Hello everyone,
I am looking for an all-in-one 8-channel DAC solution for a low-power application. Four pairs of channels will be fed from the DAC to stereo amplifier boards with <15 W of RMS power @ 8 ohms.
The only product I have found so far is the following:
8 Channels 384kHz 32bit ES9016 PCM DXD DSD Audio DAC - DIYINHK
However, I don't need such a high resolution DAC—96 to 192 kHz and 16 to 24 bit is sufficient.
Sound quality is still important, but I am hoping that someone knows of a cheaper DAC + board setup. I need USB-C (Windows compatible) to 8 channels of analog line-out.
Thank you!
I am looking for an all-in-one 8-channel DAC solution for a low-power application. Four pairs of channels will be fed from the DAC to stereo amplifier boards with <15 W of RMS power @ 8 ohms.
The only product I have found so far is the following:
8 Channels 384kHz 32bit ES9016 PCM DXD DSD Audio DAC - DIYINHK
However, I don't need such a high resolution DAC—96 to 192 kHz and 16 to 24 bit is sufficient.
Sound quality is still important, but I am hoping that someone knows of a cheaper DAC + board setup. I need USB-C (Windows compatible) to 8 channels of analog line-out.
Thank you!
you want cheaper than DIYINHK? this is what is driving bad design in audio. DIYINHK already do not meet the specifications they claim, due to cut corners and less than optimal PCB design. also, we are talking about a DAC here, your 15W spec is what you will ask from your amplifiers and has nothing to do with the dac.
Thank you for your input. Do you think it's not wise to order from DIYINHK? What other online stores or products would you recommend?
I am a complete beginner when it comes to DIY audio, so I included the power requirements in case it was a factor to consider.
There are not many/any other options really for multichannel dacs if you are just talking bare PCB. the DIYINHK is what it is and they are priced very low. they appear to do the job, but dont count on them to match the spec they quote. there are many options if you include soundcards and recording interfaces for proaudio, but if you want something cheaper, then good luck with that..
may I ask where I can verify this, I have this board so I am very interested. Have any measurements been done?
every single board they do ... They just take the headline datasheet numbers for the parts and say thats what their board is capable of. ASR has some measurements and i'd say there are some here too. when a board looks like this and quotes exactly datasheet performance, but doesnt have any measurements on their site, you know they never made any. Talking from experience, you dont match datasheet performance of -115-> -120dB THD+N while cutting corners in most parts on the board, without making any measurements.
The cheapest 8ch is one of the many USB1.1 highspeed devices from amazon/ebay/... Of course only 16bit, 8ch only at 44.1/48kHz.
https://www.ebay.com/itm/353790400940https://www.amazon.com/StarTech-com-7-1-USB-Sound-Card/dp/B002LM0U2S
https://www.ebay.com/itm/353790400940https://www.amazon.com/StarTech-com-7-1-USB-Sound-Card/dp/B002LM0U2S
Agree with what the other guys have said. Its not that easy to make a good 8-channel dac, one that meets published dac chip specs. Its even harder to make it subjectively sound great while still measuring well. IME it can't be done for super cheap. What can be done at lowest cost are devices like phofman suggests; they are basically very low cost dac circuits similar to what might found built into a very low cost PC motherboard (maybe one from several years ago).
Subjective impression of sound doesn't need measurement or higher budget. It's all in one's head. That's why it's calle subjective.
For little more than DIYINHK 8 channel DAC kit you can get ESI Gigaport eX 8 channel DAC https://www.thomann.de/gb/esi_gigaport_ex.htm
what is ready to run device.
what is ready to run device.
I thought this was a DIY audio forum. I really don't understand people bashing the DIYINHK DACs. They don't claim any specs at all, and most of the important stuff is left op to the diy-er: power supplies, separate IV stages, you name it. The layout of the 4 layer board is fine, a solid ground and short traces where it matters. I'm using it with Twisted Pear IVY-III stages. Dead silent, and great sound. The ESI you recommend is very noisy, though.
I find myself spending less and less time reading diyAudio, very opinionated posters that do claims without any data to back it up.
I'm in no way associated with DIYINHK, but must say they have great support and deliver what they claim, no more, no less.
It's up to the diy enthousiast to do the implementation.
InspectorGadget's post are a clear example of opinionated posts: there are no measurements of this DAC here, nor on the ASR forum.
If you have nothing to say from experience or measurements, don't post.
ha!!! Thats kinda hypocritical dont you think? Do you want to know who makes baseless attacks? I couldnt be bothered the other day, because it should have been very easy for someone to do it themselves. I just now literally typed 'DIYINHK 9016 DAC measurements' into duck duck go and the first hit is ... drum roll .....
https://www.audiosciencereview.com/forum/index.php?threads/diyinhk-dac-measurements.23629/
These for 9016 are the BEST measurements ive seen from one of their boards. They are not 100% fair perhaps, because a MOTU was used for the measurement and that will impact noise performance, so keep that in mind when looking at anything that is impacted by noise. like SNR or THD+N. pure THD should not be overly effected. Selling a DIY dac and not providing reference measurements, so that your customer knows if theyve implemented it properly, is lazy at best.
ive seen their designs for years and initially yes, they were 2 layer monstrocities, but they arent so much better after going to 4 and they very much did used to claim datasheet performance, word for word. Perhaps they stopped after they were shown to be wrong, or because of websites like ASR existing. The IV stage and clock layout is of particular note, but i'm not here to rag; it was just a comment to be wary of what you expect from a cheap dac board. they also used to claim that you could just drop in the 9038PRO on their earlier 9018pro boards ... on the page linked above, you can read how the (more expensive) 9038PRO board measured considerably worse than the 9016 board, with each channel measuring differently and that had nothing to do with his measurement setup and probably much more to do with the layout and the amount of current the 9038 puts out (and thus how much the IV stage opamps have to deal with) and them not doing anything to account for that ... I could go on and list whats wrong with the layout, but I wont.
Now, the 9016 problems are not likely to be audible, unless you really push it wrt your gain structure, or your digital crossover settings. The 9038, not so much.
https://www.audiosciencereview.com/forum/index.php?threads/diyinhk-dac-measurements.23629/
These for 9016 are the BEST measurements ive seen from one of their boards. They are not 100% fair perhaps, because a MOTU was used for the measurement and that will impact noise performance, so keep that in mind when looking at anything that is impacted by noise. like SNR or THD+N. pure THD should not be overly effected. Selling a DIY dac and not providing reference measurements, so that your customer knows if theyve implemented it properly, is lazy at best.
ive seen their designs for years and initially yes, they were 2 layer monstrocities, but they arent so much better after going to 4 and they very much did used to claim datasheet performance, word for word. Perhaps they stopped after they were shown to be wrong, or because of websites like ASR existing. The IV stage and clock layout is of particular note, but i'm not here to rag; it was just a comment to be wary of what you expect from a cheap dac board. they also used to claim that you could just drop in the 9038PRO on their earlier 9018pro boards ... on the page linked above, you can read how the (more expensive) 9038PRO board measured considerably worse than the 9016 board, with each channel measuring differently and that had nothing to do with his measurement setup and probably much more to do with the layout and the amount of current the 9038 puts out (and thus how much the IV stage opamps have to deal with) and them not doing anything to account for that ... I could go on and list whats wrong with the layout, but I wont.
Now, the 9016 problems are not likely to be audible, unless you really push it wrt your gain structure, or your digital crossover settings. The 9038, not so much.
Last edited:
I'll mention one, in the hope that DIYINHK may take note. in my experience the most significant thing with ESS dacs (and it is mentioned in the app notes too), is to make sure you have a direct, low impedance connection between AVCC supplies, the DAC and the IV stage. Not having at least the AVCC regulators on the DAC PCB (there are no regulators on the PCB at all) will be the biggest contributor. You can see how mdsimon2 shortened the wiring on his AVCC regs and this improved performance every time he did so. This is plain as day in the ESS app notes and datasheets, stretching back to the ES9018S, yet they either ignored it, or didnt read it. IV stage and AVCC regs should be on the DAC PCB. an example of how you can improve performance without any additional cost whatsoever, just layout.
Probably they—rather cynically—removed the analogue regs from the board, because modders usually replaced them with something decent; so they just didn't put them on at all and sold regs separately, instead of putting something decent on there themselves. Their lt3045 boards are not perfect examples of that layout either, but I digress. Because they either don't do (or just don't publish) measurements, there was no way for people to see they had made datasheet level performance impossible with this move.
Funnily enough, I hadnt spotted it before, but the only (tiny) remaining regulator on the board is the least important, for digital logic supplies. in other words, the only reg that probably could have been left off the board, was left on there and the ones that should have stayed, were removed. I guess its also a way for them to make a bit more bank from each sale, by selling reg boards separately, i'm sure they can make a bit more profit than selling a complete board. with all the bits needed, its actually not that cheap and you would IMO do better buying a motu M4 or similar pro interface (depending how many channels you actually need). Pretty sure you could get something rather nice on the used market for that money.
Yes @Markw4 , we certainly don't disagree all the time
On that note. I havent got the board, so i'm just guessing these problems are not audible, but there are certainly situations where it could be.
Funnily enough, I hadnt spotted it before, but the only (tiny) remaining regulator on the board is the least important, for digital logic supplies. in other words, the only reg that probably could have been left off the board, was left on there and the ones that should have stayed, were removed. I guess its also a way for them to make a bit more bank from each sale, by selling reg boards separately, i'm sure they can make a bit more profit than selling a complete board. with all the bits needed, its actually not that cheap and you would IMO do better buying a motu M4 or similar pro interface (depending how many channels you actually need). Pretty sure you could get something rather nice on the used market for that money.
Yes @Markw4 , we certainly don't disagree all the time
On that note. I havent got the board, so i'm just guessing these problems are not audible, but there are certainly situations where it could be.
Well aware of these posts, I have made comments there and had a personal conversation with the author, Michael, after he posted. So no hypocrisy here, I did my work. The post you referenced is about the ES9016, though, not the ES9038. No measurements from this board anywhere, just some preliminary data and him saying he is so satisfied with the ES9016 that there is no reason to go to the ES9038. Michael told me he did some more -unpublished- measurements on the ES9038, and it was - with some more care - easily measuring better SN than the ES9016. But this is besides the point: the ES9038 implementation was criticized without proper data.
I don't understand your criticism about the modular approach DIYINHK uses; it's ideal for the DIY-er we are supposed to be here. Proper implementation (short low impedance wiring, etc) is what DIY is all about IMHO.
But I should not have singled you out, for that I apologize. My disappointment on the quality of the posts and the level of prejudice and non-informed opinions got to me. Especially when my precious dac under attack.