Unity horn, celestion + bms

Status
Not open for further replies.
Finally I took the step into making my own unity horns. Driven by all the info I found on this forum I made a 'clone' of the horns Jmorken has made.

I used the BMS4550 cd and the Celestion TF0410MR.

I am using a Minidsp 2x4 for crossover and so far so good. There both up and running and now I am mostly busy getting the crossover good.

I'v made an adaptor plate which goes from 7º to 30º and 3d printed some parts that are now covered in paint en putty to make an smooth transition from cd to square horn. The entry ports are also cnc'd.

I'm using a little delay on the mids since the path length of the bms was a bit longer then expected (using an annular diaphragm)
But I still think I can get the crossover better

I far it is possible in the space where I'm testing they sound pretty good, but I think I can get more out of the dsp.

pictures you can find here:

fotsi_nijts's Recent Uploads | Photobucket
 

Attachments

  • IMG_0988.jpg
    IMG_0988.jpg
    681.4 KB · Views: 746
  • IMG_0931.jpg
    IMG_0931.jpg
    362.7 KB · Views: 685
  • IMG_1125.jpg
    IMG_1125.jpg
    405 KB · Views: 670
Last edited:
response of the mids and cd after eq

Beautiful cabinetry work!

From this graph it looks as if you have done nothing to compensate for the mass rolloff of the compression driver?

Most designs using a compression driver use a first order filter at about 20Khz or so to flatten out the response. Still has the BMS dip and then rebound at the highest frequencies.
 
OP, will you share what crossover settings (if active) you are using or experimenting with? I am the owner of the only Unity still sold (?), the Yorkville U15. I made mine active mainly because the used pair I bought had blown crossovers 🙂 if anyone cares, I used the "greatest null" method and came up with:

Mid-Tweet x-over 48 db/oct, 1000 Hz, 0.12 mSec.
Bass-Mid x-over 48 db/oct, 300 Hz, 0.58 mSec.

The Yorkville uses same "tweeteer" as yours (but "4550-OEM"), the midranges are Celestion TF0510MR which may be similar to your 4" models?
 
No reply? Well, I am trying a longer delay on my mids now. Instead of the above method, I looked for the usual mid notch. Mine is at 1.91KHz. This equates to a period of about 0.53 msec, so i am trying that in the EQ/delays.

Don't think it will work that way, there will already be a delay due to the lowpass crossover to the mids. Delay is more than distance, in a speaker system. In Synergies I made, the tweeter actually needed delay because all the head-start the midranges got was lost in the midrange response delay.

The best way to figure your delay setting is get a measurement setup that can show you an impulse response (REW, HolmImpulse, or my Omnimic) and then you can see the delay offsets directly and tweak them in real time. You can also see what you might need to do to the rolloffs of the drivers to get the impulse response nice and tight while keeping the frequency response flat. Nice thing about a concentric driver arrangement like Synergy is you don't have to worry about off-axis effects, you can do whatever is needed crossover shape and delay-wise to make the speaker behave on-axis and your off-axis will track.
 
Status
Not open for further replies.