Thanks, that was in my mind.
Yes but there are trade-offs and they aren't simple.
I want to keep the horn shallow but I want to have a wide mouth for pattern control and the sidewall angles are fixed to achieve specified cover, damn!
I actually hope to keep the crossover not so steep so I can use the extra width of the horn + woofer to maintain the vertical pattern control a bit lower in frequency.
This is not an idea I have seen discussed much, a bit by Wayne Parham.
It puts further constraints on the system, not sure how much is achievable.
Yes, I have JBL 2226 for the 15" woofers and they are pistonic to decent frequencies.
670 Hz crossover is my baseline, not least because wave is half a metre, easy to calculate for approximations.
The QSC is probably twice that price in Australia, in the rare event there is one for sale.
So I'd balk a bit at that, but mainly for the reasons already stated.
Also QSC have some proprietary network IIRC.
Maybe an RME Fireface on USB 3.
Best wishes
David
Yes, DIYSG isn’t really structured like a business, it’s structured more like a massive group buy that folks can opt in on at their leisure. Unfortunately, their offerings aren’t constantly in supply and a lot of this is due to sourcing issues. I have ordered a lot from them and still have a handful of SEOS waveguides. I’ve built many of their speaker kits and was especially fond of the Fusion 12 loudspeaker (as was my wife).
Bill Waslo (who posts here) was fairly involved with DIYSG before he moved across the country, he also uses the DNA-360 in his printed unity horn.
Hopefully the DNA-360 and the SEOS-15 will both be available again soon.
I guess I didn't make myself clear
For that I apologize nor did I intend to insult you! "I've been on a quest to build my own own reference monitor." - I assumed that most peoples take on a reference monitor, in a studio setting, as a one mans show. Ether you are in the triangle or you aren't kinda thing. Especially with intentions of using tractrix/jmlc horn. My fault for assuming.
ignoring circular waveguide/horns for a second, what merit is there for running traditional PA designed waveguides/horns that are usually setup horizontally, vertically ? especially in a domestic/home environment setup? Would the soundwaves 'blend'/integrate better with the woofer and tweeter(if there is one) as dispersion is wider on the y-axis allowing for a more suitable near/mid-field listening?
It depends on the manufacturer.
The curves in the specsheets of P.Audio and 90% of (Asian) B-brands bear little resemblance to reality, as Dr. Geddes stated.
Previously in this thread, it was shown the on-axis response curves can provide usable clues. The breakup of some 3"diaphragm B&C drivers is blatantly obvious in the specsheet. You can also make an assessment of the impact on sound quality. In the response plot 3 major spikes are visible, one of which is very high Q. This suggests multiple badly damped resonances. These will affect performance and cannot be fixed.
In the spectrogram of this driver (posted previously) the breakup (ringing) is clearly visible. Conclusion: the B&C plot is reliable in this respect.
Now if we take a look at the plot of the Faital Pro HF20AT, one medium Q spike is visible, despite the 10dB scale
This suggests the diaphragm is better damped compared to the B&C.
This is confirmed in the spectrogram. While the TAD TD4001 is superior in this respect, the Faital isn't too far behind.
The Faital HF10AK has been referred to repeatedly. Specsheet and measurement data with different horns were posted.
The 'valley' between 1800 and 3500Hz is visible in the specsheet as well as in the individual measurements.
Conclusion: Faital specsheets seem pretty reliable as well.
What's the cause of the dip?
In Faital's archive you'll find a few drivers with the same 21° conical exit section and these al show the same valley.
The dip can be fixed, but it is driver related.
I know a few professionals who are quite obsessed with the curves in the specsheets and pretty much make their buying decisions based on the presented data.
Evidently, the waveguide largely determines the off-axis response.
Sorry of topic: Regarding the HF10AK dip, isn't "the BBC dip" around the same frequencies, a harshness dip? Maybe it is intentional and part of why the driver is mentioned to sound good? Designed for smoother power response transition if used with smaller mid cone and crossed over ~2kHz? Anyone having this dip in their horn speakers eq:d or in the crossover?
Last edited:
I have found my measurement of the HF10 AK on an XT1086 again:
2-Way DiY Hifi speakers eighteensound
I did not have the opportunity to build the box that I inteded to use it in. But I think I would not lift the area of that dip but lower the bump above it and maybe take care of the one below in the crossover. And the rise at the upper end I would leave probably untouched.
Regards
Charles
2-Way DiY Hifi speakers eighteensound
I did not have the opportunity to build the box that I inteded to use it in. But I think I would not lift the area of that dip but lower the bump above it and maybe take care of the one below in the crossover. And the rise at the upper end I would leave probably untouched.
Regards
Charles
On closer look these are most likely custom built.
The finish is rather clumsy, as is the design.
The horn looks similar to Zingali's Omniray.
Zingali is quite popular in Asia.
This is a (shameless) clone of the horn:
Attachments
Last edited:
Thanks Charles, interesting comments on the 18Sound drivers in the linked thread.
It is a "castrated" subtractive delay crossover. I.e. I subtract the lowpass from a second order allpass (instead of a pure delay) with a low Q. The overall group delay will then become that of the allpass function. I.e. the group delay is not flat but it is not very high either and it does show a very gentle down-slope. I know that someone on this forum has also experimented with crossovers of this kind but can't remember who it was. I once used this type of crossover with a prototype using a Beyma coax. Because of the shape and amount of the group delay I could use a phase equaliser of only 10th or 12 th order in front of the crossover in order to achive very flat group delay up to its upper cutoff frequency. With my large MTM the crossover frequency is 650 Hz. This amounts to a crossover-induced group delay of +- 225 microseconds. It would need a phase equaliser of quite a high order to correct that so I just leave it. While this is not perfect - the amount and shape of group-delay achieved this way is below the thresholds shown in a recent AES paper.
Regards
Charles
Last edited:
I "abused" my group delay EQ spreadsheet in order to show the difference in group delay behaviour between my crossover and a LR4 crossover both at the same x-over frequency (at data point 138 on the x-axis). The frequency scale is only relative (and linear) but it is still interesting IMHO. The blue trace has no meaning in this comparison. The cyan one is my crossover and the green one is the LR4. The slopes of my crossover are less steep however with 2nd order lowpass and 3rd order highpass. But a 2nd order LR2 for instance would not be an alternative due to the even less steep slopes and the needed signal inversion of one of the filter branches.
Regards
Charles
Edit: Forgot to Mention: The Y axis is also only relative of course. For the Xover frequency of 650 Hz that I use the max GD of my topology is 450us at 0 Hz and for the LR4 it is 690us at 0 hz.
Regards
Charles
Edit: Forgot to Mention: The Y axis is also only relative of course. For the Xover frequency of 650 Hz that I use the max GD of my topology is 450us at 0 Hz and for the LR4 it is 690us at 0 hz.
Attachments
Last edited:
All of that is easy to say in hindsight but you never know in advance. The details are just missing, it can be all over the map and unfortunately the horns used tend not to be particularly good. I don't really look at that much and when there is only a one curve, I don't really give a f***.
I'm sorry if it seemed as a generalization aimed toward you...no such intention
I try not to read into what folks say, just take statements as said, and not extend them into generalizations. Communication can be a bitch, eh?
Well I’ve made some progress, whenever the board ignores my inquiries I assume either I’m asking redundant questions or questions that I can answer myself with enough research. So I dig. I emailed the creator of xdir about the spacing setting. The drivers are assumed to have 0 diameter. So it’s edge and center....thus it will show you the polar from multiple areas of a driver pair. You can see from center to center or edge to edge, it’s all valid.....thus, my assumption of the polar was incorrect about how narrow it would be in total...Which is a good thing.
I’ve been back on Winisd looking at excursion, in light of the discussion about sq and sealed woofers....3 td15ms per side? Sealed? Naw that would be too much...hmm
I’ve been back on Winisd looking at excursion, in light of the discussion about sq and sealed woofers....3 td15ms per side? Sealed? Naw that would be too much...hmm
Last edited:
@Camplo, a lot of points in post #2284, not as easy to follow where you're at.
A thought.. An MTM is supposed to take care of vertical dispersion at the crossover. The spacing between the tweeter and one woofer creates nulls which are tilted by the phase difference in the crossover, and are filled in by the other woofer's lobes.
Go down an octave and the woofers are spaced at twice the above distance and you're in the same boat without a crossover. Is it just a matter of which region gets the lobing, (or neither). Is the benefit of MTM the woofers better contribution to lower frequency room modes (not bad).
A thought.. An MTM is supposed to take care of vertical dispersion at the crossover. The spacing between the tweeter and one woofer creates nulls which are tilted by the phase difference in the crossover, and are filled in by the other woofer's lobes.
Go down an octave and the woofers are spaced at twice the above distance and you're in the same boat without a crossover. Is it just a matter of which region gets the lobing, (or neither). Is the benefit of MTM the woofers better contribution to lower frequency room modes (not bad).
JBL 2352 horn + BMS 4595ND, 2x B&C 15TBX100, XO @ 500Hz with 8th order slopes:
An externally hosted image should be here but it was not working when we last tested it.
Attachments
Last edited: