lingDAC - cost effective RBCD multibit DAC design

FR response plot

I created a white noise test waveform in Audacity at 96kHz sample rate and played it through PhiDAC, recording the output with my Sony PCM-M10. Here's the FFT of the output displayed in Audacity. The lift in the FR at HF is there because the DAC's equ'd for 44k1 which requires more droop correction than 96kHz. I'll make another recording using 44k1 sample rate to show how effective the droop correction is.

Incidentally I suspect the slight LF roll-off might be in the recorder, I can't of any mechanism in the DAC to cause it.

<later - added the 44k1 plot. The FR is pretty flat to just above 17kHz, then rolls off to -2dB @ 20kHz. The chasm just above 20kHz might be because I downsampled the 96kHz white noise file to create this stimulus.>
 

Attachments

  • PhiDAC_FR.png
    PhiDAC_FR.png
    43.2 KB · Views: 377
  • PhiDAC_44kFR.png
    PhiDAC_44kFR.png
    42.6 KB · Views: 347
Last edited:
That depends on how high in frequency your hearing goes - mine definitely cannot even reach 15kHz nowadays so I can't see the benefit from extending the response to 22kHz. That will have a downside of increased images making it through. However, the schematic's open so anyone can fiddle to see if a wider bandwidth helps.
 
I've been trying several cut-off (-3dB) frequencies, including 20Khz, 30Khz, 50Khz, 96Khz... etc. My source is fixed at 192Khz. They all sound different to me. With lower cut-off, it sounds more smooth. On the other hand, with higher cut-off, the sound of violin is more airy. After some listening test, I'm now settled with 27Khz. I guess this will highly depend on your speakers.

Poting
 
Last edited:
Pretty amazing that even various cutoff frequencies at or above 20KHz can sound different.

We usually look at log/log graphs of frequency vs amplitude, since that is believed to reasonably approximate how humans hear. However, if you look at what a filter is doing with respect to the so-called cutoff or corner frequency in lin/lin format you would see that things actually change much more gradually than you might imagine. If it can sometimes be true that audibility doesn't follow the log/log model as close as we may assume, then maybe it wouldn't seem quite as 'amazing.'
 
Digital

DIGITAL filters sound "best".
'Course, toots would find that "unbelievable."
BTW ... say one were to precede the Ling (or the Phi or the ... aw, hell, I've lost track of all of Toots' dime-store projects) with a digital filter ... how would that change the parameters of the "stock" NOS reconstruction filter?
 
Wow. I THINK I can hear phase shifts from the different 1387 droop correction filters I have heard (Audiophonics Hat and LingDAC prototypes). Pretty amazing that even various cutoff frequencies at or above 20KHz can sound different.

Sorry, I forgot to mention that my DAC is based on PCM63, not this particular lingDAC. However, they are all current output DACs. The idea of LPF filter before IV works great for me.
 
Updated PhiDAC schematic for v1 kits

I've made one or two minor tweaks to component values for this version, based on measurements of the first three v1 PhiDACs. This is the schematic I'll be putting together kits for (barring any unforeseen howlers of course :p). I'll aim for 60 or 70 boards in kitting so there will be one or two spare based on those who've indicated interest to date. If your interest has waned or you'd like to join the PhiDAC building fraternity do please step forward in the next week as I'll be ordering parts for kits as soon as Taobao shops re-open after the Chinese New Year (in the week beginning 11th Feb).

BOM for kits to follow shortly. Mouser BOM will be along after that.
 

Attachments

  • Schematic_phiDAC-v1_revG.pdf
    70 KB · Views: 157
Richard, can you yet say which electrolytic capacitor the BOM will specify for the 1,800uF output coupling units, C21 and C22? These are unusally large in value for line coupling. Perhaps, headphone coupling is also being anticipated. Or, maybe, you simply find the chosen 1,800uF units to sound particularly good in this application. Just curious.

Thanks.
 
Last edited:
Hi Ken - interesting you picked up on that one. I fitted 1800uF because I already have one doing filtering duty on the PCB and was too lazy to add another line item. But I will do more listening to see which kind sounds best - at present I'm veering towards replacing those (labelled as Sanyo but likely just badge engineered) with some Panasonics. The main reason they're large is I couple this DAC directly to a classD amp which has an input trafo - the capacitance and primary inductance together form a very low frequency resonance and I want to be sure its well below the audio band.

The ones I've fitted look like this : 6.3V1800UF