@hifijim , sorry for offtopic posts, though they are partially ontopic and valid generally.
@mV8, if drivers allow and there are no cencerns regarding breakups, distortion, offaxis irregularities, then I am all in favor of the simpler crossovers, not only for practical and budget reasons, yet also for the good sound. Of course, overly simple crossovers, not addressing drivers issues or resulting in on axis or off axis strong irregularities, are not what I meant. LR2 acoustic, it is what I find well balanced crossover overall.
@mV8, if drivers allow and there are no cencerns regarding breakups, distortion, offaxis irregularities, then I am all in favor of the simpler crossovers, not only for practical and budget reasons, yet also for the good sound. Of course, overly simple crossovers, not addressing drivers issues or resulting in on axis or off axis strong irregularities, are not what I meant. LR2 acoustic, it is what I find well balanced crossover overall.
I fully agree.
Unfortunately, too many times I have had situations where the simulation showed great performance, the speakers used had very low distortion, but the sound did not suit me at all.
https://david-crystal.blogspot.com/2009/12/on-being-tragic.html
I would agree with the first part. The DI can be flat or rising with a waveguide (or anywhere in between) depending on the curve used and the baffle and edge treatment used. With very wide directivity drivers like the Bliesma Aluminium the baffle has a big influenece on the amount and slope of the directivity. A baffle like Jim's prototype with dual slopes will create some 2pi like load and cause the directivity slope to rise less quickly than it would if the edge of the cabinet was much closer to the driver.
T25A matches the inherent directivity of the M74A the closest if the target is to create the widest directivity to the highest frequencies. That is very different to the Satori tweeters that have a much stronger natural directivity at the highest frequencies whatever you do to them. The Vituix graph I posted before is a waveguide simulation for an SB29TX. It is small, the same size as the faceplate and does a better job of creating a smoothly rising directivity than the factory faceplate does, the slope is still quite steep which is hard to avoid given a shallow dome and wide surround.
The T25B is somewhere in between, more directional than the T25A less that the Satori.
All of them can be made to match quite well and are of a similar quality.
I think the choice would come down to how wide the highest frequencies are desired to be.
I imagine that would probably still be too directive to match the M74 perfectly at a higher crossover frequency. I had to use a 5mm deep waveguide to match T25B and M74B for someone else and the M74B is more directive than the M74A as frequency rises.
For a really good match yes. Small and shallow can be relative terms but anything designed to match with a larger driver is probably too much.
Any tweeter can be made to be slightly better behaved than a bare driver with a replacement face plate or additional one. It does need to be simulated in the proposed baffle. Without the ability to simulate it properly or testing multiple prototypes it may be better to just go with a bare driver in a contoured baffle and be done.
Yes, I can't wait to use this with regards to all my DIY speaker tragic friends.
The online dictionary is paywalled; so you’ll have to read an excerpt I snapped here from my
Macquarie Dictionary 7th Edition (2017) Volume 2 (L-Z):
Macquarie Dictionary 7th Edition (2017) Volume 2 (L-Z):
Last edited:
I was recently exposed to a similar colloquialism by a younger relative. We were talking about the new car, a 2024 Corvette Z06, and he said "that super charged Z06 engine is sick, and the tires and brakes are just stupid"... Without the context of the conversation, I would have misunderstood his meaning, but based on his tone of voice and his general admiration for the sports car, I understood the "sick" and "stupid" now mean "amazing" and "outstanding". Word meanings change over time...
Not MY new car... It belongs to a friend who is corvette tragic 🙂
I completed the prototype cabinet (version 1) for testing the midrange driver. This cabinet represents the outer mold line shape of the current design. The baffle is plywood, and the sides are XPS foam board. The back is open and stuffed with foam. The purpose of this prototype is to collect polar response scans of the midrange driver to assess diffraction and baffle step.
I experimented with a tapered bevel to transition from the square-edged lower box to the large-beveled upper box. I tried several methods to cut this, and the one I liked the best utilized the band saw. I think with practice I can do a better job and generate a softer transition.
My current concept is the upper baffle and the transition block will be solid wood, the rest of the cabinet will be plywood with veneer.
I should have some measurements tomorrow.
I experimented with a tapered bevel to transition from the square-edged lower box to the large-beveled upper box. I tried several methods to cut this, and the one I liked the best utilized the band saw. I think with practice I can do a better job and generate a softer transition.
My current concept is the upper baffle and the transition block will be solid wood, the rest of the cabinet will be plywood with veneer.
I should have some measurements tomorrow.
As I have said several times before, making good measurements is hard. By “good” I mean measurements which are repeatable and correctly performed so that they can be used in downstream software. By “good” I mean measurements which have an acceptable level of precision and accuracy to allow good design decisions.
With that in mind, I am going to present some measurements which have an error, to illustrate what can go wrong. I did not notice the error until I was documenting the results, and some data trends did not make sense.
First I will show the measurements I got right: The near field scan, and the the near field and far field burst decay plots.
These plots above indicate that the driver is functioning properly, and the prototype cabinet is not introducing any gross resonances which might pollute the data.
Now I will show some data that does not make sense.
I made horizontal polar scans in my large room, with the driver elevated to 146 cm, and a distance of 100 cm. This height gave me a window of 5.7 ms. I made scans in 15 degree increments, 0 – 180 degrees, and I also made scans in the other direction, from -15 to -165 degrees (24 total FR scans).
Later, while I was assimilating and validating the data, I noticed the positive angular scans to the right were different than the negative angular scans to the left. The shape of each curve was slightly different, which is not too unusual, but the overall levels were different as well. There is nothing about the driver, the baffle, or the cabinet that cause me to expect this kind of difference. I would expect left and right to be the same.
This is most easily noticeable in the polar map plot. There is an obvious non-symmetry for positive and negative angles.
The impact of this error is most easily seen in the Power and DI plot. The first plot uses both right and left rotations. The second plot uses only the left rotation (negative angles). There is a clear difference in power response and DI curve.
What caused this error? Obviously I need to re-do the measurements, but I need to understand what I did wrong in order to avoid a repeat. I am still investigating it, but I believe a contributing factor was that the speaker was tilting during rotation, and the amount of tilt was different for left and right rotations. When I examine the impulse responses, I see that the 90 degree right rotation arrives 0.042 ms before the 0 degree scan (14.4 mm). The 90 degree left rotation arrives 0.115 ms ahead of the 0 degree scan (39.4 mm). I blame the soft carpeting on the floor for the variable tilt. I will re-do the scans on a hard floor, and I will check vertical alignment at multiple rotations. There may be other factors at play.
j.
With that in mind, I am going to present some measurements which have an error, to illustrate what can go wrong. I did not notice the error until I was documenting the results, and some data trends did not make sense.
First I will show the measurements I got right: The near field scan, and the the near field and far field burst decay plots.
These plots above indicate that the driver is functioning properly, and the prototype cabinet is not introducing any gross resonances which might pollute the data.
Now I will show some data that does not make sense.
I made horizontal polar scans in my large room, with the driver elevated to 146 cm, and a distance of 100 cm. This height gave me a window of 5.7 ms. I made scans in 15 degree increments, 0 – 180 degrees, and I also made scans in the other direction, from -15 to -165 degrees (24 total FR scans).
Later, while I was assimilating and validating the data, I noticed the positive angular scans to the right were different than the negative angular scans to the left. The shape of each curve was slightly different, which is not too unusual, but the overall levels were different as well. There is nothing about the driver, the baffle, or the cabinet that cause me to expect this kind of difference. I would expect left and right to be the same.
This is most easily noticeable in the polar map plot. There is an obvious non-symmetry for positive and negative angles.
The impact of this error is most easily seen in the Power and DI plot. The first plot uses both right and left rotations. The second plot uses only the left rotation (negative angles). There is a clear difference in power response and DI curve.
What caused this error? Obviously I need to re-do the measurements, but I need to understand what I did wrong in order to avoid a repeat. I am still investigating it, but I believe a contributing factor was that the speaker was tilting during rotation, and the amount of tilt was different for left and right rotations. When I examine the impulse responses, I see that the 90 degree right rotation arrives 0.042 ms before the 0 degree scan (14.4 mm). The 90 degree left rotation arrives 0.115 ms ahead of the 0 degree scan (39.4 mm). I blame the soft carpeting on the floor for the variable tilt. I will re-do the scans on a hard floor, and I will check vertical alignment at multiple rotations. There may be other factors at play.
j.
Rotation axis was not at acoustic center and speaker tilted too as you suppose?
Anyway difference is within error marginal and can be ignored IMO. Peaks and dips happen at same freq, only amplitude is different.
Anyway difference is within error marginal and can be ignored IMO. Peaks and dips happen at same freq, only amplitude is different.
Well, the differences between the Left rotation measurements and the Right rotation measurements are significant enough to influence the tweeter selection. In one case the directivity index from 2k-5k rises from 5 dB to 8 dB. In the other case it rises from 4 dB to 5.5 dB. That is a big difference in terms of matching this driver up with a tweeter.
j.