DSP Xover project (part 2)

To get lower DSP load I made new setup with 48 kHz sampling, but get other problem with it. Seems Najda asynchronous sampling rate converter is not very asynchronous.
48 kHz setup works correctly only with 48 kHz SPDIF input, with 44k or 96k I get strange holes. Measurements made 1m from speaker, brown is 48k signal blue is 44k. Fall on 20 kHz is because mic was not perfectly on axis.
With old 96kHz sampling setup I have not seen this kind of holes.

48kvs96k.PNG
 
Last edited:
I have a similar (?) problem on one of the Najda channels, when doing a measurement sweep there is clear aliasing or clipping problem at around 10 kHz as I remember it resulting in a deep dip in measured response. It is wery audible when playing the shirp but I can't say that I have heard any problem when playing music. It sounds like the normal sweep up to around 10k and then some Tj Tj Tj Tj sounds, continued by the end of the sweep.
 
I have also had the same problem. I noticed that it was possible to remove the problem by switching the setup file and then rereading the one I was using. At some point, I also was able to deal with the problem by shooting down one of the channels.
I have not however been able to reproduce those so the problem seems to be random.
My Najda is boxed and so built by WAF. I have the USB in card aslo original fit.
 
Does anyone know if ant atx power supply could be used with the Najda?
If the voltages match Najda's requirements, it should be fine. I salvaged a power supply from an old Rotel DVD player and it worked fine. There might be some concerns about the noise levels on a power supply that was not designed for audio purposes, but the only way to really know if it is an issue is to try it out.
 
Najda's power supply requires +5V 1A, +12V 1A and -12V 0,3A. Any ATX power supply will do this even the original IBM PC. But to be sure check out this link.
http://gcctech.org/cec/powersupply/All about the various PC power supply cables and connectors.htm

I think much of the talk about noisy power supplies goes back to the era when we did build all amplifiers from discrete components. The noise-rejecting circuits were not at the same level as in today's circuits.
There are several YouTube videos where it has been shown that the noise from switching power supplies is negligible.
But I understand the HiFi need for perfect components so everybody must choose his or her own level of commitment to the perfect sound.
 

Attachments

  • 1-IMG_6073-001.JPG
    1-IMG_6073-001.JPG
    181.2 KB · Views: 141
  • 2-IMG_6089-001.JPG
    2-IMG_6089-001.JPG
    121.7 KB · Views: 113
  • 3-IMG_6086-001.JPG
    3-IMG_6086-001.JPG
    139.8 KB · Views: 135
  • 4-IMG_8599-001.JPG
    4-IMG_8599-001.JPG
    218.5 KB · Views: 134
Last edited:
  • Like
Reactions: Shaun and Julf
Another dumb question. I was about to purchase some hypex ucd250 modules for a multichannel amp to use with the Nadja.
I just read that they need an input buffer and the input sensitivity is about 6.5v. Does this mean I can't connect them directly to the Nadja? From what I read it can deliver up to 6v output or is this unrelated?
 
Don't know if CS3318 is directly fed to output or some kind of output stage is in place in Najda. Lowest load for CS is 2k ohm and hypex ucd250 is 1.8k ohm so it's lower. So some buffer may be needed if CS is directly fed to output. If not then it will depend how this output stage looks like.
 
Below 2k distortion will rise and CS may heat bit more, probably it will not damage but for sure it's not optimal. Evening I could measure resistance from CS output to RCA socket. This could tell us if there is anything on cs output that could help with low hypex ucd250 input impedance.