Welcome, Guest
Username: Password: Remember me
  • Page:
  • 1
  • 2

TOPIC:

"Sharc Stack" FIR project 8 years 1 month ago #21342

  • planetti
  • planetti's Avatar
  • Offline
  • Senior Member
  • Senior Member
  • Posts: 48
  • Thank you received: 8
Hi Winfried,
as you already envisaged correctly:
- when driving a first OpenDRC-DI by the 2x2-plugin, you will be limited to 48kHz, definitly. :pinch:
- when driving a first and a second OpenDRC in series via SPDIF, the digital signal will be processed through all two ASRC. I do not know if some detrimental effect is reduced when running the second OpenDRC on the plug-in 4x8 on 96kHz , but then the 3400 taps on the 8 channels could be too less for steep slopes. You may try some Xover calculations in rePhase for your dedicated 4way speaker before estimating the number of taps needed and then deciding on the plug-in fitting the requirements :unsure:

Please Log in or Create an account to join the conversation.

"Sharc Stack" FIR project 8 years 1 month ago #21343

  • leoman
  • leoman's Avatar Topic Author
  • Offline
  • Premium Member
  • Premium Member
  • Posts: 92
  • Thank you received: 32
This is a repost of the photos which have mysteriously disappeared from the original thread posting.








Attachments:

Please Log in or Create an account to join the conversation.

Last edit: by leoman.

"Sharc Stack" FIR project 8 years 1 month ago #21359

  • Hoffmann
  • Hoffmann's Avatar
  • Offline
  • Senior Member
  • Senior Member
  • Posts: 59
  • Thank you received: 8
Hello John,

thanks a lot for the effort to repost the pictures! Yes, they make several things quite obvious!
I'm really still rather undecided. There are sevaral things which enable my 4-way project, but e.g. my priority on 96 kHz sample rate reduces the amount of Taps for FIR quite drastically...

Hello planetti,

Hi Winfried,
as you already envisaged correctly:
- when driving a first OpenDRC-DI by the 2x2-plugin, you will be limited to 48kHz, definitly. :pinch:
- when driving a first and a second OpenDRC in series via SPDIF, the digital signal will be processed through all two ASRC. I do not know if some detrimental effect is reduced when running the second OpenDRC on the plug-in 4x8 on 96kHz , but then the 3400 taps on the 8 channels could be too less for steep slopes. You may try some Xover calculations in rePhase for your dedicated 4way speaker before estimating the number of taps needed and then deciding on the plug-in fitting the requirements :unsure:

I actually do keep forgeting that OpenDRC-DI is only supporting 48 kHz sample rate, so for my case two OpenDRC-DA8 in series, where the first has the DA8 substituted by a DIGI-FP (if this works...) for SPDIF connection to the second OpenDRC-DA8.

Well, I strongly agree to your suggestion to do some serious "Tap-budgeting" before buying, to see if the 6800(?) Taps (both units combined) will suffice (How many Taps a DEQX PDC 2.6 provides I do not know...). So actually, for that I'll have to make myself familiar with rePhase or another FIR Filter design software....

Thanks forks for your engaged feedback!
Regards,
Winfried

Please Log in or Create an account to join the conversation.

Last edit: by Hoffmann.

"Sharc Stack" FIR project 8 years 1 month ago #21361

  • leoman
  • leoman's Avatar Topic Author
  • Offline
  • Premium Member
  • Premium Member
  • Posts: 92
  • Thank you received: 32
Just my 2c, I would definitely prioritize the greater tap number over the sample rate for my crossovers. I'm able to get 192db/oct at 380hz and the rest are all 512db/oct. Unless taps work more proficiently at higher sample rates, I'd think the number at 96k is too low. Perhaps someone who knows otherwise will chime in.

Please Log in or Create an account to join the conversation.

"Sharc Stack" FIR project 8 years 1 month ago #21376

  • Hoffmann
  • Hoffmann's Avatar
  • Offline
  • Senior Member
  • Senior Member
  • Posts: 59
  • Thank you received: 8
Thanks for the recommendation! Definitely worth a thought! While I will admit, that until now the steepest linear phase filter slopes I have used were 96dB/Oct. :unsure: One potential issue I see with my set of drivers would be sudden discontinuities in the polar response at the cross over Frequencies (e.g. from a 100 mm to a 38 mm Driver), leading to coloured sounding reverberation in the room. Actually, optimising/smoothing polar response is one of my design goals in addition to phase-/frequency-response linearization.

Although, the good news (for me anyway ;) ) is that the Sampling Rate is subject to relatively affordable plug-ins, thus enabling comparisons of different solutions with differing sampling rates and filter architectures.

Regards,
Winfried

Please Log in or Create an account to join the conversation.

Last edit: by Hoffmann.

"Sharc Stack" FIR project update 2 8 years 1 month ago #21430

  • leoman
  • leoman's Avatar Topic Author
  • Offline
  • Premium Member
  • Premium Member
  • Posts: 92
  • Thank you received: 32
I had promised to show the result of an attempt to build a single-housing 6-amp array to power my very non-power-hungry horn array, so here are some photos. The box once apparently housed a SA DSP of some sort. The operative controls are the power switch (and led) and the 3 (dual/stereo) input level pots on the front, plus the barrier strip and 3 RCA inputs on the back. Results, despite my scopes' difficulties with compatibility (I'm learning a lot about class D), are audibly good enough that I'm now going to try bridging a few of these little amps (based on STMicro's TDA7492 amp chip) in another box to run my not-terribly-powerhungry Altec woofers. The box shown here replaces 3 big amps (and has no fan noise).

Note: a look with the scope (when it doesn't induce oscillation) shows what to me was a horrifying 250khz 'noise' (switching) signal on the outputs, varying from 20 to 80mv depending on the board, but a quick check of a different class D amp showed even worse, so I guess I'm not going to become a phased(?) array longwave transmitter, at least not a big one.

Here are shots of the front closed and open, and the rear open. Note that the cover is shimmed a bit to accommodate the also single-rack sized power supply.










Here is a shot of the left-side speaker array. The little short-horn ribbon tweeter is not connected; the small CD horn next to it is. I doubt the ribbon is capable of putting anything like 250khz into the air, but the CD seems a better match anyway.



You probably can't see it, but the woofer cone has a terrible gash on one side (caused years ago by me), 'repaired' with automotive RTV about 20 years ago! I was about to send it out for reconing when my tests showed excellent response with zero buzzing or audible/visible distortion well past its intended band. The other side has a matching (but reconed) Altec 421-8H. They sound essentially identical, BUT the recone is a good 4-5 dB less efficient than the repaired original! Easily compensated; just some info for those pondering a recone.

I am unable to get phase measurements for my amps, but a look at most generic class-D writeups/tests shows up to 50* shift at 20Khz, a shift beginning at near-zero around 1Khz. It looks alarming, but as long as the amps are fairly well matched the crossover regions and the stereo imaging shouldn't suffer as their slight phase shifts would be nearly identical at any frequency. Audibly nothing jumps out at you. Besides, the fan-cooled amps being replaced are class D anyway.
Attachments:
The following user(s) said Thank You: devteam

Please Log in or Create an account to join the conversation.

Last edit: by leoman.

"Sharc Stack" FIR project 8 years 1 month ago #21473

  • Hoffmann
  • Hoffmann's Avatar
  • Offline
  • Senior Member
  • Senior Member
  • Posts: 59
  • Thank you received: 8
Hello John,

thanks a lot for the thorough insight into the implementation!

After doing a few cross-over simulations with rePhase I'm slightly discouraged from implementing my 96 kHz 4-way FIR Sharc Stack project due to a lack of number of filter taps :(

Well, having said that, my Speakers do have a 4-way analog LR subtractive cross-over with delay compensation, so maybe my project morphs into a miniSHARC Stack with an I²S Stereo DAC, to implement an overall linearization "only". I'll probably start a separate Topic on that.

Thanks again for all the guidance!
Regards,
Winfried

Please Log in or Create an account to join the conversation.

Last edit: by Hoffmann.

"Sharc Stack" FIR project update 2 8 years 1 month ago #21500

  • devteam
  • devteam's Avatar
  • Away
  • Administrator
  • Administrator
  • Posts: 9231
  • Thank you received: 1514
Very cool update @ Leoman! Thanks for keeping us updated on your project. Quite a big endeavor you've got here.. :-)

All good if we post some update on the Project Gallery?

DevTeam
miniDSP, building a DSP community one board at a time.

For any official support, please contact our technical support team directly @ support.minidsp.com/support/home

Please Log in or Create an account to join the conversation.

"Sharc Stack" FIR project 8 years 1 month ago #21505

  • leoman
  • leoman's Avatar Topic Author
  • Offline
  • Premium Member
  • Premium Member
  • Posts: 92
  • Thank you received: 32
Of course! Please do. No need to ask.

Should have a bridged bass amp in a few days with similar components and enclosure. I'll post about it if it works.

Question: what is the lowest load impedance the miniDAC8 outputs will tolerate? I ask because I'll have to split each bass output channel to two amplifier inputs to bridge. I believe my amps are something like 60k input impedance, but if the boards have 'support circuitry' that knocks it down to say 10k (which I doubt), then would 5k present a problem to the miniDAC8? I don't want to blow that! The amp boards I can afford to blow :)

Please Log in or Create an account to join the conversation.

"Sharc Stack" FIR project 8 years 1 month ago #21549

  • leoman
  • leoman's Avatar Topic Author
  • Offline
  • Premium Member
  • Premium Member
  • Posts: 92
  • Thank you received: 32
And here is the bass amp. I didn't shoot the inside because it was already together and it looks roughly like the previous multi-horn amp but with 2 amp boards and slightly different wiring.

I discovered to my dismay that bridge mode doesn't work very well, about 6v of 250khz switching wave and huge distortion for some reason. So I used just one side of a board (either channel since I had wired the respective input to both L and R on the board, expecting to bridge) and found almost zero noise! To add to my joy, with an unmounted 15" 8 ohm woofer acting as a load, I saw an undistorted waveform at 350hz all the way up to 30V, where I quit. That's a good 100w peak or so, way more than I ever intend to use! Thus each dual amp got one channel used, the other left open. Optimal? No idea. Don't care as it exceeds my expectations even for bridging. I also found that my commercial amps (also class D) were churning out 4v of longwave RF switching signal all along! Anyway...

The front (pardon the garish lettering; dry-transfers are getting hard to find like everything else around here):





The back. Note this time I used Neutrik-type connectors since I had speakon bass cables.





The new pile of stuff. Six original boxes have become four, with much-reduced height.


Attachments:

Please Log in or Create an account to join the conversation.

  • Page:
  • 1
  • 2
Moderators: devteam