more transformers, more power supplies and more PCB's with the ES9016....and more money
Attachments
Last edited:
Good luck with the ES9016 design, it should be a better chip so I am curious how things will turn out sound wise. It will be harder to get things right ( I just read some threads over here) but at least you won't have the charge pump
You will need a larger case, more transformers, more power supplies and more PCB's with the ES9016....and more money
Thanks.
It would be interesting to see how the ES9023 would perform with the ultra low noise shunt supply, 30 dollar crystek clock and nice PCB, I'm planning to use for the ES9016. But it feels a bit strange to spend that amount of money for parts that surround a 2 dollar chip.
I'll make a post on the ES9016 design before I'll make the PCB. But that will take some time, first have to finish the USB to I2S interface.
You're welcome to drop by for a demo when I'm done!
The cost of the both chips won't differ very much (in production). I think it is quite safe to assume that at least some of the technology used in the higher priced chips is equally used in the lower priced chips. It would not make sense to invent a complete new die nowadays. That is what ESS tells us anyway. Also the cheaper chips will be produced in larger numbers than the expensive ones so a manufacturer can't skip on quality.
We have a nice PCB design but no plans for using shunt supplies and 30 $ XO's but we'll see how V3 will be received.
Offer accepted but although we're not the fastest on earth I think our V3 will be finished earlier than your ES9016 design
We have a nice PCB design but no plans for using shunt supplies and 30 $ XO's but we'll see how V3 will be received.
Offer accepted but although we're not the fastest on earth I think our V3 will be finished earlier than your ES9016 design
Last edited:
Last Friday, I experimented some with stacking of ES9023. Don't whether there are any posts on this topic on Diyaudio.
I just soldered two ES9023 on top of each other, leaving the CP/CN pins unconnected. The MSCLK pins are not directly connected, both have their own 100 Ohm resistor. I left the no-audio pin disconnected for the top DAC chip. All other pins are directly connected. The chips get a symmetrical supply (-3.6V and +3.6V)
It works without any problems.
Initial impressions are:
- Larger soundstage (perhaps bit too large, "virtual surround" like)
- Less bass
but the second chip is brand new, so I must repeat tests after 10 hours of burn-in.
Any comments on the way to merge the output signal of the two ES9023 dacs? Like giving both dacs their own passive LPF, or even separate buffer stages.
Or does someone know of a simple way to split L/R signals in the I2S stream? That would definitely improve the cross-talk performance.
I just soldered two ES9023 on top of each other, leaving the CP/CN pins unconnected. The MSCLK pins are not directly connected, both have their own 100 Ohm resistor. I left the no-audio pin disconnected for the top DAC chip. All other pins are directly connected. The chips get a symmetrical supply (-3.6V and +3.6V)
It works without any problems.
Initial impressions are:
- Larger soundstage (perhaps bit too large, "virtual surround" like)
- Less bass
but the second chip is brand new, so I must repeat tests after 10 hours of burn-in.
Any comments on the way to merge the output signal of the two ES9023 dacs? Like giving both dacs their own passive LPF, or even separate buffer stages.
Or does someone know of a simple way to split L/R signals in the I2S stream? That would definitely improve the cross-talk performance.
Hi, information on balanced use of your Subbu Dac is either hard to find, or very sparse. But from what I found, I conclude that for the second dac, the SD line of the I2S signal is inverted. Can you explain what this does for the analog output signal?
I'm interested in making two mono versions of the ES9023, one for left and one for right.
So, I either mute the SD line, when LR is high. Or even nicer, I buffer the SD data when LR is high, and playback this buffer when LR is low. But my knowledge of low level digital electronics is too limited to that.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.