D10S still takes the cake for worst sounding DAC I've heard. mp3 players, various mobile devices (laptop, phones) with the most basic DS sound chips, all diy projects and a few other commercial DACs (including topping D50) ... all clearly better in the fundamental areas.
I suppose slightly it is clearer and 'hi fi' sounding than your average DAC but sounds completely unnatural and rapidly induces fatigue, more than likely it is fake 'detail'.
small, battery powered devices always had huge leg up in listenability, musicality. Throw any DAC at all into an isolated, very low noise environment and you'll get decent results it seems.
Ernest, I‘m hoping to get some clarification regarding your MSB-trim distortion reduction test results. I presume that your chart shows the DAC’s distortion AFTER the nearly 20dB of reduction due to MSB trimming. So, the AD1862N un-trimmed distortion was nearly 20dB higher. Do I have that correct? Thanks.
Then must be something wrong with the chip, or the distortion analyser.
When using K grade, we have had practically no gain in trimming.
You can compare K grade spec to yours (whatever grade they are) in datasheet.
And only changes H3,5,7, ... etc.
Cheers,
Patrick
It can be improved a lot by external power supply and inserting Sparkos SS3602 instead of LME49720. But the investment exceeds the value of the original device, and again it is not at the level of, say, the very simple Pavouk AD1865 DAC+ JLS I2SoverUSB.
Let's say what can be done for around 250eur without the box. I packed this as is for fun, just to hear if it plays well. Pavouk AD1865 with I2S input, JLS I2S over USB board, AK4113 SPDIF/I2S board from ebay, power supply with two LM1117T-5.0 for JLS and AK. So I have two SPDIF inputs, TOSLINK and USB inputs and a small display that shows the active input, frequency and bitrate.
The AK board has an I2S input that is selectable by a button (key 1), and I2S output. SPDIF/TOSLINK inputs are selected automatically according to which input is active. The board should be able to be set so that the SPDIF/TOSLINK inputs are selected with a different button (key 2), but I haven't tried it. Powered by +5V - 50mA.
https://www.ebay.com/itm/284351999799
The sound is not at the level of Miro DACs, but the price is really low and it can be listened to without getting annoyed.
I also got a Bluetooth 5.0 module on ebay with I2S output, to try how it sounds (I will install it instead of JLS). The plan is to build a cheap Bluetooth only DAC (or AK4113+Bluetooth 5.0) with the Pavouk AD1865 DAC board.
https://www.ebay.com/itm/284170951663
The AK board has an I2S input that is selectable by a button (key 1), and I2S output. SPDIF/TOSLINK inputs are selected automatically according to which input is active. The board should be able to be set so that the SPDIF/TOSLINK inputs are selected with a different button (key 2), but I haven't tried it. Powered by +5V - 50mA.
https://www.ebay.com/itm/284351999799
The sound is not at the level of Miro DACs, but the price is really low and it can be listened to without getting annoyed.
I also got a Bluetooth 5.0 module on ebay with I2S output, to try how it sounds (I will install it instead of JLS). The plan is to build a cheap Bluetooth only DAC (or AK4113+Bluetooth 5.0) with the Pavouk AD1865 DAC board.
https://www.ebay.com/itm/284170951663
Attachments
Last edited:
Hi Ken, you are perfectly right. Yet my nearly 20 db for sure was slightly exaggerated. Initially it was slightly more than -90 db and after the procedure it went down to what you see in the attachment. K3 might be at -107, K5 just below -110 db. So probaby "went down by 15 db" would have been more correct, yet I was so happy by what I saw. I will follow up this in the answer to Patrick. Cheers, Ernst
Patrick, a few things I want to mention:
The distortion analyser was REW, interface was Focusrite 2i2. A second try was done with ADI2PRO and it gave roughly the same result.
I did not measure the DAC without the resistors and the trimmer. So the naked DAC might have measured better before.
The chips were used and did not come from rochester - I had ordered them via ebay, as long as they were available well below 10 Euro. I had ordered 4 and 3 of them were working examples, even they were not look promising.
My chips from Rochester I had put into a design, which does not allow trimming. Their look was much more trustworthy.
After the trimming - apart from PSU noise - the DAC is close to the spec in my eyes.
Cheers, Ernst
Last edited:
This is exactly what I observed with some basic measurements earlier in this thread, connecting the trimmer to the pins on DIYinHK board caused a huge increase in distortion until it was adjusted into a certain range, but I had absolutely zero trust in those measurements, it was using line in on PC MoBo and I was figuring out the software as I went along, the distortion was below the noise floor of line in without trimmer so I couldn't get any improvement.
I guess this confirms it was accurate
Trimmer wrongly adjusted will cause 20~30dB more odd-order harmonics easily.
Unless you have proven distortion measuring equipment, best not to solder in the trimmer.
Gain is marginal. Potential detriment huge.
Easiest solution is to buy K grade. 🙂
Patrick
Unless you have proven distortion measuring equipment, best not to solder in the trimmer.
Gain is marginal. Potential detriment huge.
Easiest solution is to buy K grade. 🙂
Patrick
Yes but on Ali you can personalize your dac ic...I got an AD1862 ic with one K; a double crown...and also a Y...
unluckilly one leg was broken from the inside...so I will never know nore you, how is sounding an AD1862 K double crown Y ic....😢
Lesson learned... now I source from Paddy.
unluckilly one leg was broken from the inside...so I will never know nore you, how is sounding an AD1862 K double crown Y ic....😢
Lesson learned... now I source from Paddy.
Okay, thanks. I must say that I'm surprised that there was such a significant improvement for -3dBFS test signal, as the data sheet suggests there would be little to none for a test signal so close to 0dBFS. Do you happen have the distortion improvement charted for a -60dBFS test signal, as prescribed in the data sheet for MSB-trimming? Also, were the test signal samples 16-bits, or 20? Thanks, again.
@diyiggy I'm so sorry, but that's how it is when you buy on Aliexpress/ebay. It happened to me a couple of times for some less expensive things, I protested and got my money back. We had a lot of luck with the AD1865 N-K on ebay. I think we took a total of 6 pieces. They have never been used and play so well that they are certainly not fakes.
I also bought the PCM1794A on ebay, for DDDAC, a lot of pieces, and they were all without error.
I also bought the PCM1794A on ebay, for DDDAC, a lot of pieces, and they were all without error.
To make a fair comparison, you must first remove the trimmer, and measure H2, H3.
Then you install the trimmer and trim to best you can measure.
If trimming can get you 15dB, then it is very poor manufacturing.
Does not happen on BB DACs.
Cheers,
Patrick
Then you install the trimmer and trim to best you can measure.
If trimming can get you 15dB, then it is very poor manufacturing.
Does not happen on BB DACs.
Cheers,
Patrick
Ken, as Patrick pointed out, I should have done a measurement before soldering the trimming circuit in order to do a meaningful comparison. My bad. Yet a good day for learning something new 🙂 The same could be said for the -60 db signal level, which I didn't respect - haven't read - didn't know. Will do next time - either with lifted resistors legs or with a newly built DAC. For me at least the figures and the sound of my DAC are good enough ;-)
About the bit depth. I think, it was 16 bit, as I used the java driver within REW. At this point of time I wasn't aware of FlexAsio. And after my info, the Java driver is just 16 bit capable.
I will have a look into that later this year, when the soldering heat is welcome again in my room.
Cheers, Ernst
At iso voltage output these two ic do not need the same resistors values, hence the difference. On the PSU 1 or 2 you can of choose two different symetric outputs with the rigth resistors values plus the rigth trimming pots.
E.g, : 2 x +5/-5 V ; 1 x -3/-3 V and 1x + 9v and -6V as there is a zero ref and the ref is shared with the load. Symetry or not is up to you and you can feed a non symetric load, each lt19 and lt30 is independant and share just ground with the others pairs.
With the PSUII miro's bom you have a margin of around 1v with the trim iirc. For different voltage see the datasheet of the ics. Is it what you are asking ?
About the Ali ad1862 : it was when the guys sold it at less than 10 euros, Miro's tip, before it was indexex on gold. So on a pair just one worked. No problemo I made a therapy and no nigthmare anymore. Also I had yet a pair I bougth at Rochester long time ago. 🙂. Since, Paddy is my corner dealer.
E.g, : 2 x +5/-5 V ; 1 x -3/-3 V and 1x + 9v and -6V as there is a zero ref and the ref is shared with the load. Symetry or not is up to you and you can feed a non symetric load, each lt19 and lt30 is independant and share just ground with the others pairs.
With the PSUII miro's bom you have a margin of around 1v with the trim iirc. For different voltage see the datasheet of the ics. Is it what you are asking ?
About the Ali ad1862 : it was when the guys sold it at less than 10 euros, Miro's tip, before it was indexex on gold. So on a pair just one worked. No problemo I made a therapy and no nigthmare anymore. Also I had yet a pair I bougth at Rochester long time ago. 🙂. Since, Paddy is my corner dealer.
Last edited: