A Summing Amp I Built

Help Support GroupDIY:

Chrisfromthepast

Active member
Joined
Mar 5, 2018
Messages
35
White and yellow are my picks for out of box.
If those are the ones, I think I might be a convert. Plugins don’t do such a nice job of rounding things off. The tracks sound together, but not necessarily veiled. Of course, for some source material, blue might be the right choice.
Im generally curious if we could tell them apart if A was a loud 2 track mix into the summing mixer and B was summing 9 stereo mixes that hit the 990 with the same input V p/p.
Know what im saying? Is active summing the source of the sound? Is avoiding protools math the magic? Is it that all of the D/A’s working at lower levels that provide more headroom or less intermod, or more quantization distortion or something giving us a sound we are more used to? I personally doubt these claims, and think the transformer is the principal source of nonlinearities.
But then again, Im wrong ALOT, and Im comfortable with that fact, because it keeps me learning.
 
Last edited:

bjosephs

Active member
Joined
May 10, 2021
Messages
42
Location
Massachusetts
I’m at work and honestly can’t recall which was which (intentional, wanted to make it hard to listen with expectations) but my impression was that the biggest change was in the low end. The thumpy bass is more muted in the analog summed versions than the digital. Those cinemags are CMOQ2 and have pretty high handling at 20Hz (and were talking about more like 40Hz anyway) but they may be starting to saturate with all the gain I added. I should do the math on the signal levels tonight and see what I’m really hitting it with. Add gain controls to the feature list for version 2...

https://cinemag.biz/output/PDF/CMOQ-2.pdf
 
Last edited:

bjosephs

Active member
Joined
May 10, 2021
Messages
42
Location
Massachusetts
Attached are screenshots of the outputs of my summing mixer as monitored in my DAW with a spectrum analyzer plugin. I bought the parts and designed (mostly) the board to make a smaller, EMI shielded supply.


Right Channel.pngLeft Channel.png
 

bjosephs

Active member
Joined
May 10, 2021
Messages
42
Location
Massachusetts
In terms of levels hitting the transformer, is this thought process correct? The bass track is peaking around -12dB, this is reference to full scale. If the outputs are calibrated to +4 that would make the analog signal -8dB; I'm adding 12 db at the mix amp so I'd be hitting the transformers with a +4dB peaks at 41Hz (Low E on bass)? Something like that? There's some more energy in that range from the kick but am I roughly on the right track?
 

bjosephs

Active member
Joined
May 10, 2021
Messages
42
Location
Massachusetts
And here is a quick screenshot of the board for the DC-DC converter board I'm going to try next. This has been on my bucket list anyway so now's a good time to try it. It will be 9-36vDC in to +/-24vDC out using a PDQE15-Q24-D24-D from CUI. It may be easy to follow without a schematic due to simplicity but I'm implementing the EMC recommendations on the input and output, a ground plane on the bottom for shielding, and 10uH chokes on the output. I'm planning to 3D print a little box that will be lined with copper foil tape for a top shield if needed.

It's an isolated converter being fed from a non-grounded power supply (probably a standard wall wart for initial testing) so I have to think through grounding a bit. The output side on the left will get referenced to audio ground at the mixer board, that's easy. I wanted to use a ground plane for shielding AND for a low impedance return for the caps so I tied the low side of the input to the plane. I'm debating whether this will provide a shielding effect if the PS isn't grounded so option 2 is to then tie the ground plane to chassis and let audio ground facilitate shielding.... I think this would be OK because no current from the incoming DC supply should flow through this ground (As long as the supply is transformer isolated). Option 3 would be to run return traces back to the 0v input terminal and use the plane only for a shield tied to chassis which, again, will get ground referenced by the audio jacks. Maybe a ground stud on the chassis as an option?


Screen Shot 2021-05-29 at 12.51.14 PM.png
 

bjosephs

Active member
Joined
May 10, 2021
Messages
42
Location
Massachusetts
Power supply board came out great and works as intended. There’s about 5mV of high frequency ripple on the outputs but I’m not assuming that will be a problem. Need to make a new harness for PS wiring and then I can test it in the mixer.
 

Attachments

  • C55F3015-A80C-47C9-BFFA-7F6E8381A58E.jpeg
    C55F3015-A80C-47C9-BFFA-7F6E8381A58E.jpeg
    112 KB · Views: 6
  • C89EDE88-D049-4407-AFCF-F2B180B42790.jpeg
    C89EDE88-D049-4407-AFCF-F2B180B42790.jpeg
    94.5 KB · Views: 6

Latest posts

Top