First transportable Ozone incarnation
I took this to the park this morning, it survived the journey in my backpack but it exposed a dodgy RCA connector. The case I used (temporarily) for it is a cylindrical tea carton - the red 'tower' behind. Battery power comes from 3 NiMH AA cells giving around 10hrs of listening.
I've just done a mental estimate of the BOM cost of this DAC (labour cost is way dominant, but this is DIY right?) and it turns out to be dominated by the cost of the caps in the top three tiers (the small black ones). These are real Japanese Rubycons, chosen for their diminutive height (16mm) vs the Chinese ones at 25mm height. So if you didn't mind adding 3cm to the overall neight you could build this with all Chinese caps and save a few dollars. As it is the total parts cost is under $30 including batteries. If you wanted S/PDIF input (currently its I2S) you'd add a couple more $.
I've just done a mental estimate of the BOM cost of this DAC (labour cost is way dominant, but this is DIY right?) and it turns out to be dominated by the cost of the caps in the top three tiers (the small black ones). These are real Japanese Rubycons, chosen for their diminutive height (16mm) vs the Chinese ones at 25mm height. So if you didn't mind adding 3cm to the overall neight you could build this with all Chinese caps and save a few dollars. As it is the total parts cost is under $30 including batteries. If you wanted S/PDIF input (currently its I2S) you'd add a couple more $.
Total Comments 13
Comments
-
Was this to allow it to relieve itself of excess "electrolyte"? ... ;)
And I hope you had it on a lead ...Posted 27th April 2014 at 11:35 PM by fas42
-
You've given me an idea for a horizontal version - the 'Ozone dachshund'
Posted 28th April 2014 at 12:28 AM by abraxalito
-
Related to DACs, possibly interesting new thing around, the Lumin network music player - credited with high grade 'analogue' sound, using single Wolfson chip per channel.
Parts of its bag of tricks are using Lundahl transformers in the output circuitry, and a very solid aluminium "cage" around the bits - would be interesting to dissect it ...Posted 28th April 2014 at 01:36 AM by fas42
-
Saw your mention of the LMH6738's OLG so went to investigate. Its not quite as you imagined it as the OLG vertical axis is not dB, rather dBohms for a transconductance. This being a CFB part which actually have lower (and flatter) OLG plots than VFB. So its not king of the hill for OLG after all.Posted 5th May 2014 at 03:00 AM by abraxalito
-
Well caught! This was quite some years ago, when I went on a bit of frenzy, chasing down every high speed opamp to see if genuine high performance could be got at the highest audio frequencies, at least by nominal DS numbers. I'm not so fussed about it these days, especially since I found that subjectively decent results can be obtained from very unpromising parts.Posted 5th May 2014 at 12:53 PM by fas42
-
I was looking at the ADA4899 this morning perusing that thread and wondering it it would make a top-notch voltage regulator. The output impedance was nicely flat at 4mohms through the audio band. 1/f noise corner lets it down rather though.= getting it stable into a capacitive load calls for a very low DCR inductor so as not to compromise the Zout....Posted 5th May 2014 at 01:39 PM by abraxalito
-
In that vein, I was looking at some of the other interesting opamps I came across previously, last night, and the OPA847 is interesting: better than 3mohms up to 100KHz. And, the Jung regulator as a discrete build would take some beating, better than 100uohm over nearly the whole audio range, can nudge 1uohm.Posted 5th May 2014 at 11:10 PM by fas42
-
I used current sensing (aka positive fb) on my AD815 reg to lower the output impedance to the mohm range - still didn't sound as good as a cap bank.Posted 5th May 2014 at 11:37 PM by abraxalito
-
You should have a bit of a handle now on the comparative qualities of the different ways of doing the power supplies, would you be able to put some very rough, subjective numbers, say out of 100, on how they come across - starting from an absolutely basic, technically correct arrangement, right up to your current best effort?Posted 6th May 2014 at 01:46 AM by fas42
-
Nope, not yet - power supplies rather depend on what's being powered. The DAC chip I'm using is more sensitive to power than the AD815 for example. What I can say though is passive definitely beats active - just throwing caps across the supply hasn't been beaten by any reg so far. It would be nice though to have a much less bulky solution and I'm optimistic that this can be achieved using DSP.Posted 6th May 2014 at 04:04 AM by abraxalito
-
Do you feel you have reached the limit of potential improvement, with the DAC then, in terms of amount of capacitance used - if you used, say, half of the capacity how much would be lost, where does Dimishing Returns kick in?
Certainly agree with passive over active, especially in close proxity to the DAC - perhaps a hybrid, with well thought spacing from and isolation of the active section.Posted 6th May 2014 at 04:22 AM by fas42
-
You do love the tough questions - very hard to quantify qualitative changes
In terms of the amount of capacitance on the DAC chips (TDA1387s) I won't be going beyond 2F. The amount of change going from one tier (0.65F) to two (the purple electrolytics) was small. Ditto going from 1F to 2F with the black 10,000uFs. So I will probably stick with just a single tier, though it might be worth playing with putting individual DACs on their own smaller hexacaps, then paralleling rather than stacking them as I'm doing now.
There's another strand to this which is the physical impracticality of getting lower impedance because of the physical size of caps and lengths of interconnects. Getting below 1mohm at the DAC is hard because of the resistance of practical lengths of copper. Then inductance and skin effect start to bite. So it could be that no further improvement is being noticed because in practice adding more caps is making next to no reduction in impedance at the DAC itself. Somewhere between 1kHz and 10kHz the hexacaps turn inductive. Adding more caps is reducing the LF impedance though, I'm not totally sure which frequency range is most critical in terms of the subjective SQ. So lots of handwaving because I simply don't know
What gets lost as I reduce the capacitance is the acoustic background ambience at lower freqs. Its still there but its muddled - at first listening a DAC with fewer uF on the supply can sound more 'impressive' but impressiveness turns out on further listening to be a colouration - bass 'bloat' if you will - because the brain can't make sense of the LF ambience cues. I think this is what's meant by 'hi-fi sound'. More uF makes for a so much more relaxing listen.
Then there's the difference uF makes on the AD815 - its a difference in danceability, or snappiness, or PRaT rather than in ambience. Too hard to quantify this but on balance I'd rather give up snappiness than ambience so I tilt the allocation of caps towards the DAC.Posted 6th May 2014 at 07:26 AM by abraxalito
-
Just here to help ... :D
Thanks for that, it helps me get a better handle of how it's progressing: the ambience thing strikes me as better retrieval of low level detail, the DAC is behaving "more correctly" with reproducing the the finer, more subtle information captured in the recording - you can hear 'deeper' into the space where the recording was made, get a better sense of the environment; whereas the snappiness is cleaner reproduction of the treble, high harmonic content - in the optimum state both of these aspects are able to be fully resolved, it can only get better ...
Cheers ... :)Posted 6th May 2014 at 12:55 PM by fas42
Updated 6th May 2014 at 01:15 PM by fas42
