Let me clarify, I used the RCA phono outputs on the ESS board to connect to my amp, I used an interconnect cable into the RCA phono sockets marked DVD/LD as that one happened to be unused, it is a normal line level input, not a phono stage for a turntable.
I played music from foobar2000 into the ESS board via a TOSlink cable. I didn't try adjusting the volume of the PC's output.
I use my CS4344 DAC board in exactly the same way and it sounds good. My usual setup is TOSlink from the PC into the onboard DAC in my Yamaha amp, I don't touch the volume output on the PC, I just use the volume knob on the amp.
I played music from foobar2000 into the ESS board via a TOSlink cable. I didn't try adjusting the volume of the PC's output.
I use my CS4344 DAC board in exactly the same way and it sounds good. My usual setup is TOSlink from the PC into the onboard DAC in my Yamaha amp, I don't touch the volume output on the PC, I just use the volume knob on the amp.
Well, you should try it. This dac works differently than your old dac. 4dB down in the PC should be enough for whatever good it may do.
Folks may recall that I have strongly suggested modding a $32 LME49600 headphone amplifier board in order to be able clearly hear how well your dac is performing. So far, I am not aware of anyone doing that. If anyone would rather purchase a finished LME49600 HPA, a well constructed one can be found here: HP-1: Ultra-High End Headphone Amplifier
Last edited:
After living with this DAC using the ES9038Q2M and toyed with the idea of updating to an ES9028PRO DAC in the last month, I decide that the sound quality of this DAC is sufficient for the TV audio, DVD and CD digital sources that I use. I built another one using dual voltage internal PSU. It requires a deeper (180mm) case. The original factory built DAC and the home built DAC is shown side by side below.
ES9028 DAC DSD Decoder Board Supports IIS DSD 384KHz Coaxial Fiber+TFT LCD | eBay
length 180mm DAC chassis Aluminium case for ES9038Q2M DAC with display window | eBay
It is the same VR1.07 SMPCE Chinese board. The new case uses the onboard RCA jacks instead of off-board ones. The new home built DAC is working fine as is, but I am waiting for the power on-off and volume knob.
The transformer is from Antek with a dual 15VAC windings. The PSU has dual full bridge rectifier with LT1083CP regulators set at +/- 15VDC.
The schematic of the PSU is notional, not the actual values of components that I used.
I like the home built DAC very, very well and may sell the factory built DAC soon. If you are interested, PM me.
Last edited:
@keilau, Thank you for sharing. We definitely like to keep in touch and see pics of what people are doing with their Q2M DACs.
Don't know if you have been following the thread lately, but occip found a way to reduce distortion and improve sound quality of the Chinese ES9038Q2M DACs that are still using voltage mode output stages.
It does require hooking up an Arduino, RPi, or other MCU to set the H3 compensation register in the DAC, but its pretty easy to do. An Arduino 3.3v Trinket would easily fit in your DAC case.
It helps a lot of you do some kind of AVCC mod too, and I think one could easily fit under the board in the space available inside the case you have.
With a low-profile AVCC supply, harmonic distortion compensation, and keeping the voltage mode output stage you have, you could have a much nicer sounding DAC. You might be surprised how much you would like it once you heard it, and not much work involved either.
Don't know if you have been following the thread lately, but occip found a way to reduce distortion and improve sound quality of the Chinese ES9038Q2M DACs that are still using voltage mode output stages.
It does require hooking up an Arduino, RPi, or other MCU to set the H3 compensation register in the DAC, but its pretty easy to do. An Arduino 3.3v Trinket would easily fit in your DAC case.
It helps a lot of you do some kind of AVCC mod too, and I think one could easily fit under the board in the space available inside the case you have.
With a low-profile AVCC supply, harmonic distortion compensation, and keeping the voltage mode output stage you have, you could have a much nicer sounding DAC. You might be surprised how much you would like it once you heard it, and not much work involved either.
Last edited:
Markw4, thank you for your interest and encouragement. I was on vacation the first 2 weeks of July. I tried to catch up when returning. But did not quite fully digest the conversation among you, occip, Mikett and others yet.@keilau, Thank you for sharing. We definitely like to keep in touch and see pics of what people are doing with their Q2M DACs.
Don't know if you have been following the thread lately, but occip found a way to reduce distortion and improve sound quality of the Chinese ES9038Q2M DACs that are still using voltage mode output stages.
It does require hooking up an Arduino, RPi, or other MCU to set the H3 compensation register in the DAC, but its pretty easy to do. An Arduino 3.3v Trinket would easily fit in your DAC case.
It helps a lot of you do some kind of AVCC mod too, and I think one could easily fit under the board in the space available inside the case you have.
With a low-profile AVCC supply, harmonic distortion compensation, and keeping the voltage mode output stage you have, you could have a much nicer sounding DAC. You might be surprised how much you would like it once you heard it, and not much work involved either.
I believe that the newer version Chinese board has a reasonable noise floor and the THD is not too bad. But the SQ can be greatly improved if the third harmonic distortion can be reduced. Your suggestion is greatly appreciated. I will see if I can implement something meaningful and report it here.
perhaps an earth (loop) Problem, lift up the earth from the amp and test. this was the key here.
Last edited:
Latest forecast on Allo Katana ES9038Q2M dac is on their website: Katana DAC
It would seem the thinking may be that there could be some benefit from reducing the bandwidth of the analog low pass filter integrated into the I/V and differential output stages. Apparently, ESS suggests a corner frequency for analog filtering up at a few hundred kHz, and some people feel something lower, maybe 90kHz would be better. Why one or the other? As is very common with electronic design, sometimes design trade off choices are necessary. A reason for making analog LP filter corner frequency higher is that even though the so-called corner is high, lower frequencies can be affected somewhat too. It is common for preamps and power amps to to have a flat frequency response up to a few hundred kHz because if the corner frequency is set too low it can audibly affect phase shift of frequencies up around 20kHz (assuming one can hear high enough to notice). However, the analog filters in dacs can be designed to have a sharper cutoff and thus be set closer to the audio band, but there is still a need to avoid getting to close to that band. It can also be important to use amplifiers in dac output stage circuits designed to be capable of operating at frequencies well above the audio band. They may be more likely to successfully implement analog filtering without succumbing to their own RF induced distortion while doing it. This is one possible reason high speed opamps such as LME49720 may be used.
With delta-sigma dacs the engineering decision about analog filter corner frequency is also affected by low level RF noise coming out of the dac analog outputs. While such frequencies are too high for humans to hear, if allowed to reach sufficient amplitude they can get rectified in semiconductor junctions in amplifiers and cause internal amplifier bias voltages to shift. This can cause audible distortion to varying degrees. Also it may be possible for noise just above the audio band to intermodulate and directly introduce distortion into the audio band. The exact mechanisms can be complex and hard to pin down when they occur inside circuitry, but the result can be audibly degraded sound quality.
Not only can circuits inside a dac be affected, but preamps and power amps, and sometimes even speakers can affected by HF artifacts above the audio band.
Regarding opamps used in dac output stages, there may also be some trade offs available to the designer in terms of using integrated opmaps or discrete class-A opamps. Although discrete opamps have many very desirable features that may be hard to match in a discrete opamp, one typical characteristic of integrated opamps is that the distortion the produce depends on output level. At very low output levels the output stages may be mostly operating in the cross-over region where amplifier negative feedback works to linearize the opamp output stage. At higher output voltages the output is naturally more linear, so overall distortion can be less. Data sheet graphs of integrated opamp distortion vs output level generally tend to show an effect of decreasing distortion as output voltage level goes up.
Class-A discrete opamps don't have an output cross-over region like integrated opamps do. Discrete opamps may be worse in other regards (although discrete input stages are sometimes useful even when used in conjunction with integrated opamps) , but as always there are design trade offs and output stage intrinsic distortion in the absence of overall feedback may be made lower at low output voltages. Depending on how people hear and how opamps are designed and selected for particular uses, there may be preferences that vary from person to person for what is liked best.
All the foregoing is necessarily a quick, simplified explanation to try to help people start developing a little insight into some of the complexities of dac design and engineering trade off decisions. As always, all IMHO. Also, the proof of how well a particular design works will be seen in measurements and by listening, not by claims as to the technology used.
It would seem the thinking may be that there could be some benefit from reducing the bandwidth of the analog low pass filter integrated into the I/V and differential output stages. Apparently, ESS suggests a corner frequency for analog filtering up at a few hundred kHz, and some people feel something lower, maybe 90kHz would be better. Why one or the other? As is very common with electronic design, sometimes design trade off choices are necessary. A reason for making analog LP filter corner frequency higher is that even though the so-called corner is high, lower frequencies can be affected somewhat too. It is common for preamps and power amps to to have a flat frequency response up to a few hundred kHz because if the corner frequency is set too low it can audibly affect phase shift of frequencies up around 20kHz (assuming one can hear high enough to notice). However, the analog filters in dacs can be designed to have a sharper cutoff and thus be set closer to the audio band, but there is still a need to avoid getting to close to that band. It can also be important to use amplifiers in dac output stage circuits designed to be capable of operating at frequencies well above the audio band. They may be more likely to successfully implement analog filtering without succumbing to their own RF induced distortion while doing it. This is one possible reason high speed opamps such as LME49720 may be used.
With delta-sigma dacs the engineering decision about analog filter corner frequency is also affected by low level RF noise coming out of the dac analog outputs. While such frequencies are too high for humans to hear, if allowed to reach sufficient amplitude they can get rectified in semiconductor junctions in amplifiers and cause internal amplifier bias voltages to shift. This can cause audible distortion to varying degrees. Also it may be possible for noise just above the audio band to intermodulate and directly introduce distortion into the audio band. The exact mechanisms can be complex and hard to pin down when they occur inside circuitry, but the result can be audibly degraded sound quality.
Not only can circuits inside a dac be affected, but preamps and power amps, and sometimes even speakers can affected by HF artifacts above the audio band.
Regarding opamps used in dac output stages, there may also be some trade offs available to the designer in terms of using integrated opmaps or discrete class-A opamps. Although discrete opamps have many very desirable features that may be hard to match in a discrete opamp, one typical characteristic of integrated opamps is that the distortion the produce depends on output level. At very low output levels the output stages may be mostly operating in the cross-over region where amplifier negative feedback works to linearize the opamp output stage. At higher output voltages the output is naturally more linear, so overall distortion can be less. Data sheet graphs of integrated opamp distortion vs output level generally tend to show an effect of decreasing distortion as output voltage level goes up.
Class-A discrete opamps don't have an output cross-over region like integrated opamps do. Discrete opamps may be worse in other regards (although discrete input stages are sometimes useful even when used in conjunction with integrated opamps) , but as always there are design trade offs and output stage intrinsic distortion in the absence of overall feedback may be made lower at low output voltages. Depending on how people hear and how opamps are designed and selected for particular uses, there may be preferences that vary from person to person for what is liked best.
All the foregoing is necessarily a quick, simplified explanation to try to help people start developing a little insight into some of the complexities of dac design and engineering trade off decisions. As always, all IMHO. Also, the proof of how well a particular design works will be seen in measurements and by listening, not by claims as to the technology used.
Last edited:
It's not just the ESS Sabre chips, modern delta-sigma dacs that are commonly used for high end audio have similar types of characteristics as far as HF noise coming out of dac chip analog outputs.
A big part of what sets Sabre chips aside or makes them unique is the way they help to reduce jitter effects. They have other things to they do too, some of which are described in patents and some of which are trade secrets.
AKM has it's tricks and features too, and designing a very, very good AK DAC is also possible, although could be it might take more chips on a board to do it.
A big part of what sets Sabre chips aside or makes them unique is the way they help to reduce jitter effects. They have other things to they do too, some of which are described in patents and some of which are trade secrets.
AKM has it's tricks and features too, and designing a very, very good AK DAC is also possible, although could be it might take more chips on a board to do it.
Now you're beginning to ask hard questions.
A preamp line stage is also simple. It can be designed with one transistor, one mosfet, multiple transistors, multiple mosfets, Fire bottles/ Tubes, and op amps ranging from $0.05 each to $25 ones as well. They will all produce some level of distortion. Yet all of the circuits will sound different in some fashion and the rating of the better ones will not always correlate to the level of distortion measured. And oh, that passive preamp will essentially measure near zero distortion but it will also sound different. So is the passive the correct preamp with no active circuitry or are we all being taken for a ride with line stage preamps? Oh that rating...of the better ones... which camp has that opinion.
You'll essentially get multiple camps and opinions. Now choose which one you want to sit in. There is no correct answer at the end of the day. Just what you prefer in reality. So don't get too hung up on the best. Yes, there is a current best depending on the camp you reside but do note that there are different opinions as well which will produce different bests. Some prefer an accurate tight sound. Some prefer a sweet relaxed sound and some........they don't know but always rely on what the leader of a camp says and there are different camps...multiples.
Music is for you to enjoy. Better equipment might allow you to enjoy more of it but know what aspects are important to you. Then and only then will you be able to know based on your wants, which is better.... Some nice DACs also use transformers at their output and shun active stages. And some use tubes. and some use op amps and some use discrete op amps. They are all correct.
And to answer your question. No it is not. But it comes back to the question of a line stage.... A low pass filter can be designed without feedback and active circuitry, otherwise known as passive, passive can also be integrated with active circuitry and then there is the use of a transformer whose operation and method of working will inherently be a passive LP filter at the same time. Each again will sound different. At the beginning of this thread way way back, some were advocating the use of transformers rather than active IV stages. So choose your camp.......
A preamp line stage is also simple. It can be designed with one transistor, one mosfet, multiple transistors, multiple mosfets, Fire bottles/ Tubes, and op amps ranging from $0.05 each to $25 ones as well. They will all produce some level of distortion. Yet all of the circuits will sound different in some fashion and the rating of the better ones will not always correlate to the level of distortion measured. And oh, that passive preamp will essentially measure near zero distortion but it will also sound different. So is the passive the correct preamp with no active circuitry or are we all being taken for a ride with line stage preamps? Oh that rating...of the better ones... which camp has that opinion.
You'll essentially get multiple camps and opinions. Now choose which one you want to sit in. There is no correct answer at the end of the day. Just what you prefer in reality. So don't get too hung up on the best. Yes, there is a current best depending on the camp you reside but do note that there are different opinions as well which will produce different bests. Some prefer an accurate tight sound. Some prefer a sweet relaxed sound and some........they don't know but always rely on what the leader of a camp says and there are different camps...multiples.
Music is for you to enjoy. Better equipment might allow you to enjoy more of it but know what aspects are important to you. Then and only then will you be able to know based on your wants, which is better.... Some nice DACs also use transformers at their output and shun active stages. And some use tubes. and some use op amps and some use discrete op amps. They are all correct.
And to answer your question. No it is not. But it comes back to the question of a line stage.... A low pass filter can be designed without feedback and active circuitry, otherwise known as passive, passive can also be integrated with active circuitry and then there is the use of a transformer whose operation and method of working will inherently be a passive LP filter at the same time. Each again will sound different. At the beginning of this thread way way back, some were advocating the use of transformers rather than active IV stages. So choose your camp.......
Last edited:
While this is an ESS thread, do be aware that there exists other kinds of DAC, for example PS Audio also offers three DACs, the least expensive one is based on an ESS chip with all active circuitry, Stellar Gain Cell DAC | PS Audio
Then their top of the line is based on an FPGA rather than a dedicated DAC chip. So all the functions are based on their design from scratch....and uses transformers. DirectStream DAC | PS Audio
Then there is also a lesser one that sits on top of their ESS offering but below their top.
It is also based on their FPGA but uses active circuitry for output rather than transformers.DirectStream Junior DAC | PS Audio
I am sure all three sound different, but it goes to show you that there are more ways to reach a goal. Yet essentially all the equipment that they offer are designed by the same engineers. So it is a nice lineup only to illustrate that the same engineers having the same knowledge and experience will come up with different approaches and end results. I am sure they did not knowingly design anything incorrectly at any stage...just that they had cost limitations to each design that forces them to make decisions as to what they could offer...with different end results.
But bear one thing in mind. This company does not make equipment for recording studios which is actally one camp i.e Crane Song and Benchmark and others. They are a different camp...entirely. To them at the end of the day musicality is what counts not necessarily objective accuracy, which may or may not be your cup of tea. So even in one camp their exists subcamps or preferences. Then what is musicality? Is it accuracy? and then how is that measured.
Then their top of the line is based on an FPGA rather than a dedicated DAC chip. So all the functions are based on their design from scratch....and uses transformers. DirectStream DAC | PS Audio
Then there is also a lesser one that sits on top of their ESS offering but below their top.
It is also based on their FPGA but uses active circuitry for output rather than transformers.DirectStream Junior DAC | PS Audio
I am sure all three sound different, but it goes to show you that there are more ways to reach a goal. Yet essentially all the equipment that they offer are designed by the same engineers. So it is a nice lineup only to illustrate that the same engineers having the same knowledge and experience will come up with different approaches and end results. I am sure they did not knowingly design anything incorrectly at any stage...just that they had cost limitations to each design that forces them to make decisions as to what they could offer...with different end results.
But bear one thing in mind. This company does not make equipment for recording studios which is actally one camp i.e Crane Song and Benchmark and others. They are a different camp...entirely. To them at the end of the day musicality is what counts not necessarily objective accuracy, which may or may not be your cup of tea. So even in one camp their exists subcamps or preferences. Then what is musicality? Is it accuracy? and then how is that measured.
Last edited:
The nature of electrical engineering is that the devices we have that can used to make amplifiers and filters, anything really, are not ideal. They are what they are and we mostly learn how to manage the limitations as needed. Need, by the way, includes constraints like cost and time spent.
As far as making a good enough filter/amplifier for dacs, it depends on how good you want. Also, many things can seem difficult until they are fully understood. There may be some relevant characteristics of components that are not described in data sheets or guaranteed by manufacturers. A certain amount of R&D may be needed to improve this or that part of a system.
In other words there is no one simple answer to your question. The real world is messy and complicated.
Speaking of those last two things, I'm not sure about what some of what Mikett was meaning to say other than some of it sounded more on the philosophical side of engineering. Also, don't know about his choice of alternate technology DACs, but I think a very interesting one might be this one: Mola Mola As far as distortion goes, it seems to be mostly limited by the analog electronics, no surprise there I guess. With the preamp they say they can get distortion down to around -130dB, but detailed information about how that was measured is not available. Probably doesn't matter though, it looks like one of those things where if you have to ask how much it costs, probably means you can't afford it.
It rather depends on which kind of distortion you're trying to avoid. An opamp based RC active filter isn't hard to implement and will likely give impressively low THD+N if done with low enough impedances for the passives.
On the other hand, a passive component LC filter will likely give higher THD due to slight inductor non-linearities but most likely better IMD when fed with well OOB (out of band) signals in addition to test tones.
Last edited:
Thankyou for the detailed explanations I am paying attention and finding this very informative and useful.
"impressively low THD+N if done with low enough impedances for the passives"
This means metal film types for the resistors, correct? What sort of caps would be the best choice?
I'm putting much thought into an output stage at the moment, I have built a simple one but feel it is the area where I should put more effort into the design and component selection to get meaningful increases in SQ.
"impressively low THD+N if done with low enough impedances for the passives"
This means metal film types for the resistors, correct? What sort of caps would be the best choice?
I'm putting much thought into an output stage at the moment, I have built a simple one but feel it is the area where I should put more effort into the design and component selection to get meaningful increases in SQ.
Thin film resistors and C0G capacitors. The information about that as suggested by ESS can be found here: http://www.esstech.com/files/4514/4095/4306/Application_Note_Component_Selection_and_PCB_Layout.pdf in the section entitled, "Component Selection."
Cheers, I shall study that document.
I looked at C0G and the vast majority seem to be SMD types. The through holes ones are rather expensive, also, the values are all in the pico farad range, I am beginning to see some of the less obvious difficulties in designing output stages, especially for current output DACs.
I looked at C0G and the vast majority seem to be SMD types. The through holes ones are rather expensive, also, the values are all in the pico farad range, I am beginning to see some of the less obvious difficulties in designing output stages, especially for current output DACs.
C0G is essentially the same as NPO. They are available up to maybe around 0.33uf or so that I have seen. They are really good caps, but not the cheapest. In addition, SMD is not all that bad except the super small sizes. I am generally okay with hand soldering many of the SMD parts, although it seemed kind of daunting until I started practicing a little first.
It turns out there are practice soldering kits for SMD parts and also other things to help you such as very small diameter solder iron tips, small diameter solder wire, head mounted magnifiers, adjustable spot lighting, and small tweezers and other tools and devices to make it all pretty doable. Just part of modern electronics. You can probably do it just like most other people can.
It turns out there are practice soldering kits for SMD parts and also other things to help you such as very small diameter solder iron tips, small diameter solder wire, head mounted magnifiers, adjustable spot lighting, and small tweezers and other tools and devices to make it all pretty doable. Just part of modern electronics. You can probably do it just like most other people can.