DSP Xover project (part 2)

The delays are already implemented. The bass requires 512 samples and the remaining 256 samples. The delay is thus 5.333 ms.

Oops indeed, sorry.

What actually I don't like too much, that's the dip at roughly 350 Hz, 6 dB deep.
Looking again at your settings, that's the transition between the woofer and the mid-low. The problem seems to come from the FIR approximation.

If you check the rephase screenshot, you have in red the frequency response of the FIR that you have implemented. The deviation is significant in the highpass transition. This happens because you have allocated only 511 coefficients to this filter and that doesn't seem to be sufficient.

On the other hand, you have also allocated 511 coefs to the tweeter, which is much more than required. I suppose that roughly 100 coefs would be more than enough here.

My suggestion: make the tweeter filter shorter and move the saved resources to the mid-low filter.

Another suggestion: you're using digital gains to boost the signals after they have been filtered. Try as much as possible to keep the post-gains at 0 dB and use the analogue gains instead (Analogue Outs Configuration dialogue).

Thanks for the quick reply in connection with my display. Unfortunately, it still does not work. Do I need to completely reinstall the control board (LEDs, resistors, etc.). Can I order accessories necessary with a new control board?

Displays are pretty sensitive to wrong connections. If you ever turned the system on with the display not properly wired, then chances are that it's fried.
Please contact me by email if you need parts.

Nick
 
fbu,

One thing that may cause your display to not work is that some lcd displays use pin 15 and 16 interchangeably. I had to swap pins 15 and 16 around to get the display to function in a remote controlled preamp I built with another kit previously.

If you can get a data sheet for the specific model you have then that should confirm the correct pinout. May be worth a try if you can't get it to work any other way.
 
@fluid

It is an OLED Display with WS0010 driver. I'll have a look in the manual. many thanks for your help.

@nick

now I have changed the samples in rephase (tweeter 100 and mid 1023) and let the post-gains at 0 dB. I was surprised after the first measurement with Arta, that I had to make no change in the analogue gains. It was a good hint from you.

new measurement:

arta2.jpg
nuc.jpg
arta.jpg


Good night
Frank
 
Hi Nick!
Here are screenshots. It seems to me that I do not have enough output level to drive my amplifiers. Before Najda my crossover was based on PC (JRiver + RME HDSpe AIO sound card) and I never had a problems with the gain structure. The analog outputs of RME were connected to amps, as I have now with Najda and I can drive amps to very loud level.
 

Attachments

  • Set_1.jpg
    Set_1.jpg
    125.1 KB · Views: 433
  • Set_2.jpg
    Set_2.jpg
    215.6 KB · Views: 429
Ahhh, I'm so stupid :)
Oops indeed, sorry.
My suggestion: make the tweeter filter shorter and move the saved resources to the mid-low filter.
Nick

It's what I need! I missed this window. It was 1V, now I set 3V and have a lot power :)

My last settings:
 

Attachments

  • set_5.png
    set_5.png
    81.2 KB · Views: 417
  • set_4.jpg
    set_4.jpg
    125 KB · Views: 414
  • set_3.jpg
    set_3.jpg
    225.6 KB · Views: 423
Last edited:
now I have changed the samples in rephase (tweeter 100 and mid 1023) and let the post-gains at 0 dB.

Thanks for posting that, Frank! Looks really good!

It's what I need! I missed this window. It was 1V, now I set 3V and have a lot power :)

My last settings:

Looks like you've fixed yourself your clipping issue, right? ;)

Cheers

Nick
 
Nick, my quote in previous message was wrong. It should be:
Another suggestion: you're using digital gains to boost the signals after they have been filtered. Try as much as possible to keep the post-gains at 0 dB and use the analogue gains instead (Analogue Outs Configuration dialogue).

Nick

I did not try CDP yet, but analog (vinyl) and optical (HTPC) inputs are OK now. What about LEDs brightness? How I can reduce it?
 
Last edited:
Hello Mike,

I attached a screenshot with the settings for rephase.

Regards, Frank

View attachment 432230

OK Frank! I do the same, but if I'm trying to load filter from file in NUC, I've got "Invalid file" message.
Look how rephase fili looks like inside:

rePhase settings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 
I should be getting the slowest Najda build award I think. :)

But since Nick likes photos, I thought I could post the current status:
Power supplies in place, SE->BAL boards working, Arduino in place to switch between four spdif inputs (Twisted Pear Mux doing that) and display LED status.

Todo: Bring out a trigger for amps, install a USB->I2S interface and clean up the wiring a bit.

attachment.php


attachment.php
 

Attachments

  • photo-5.jpg
    photo-5.jpg
    356.9 KB · Views: 1,110
  • photo-7.jpg
    photo-7.jpg
    103.1 KB · Views: 1,086
Hello,

you generate 1024 taps. Nuc is only able to work with max 1023 taps.
Channel 3 or 4 is limited to 512 taps.

Best regards
Frank
Hi,
You can open the file in a text editor and delete the last line to get 1023 taps.


I should be getting the slowest Najda build award I think. :)

But since Nick likes photos, I thought I could post the current status:
Power supplies in place, SE->BAL boards working, Arduino in place to switch between four spdif inputs (Twisted Pear Mux doing that) and display LED status.

Todo: Bring out a trigger for amps, install a USB->I2S interface and clean up the wiring a bit.
Wow, looks nice and tidy

best,
Paal
 
Hello,

you generate 1024 taps. Nuc is only able to work with max 1023 taps.
Channel 3 or 4 is limited to 512 taps.

Best regards
Frank

Hi Frank! Thank you very much! I found this information in the Manual :) Now I'm able to load file from rePhase to Najda. Now another question - because "Settings" are not available in FIR engine, what to do with Analog Outs Configuration to increase max. output voltage from 1V to 3V?

Regards, Mike
 
Hello Mike,

I have an output voltage of 2 volts chosen. This is roughly equivalent to the output of a CD-player on analog-out. I've just have not managed to save the settings in the long term. When I reboot my pc NUC forgets the selected settings for analog out and bass/treble Settings.

You can also edit in rephase the frequency response. For that you are using "Paragrphic Gain EQ" in rephase and import the result into the input processing engine of NUC.
Unbenannt.jpg


regards
Frank