Yeah I am interested in getting something decent.
I feel the ELAC discovery streamer is better than the oppo but I am sure denafrips and others would be better.
Jeff,
I was referring to your schematic in post #316.
Back to the new schematic Q5 and Q8 probably do no need heat sinking as there is not much of a voltage drop across them.
Looking good.
Jam
I was referring to your schematic in post #316.
Back to the new schematic Q5 and Q8 probably do no need heat sinking as there is not much of a voltage drop across them.
Looking good.
Jam
Pico,
Mark has decreed that you visit us the next time you are in the vicinity......
You will be compensated with required quantity of Vegemite.
All hail Mark! 😀
Jam
Mark has decreed that you visit us the next time you are in the vicinity......
You will be compensated with required quantity of Vegemite.
All hail Mark! 😀
Jam
Attachments
Last edited:
Kind of depends on a number of factors. DIY? Sigma-delta? Ladder dac? Budget? Sound Quality goals? Computer-based DSP?
Mark,
I'd also be very interested in a DAC. I'm quite happy with my Metrum Onyx, but at some point it'll become the only thing in my system which I didn't build.
Jam,
Yeah, note that I don't list a power for Q5 and Q8 on the schematic as it's quite low. I stuck them on the backs of Q6 & Q8's heatsinks because it was a convenient place to put them. Is there any reason not to do that?
Yikes, the feedback wire going to U2 pin 2 got two junction dots where it shouldn't have. I hope I haven't found a Kicad bug....
Pico,
I must look into your suggestions again. I did try loading the VAS, but it was back when I was struggling to get the cascode working and all the other experiments fell by the wayside....
Cheers,
Jeff.
I'd also be very interested in a DAC. I'm quite happy with my Metrum Onyx, but at some point it'll become the only thing in my system which I didn't build.
Jam,
Yeah, note that I don't list a power for Q5 and Q8 on the schematic as it's quite low. I stuck them on the backs of Q6 & Q8's heatsinks because it was a convenient place to put them. Is there any reason not to do that?
Yikes, the feedback wire going to U2 pin 2 got two junction dots where it shouldn't have. I hope I haven't found a Kicad bug....
Pico,
I must look into your suggestions again. I did try loading the VAS, but it was back when I was struggling to get the cascode working and all the other experiments fell by the wayside....
Cheers,
Jeff.
It would be good to find out if your and Pico's DAC interests point in the same direction. Budget too. If you go very low level DIY, say to PCB design (4-layer minimum) there is RF to deal with. Clock stability (jitter) can matter a lot with some types of DACs, less of a concern with other types. Lowest distortion we can do today is around -120dB to -125dB, somewhere close or within that range. It requires very careful design and layout to get down to those numbers. Also, even DIY, it can get sort of expensive to do well (say, ~$1k+). Cheap can be done too, but sound quality will suffer in the range of some to a lot depending on particulars.
Maybe Jam would be willing to express some feelings about DACs, we listen to some various ones here and he has good ears for what people tend to like (in perceptual terms). Then again, maybe we should take this subject over to one of the digital forums (digital line level).
Last edited:
I like the idea of NOS ladder DACs better (because I can understand them). I don't actually know if I like their sound better as I've never owned a good delta-sigma. I should probably pick up a Schiit Bifrost 4490 to compare first....
As for budget sort of expensive is fine with me.
As for budget sort of expensive is fine with me.
Yeah I like the idea of ladder dacs like Denafrips and others.
Chords implementation of fpga really interests me from a technical standpoint, I have no idea if sonically this is the right approach though.
Chords implementation of fpga really interests me from a technical standpoint, I have no idea if sonically this is the right approach though.
Once you work this out, these are just all options that can be added in or left out, but it will make the pcb extremely versatile.
This is a good thing to have.
Jeff,
My thoughts were if the cascode device needs to dissipate mote heat, sharing might not be so good but might work in this case but 36 mils ia quite a bit of current....
My thoughts were if the cascode device needs to dissipate mote heat, sharing might not be so good but might work in this case but 36 mils ia quite a bit of current....
Last edited:
Ladder DAC or FPGA (like Chord implementation).
Probably have a preference for Ladder though.
Take a look at this...
https://kitsunehifi.com/product/holo-audio-spring2-dac-kitsune-tuned-edition/
Jam
https://kitsunehifi.com/product/holo-audio-spring2-dac-kitsune-tuned-edition/
Jam
10K loading to upper and lower bias:
Whacks the slew rate. SixPack has slew rate to burn though, so not an issue.
Adds about 10dB of distortion, slightly weighted toward H2. Resulting H2/H3 spread increases from 10 to 12dB.
Is this about what I should expect, or did I do something wrong?
Whacks the slew rate. SixPack has slew rate to burn though, so not an issue.
Adds about 10dB of distortion, slightly weighted toward H2. Resulting H2/H3 spread increases from 10 to 12dB.
Is this about what I should expect, or did I do something wrong?
Yeah this is probably right.
What we are trying to achieve is producing distortion in the realms of 0.02% to 0.07% at 1W to emulate the first watt sound, just to get some sonic quality people may like/prefer.
You could try it both ways with and without to see what you like.
Yep. Apparently current king of the ladder dacs. They use proprietary error correction, perhaps nearing as complicated as basic SD. Distortion is higher than good SD.
By the way, one would typically not design one's own SD modulator, etc. It they did, it would usually be implemented in an FPGA which has its own learning curve. What most people would probably consider as scratch is starting with a good dac chip, typically a Sabre by ESS, or an AKM, say, AK4499. Like for example in another thread: AK4499 DAC Design ...in that case the exact inner workings need not be understood in detail, just how to design an implementation around the chip (which is plenty of work in itself). If a discrete modulator is wanted, there is a thread and Linear Article about that, but performance lags the best of chip dacs, IMHO.
Last edited: