This project is about the ESS Sabre ES2093 DAC chip. This one is the least expensive DAC chip from the whole line up of DAC’s manufactured by ESS Technology. What i wanted to do here is to make something like an evaluation board which can be almost universal.
ESS Technology might not be quite a famous among audiophiles or electronics fans. They have this kinda odd marketing policy that puts them on the edge of OEM and retail markets. However thy are manufacturing some very high specs chips (DAC’s and ADC’s) and supply them to many manufacturers of high quality AV components like OPPO and many others.
The famous Buffalo DAC is built around an ESS 32-bit DAC. More info on this one can be found here:
Of course with this project I’m not aiming on beating a complete high-end DAC’s like the Buffalo. What I want is to have this board as an upgrade to any low-end delta-sigma DAC. Besides the fact that it is a low-cost chip does not mean that it’s not worth a try. I believe that a proper designed board and a good power supply can make a low-cost DAC like the ES9023 sound very good. The history knows some good examples of CD players using low-cost DAC’s but sounding very good indeed.
More info about the ES9023 DAC chip can be found here:
Unfortunately this is just a product brief but it gives some basic information. Due to the ESS Technology policy I can post the complete technical datasheet. If you want to have it you would need to contact some of the local ESS Technology distributors.
As you can see this little DAC offers a complete solution. It includes digital filtering, DAC, and output stage. However I’m including a buffer option on the PCB as I don’t really know the driving capabilities of the DAC alone. Some fancy cables actually have huge parasitic capacitance and can make the life hard for the little DAC.
So this is what the schematic looks like:
You might note that the DAC is powered by 3.6V (and not the standard 3.3V). This is due to the output swing hitting the supply rails. The chip has an internal charge pump that generates the negative rail so that the output signal is centered around the 0V (no coupling capacitor required). The manufacturer had increased the supply voltage to 3.6V to prevent the output stage from clipping.
Few words about the power supply circuit. The output buffer formed bu IC6 and IC7 is powered by a simple voltage regulators 7815 and 7915. Nothing interesting here. The DAC chip is powered by a low noise low dropout voltage regulator with a series pass element. Q2 and IC4 form a voltage reference. Its output is filtered by a low-pass filter formed by R2 and C6 and fed to the non-inverting pin of a low-noise opamp IC5. The output voltage is set by the two feedback resistors R3 and R4. C7 is placed across the opamp to improve stability.
The overall noise performance of this regulator is outstanding. The opamp’s noise is critical. For lower cost version you can use NE5534 but some additional compensation is required. Don’t forget to include a 33-68pF capacitor between pins 5 and 8.
I have included an input for an external clock source. This is implemented with one SMB connector. The solder joint SJ1 serves as a switch point. If you are using the external clock SJ1 must be desoldered.
The jumper JP1 is there to set the input format. It can be either left justified or Philips – I2S. The JP1 settings a re the following:
JP1 – CLOSED – I2S
JP1 – OPEN – LEFT JUSTIFIED
The output buffer stage is formed by two unity gain opamps. Those are subject to a personal taste of course. I have made the PCB with sockets to make it easier to switch the opamps.
This is how the pcb looks like:
The PCB has a small prototyping area just in case. Many modifications might be needed so its always a good thing to have some place for additional components.
I’m including a complete project manufacturing files available for download HERE
EAGLE BRD file – HERE
Update 09.12.2012 – I have made some changes to design. Revision B is now available for download.
Update 22.03.2013 – Added EAGLE *.brd files for downloading
THIS PROJECT IS AVAILABLE FOR PERSONAL USE ONLY. COMMERCIAL USE OR MANUFACTURING PCB’S FOR SALE IS NOT ALLOWED.
Why did you use opamps as output buffer, the ES9023 already has a buffer stage build in, it can even drive headphones without any problem!
Do you also have the eagle files for download?
The opamps are just optional. There are unbuffered outputs as well. Basically I don’t trust anything that is powered by +3.6V to have enough driving capability especially with high capacitance cables.
Only gerber files and pdf files are available online. You need to email me for the pdf files.
Chip specsheet says its outputs can be attenuated by external resistors, how exactly it can be done?
Hello. Thanks for your question. I asked myself the same thing. It turned out that in the early version of the DAC – ES9022 there had been a signal clipping at the output due to the low supply voltage(3.3V). In this version(ES9023) they increased the supply rating to 3.6V to eliminate the clipping problem. However the attenuating resistor should be used only when the DAC is supplied with 3.3V. This is to limit the output voltage swing and prevent the output from clipping, but not to be used as a volume control etc.
Managed to obtain the datasheet meanwhile. Have you (or someone) ever tried to connect pot in parallel with C13? My thoughts are: pin 6 is marked as Vref output (for cap buffering) but they showed how to pull it down it to lower output for 3.3V supply. Moreover, chart on p.10 of manual states this resistor MUST be added for both 3.3 and 3.6 supplies to avoid clipping, and it should be even less with higher supply! This implies Vref is not regulated so likely any voltage reference can be set for the DAC and therefore could do the trick.
C16 and C15 look huge at 470n. Do you have details of the internal resistor and the -3dB point of the post filter? The ODAC has an external 220R and 2200p to ground. I increased the resistance to 1k. This gave <1dB reduction at 20 kHz but a big increase in clarity and smoothness when used with my O2.
Hi, the external filter values are the suggested in the datasheet. The two caps are 1210 package. That’s because those are film capacitors:
Hi. You are quite right. The schematics shows the wrong value. Those really must be 4.7nF
I’ll make the necessary corrections.
Wills Sabre DAC suggests that 4n7 is the correct value when used with 220R. For the same bandwidth this suggest that the internal resistance is 25R. I think the internal amp is going to be working very hard driving 25R in series with 470n!
I misread the bands on the resistors. They are 22R.
Things are getting clearer now. Do you know the output impedance of the DAC? I have a feeling that I read it was around 200R somewhere.
Have you had time to evaluate the sound of it? How is it?
Looks like an interesting project to try out.
I have been playing with mods to the Hifimediy UAE23 Es9023 setup. I still get clipping if I do not run the PC gain at around 68%. Does the Vreg pin resistor and cap to ground impact the buffer or would it help me with the digital ripping/clipping I hear when the PC gain slider is up? ODAC uses 118K there and I think 10UF. I confess to adding 3.3V regulator for the DAC IC since I have no 3.6. ODAC uses ~22Uf on the NEG pin. I use Fox 924B 3.3V TCXO on pin 37 of TE7022L USB receiver driving the ES9023. I have made many tweaks and get wonderful sound! But only with the PC gain slider down some. Even with 100K and 10uF on the Vreg pin to ground, I get some crackle if I don’t back off the slider some. Thank you for sharing Eagle files, I would like to retrofit ES9023 into a few DAC and need to make a board.
Have you checked with the oscilloscope at what level the clipping occurs?
Very interesting project.
Any chance you could design a board with balanced outputs?
Hi. There’s no problem in designing a board with balanced outputs. You could also use the board described here:
Hi. I have an idea of how to implement two of these DAC’s in dual differential mode. This is a subject to experimentation before releasing any schematic online.
I’d love to have a nicely implemented balanced, using 2x ES9023, board with I2S input (using u.fl connectors) and pads for outboard IV stage. This design seems very nice, but only has SE output http://www.diyaudio.com/forums/digital-line-level/151846-anybody-using-new-ess-vout-dac-es9022-51.html#post3358436
Awesome. Can’t wait to see your new implement. 🙂
Oh, I forgot one more thing, did you take a look at TI ‘s new lowest noise regulator, the tps7a4700? I suspect it would a sweet improvement from the current 7815 and 7915.
the EAGLE files will be shared shortly.
I haven’t seen the NJM regulators. However you should be careful with those. I’ve seen some Japanese 78x and 79x regulators that don’t use the standard pinout.
Can I get the Eagle files? I want to try placing the Joachim Gerhard filter before an ouput op amp like LME49710 or LMe49720
BTW have you seen the NJM7815FA and NJM7915FA? And for 3.6V a simple Micrel 5205? 3.6V
Eagle files added.
Nice design, tried to find your email adress to contact you.. is that possible? Where did you find the ES9023 IC, I can only have it by âying 20euro of shipping & handling!
All the best
me personal email is firstname.lastname@example.org
I was able to obtain some chips trough a private channel. However I believe that it’s way easier to use the alternative provided by TI’s DAC.
thanks for publishing this design. I’ve had the boards made and am going to have a play. Do you have a BOM available to save me clicking through the Eagle BRD file to pull out component values?
I’ve just found the ‘export partlist’ function in Eagle so have my BOM. Thanks.
That’s why complain about bad sound or oscillation with es9023. poor pcb …
I realize that this is a bit old, but do you know if an external digital filter chip can be used with an ESS DAC?
I would like to try using a PMD-100.
Great post! Have nice day ! 🙂 dybas
Hello! Impressive design and good work!
you can send a list of components on the board, plese
Why did you use Q2 + IC4? I guess it equal 2.5V + 0.6V = 3.1V. But it was the worst temperature drift. Why can’t use only IC4? Thank you from RUS )
Hey! I stumbled upon your project and it’s very nice!
I myself would like to include this DAC in my project, but had some issues with the audio being distorted a lot, and the IC probably not working correctly.
I wanted to try and explore your project files and figure out my issue, but your download link no longer works. Would you be able to add a new download link by any chance?
Thanks a lot! Have a nice day!
I really need project files. Could you send them to me? The ess_technology.lbr file is really needed.