Goodevening all! After much trail and error i think i may possibly have come up with an almost (if not completely) finalised theoretical loudspeaker design. I've used the Scanspeak 18W/4531G00's as twin bass units crossed to a Seas MCA15RY midrange unit then onto a Scanspeak D2905/950000.
Ive modelled cabinets, baffle design and shape etc and combined everything together with FRD tools and used all that data in speaker workshop to come up with this calculated crossover design. Ive followed the tutorial on rjbaudio website and read around extensively on the subject to try and come up with something good here. I'm hoping this design which is attached below should be reasonable, needing only minor tweeks or none at all. I also realise in the real world, results may differ slightly due to a lot of variables but im hoping this should be quite a close representation.
Now please can some of you wise people have a quick look at the data ive given here and see if you can spot anything out of the ordinary that may potentially cause problems! Im asking this because i havnt actually done this before and followed it through, but i do intend to buy these drivers and build these speakers if i think the design is reasonable enough!
Any help or advice at all would be very much appreciated.
Thank-you for reading & hope some of you will have some good words of wisdom or helpfull advice!
CABINET DESIGN:
BAFFLE DIFFRACTION SIM DATA:
CROSSOVER DESIGN, CALCULATED FREQUENCY RESPONSE, IMPEDANCE & PHASE INFO:
Ive modelled cabinets, baffle design and shape etc and combined everything together with FRD tools and used all that data in speaker workshop to come up with this calculated crossover design. Ive followed the tutorial on rjbaudio website and read around extensively on the subject to try and come up with something good here. I'm hoping this design which is attached below should be reasonable, needing only minor tweeks or none at all. I also realise in the real world, results may differ slightly due to a lot of variables but im hoping this should be quite a close representation.
Now please can some of you wise people have a quick look at the data ive given here and see if you can spot anything out of the ordinary that may potentially cause problems! Im asking this because i havnt actually done this before and followed it through, but i do intend to buy these drivers and build these speakers if i think the design is reasonable enough!
Any help or advice at all would be very much appreciated.
Thank-you for reading & hope some of you will have some good words of wisdom or helpfull advice!
CABINET DESIGN:
BAFFLE DIFFRACTION SIM DATA:
CROSSOVER DESIGN, CALCULATED FREQUENCY RESPONSE, IMPEDANCE & PHASE INFO:
Any thoughts? Im particually interested in any comments on the phase info. I think it looks ok to me, but id appreciate it if anyone could comment. Im very happy with how the bass drivers and mid blend into each other in freq and phase but Im finding it difficult to cross between the mid/tweeter with the same degree of success. Their reletive phase is little out atm and seems to be cancelling a bit, hence the shallow dip also in freq response around 7k. Ive tried 2nd, 3rd and 4th order crossovers and so far this is my best result. Anyone got any idea how to improve this section?
Thanks,
Andrew.
Thanks,
Andrew.
Hi Andrew,
I'm by no means an expert with 3 way speakers but I'll make a few comments and hope to be corrected if they are wrong
The bass - mid crossover point looks to be around 1kHz. This is fairly high - it might be more natural to let the mid handle from 300 Hz up - this would cover the vocal range where the ear is very sensitive, with one driver.
Next, the mid to tweeter crossover also appears high - about 4kHz but maybe higher. The scan tweeter would be more than capable of crossing at 2.5kHz or 3kHz, and this would give a more even dispersion as the MCA15 seems to start beaming between 2 and 3kHz.
Actually If I was building this speaker I'd probably be tempted to forget about the mid range and cross at just over 2kHz. I think the midrange qualities of the scan drivers might exceed the seas midrange - this is just a gut feeling based on the mid-bass quality of other scan drivers that I have used.
Cheers,
Nick
I'm by no means an expert with 3 way speakers but I'll make a few comments and hope to be corrected if they are wrong
The bass - mid crossover point looks to be around 1kHz. This is fairly high - it might be more natural to let the mid handle from 300 Hz up - this would cover the vocal range where the ear is very sensitive, with one driver.
Next, the mid to tweeter crossover also appears high - about 4kHz but maybe higher. The scan tweeter would be more than capable of crossing at 2.5kHz or 3kHz, and this would give a more even dispersion as the MCA15 seems to start beaming between 2 and 3kHz.
Actually If I was building this speaker I'd probably be tempted to forget about the mid range and cross at just over 2kHz. I think the midrange qualities of the scan drivers might exceed the seas midrange - this is just a gut feeling based on the mid-bass quality of other scan drivers that I have used.
Cheers,
Nick
Bank Account Reality: Designing a passive crossover requires about $300 USD in misc R, L, and C to obtain the best crossover and tune for your room and ears. If you want an education, then this is part of the tuition. If you want great speakers for a fixed budget, you can purchase better speakers for the same total cost if you copy a proven design.
If you want to stay with the Scanspeak 18W/4531G00 woofers, then you could obtain better directivty control and upper midrange detail with a 4" midrange like the Scan Speak Discovery 10F/4424G00. A 4" is better matched to a 2.5Khz Xover. Keep the baffle around the rear of the speakers well "cut-out" to minimize side cone reflections. Put the midrange in a larger than required box and stuff to absorb rear wave. A "V" rear diverter behind the midrange center to breakup the rear wave is also common. Keep the tweeter as physically close to the midrange as possible. LR4 (frequency)crossovers would be my first design attempt. LR2 my second design attempt since the speakers have controlled SPL.
You may want to model putting the upper Scanspeak 18W/4531G00 in a sealed box for midbass transients, and the lower Scanspeak 18W/4531G00 in a ported box for deep bass.
If you want to stay with the Scanspeak 18W/4531G00 woofers, then you could obtain better directivty control and upper midrange detail with a 4" midrange like the Scan Speak Discovery 10F/4424G00. A 4" is better matched to a 2.5Khz Xover. Keep the baffle around the rear of the speakers well "cut-out" to minimize side cone reflections. Put the midrange in a larger than required box and stuff to absorb rear wave. A "V" rear diverter behind the midrange center to breakup the rear wave is also common. Keep the tweeter as physically close to the midrange as possible. LR4 (frequency)crossovers would be my first design attempt. LR2 my second design attempt since the speakers have controlled SPL.
You may want to model putting the upper Scanspeak 18W/4531G00 in a sealed box for midbass transients, and the lower Scanspeak 18W/4531G00 in a ported box for deep bass.
The bass-mid f_c looks to be slightly left of 1kHz, which should be enough for the ear not to notice any directivity issues, provided the phases are aligned properly. I expect the mids to need a slight delay w.r.t. the woofers.
BTW, there are good-sounding speakers where the X-over point is exactly there where the ear is most sensitive. It's not necessarily a problem.
Since you have two bass drivers, their added efficiency will be slightly higher than what you expect based on datasheets alone. In a real implementation, this means you might benefit from lowering the bass drivers' cutoff freq slightly w.r.t. what you have calculated here.
--
Greetz,
MatchASM
Hi,
Why don't you use a revelator midrange ?
Scan-Speak
You could have a mismatch between the drivers.
A 10F/8424G00 could be a cheaper alternative.
Can you offset the mid-tweeter ?
The crossover is Ok for me if you use a 4", 3kHz is better for a 5".
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.