Also, looks like AVCC might be powered by a 3.3V 3-terminal regulator that also provides some power for the 100Mhz oscillator. Appears that the 3.3 volt bus is RC filtered then used for AVCC.
In addition, found another post from back in 2009 saying the ESS DACs outputs can either run into a Hi-Z input buffer buffer for voltage mode, or into a virtual ground IV for current mode operation. The DAC doesn't care which, but SQ is better with IV.
In addition, found another post from back in 2009 saying the ESS DACs outputs can either run into a Hi-Z input buffer buffer for voltage mode, or into a virtual ground IV for current mode operation. The DAC doesn't care which, but SQ is better with IV.
Last edited:
Oh bummer! Well, it's clear where your chip belongs. And it's a good start to look at what JensH and vicnic have done for I/V, not that either is too crazy. Then lots to do with filtering and bypass, but gets trickier to know which rails are misbehaving. Definitely worth getting the I/V on it's own supply.
Last edited:
Hey guys, I've been cliking through this forum but can't seem to find this so I thought I'd ask:
Can someone post some measurements of this DAC? (noise levels, freq response, basically RMAA stuff). I'd like to understand how it compares to something like the Odac, objectively.
Also, if you wouldn't mind linking to the DAC board you used along with the PSU (I know there are a couple of versions of the dac board menitoned in this combo and I just want to know I'm looking at the right one).
Thanks so much!
Can someone post some measurements of this DAC? (noise levels, freq response, basically RMAA stuff). I'd like to understand how it compares to something like the Odac, objectively.
Also, if you wouldn't mind linking to the DAC board you used along with the PSU (I know there are a couple of versions of the dac board menitoned in this combo and I just want to know I'm looking at the right one).
Thanks so much!
Yeah I just compared it to the OLDAC by JDS LABS, and this site pretty much on par with it whenthe chip should be able to perform much better.
Would you be able to give me a comprehensive guide perchance to the I/V modifications you have done (or would suggest I do to mine if I buy)? I have some basic understanding of these circuits but not enough that I'd really feel comfortable winging it from the info I've read on the forum so far (I've now gone thorugh this entire thread)
Sorry, as of right now I'm gearing up to buy my own boards and figure out what needs addressing myself. But ESS's own documentation tells us that their DACs act as mediocre voltage sources and great current sources, and the before/afters of JensH/Victor show the marked improvement in these systems by loading the DAC with an opamp I/V. As shown in Jens's plots, there's still some distortion to worry about (even if we need less Vout), but the challenge gets tougher as we're talking >100 dB. I'd be looking at rails that are being shared that shouldn't, contamination of grounds, proper bypass as the next steps. External noise injection into the circuit. Basically addressing the care and feeding of the chip to get close to DS performance.
At a certain point though, the noise/distortion will be sufficiently low that you're throwing the book at it, and there's probably bigger fish to fry. And that's well after the point of wondering if there's any audible improvements to be made.
At a certain point though, the noise/distortion will be sufficiently low that you're throwing the book at it, and there's probably bigger fish to fry. And that's well after the point of wondering if there's any audible improvements to be made.
It looks to me like the first thing to do might be to fix AVCC+. It also looks like the ES9038Q2M is a reduced pin count package of a ES9038PRO, maybe with 4 channels rather than 8, based on the appearance and locations of pins. That being the case, one op-amp ought to be able to supply what looks to be two AVCC+ pins, instead of 4 on the PRO version. Right now AVCC+ is supplied from the 3.3V supply for the digital circuitry, using only a single pole RC filter for isolation. Great for a critical analog signal, and they saved a whole 30 cents on not having buy another regulator!
I have an idea how to implement the ESS recommended circuit, by removing a few parts and mounting a little proto board under the bottom of the DAC PCB, where there is a very nice and convenient ground plane. I'm guessing it should reduce distortion and improve PSRR at the same time.
A bit more involved mod could provide current mode outputs ala Victor. Also, thinking of mounting it on the bottom.
If lucky, the cost of parts for the mods might possibly be made up by being able to use a slightly lesser quality power supply than a Silent Switcher (which costs twice as much as the DAC board).
I have an idea how to implement the ESS recommended circuit, by removing a few parts and mounting a little proto board under the bottom of the DAC PCB, where there is a very nice and convenient ground plane. I'm guessing it should reduce distortion and improve PSRR at the same time.
A bit more involved mod could provide current mode outputs ala Victor. Also, thinking of mounting it on the bottom.
If lucky, the cost of parts for the mods might possibly be made up by being able to use a slightly lesser quality power supply than a Silent Switcher (which costs twice as much as the DAC board).
Got around to looking at the ES9038Q2M board on a spectrum analyzer. A 1kHz test tone produces a lot of LF aliasing at 48kHz sample rate and using the default reconstruction filter. It looked virtually fixed when the ADC sample rate was upped to 96kHz. It also sounded much cleaner using some headphones, fixing a large amount of the distortion I said sounded really awful in a previous post. Guess that may have a fair bit to do with why Benchmark went with pre-DAC SRC to 210kHz.
That left primarily some 3rd, and a bit less 2nd harmonic distortion. Those would be expected from voltage mode operation.
It is still running on a Silent Switcher PS. PSRR should probably be further investigated, although it may always need very clean power. Have to see about it later, I guess.
That left primarily some 3rd, and a bit less 2nd harmonic distortion. Those would be expected from voltage mode operation.
It is still running on a Silent Switcher PS. PSRR should probably be further investigated, although it may always need very clean power. Have to see about it later, I guess.
Anybody ever try one of these things with one of these DACs? miniDSP Kits : miniDIGI
May grab one and do some experiments with it do see if I can get a workable solution using the ES9038 slow transition reconstruction filter. The biggest drawback may be the shipping and handling fee which is apparently $20 for a $30 item. Ouch!
May grab one and do some experiments with it do see if I can get a workable solution using the ES9038 slow transition reconstruction filter. The biggest drawback may be the shipping and handling fee which is apparently $20 for a $30 item. Ouch!
Hello,
I'm a newbie in this matter, so, my questions, probably, are stupid. Sorry if it true.
I have the "old version" (1.04) of the board and I would like to improve power lines on it slightly. Unfortunately, there is no headroom in the selected case to place big regulators boards and I plan to:
- replace AMS1117-3.3 regulator by LT1963-3.3 in the same case
- replace 7808 regulator by something like LM1084IT-5.0
Therefore, the questions:
- is there the sense to do it? Or I need to add some patch pcb like this https://www.ebay.com/itm/TPS7A4700-...y-Module-1-5V-20V-3-3V-5V-12V-1A/172746164061
- is 5V instead of 8V sufficient? If yes, what is the best solution for 5V regulator in place (I mean in TO220-3 package)
Many thanks in advance for your opinion.
I'm a newbie in this matter, so, my questions, probably, are stupid. Sorry if it true.
I have the "old version" (1.04) of the board and I would like to improve power lines on it slightly. Unfortunately, there is no headroom in the selected case to place big regulators boards and I plan to:
- replace AMS1117-3.3 regulator by LT1963-3.3 in the same case
- replace 7808 regulator by something like LM1084IT-5.0
Therefore, the questions:
- is there the sense to do it? Or I need to add some patch pcb like this https://www.ebay.com/itm/TPS7A4700-...y-Module-1-5V-20V-3-3V-5V-12V-1A/172746164061
- is 5V instead of 8V sufficient? If yes, what is the best solution for 5V regulator in place (I mean in TO220-3 package)
Many thanks in advance for your opinion.
very very good price and SQ...1 ES9038ES ... NX4 2017 version - portable chip ES9038Q2M
NEUE Richt NX4 DSD XMOS XU208 Chip DAC ES9038Q2M Chip Tragbare USB DAC DSD Decoder Verstarker Kopfhorer AMP Verstarker in NEUE Richt NX4 DSD XMOS-XU208 Chip DAC ES9038Q2M Chip Tragbare USB DAC DSD Decoder Verstarker Kopfhorer AMP Verstarker aus Kopfhorer Verstarker auf AliExpress.com | Alibaba Group
There is no SPDIF, only USB of some type? Don't know if there is even a USB driver for a PC.
They specify the output as line out, but expect it to be used with headphones or earbuds? Looks like opamp might be driving output, not proper headphone amp.
They do show a picture of the circuit board. Also, there are some performance graphs and THD table.
If you want a portable, battery operated, line output DAC, maybe okay. But what happens when the battery stops holding a charge?
Personally, I would pass on this one. But, that's only me.
i understand you completely...it was just an example (i don´t know if its possible for mods..)
I haven't had a chance to look at that version of the board. It is possible for most of the electronics to run on 3.3v, and more volts than that would probably only be needed for the op-amps. They probably work best with +- 15 volts, but maybe they are running them on a single-sided 8 volt supply? Do you know how to measure the pins?
As far as reducing power supply related hum and noise, a fix to AVCC+ 3.3v power may enough. The digital circuitry isn't so sensitive. Op-amps would like good power, but they do have some PSRR capability.
Hi Adson,
To my understanding the 7808 takes the +15V and makes a first step voltage reduction 15V->8V to not overload and maybe have better PSRR for the 2nd step 3,3V AMS1117 reg. Also 5V should work here to feed the AMS1117 reg.
I think you could also directly feed clean 3,3V to the output of the AMS1117 after removing this one.
I would also be interested for some better alternatives for 7808 and AMS1117...
To my understanding the 7808 takes the +15V and makes a first step voltage reduction 15V->8V to not overload and maybe have better PSRR for the 2nd step 3,3V AMS1117 reg. Also 5V should work here to feed the AMS1117 reg.
I think you could also directly feed clean 3,3V to the output of the AMS1117 after removing this one.
I would also be interested for some better alternatives for 7808 and AMS1117...