Hello and thanks for tuning in to this thread. This is my first speaker build and post on the forum, so please excuse my ignorance of some of the finer points of speaker design.
This prototype build is a tri-amp clone of the Danley Sh-50 with active DSP processing. My goal is to get 60hz to 16khz with a pretty flat response and operate around 130dBc @ 1m.
So far everything is working great at the desired volumes but I would like to improve the phase of the speaker and maybe get some feedback on ways I can improve this build before I commit to building the final pair in 3/4 Birch. Any comments are appreciated!
***
Attached is a picture of the SPL/ Phase response in REW, measured at the center of the horn opening with calibrated levels and RTA mic.
XO: CD (HP -48 Linkw @1350hz), MIds (-12db Butter Low Pass @ 1350hz and HP -12 DB Butter @ 350hz), Sub (HP -48 Linkw @ 60hz and LP -48 Linkw @350hz)
EQ: Peak EQ boost at each XO q=1 +8db
Build parts list:
3/4 MDF
Ashly Protea SP 4.8 DSP
PRV 290PYS
4x Pyle 5" Sealed Mids
2x Faital 12FH510
Amps:
EV CP1200 - CD
EV CP2200 - Mids
EV CP3000 - Sub
This prototype build is a tri-amp clone of the Danley Sh-50 with active DSP processing. My goal is to get 60hz to 16khz with a pretty flat response and operate around 130dBc @ 1m.
So far everything is working great at the desired volumes but I would like to improve the phase of the speaker and maybe get some feedback on ways I can improve this build before I commit to building the final pair in 3/4 Birch. Any comments are appreciated!
***
Attached is a picture of the SPL/ Phase response in REW, measured at the center of the horn opening with calibrated levels and RTA mic.
XO: CD (HP -48 Linkw @1350hz), MIds (-12db Butter Low Pass @ 1350hz and HP -12 DB Butter @ 350hz), Sub (HP -48 Linkw @ 60hz and LP -48 Linkw @350hz)
EQ: Peak EQ boost at each XO q=1 +8db
Build parts list:
3/4 MDF
Ashly Protea SP 4.8 DSP
PRV 290PYS
4x Pyle 5" Sealed Mids
2x Faital 12FH510
Amps:
EV CP1200 - CD
EV CP2200 - Mids
EV CP3000 - Sub
Last edited:
Very nice proto! Methinks you are ready to build the finals.🙂
I like the way you made an outer shell, ala Danley style. I think that's necessary to get the rear volume needed for the bass-reflex ports to work.
I've never been willing to do that, but solely due to wanting to avoid weight. What do estimate you builds will weigh?
I like the way you made an outer shell, ala Danley style. I think that's necessary to get the rear volume needed for the bass-reflex ports to work.
I've never been willing to do that, but solely due to wanting to avoid weight. What do estimate you builds will weigh?
Your tri-amped SH-50 clone looks like the actual SH-50 in terms of its phase response (including the group delay rise below 200 Hz due to its bass reflex design , and your SPL response is flatter than the SH-50 in-room--comparing to my own measurements (note the vertical scale in milliseconds/microseconds):
Chris
Oooff these def are not light! I would guess they are around 140lbs easy. I dont mind the weight because this clone will replace four 2-way 15" powered speakers that weigh as much in total.
Nice work! since you have DSP as crossover you can achieve much flatter respobse. The scale of frequency response plot is crazy 5db per line, the response seems to be +/-5db and you can do a lot better with DSP I think, +/-1db.
Thanks for the tip on the axis. I only did two eq boosts so it would be easier to diagnose problems with the spectrum. I didn't really think about leveraging all the processing power to get the response super flat like that but its a good idea to try out!
share the raw driver responses?
from my experience using lots of PEQs to hammer a response flat sounds great as long as the underlying measurements are good (no reflections) the crossovers are good (large phase overlap region, meeting target slopes, time aligned) and your not trying to boost nulls due to acoustic cancellation. In a Synergy you may need to have differing high and low pass slopes to get a good phase overlap.
from my experience using lots of PEQs to hammer a response flat sounds great as long as the underlying measurements are good (no reflections) the crossovers are good (large phase overlap region, meeting target slopes, time aligned) and your not trying to boost nulls due to acoustic cancellation. In a Synergy you may need to have differing high and low pass slopes to get a good phase overlap.
Yeah the differing high and low pass slopes at crossover points seemed to improve the magnitude response significantly. My measurements are not perfect either, and could def benefit from an anechoic chamber at times!
Agreed. I'm not new to this, but I was just saying that the field I measure in is a free anechoic chamber.
That's the one technique that has never worked for me.
Never measured any type speaker , subs, mains, whatever, that i could lay on the ground pointing up and not have major ground reflection whacko stuff appear.
Maybe i just need to dig a pit, to get the speaker flush mounted with the ground,...... like folks say...sigh, not gonna happen ...
And honestly, why gate outside? The whole purpose of going outside is to get away from needing to gate, imho.
I used to have some old futons and besser/cinder blocks to get them where I wanted on the sides of the enclosure under test.
haha yes, me too !
I've tried a bunch of stuff like that...and always ended up saying these measurements stink,
compared to normal ground plane measurements (speaker on ground and mic on ground)
When it's easier to do, it can be easier to do it without appreciating what should and shouldn't be made flat.
Worked fine for me as it was only for my education, not an AES paper.
I have a huge field to do the next lot of tests in and for my Synergys, I think I might put them chest height and angled up so the radiation axis shouldn't illuminate the ground near where I'm measuring. That's only if I need to though.
Certainly, lets see how the thread evolves, nugget of info at a time 🙂 Measuring at least listening window and perhaps even 0-180 would help. It is symmetric device so full set of measurements is just 0-180, input them to VituixCAD and bam, graphs to see what should and what should not be EQd.