Hello guys, want make dac on this board.
Maybe some one has experiense of building?
I have some question, maybe someone can give me a advise:
1. Capacitors around chip, NP0 ceramic will be ok?
2. Oscillator NDK NZ2520SD will be ok?
3. KEMET Aluminium Organic Polymer Capacitor will be on on digital supply?
On analog supply will be elna silmic II.
4. Power stabilizers LT1963? Or need extreme low noise LT3042?
5. ES9038 have high output current in mono mode, which op amplifier which
one to prefer?
OPA1688 ：75mA, OPA1612：55mA, OPA1622：145mA
6. otput stade capacitors xicol polystyrene, dale resistors?
7. I think don't used a I2C controller because PCM/DSD and I2S/SPDIF
switch automaticaly?
[/url][/IMG]
[/url][/IMG]
[/url][/IMG]
Maybe some one has experiense of building?
I have some question, maybe someone can give me a advise:
1. Capacitors around chip, NP0 ceramic will be ok?
2. Oscillator NDK NZ2520SD will be ok?
3. KEMET Aluminium Organic Polymer Capacitor will be on on digital supply?
On analog supply will be elna silmic II.
4. Power stabilizers LT1963? Or need extreme low noise LT3042?
5. ES9038 have high output current in mono mode, which op amplifier which
one to prefer?
OPA1688 ：75mA, OPA1612：55mA, OPA1622：145mA
6. otput stade capacitors xicol polystyrene, dale resistors?
7. I think don't used a I2C controller because PCM/DSD and I2S/SPDIF
switch automaticaly?
Last edited:
which dac chip are you using? the board says its for 9018/28 and 38 ... lol. these chips all have different requirements, so this board is a bit dubious. the pcb also only has DIP8 opamp type placement and this is not compatible with opa1622 or opa1688. the opa1612 will work on an adapter, but the only dip adapters ive seen with 1622 and 1688 on ebay have been stupidly expensive and dont have a layout that suits the chips; neither does this board, its not as simple as just plugging it into an existing circuit.
I have the full datasheet. What I need to know is if YOU are running it mono, because you mentioned mono mode, but the PCB is a stereo build. in a stereo build, if you were running current mode (you posted the voltage mode schematic and honestly, knowing this seller, it's a voltage buffer, not an IV stage ..
in stereo current mode, its around 61mA peak to peak. so the opa1612 could almost handle it. it would probably be ok, although I might want a bit more margin. That being said, if its a voltage mode output, it doesnt matter.
regarding the bypass caps. for the analogue supply pins, I would use films. the rest, id use x7r rather than npo. they will be better damped and less likely to ring in this haphazard layout. The kemets are fine for bulk decoupling. use whatever polystyrene or polyprop you wish in the right values.
the clock. if you can get the ndk in 90.3168MHz, or 100MHz, yes. depends on what the highest sample rate you are targeting is.
I would use a controller, there are too many options you wont be able to use otherwise.
I would use lt3042 or similar high quality linear reg for AVCC supplies and clock. the rest are not as important
in stereo current mode, its around 61mA peak to peak. so the opa1612 could almost handle it. it would probably be ok, although I might want a bit more margin. That being said, if its a voltage mode output, it doesnt matter.
regarding the bypass caps. for the analogue supply pins, I would use films. the rest, id use x7r rather than npo. they will be better damped and less likely to ring in this haphazard layout. The kemets are fine for bulk decoupling. use whatever polystyrene or polyprop you wish in the right values.
the clock. if you can get the ndk in 90.3168MHz, or 100MHz, yes. depends on what the highest sample rate you are targeting is.
I would use a controller, there are too many options you wont be able to use otherwise.
I would use lt3042 or similar high quality linear reg for AVCC supplies and clock. the rest are not as important
thank you very much, I mean mono: stereo mode not 8 channels
maybe i can recalculate resistors values for another op? maybe for opa604?
the one guy is telling about it there:
y2blog >> DIYINHK ES90x8PROリファレンス基板
maybe i can recalculate resistors values for another op? maybe for opa604?
the one guy is telling about it there:
y2blog >> DIYINHK ES90x8PROリファレンス基板
Dont follow his advice about adding a buf634. adding this doesnt make it useable, well hard to see without his schematic using it. no, opa604 isnt suitable, it doesnt have large enough current drive. his result of saying its 70ma? ... he is making an error somewhere in his calculation. most likely he is not allowing for losses and hasnt got (or hasnt read) the datasheet ... doing what I assume is his simple calculation of 3.3v (AVCC voltage) / 50Ω output impedance *1000 = 66mA, when the actual result, using 3.3V * 0.924/50Ω = 0.060984 * 1000 = 60.984mA.
actually looks like you can get opa1688 on soic8 adapter from diyinhk, so i'd just go with that until you get it working and maybe try some better chips later.. I didnt think it was available in soic, rather much smaller DFN package.
actually looks like you can get opa1688 on soic8 adapter from diyinhk, so i'd just go with that until you get it working and maybe try some better chips later.. I didnt think it was available in soic, rather much smaller DFN package.
I know that DIYINHK board. Had one myself to play around with. Sketched up the schematic, if I can find it somewhere. Overall though, I think a 2-layer board for ES9038PRO is probably a mistake. Also, these days ESS is designing their ES9038PRO evaluation board more like AKM designs theirs. For one thing, there is the option for more filtering before the differential summing stage. Differential summing doesn't have to be of multi-feedback filter topology either since that is off-board and up to the designer.
For any of the analog circuitry bypass caps, I would suggest using good quality .1uf film caps. It will probably sound better than X7R. Also, the clock should be considered to be 'analog circuity.'
For any of the analog circuitry bypass caps, I would suggest using good quality .1uf film caps. It will probably sound better than X7R. Also, the clock should be considered to be 'analog circuity.'
yeah. I would just avoid it too, tbh. I think it would end up costing you a bit to build, for a substandard result. maybe if you had some parts lying around already, but if you are buying and shipping everything I think you would end up spending topping money, for a MUCH worse result. I know its probably not what you want to hear, but losing $15 is much better than losing out $600. The rest of my advice stands, apologies, didnt spot 1688 was a dual, I havent used it. If its all about the experience ... sure, but this board wasnt a great design 10 years ago, for the 9018, let alone trying to say the same ordinary board should be used for 9038 ...
I didnt suggest X7R for analogue. I suggested film, but suggested X7R instead of C0G for decoupling digital around the chip, as C0G can ring more easily if you arent applying a tuned, measured value.
I didnt suggest X7R for analogue. I suggested film, but suggested X7R instead of C0G for decoupling digital around the chip, as C0G can ring more easily if you arent applying a tuned, measured value.
Last edited:
Not sure if its automatic. The default settings of the I2C register bits shown in the datasheet should tell you what it will do. If you want anything other than the defaults then you have to program the registers.
If you need help deciphering the data sheet register info, I probably have some links I could post that would help. ESS uses nomenclature that looks like Verilog style to me. Other dac manufacturer data sheets don't seem to me to assume so much that designers will be familiar with information presented that way.
EDIT: @InspectorGadget, I didn't read the whole thread carefully before adding my two cents. Looks like we agree on bypass caps.
You cant and shouldnt. this board has only a single clock/single speed and no clock buffer. amanero expects a clock to match whatever samplerate its playing. you would be better off running amanero as master, but its clk isnt isolated. so just leave amanero as its own master and leave the 9038 running on its own clock.
If using Amanero, I would suggest to use the power supply modification to run it from clean 5v power rather than USB power. Its clock jitter will be less, and regardless of ESS marketing claims, the ASRC inside the dac chip should sound better if there is less incoming jitter, especially so if you get into the registers and reduce DPLL_Bandwidth.
Sitime low jitter clocks are not ideal for audio because of higher close-in phase noise than good crystal clocks. NDK SDA are still very good if used appropriately.
EDIT: Also, the I2S interconnections between the Amanero and the dac board in the pic are not ideal for carrying I2S signals. Better to keep the wires shorter, and twist each I2S signal wire loosely with its associated ground wire. Keep some space between the twisted pairs so they don't couple with each other more than necessary.
EDIT: Also, the I2S interconnections between the Amanero and the dac board in the pic are not ideal for carrying I2S signals. Better to keep the wires shorter, and twist each I2S signal wire loosely with its associated ground wire. Keep some space between the twisted pairs so they don't couple with each other more than necessary.
Last edited:
A 'standard' grade clock isn't so good for audio either. NDK SDA are the lowest cost audio clocks I am aware of that I would consider good quality.
For NDK SDA, diyinhk is charges about $8 each, which IMHO is a reasonable price. Also, those clocks may be hard to get at some point when existing stocks run low. According to one dac manufacturer who uses them, NDK SDA aren't being made right now due the AKM factory fire.
NDK NZ2520SDA 3.3V 49.152Mhz 45.1584Mhz 24.576Mhz 22.5792Mhz Ultra low phase noise oscillator - DIYINHK
For NDK SDA, diyinhk is charges about $8 each, which IMHO is a reasonable price. Also, those clocks may be hard to get at some point when existing stocks run low. According to one dac manufacturer who uses them, NDK SDA aren't being made right now due the AKM factory fire.
NDK NZ2520SDA 3.3V 49.152Mhz 45.1584Mhz 24.576Mhz 22.5792Mhz Ultra low phase noise oscillator - DIYINHK