Hey guys, I wonder if I might ask a question of you which is, has anyone at all ever lifted the the I2C bus pins of the microcontroller, or otherwise rigged up the board to allow Arduino control of the DAC chip? I know at one time a few people expressed curiosity to see how I did it, but after I posted some pics showing the shield and about lifting the pins of the microcontroller than that was the last I ever heard of any interest in it.
It would be understandable, I guess, if people felt there was some risk of bricking the DAC if any pins broke off the microcontroller or something like that. Maybe another possibility could be that people were just curious to see how an Arduino might be hooked up just in case they ever wanted to do it at some point in the future, but so far haven't wanted to.
Regarding risk of breaking off microcontroller pins, as long as they aren't bent back and forth enough times to fatigue them off I haven't found them to be excessively delicate, and even if one did break off, some more experimentation shows it ought to be repairable without too much trouble.
Anyway, it would be of great interest to me and maybe to others following the thread to know if anyone has done it, or if not, why not? If anyone, or even better a few people would like like to say something about it, it would be much appreciated.
Thanks,
-Mark
EDIT: @system54, looks like we cross posted. Looking forward to seeing your buffer.
It would be understandable, I guess, if people felt there was some risk of bricking the DAC if any pins broke off the microcontroller or something like that. Maybe another possibility could be that people were just curious to see how an Arduino might be hooked up just in case they ever wanted to do it at some point in the future, but so far haven't wanted to.
Regarding risk of breaking off microcontroller pins, as long as they aren't bent back and forth enough times to fatigue them off I haven't found them to be excessively delicate, and even if one did break off, some more experimentation shows it ought to be repairable without too much trouble.
Anyway, it would be of great interest to me and maybe to others following the thread to know if anyone has done it, or if not, why not? If anyone, or even better a few people would like like to say something about it, it would be much appreciated.
Thanks,
-Mark
EDIT: @system54, looks like we cross posted. Looking forward to seeing your buffer.
Last edited:
Well... Looks like we are into Thursday out here where I am. No word on shipping yet. Hopefully an update will come before too long. Interesting to hear what it sounds like after the firmware upgrade.
Last edited:
Hi Mark
today its on the way , you will get the tracking number (today). Will be interesting to see how it will stack up vs DAC3 and modded far east boards.
I am listening non stop to the unit and cannot stop smiling. ;-)
Sound quality is good from first day , opens after a day , and 3 days in it reaches the summit (if not yet the top).
I listened to it to many genres of music , and on classical I can hear distinctive string instruments being "pinched"...but I am not a "golden ear"
Seems that you can (you found the issue at HF on old firmware by listening to it) so it will be an interesting week.
Regarding PSU , from a pure THD+N we cannot see a difference powering using SMPS of LPS...however we can certainly hear one . I recommend a SMPS for RPI , IFI (or any low noise) for Katana , SMPS (or low noise) for Microcontroller
The way that we implemented filtering is pretty advanced in the industry(thats well know for filtering) and it should not make a difference..and yet it does. We cannot measure the difference in SQ but we can certainly hear it. Every rail (and I mean every rail) is filtered. A lot of attention was given to the bypassing caps position and grounds. LDOs are NOT sufficient. In most rails we use a minimum of LC filters + LDOs but in critical paths we use "ripple eaters" + LDOs (some times 2 of them in series) + LC filters that use damping..
Try reading the noise on (any) specific rails and you might be at noise floor of you test equipment .
I hope that you will enjoy your testing. 🙂
today its on the way , you will get the tracking number (today). Will be interesting to see how it will stack up vs DAC3 and modded far east boards.
I am listening non stop to the unit and cannot stop smiling. ;-)
Sound quality is good from first day , opens after a day , and 3 days in it reaches the summit (if not yet the top).
I listened to it to many genres of music , and on classical I can hear distinctive string instruments being "pinched"...but I am not a "golden ear"
Seems that you can (you found the issue at HF on old firmware by listening to it) so it will be an interesting week.
Regarding PSU , from a pure THD+N we cannot see a difference powering using SMPS of LPS...however we can certainly hear one . I recommend a SMPS for RPI , IFI (or any low noise) for Katana , SMPS (or low noise) for Microcontroller
The way that we implemented filtering is pretty advanced in the industry(thats well know for filtering) and it should not make a difference..and yet it does. We cannot measure the difference in SQ but we can certainly hear it. Every rail (and I mean every rail) is filtered. A lot of attention was given to the bypassing caps position and grounds. LDOs are NOT sufficient. In most rails we use a minimum of LC filters + LDOs but in critical paths we use "ripple eaters" + LDOs (some times 2 of them in series) + LC filters that use damping..
Try reading the noise on (any) specific rails and you might be at noise floor of you test equipment .
I hope that you will enjoy your testing. 🙂
Haha, no golden ears here. Learned and long practiced listening from mixing records is where skills seem to come from.
Regarding burn-in or opening-up, in my experience there can be two reasons for that. They are, (1) the equipment needs to stabilize for some hours, and, (2) the listener's System 1 brain processes learn to interpret the sound as normal. (In this context, System 1 refers to the two system model of the mind sometimes used by cognitive psychologists, wherein System 2 processes constitute the 5% of brain activity that make up conscious awareness, and System 1 processes constitute the other 95% of brain activity that is not directly observable by conscious awareness. Please see Kahneman 2012 for more information.)
Regarding burn-in or opening-up, in my experience there can be two reasons for that. They are, (1) the equipment needs to stabilize for some hours, and, (2) the listener's System 1 brain processes learn to interpret the sound as normal. (In this context, System 1 refers to the two system model of the mind sometimes used by cognitive psychologists, wherein System 2 processes constitute the 5% of brain activity that make up conscious awareness, and System 1 processes constitute the other 95% of brain activity that is not directly observable by conscious awareness. Please see Kahneman 2012 for more information.)
I haven't go golden ears either, but what started me off on the quest to improve my hifi was my sister kicking her longtime partner out, an moving house to somewhere smaller for her and the dog. He left behind his Meridian separates and my sister is the type who can't wire a plug without help, so called me to go plug it all in for her, which I did. Anyways, there I am, plugged it all in, positioned the fancy electrostatic speakers where I think they should roughly be and I grabbed a CD to test it - Nirvana Unplugged, it just so happened to be, a recording I have heard a very great number of times. Well, I was amazed, I could hear things I never knew were there, bits of fret buzz from Pat Smear's guitar, subtleties in Dave Grohl's percussion I never noticed etc.
Now, I'll never own such expensive high end kit myself, but I want to try to get my setup to play Nirvana Unplugged like I heard it that day.
Now, I'll never own such expensive high end kit myself, but I want to try to get my setup to play Nirvana Unplugged like I heard it that day.
The sad part about today is the inability to hear a real Hi Fi easily. We're not talking Best Buy or those big box stores. Equipment has gotten better and the ability/access to initially experience it has decreased. If your circle of friends is removed from that arena there is so little opportunity unless you splash out in big purchases.
Back in the day, you could still visit some boutique stores and they would try to hook you. My real experience was when I went went to my high school friend's apt in 1980 and he introduced me to a Linn, Naim and B&W 801 setup. I went back home and could not stand what I had anymore.
Then I went through at least six turntables in 2-3 years. Equal number of tonearms and cartridges. Listened to so many amps, from Bedinis Classe Bryston Threshold Krell McIntosh. Each had its own character that sometimes clash with the system. Synergy is a big part of a system for me now. It's all in there and beyond a certain level of performance, like CDS games has commented, The numbers don't tell the real story. Controversial but wire direction can make a difference if you are inclined to discover it. To me now, it is too much work to optimize but it is there provided your system can resolve it. Yeah I pooh poohed that until I started to hear it and then once you familiarize yourself with a track and system, you could pick it out. However randomly using a track and system you're not familiar with, I'd say it is near impossible to discern.
Then one day, Nelson Pass came through Vancouver in the early 80s, visited out group and foretold of the Class D amps to come. We could not grasp/visualize how small they could be. Here we are.
Back in the day, you could still visit some boutique stores and they would try to hook you. My real experience was when I went went to my high school friend's apt in 1980 and he introduced me to a Linn, Naim and B&W 801 setup. I went back home and could not stand what I had anymore.
Then I went through at least six turntables in 2-3 years. Equal number of tonearms and cartridges. Listened to so many amps, from Bedinis Classe Bryston Threshold Krell McIntosh. Each had its own character that sometimes clash with the system. Synergy is a big part of a system for me now. It's all in there and beyond a certain level of performance, like CDS games has commented, The numbers don't tell the real story. Controversial but wire direction can make a difference if you are inclined to discover it. To me now, it is too much work to optimize but it is there provided your system can resolve it. Yeah I pooh poohed that until I started to hear it and then once you familiarize yourself with a track and system, you could pick it out. However randomly using a track and system you're not familiar with, I'd say it is near impossible to discern.
Then one day, Nelson Pass came through Vancouver in the early 80s, visited out group and foretold of the Class D amps to come. We could not grasp/visualize how small they could be. Here we are.
Last edited:
Thank you for the pictures. When you install it, very much recommended to keep the ground and AVCC wires to the dac board as short as possible. Shortest way is probably from the bottom side of the dac board to the existing AVCC cap pins. Just need to cut the traces or remove the inductors going back to the dac board 3.3v supply.
yes i want to remove the resistors and avcc caps of the dac board. so the wire lenght of the avcc reg polymer cap to the dac board is about 15mm
Last edited:
The existing caps don't hurt anything and might help. If attaching to the leads from the bottom of the board there is no need to remove them. The 'resistors' actually seem to be some kind of inductors or ferrites. They would make better filters than resistors at HF, yet pass DC and LF more readily than resistors. That would be the idea, anyway.
System54,
So you use +15V - 0V reg. to feed the avcc opamp power rails? Where will you get the reference voltage of 3,3V from?
The buffer you mentioned before is placed after an I/V converter to run headphones?
So you use +15V - 0V reg. to feed the avcc opamp power rails? Where will you get the reference voltage of 3,3V from?
The buffer you mentioned before is placed after an I/V converter to run headphones?
I desolder the caps in order to fit my pcb. after that i will perhaps solder better poly caps.
Regarding a voltage reference for AVCC, to my way of thinking it would make sense to use a separate 3.3v regulator or reference for AVCC rather than using the same 3.3v supply for all the digital circuitry. The 10k and 10uf filter before the AVCC opamp can attenuate noise from the 3.3v reference before it goes into the AVCC supply opamp, but the attenuation is only so much. Up at frequencies where digital circuits put noise on the power rails, electrolytic filter caps may not filter very effectively. At least that was part of the thinking that led me to use a dedicated 3.3v reference for my AVCC supply, and I have already outlined reasons for putting on the back side of the board and right against the ground plane. It may seem like common sense for somebody who has been doing this stuff for a very long time, but not so obvious at all to somebody with a different background and different experiences. My recommendation as always would be to make AVCC as clean, pure, well shielded, and close to the dac pins as possible.
i use the recommended ess avcc circuit but without dvcc.
the lt3042 is the vref (3,3v) vin is +15v
Some interesting info recently came in by PM. Since the person sent it to me rather than posting in the forum I will just summarize what may be of interest to everyone here. If the person would like to say more about it, of course they would be very welcome to.
Anyway, it turns out someone not afraid of I2C bus just modded AVCC to work with an opamp regulator, but kept for the original voltage mode output stage. They then went into the dac control registers over I2C bus and adjusted harmonic distortion compensation to minimize output stage distortion. Turns out they were able to get the worst harmonic, the 3rd, down to -114dB at -6dB FS. The 2nd harmonic was already at -120dB and didn't need any fixing. And, yes, you heard it right, it was still using a voltage mode output stage.
If anyone wants to try it and doesn't have a way to measure distortion, the control value that minimized 3rd harmonic in this case was 0xf800. However, ESS says those registers have to be custom set for every dac, so don't know how well a one-size-fits-all setting would work for someone else.
Another thing maybe to bear in mind about dacs like Sabre is that not all the distortion they produce can be characterized by static test test tones, since modulator state space variables settle into a stable, kind of minimized condition with steady input signals. But, it sure sounds like an interesting thing to try.
Anyway, it turns out someone not afraid of I2C bus just modded AVCC to work with an opamp regulator, but kept for the original voltage mode output stage. They then went into the dac control registers over I2C bus and adjusted harmonic distortion compensation to minimize output stage distortion. Turns out they were able to get the worst harmonic, the 3rd, down to -114dB at -6dB FS. The 2nd harmonic was already at -120dB and didn't need any fixing. And, yes, you heard it right, it was still using a voltage mode output stage.
If anyone wants to try it and doesn't have a way to measure distortion, the control value that minimized 3rd harmonic in this case was 0xf800. However, ESS says those registers have to be custom set for every dac, so don't know how well a one-size-fits-all setting would work for someone else.
Another thing maybe to bear in mind about dacs like Sabre is that not all the distortion they produce can be characterized by static test test tones, since modulator state space variables settle into a stable, kind of minimized condition with steady input signals. But, it sure sounds like an interesting thing to try.
Last edited:
Happy to hear about it. I opted for voltage out mode (since the beginning of this topic), and keep recommending the Burson V5 opa as a great choice. I am going to change all the smd resistors (at the output) to melf type. Did somebody try melfs? This modd was recommended in the 9018K2M topic earlier.
Is this the explanation why I did not feel this DAC as such horrible experience even in unmodded voltage mode such as other user describe here? Maybe I was lucky with orig. distortion setting...
Further I am not sure if everybody here is aware to not drive the DAC with 0dB digital volume as Distortion (inter sample clipping caused by digital filter I guess) increases dramatically. - 4dB digital is Ok.
Further I am not sure if everybody here is aware to not drive the DAC with 0dB digital volume as Distortion (inter sample clipping caused by digital filter I guess) increases dramatically. - 4dB digital is Ok.
ESS Technology :: ES9038PRO
According to ESS, the compensation will vary with each board itself and is a function of the external components used to "glue" this DAC together. So for example does that mean each op amp used will change the required value? This sounds to be a tricky thing to nail down without a series of trial and error to home in on the value though an iterative process.
According to ESS, the compensation will vary with each board itself and is a function of the external components used to "glue" this DAC together. So for example does that mean each op amp used will change the required value? This sounds to be a tricky thing to nail down without a series of trial and error to home in on the value though an iterative process.
Yes thats exactly what it means. Each board needs THD+N compensation for that particular board. Each opamp , R tolerance at 0.1% to 5% and C tolerance at 20%)+ voltage variation means that every board has to be manually programmed for THD compensation.