Hi!
I'ts been some years now since I last designed a speaker.
This was a speaker I spent considerable time doing research on.
But in the end, I never got it to sound "right", even do I did everything "by the book" (as far as I knew).
The design was a B&C DE250 in a SEOS-12 waveguide crossed to a B&C 10HPL64 at 1600 hz to match dispersion. Crossover, baffle step and CD horn roll off was all handled by the miniDSP.
I did hundreds of measurements, and tried all kinds of possible XO slopes and XO points, but to no avail. In the end, I gave up on it and sold the whole deal.
But I'm still kind of curious how a design that theoretically should've sounded quite good just didn't.
Maybe some smart folks here on this forum could shed some theory on what I might've been the issue here, looking back in hindsight?
Picture below:
I'ts been some years now since I last designed a speaker.
This was a speaker I spent considerable time doing research on.
But in the end, I never got it to sound "right", even do I did everything "by the book" (as far as I knew).
The design was a B&C DE250 in a SEOS-12 waveguide crossed to a B&C 10HPL64 at 1600 hz to match dispersion. Crossover, baffle step and CD horn roll off was all handled by the miniDSP.
I did hundreds of measurements, and tried all kinds of possible XO slopes and XO points, but to no avail. In the end, I gave up on it and sold the whole deal.
But I'm still kind of curious how a design that theoretically should've sounded quite good just didn't.
Maybe some smart folks here on this forum could shed some theory on what I might've been the issue here, looking back in hindsight?
Picture below:
Attachments
Last edited:
It's 4 years since I sold the drivers and threw away the cabinets, so they're long gone unfortunately
This thread was meant to see if there might have been any obvious flaw I overlooked, so I can learn from it in hindsight.They sounded okey. But there was something off with them.
Kind of hard to describe, but they just sounded kind of weird.
Almost the kind of sound you get with one speaker out of phase.
Last edited:
Measurement errors, inadequate measurements (not enough angles), crossover design, build errors, or defective drivers are likely culprits. Without seeing any data who's to know?
The unterminated waveguide and surface mounted woofer may have had some effect as well. Those attributes both generally fall into the category "bad design."
Hmm, horn looks flawless.
Lack of under 100hz making it dissatifying maybe ?
61hz fs and qts .31, not much below 100hz.
I find porting above 30hz to sound "wrong", just me.
And all the boxy midrange flying out the port isn't helpfull either.
Even with eq, may not be satisfying, and woof not on floor either, maybe floor bounce dropping midbass.
I just went through that.
Perhaps you needed 24db slopes, then slide tweet forward till time aligned.
This can be a frustrating hobby initially.
And even pricy.
Lack of under 100hz making it dissatifying maybe ?
61hz fs and qts .31, not much below 100hz.
I find porting above 30hz to sound "wrong", just me.
And all the boxy midrange flying out the port isn't helpfull either.
Even with eq, may not be satisfying, and woof not on floor either, maybe floor bounce dropping midbass.
I just went through that.
Perhaps you needed 24db slopes, then slide tweet forward till time aligned.
This can be a frustrating hobby initially.
And even pricy.
This is where we get into the "art" side of things. Plus experience... Getting a short horn and 12" woofer to sing is not an easy task. Why your project did not work will always be a mystery..... However....
Suspect three things.... lack of blending .... not enough overlap. At the crossover point, your two drivers sound VERY different.... a large heavy direct radiator vs. a small very light weight, horn loaded diaphragm.
Phase issues due to different vertical and horizontal acoustic centers.
And due to very different natural roll off profiles for each driver, a VERY asymmetrical spread frequency crossover topology is required.
A lot of your dissatisfaction is typical for poorly integrated coaxial drivers, especially larger ones. Base this on working with coaxial drivers from 4" up to 18". Measurement and theory can get you moving in the right direction, but the ear, experience and a bit of luck is required to make it all work.
Suspect three things.... lack of blending .... not enough overlap. At the crossover point, your two drivers sound VERY different.... a large heavy direct radiator vs. a small very light weight, horn loaded diaphragm.
Phase issues due to different vertical and horizontal acoustic centers.
And due to very different natural roll off profiles for each driver, a VERY asymmetrical spread frequency crossover topology is required.
A lot of your dissatisfaction is typical for poorly integrated coaxial drivers, especially larger ones. Base this on working with coaxial drivers from 4" up to 18". Measurement and theory can get you moving in the right direction, but the ear, experience and a bit of luck is required to make it all work.
minidsp is one problem, I never liked the sound of processors and B&C are built for power over frequency flatness.
With a normal XO there is almost everything you can fix but it is very complicated.
I think you could have made a cardboard model before building the real thing and see if driver offsets could fix some frequency or perception problems.
Horn speakers are especially hard to XO from what I heard of them, including diy builds.
In commercial speakers they have a team of experts. send speakers to labs and test for multiple parameters with detail logs. One person cannot do that even full time for years.
If you start digging into speaker XO you will see few knows how to do it right.
so, you have 5 things which are why it went wrong:
1. Choosing PA drivers
2. Hoping DSP will fix the sound (it is more complex than DSP can do)
3. Not building a prototype before final built which can have impossible to correct flaws due to speaker shape and driver placement, offsets, cabinet type
4. Not having a proper XO modeling and log book with years of improvements
5. Not having the team and tool to measure the speakers in all parameters possible, including Z, R, time delay, port, room gain, dispersion, Frequency response.
With a normal XO there is almost everything you can fix but it is very complicated.
I think you could have made a cardboard model before building the real thing and see if driver offsets could fix some frequency or perception problems.
Horn speakers are especially hard to XO from what I heard of them, including diy builds.
In commercial speakers they have a team of experts. send speakers to labs and test for multiple parameters with detail logs. One person cannot do that even full time for years.
If you start digging into speaker XO you will see few knows how to do it right.
so, you have 5 things which are why it went wrong:
1. Choosing PA drivers
2. Hoping DSP will fix the sound (it is more complex than DSP can do)
3. Not building a prototype before final built which can have impossible to correct flaws due to speaker shape and driver placement, offsets, cabinet type
4. Not having a proper XO modeling and log book with years of improvements
5. Not having the team and tool to measure the speakers in all parameters possible, including Z, R, time delay, port, room gain, dispersion, Frequency response.
>>> any obvious flaw...
I tried to do a 10" two way using a compression driver and waveguide and had a difficult time too. The problem with my project was that I was using a cabinet that was too small. Whenever I simulated a larger cabinet I wanted to use a 12" or even a 15" which required a different crossover. Ultimately, I enjoyed the 10" two-way for a few weeks - played with stuffing and crossover parts - even moved the port from the front to the back (which reduced the boxy sound) - until finally deciding to dismantle it. The parts remain in boxes in the closet. After looking up the woofer you used, it appears to be in an 'optimal' cabinet but no matter how you use that woofer I doubt you'd get meaningful bass from it. 99db efficiency and low bass don't go hand in hand so I think you were not getting enough bass to balance out the sound. What you experienced was mostly the compression driver's sound paired with a large midrange.
>>> Kind of hard to describe, but they just sounded kind of weird.
I've never been a fan of the SEOs waveguide. Here I described the sound using the very same B&C DE250.
B&C DE250-8 Compression Driver in Various Waveguides | Speaker Projects by Zilla
I tried to do a 10" two way using a compression driver and waveguide and had a difficult time too. The problem with my project was that I was using a cabinet that was too small. Whenever I simulated a larger cabinet I wanted to use a 12" or even a 15" which required a different crossover. Ultimately, I enjoyed the 10" two-way for a few weeks - played with stuffing and crossover parts - even moved the port from the front to the back (which reduced the boxy sound) - until finally deciding to dismantle it. The parts remain in boxes in the closet. After looking up the woofer you used, it appears to be in an 'optimal' cabinet but no matter how you use that woofer I doubt you'd get meaningful bass from it. 99db efficiency and low bass don't go hand in hand so I think you were not getting enough bass to balance out the sound. What you experienced was mostly the compression driver's sound paired with a large midrange.
>>> Kind of hard to describe, but they just sounded kind of weird.
I've never been a fan of the SEOs waveguide. Here I described the sound using the very same B&C DE250.
B&C DE250-8 Compression Driver in Various Waveguides | Speaker Projects by Zilla
The Anna Karenina principle applies:
Chris
All happy loudspeakers are alike; each unhappy loudspeaker is unhappy in its own way.
In other words: in order to be happy, a loudspeaker must be successful with respect to every one of a range of criteria, including frequency response overlaps of drivers, polar coverage at the crossover regions, reserve SPL of the drivers (sensitivity matching), low distortion, bass coupling to room boundaries, the absence of box and driver resonances, etc. Failure on only one of these counts leads to unhappiness. Thus, there are more ways for a loudspeaker to be unhappy than happy.
Chris
The cabinet was sealed using Linkwitz Transform to up the low end
I tried everything from 6 db to 48 db slopes to no avail.
Might be something there. Didn't think those couple of inches would make a whole lot of difference
Tell me about it
I would've thought a 99db woofer would be dynamic enough to follow the horn.
But I see your point. How does other horn designs solve this? (if they actually do solve it)
I think I tried asymmetrical slopes, but cannot remember for sure.
So might be something there. Would've thought 48db slopes would solve this anyways tho
Was it just the bass response you were unhappy with?
Interesting! They seem to be so praised in the community in general (in my impression at least)
Yep. Although it might've been better to just do some measurements in an open field anyways? Although time gating should in theory remove the room out of the equation right? (for the midrange/high end at least)
I spent years of trial and error using different woofers to try to match up with my horns. Even with eq they all sound different. Finally I found one that just sounded right. It is my opinion that the factors that affect the Sonic character of a driver are not nearly as well understood as we like to pretend. The hard part is that it doesn't seem that one driver is always better than another but rather that the drivers in a system have to be all cut from similar cloth sonically. A woofer that sounds good with a silk dome tweeter may not mesh well with a ribbon or a horn.
(- can understand why the lowly single driver hasn't yet faded into the sunset ) DE250 can sound pretty good as used by Danley with MEH. I've run DE250 on K-tubes where there's no contouring needed due to little if any gain on the bottom and collapsing dispersion at the top. (is there such a thing as "plastic sound character" from plastic diaphragm compression drivers?)
dunno how well the transform would work subjectively with that 10. It had the ideal q for a 6th order reflex but tuning at ~60 could sound hooty. Sealed and eq-ed to 40Hz - then highpassed ??
@John Busch - would you have any general advice for making coaxial woofer/compression driver combos work well ? - that might make a good thread.
with feeble attempts and outcomes (not so much the K stuff) plus hearing mediocre commercial offerings, I sure appreciate what Henry Kloss did (in the original KLH days) with regards to making good cost effective drivers and blending them.
) DE250 can sound pretty good as used by Danley with MEH. I've run DE250 on K-tubes where there's no contouring needed due to little if any gain on the bottom and collapsing dispersion at the top. (is there such a thing as "plastic sound character" from plastic diaphragm compression drivers?)dunno how well the transform would work subjectively with that 10. It had the ideal q for a 6th order reflex but tuning at ~60 could sound hooty. Sealed and eq-ed to 40Hz - then highpassed ??
@John Busch - would you have any general advice for making coaxial woofer/compression driver combos work well ? - that might make a good thread.
with feeble attempts and outcomes (not so much the K stuff) plus hearing mediocre commercial offerings, I sure appreciate what Henry Kloss did (in the original KLH days) with regards to making good cost effective drivers and blending them.
In my opinion, there is no driver that can play bass and quality midrange at the same time. I abandoned my attempt at an Econowave a long time ago. Can you cover 5 octaves with a 12"?? In my experience, no. I have come to appreciate that the music lives in the midrange and covering 2-1/2 to 3 octaves (400 to 2400) yields the best results. Finding the 'elusive, ultimate' mid-range is my goal. One of my concerns with DIY is always chasing the next 'new' thing. I always go back to Lenard Audio articles on drivers and what they cover effectively. But it is fun and interesting to see new ideas.
Attachments
Common mistake is designing for flat at the listening location.
Next time you do a 2-way, do gated 1m measurements in situ, and start with the B&K curve as your reference. This avoids the issue of doing a quasi-anechoic, and not merging the baffle-step correctly. Even if you do use quasi, anechoic, do this as a check.
Also, use decent caps like Audyn, or Mundorf or Clarity. Even their bass line caps are decent compared to Solen/Axon/SCR.
Next time you do a 2-way, do gated 1m measurements in situ, and start with the B&K curve as your reference. This avoids the issue of doing a quasi-anechoic, and not merging the baffle-step correctly. Even if you do use quasi, anechoic, do this as a check.
Also, use decent caps like Audyn, or Mundorf or Clarity. Even their bass line caps are decent compared to Solen/Axon/SCR.
Mc Donald is very praised in the human community, i'm not sure if it is the best incarnation of the gastronomy, perhaps.
The DE250 is a good loudspeaker driver, even some of the most regarded cooking chiefs of my country have admitted that they enjoy the big mac time to time.
I you are looking for something really refined, you have to search a little more IMHO.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.