Interestingly, the 3-4 kHz bunching was there even on this foam baffle which I had made long back for studying taking measurements etc..
Would a driver like this: https://www.dibirama.altervista.org...-5h1-5cp-8-mid-woofer-5-5-8-ohm-240-wmax.html be better match with this waveguide. This one seems more directive above 2.5kHz compared to SB15CAC and about the same size diameter-wise. I happen to have two of these in hand..
. @augerpro : Thanks.. I will try out the simulation later today.
Would a driver like this: https://www.dibirama.altervista.org...-5h1-5cp-8-mid-woofer-5-5-8-ohm-240-wmax.html be better match with this waveguide. This one seems more directive above 2.5kHz compared to SB15CAC and about the same size diameter-wise. I happen to have two of these in hand..
. @augerpro : Thanks.. I will try out the simulation later today.It's not anywhere near as problematic there, the facets blend it better
It might be but I think you either need a smaller driver that can reach higher like thisWould a driver like this: be better match with this waveguide. This one seems more directive above 2.5kHz compared to SB15CAC and about the same size diameter-wise. I happen to have two of these in hand..
https://hificompass.com/en/speakers/measurements/wavecor/wavecor-wf120bd03
or get the 5" Waveguide and a router template to make the cutout bigger. Otherwise there will some significant compromise in the 3K region which can be troublesome if you are a fan of some of the less polite music there is.
Thanks fluid..
I have a pair of the wavecor wf120bd03 in hand too..
But size wise they wont fit inside this box. I really don't want to make anymore major changes to this box..
I will try to take vertical measurements, try to optimize this 2 way speaker as much as I can and leave it there.. Maybe I will use these in a casual listening setup in a bedroom or so in future.
Of late I have been liking horn type speakers more than even the smaller SICA coax. I like the LTH142 horn better even with its worse polar responses, less treble extension, and highly directive nature. There is something about them that I find addictive. I don't know if it is dynamics. What i can straight away identify is that there is some sort of a clean, big sound coming from horn-based speaker setup compared to smaller speakers partnering these woofers.. As of now, I think I might like the sound of a relatively big coax or maybe a proper ATH designed horn (something with reasonably high treble extension, maybe a good horn with a 1inch driver).
I have a pair of the wavecor wf120bd03 in hand too..
But size wise they wont fit inside this box. I really don't want to make anymore major changes to this box..
I will try to take vertical measurements, try to optimize this 2 way speaker as much as I can and leave it there.. Maybe I will use these in a casual listening setup in a bedroom or so in future.
Of late I have been liking horn type speakers more than even the smaller SICA coax. I like the LTH142 horn better even with its worse polar responses, less treble extension, and highly directive nature. There is something about them that I find addictive. I don't know if it is dynamics. What i can straight away identify is that there is some sort of a clean, big sound coming from horn-based speaker setup compared to smaller speakers partnering these woofers.. As of now, I think I might like the sound of a relatively big coax or maybe a proper ATH designed horn (something with reasonably high treble extension, maybe a good horn with a 1inch driver).
Last edited:
A 3D printed adapter ring would allow you to at least examine the differenceThanks fluid..
I have a pair of the wavecor wf120bd03 in hand too..
But size wise they wont fit inside this box.
Awesome idea.. This I will have to try out..A 3D printed adapter ring would allow you to at least examine the difference
my first contribution was just a tweak on VK's design. My usual practice is to use FIR to flatten driver response and then textbook highpass and low pass filters to form a crossover, the add PEQs to tradeoff between axial and PIR or power response. So purely out of curiosity I tried it, before seeing all your responses since my last post.
In this case, I had to do some flattening of the axial also, likely since it was a 3-way. In simulation I use Vituix mirror to emulate the FIR. I then have freedom to move the XO very easily. In this case there is a CTC issue so the lower you make it the wider the vertical window at XO. This is a nice freedom to have during listening tests to see if the XO can be set as low as I have it.
I think I did get a better result although I'm not sure how audible the difference would be. If I were going beyond simulation, I would try to reproduce this with IIR, at least using IIR for the woofers as my MiniDSP Flex doesn't have enough taps to be accurate down low.
In this case, I had to do some flattening of the axial also, likely since it was a 3-way. In simulation I use Vituix mirror to emulate the FIR. I then have freedom to move the XO very easily. In this case there is a CTC issue so the lower you make it the wider the vertical window at XO. This is a nice freedom to have during listening tests to see if the XO can be set as low as I have it.
I think I did get a better result although I'm not sure how audible the difference would be. If I were going beyond simulation, I would try to reproduce this with IIR, at least using IIR for the woofers as my MiniDSP Flex doesn't have enough taps to be accurate down low.
Attachments
The mid-tweeter CTC distance is 145 mm, and at 2k that is 0.85 wavelength. This is not ideal. However, if the crossover was 3k then that CTC distance would be about 1.27 wavelength. This might improve the DI in the troublesome region.
So, a 3k crossover may have merit for several reasons.
So, a 3k crossover may have merit for several reasons.
3k upper crossover. Now this looks like something with some real potential. I started with 3rd order, but the performance was lower than the 2k 3rd order. When I switched to 2nd order Q=.5 (i.e. LR2), things improved quite a bit. I added a bit of global EQ at 3.8k to smooth out the PIR, ER, and sound power curves. I also did some additional suppression of the SB15CAC cone resonances above 10k, since a 3k 2nd order filter does not, by itself, provide enough suppression.
The PIR line is almost flat, with a slope of -1.2 dB/octave.
The PIR line is almost flat, with a slope of -1.2 dB/octave.
I agree, the baffle shape shown in post 701 is a good match for this tweeter.
I am not ready to throw in the towel on this SB15CAC driver yet. I think it, and the cabinet, can be salvaged. My experience with it and with the larger SB17CAC35 is that they are very revealing, with low perceived distortion or harshness.
Interesting that you used FIR filters. I had lost sight of the fact that FIR may be an option open to vineeth. I am so accustomed to working with hypex fusion amps where everything is IIR.
The suggestion from Augerpro and Fluid to try 3k was a good idea.
j.
Last edited:
So I started with crossover evaluation. First tried the crossover that nc535 modified from my initial draft..
It sounds amazing, already, in mono....
This is the best system so far..
I can't wait to try out the next crossover and see/hear if there is a difference.
Just for fun, here is a clip. I know it is mostly pointless and bass is not even heard much as in real time in the recording and all other artefacts etc.. .. but still, for fun..
PS: If this can sound so good, now I also know that I have mostly screwed up somthing in the crossover implementation of the SICA coax..
It sounds amazing, already, in mono....
This is the best system so far..
I can't wait to try out the next crossover and see/hear if there is a difference.
Just for fun, here is a clip. I know it is mostly pointless and bass is not even heard much as in real time in the recording and all other artefacts etc.. .. but still, for fun..
PS: If this can sound so good, now I also know that I have mostly screwed up somthing in the crossover implementation of the SICA coax..
Last edited:
You might find that there is enough uncertainty with the merging of measurements and calibration of mics to cause the crossover simulation to be inaccurate in a negative way. The difference between horrible and great isn't always that far apart.PS: If this can sound so good, now I also know that I have mostly screwed up somthing in the crossover implementation of the SICA coax..
I'm sure that's true and a good compromise could be found. For use in this sort of three way design with the smaller waveguide I would choose the Wavecor which can fit without too much modification or the bigger waveguide can be fitted without having to do that much rework either. The cabinet doesn't need to be scrapped to accommodate either of those options. For me I see less compromise in either of those options vs trying to force the CAC driver higher, whether it is worth the hassle, much harder to say.
The beauty about having an active system and the drivers on hand means that it isn't that hard for Vineeth to test them out for himself and decide if he wants to.
Fluid - going back and rereading things, I realize I rolled past this without processing it... and now that I read it, I don't comprehend your point. Are you referring to the unevenness in the 10k to 20k region ?
No, a driver in a waveguide is rolled off compared to one on a flat baffle, this can be undone with first order filters, applying that logic in mainframe's build is here https://www.audiosciencereview.com/.../revel-m105-copy-diy-build.29465/post-1030684
Yes fluid.. Now I understand..
On top of everything, I tried to do some compensation inverse filtering thinking that the dayton Mic is completely wrong in the SICA coax crossover implementation. It looks like the minidsp mic is having some issue at the I highest frequencies.. That Mic is also 3-4 years old.
Anyway, I will have to redo the measurements with the new sonarworks mic (which is what i used in the latest measurements) to at least compare and understand what went wrong..
It can be a handy thing to try because if the capacitor is sized correctly it can be good protection for the tweeter and serve a useful part of the filtering at the same time.