I like this idea as AD1860 is in the same ballpark for pricing as PCM56. I feel sure 3 in parallel won't better one AD1862 based on the noise specs for those chips, but here (on Taobao) AD1860 is much cheaper than a third the price of AD1862, perhaps because its not a fashionable part at the moment. Thanks for the suggestion.AD1860R-J is 3 times cheaper than 1862-N . It can be purchased new from reliable sites.
It must be tested if 2 or 3 such chips in parallel can be heard as well or better than a single Ad1862-N.
Wonderful question/answer/comments here Richard, looking forward to developments from your 'new' shift of direction. 👍
Having used DAC's with both the above AD chips, the AD1862 is subjectively better in my listening. I've not heard a paralleled AD1860 so I can't comment on that.
Having used DAC's with both the above AD chips, the AD1862 is subjectively better in my listening. I've not heard a paralleled AD1860 so I can't comment on that.
Last edited:
Thanks for your listening observation Vunce - my current hypothesis is the SNR is the best determinant for SQ for DAC chips.
AD1860 has a typical 105dB SNR unweighted whereas AD1862 clocks in at 116dB. So I reckon we'd need about 13 AD1860s paralleled to get the SNR down to where AD1862 is. That's rather a lot of PCB real-estate and current consumption.
AD1860 has a typical 105dB SNR unweighted whereas AD1862 clocks in at 116dB. So I reckon we'd need about 13 AD1860s paralleled to get the SNR down to where AD1862 is. That's rather a lot of PCB real-estate and current consumption.
Although cost is a factor, it is perhaps time to bite the bullet and to move to a more expensive DAC chip straight away, one that is future proof and with already good specs to start with in single form... at least for the flagship?
That doesn't mean you can't even parallel these, but at least you should have very good sonic results (based on all your experience with the rest of the circuit) with one chip, without to have to accept the chip(s) as limiting factor...
Just thinking out loud... AD1862 could be the way forward...
Claude
That doesn't mean you can't even parallel these, but at least you should have very good sonic results (based on all your experience with the rest of the circuit) with one chip, without to have to accept the chip(s) as limiting factor...
Just thinking out loud... AD1862 could be the way forward...
Claude
Given that the spec I'm currently chasing is SNR, AD1862 doesn't have the best datasheet figure for that, PCM58 does. However I'm doubtful that PCM58's DS figure can actually be correct. But is it sufficiently in error to mean in practice AD1862 wins in the SNR stakes? I'm doubtful.
Interesting talk. I've got some PCM56 here to play with and just yesterday got delivered a pair of PCM58 'K'. My day to day DACs are Dorati and Miro's AD1862. Both in system for different duties. Between Richard and Miro I have all the DAC projects covered, present and hopefully future! Looking forward to the IV
Well, perhaps it is worth comparing the sound quality of both PCM58 and AD1862 (if it can be done reasonably and with similar layouts) before chosing thr chip?
I understand specs are important, but as at the end you beloved half will have the last say based on her ears, why not getting started on the right foot 🙂
I understand specs are important, but as at the end you beloved half will have the last say based on her ears, why not getting started on the right foot 🙂
Can you even purchase new PCM58 chips? Anyway, I have one of your new Dark I/V (catchy name) boards that I will get connected to a special Miro AD1862 boards (without opamps) today. I will report back with my subjective results as soon as I can. So don't give up on the AD1862 just yet.
Interesting Jim... how do you perceive Miro's AD1862 vs the Dorati on every day's basis?
Understanding not apples with apples but well...
Understanding not apples with apples but well...
Bought the 58s from a reputable UK seller with old military stock.
The Dorati is fed from Rpi/FIFO, and Salas regs.
1862 just fed from Miro PSU1 albeit with Ti Tpsa7 regs. And a simple AK Digi receiver. Hard for me to portray subjective listening differences. I couldn't call either way which I prefer.
The Dorati is fed from Rpi/FIFO, and Salas regs.
1862 just fed from Miro PSU1 albeit with Ti Tpsa7 regs. And a simple AK Digi receiver. Hard for me to portray subjective listening differences. I couldn't call either way which I prefer.
I got my second miro AD1862 board going today using your new dark LED I/V board. At startup it was a little harsh on the highs and a little thin, but had a lot of detail. I do not critically listen for a few days on anything because everything seems to break in with time. After a few hours of playing, it is smoothing out, the bass is getting fuller and it is really starting to sound good. It is getting better every hour. I will give it a few more days, but I also need a different transformer that I ordered to arrive, which will give the I/V PS extra headroom. Right now I am using a left over 16 VAC transformer. My thoughts are that a new DAC chip, along with all the caps, ect. need time to break in, but even so, I think maybe the biggest difference in sound comes from the PS and I/V stage.
Elcaps certainly take time for their dielectric to form and as a result their leakage current to reduce. According to various DSs I've read, DAC chips take a minute or two to warm up and deliver their rated specs but I don't know about longer-term. Glad to hear there is progress in the SQ!
The biggest difference in SQ comes from PS and I/V stage I would agree, but once those two are sufficiently optimized the differences between various DAC chips becomes readily apparent.
@ClaudeG - yes it is about subjective experience ultimately. We've done some initial listens to PCM58 vs AD1862 but I don't place much emphasis on how that turns out until we have more experiences over time.
The biggest difference in SQ comes from PS and I/V stage I would agree, but once those two are sufficiently optimized the differences between various DAC chips becomes readily apparent.
@ClaudeG - yes it is about subjective experience ultimately. We've done some initial listens to PCM58 vs AD1862 but I don't place much emphasis on how that turns out until we have more experiences over time.
PCM58 board on my bench and starting to be stuffed. This one isn't a complete DAC, it's Iout designed to be used with 'Dark LED' filter-I/V stage. I've added extra logic chips to synchronize the L and R channels which means this board can be used in balanced mode (with the I2S splitter) too, if so desired.
Excellent!
Do you plan a similar board with AD1862, so you could compare just swapping conversion boards?
Have fun!
Claude
Do you plan a similar board with AD1862, so you could compare just swapping conversion boards?
Have fun!
Claude
The comparison you mention has already been done with DAC chips lashed up on perfboards, just not quad versions, only dual. No plans for a quad AD1862 myself but one of the punters here has already made one, perhaps he'll share that at an appropriate time 🙂
If the chips are parallel will you run a special 2mA version of Dark IV.?
How do you find sourcing of PCM58? Many fakes to look out for?
How do you find sourcing of PCM58? Many fakes to look out for?
Hi Jim - Dark LED in its standard form can most probably cope with 2mA, just I haven't publicized that so far as I don't know that it'll work in all cases and at all frequencies. The 1mA rating for it is because I'm a conservative engineer 🙂 With PCM56 for example, its tolerance on output current is +/-30% so a nominal 1mA DAC chip can in theory output 1.3mA. Two of those in parallel (2.6mA) will be very likely to clip Dark LED at some frequencies. PCM58 though has its output current very tightly specified (2%).
Wifey's sourced the PCM58s so far, we've gone for 2nd hand ones. I don't know if I've gotten fakes but they're sounding rather nice....
Wifey's sourced the PCM58s so far, we've gone for 2nd hand ones. I don't know if I've gotten fakes but they're sounding rather nice....
I see , that's interesting on the tolerances, thanks. I got a couple 58s from an ex military gear UK seller but have also found a Chinese seller with them for a good price. Maybe too good!
https://www.ebay.co.uk/itm/33204630...aV1Cg3-Rh-&var=&widget_ver=artemis&media=COPY
https://www.ebay.co.uk/itm/33204630...aV1Cg3-Rh-&var=&widget_ver=artemis&media=COPY
Attachments
3 quid for a new PCM58, K grade at that? That has to be too good to be true I reckon. But if not, you certainly have a bargain there!
- Home
- Source & Line
- Digital Line Level
- lingDAC - cost effective RBCD multibit DAC design