Hi,
I have a MiniDSP U-DAC8 in my spares bin, every time I think about doing something with it, I google it, the Amir testing comes up and rinse-repeat lol.
Has anyone here tried to improve it?
Fyi, I'm comfortable hand soldering SMT, I've ordered BOMs from DigiKey and PCBs from Chinese fabs, so I'm not a complete newbie.
Cheers
Richard
I have a MiniDSP U-DAC8 in my spares bin, every time I think about doing something with it, I google it, the Amir testing comes up and rinse-repeat lol.
Has anyone here tried to improve it?
Fyi, I'm comfortable hand soldering SMT, I've ordered BOMs from DigiKey and PCBs from Chinese fabs, so I'm not a complete newbie.
Cheers
Richard
What info do you have on it? A schematic? Hi-res pics of both sides of the board so we can read part numbers?
We have sorta improved other dacs, but it can be a lot of work. Depends on how much "improvement" you want.
From a quick look around, U-DAC8 is probably not ever going to be very good. It was designed to be low-cost and have eight channels at that.
We have sorta improved other dacs, but it can be a lot of work. Depends on how much "improvement" you want.
From a quick look around, U-DAC8 is probably not ever going to be very good. It was designed to be low-cost and have eight channels at that.
Thanks, definitely not looking to made it a great DAC, just correct any daft design decisions that they used to get the price down.
I found these online, I'll take some better photos of the front/back of my PCB in the morning
I found these online, I'll take some better photos of the front/back of my PCB in the morning
I am also building my own 8 channel dac with independent ak4493 cards, https://fr.aliexpress.com/item/1005...t_main.28.26575e5bDiZjLP&gatewayAdapt=glo2fra
With 2 different xmos module, https://www.diyinhk.com/shop/audio-...high-quality-usb-tofrom-i2sdsd-spdif-pcb.html
https://docs.pawpaw.ltd/docs/docs/PXUA-XU316-KIT/PXUA-XU316-KIT_usage/
The best possible improvement must be in the power supply at the DAC chip level, which must be as clean as possible and with very low noise.
With 2 different xmos module, https://www.diyinhk.com/shop/audio-...high-quality-usb-tofrom-i2sdsd-spdif-pcb.html
https://docs.pawpaw.ltd/docs/docs/PXUA-XU316-KIT/PXUA-XU316-KIT_usage/
The best possible improvement must be in the power supply at the DAC chip level, which must be as clean as possible and with very low noise.
AK4440 needs only a couple of external capacitors so there is very little to correct. You could try better power supply but not much else. Also ASR review fails to mention that most measurements are actually within (or close to) datasheet specifications. Only big anomaly is the frequency dependent distortion.
Yes, I think, it's just assembly and small modifications to the XMOS firmware if necessary.
Thing about most low-cost dacs is there are not usually much in the way of daft design decisions. The things are carefully designed to save every penny by making every part equally bad. Its not like the designer was stupid and just changing one little part would make it sound a lot better.
Even aside from the odd spike at 25hz, THD at all frequencies is far higher than AK4440 datasheet spec. But datasheet spec is with RL>5k, and I don't know what load Amir used for that test. If the load was heavier, then the increase in THD makes sense. With heavier load, the frequency dependence seems logical for an on-chip op amp not designed for high current: thermal distortion at LF and inadequate loop gain at HF.
More or less agree...
They simply copied the AKM evaluation card...
With low noise LT1763 regulators.
I built another Chinese dac...Dual AK4493.
And it sounds wonderful...
They simply copied the AKM evaluation card...
With low noise LT1763 regulators.
I built another Chinese dac...Dual AK4493.
And it sounds wonderful...
Not exactly. AK4440 datasheet does not specify THD but THD+N only. THD+N is about 6dB worse than datasheet maximum but that may be partly due to high noise. But e.g. DR and filter stop band attenuation are according to datasheet.
Ah, I see you are correct, I read typical THD+N numbers from datasheet, not max. It seems that's where miniDSP gets their datasheet spec from. But it's still 6dB worse around 1k and significantly worse elsewhere, while datasheet spec is full audio bandwidth. As far as picking out D from N, adding up from FFT graphic: H3 -83 & H2 -86 dB gives -78.35dB. And as you note, the other specs are met spot on, which leads me to think it's related to loading or temperature.
I thought about the loading as well as that would explain the results but the 1st graph on ASR review seems to indicate that ADC input impedance was 100kohms. But the evaluation board measurements clearly show that the dac chip is capable of reaching datasheet levels so something is wrong either in the implementation or in the measurement setup.