Technically I could run two way, but I do not think it will be optimal. The mid driver has about 10% of the mass of the woofer. I did not try high crossovers yet, I guess I should not cross above 400-500 Hz.
I tried two way LR4 at 950 Hz on my previous 15" woofer, Pure audio project neo. Here I found that three way sounds better, especially at higher levels. And I want to use Harsch XO which requires a mid anyhow given the 2nd order slope.
I tried two way LR4 at 950 Hz on my previous 15" woofer, Pure audio project neo. Here I found that three way sounds better, especially at higher levels. And I want to use Harsch XO which requires a mid anyhow given the 2nd order slope.
Design from backside:
I have not mounted the diagonal wings yet, only the lower wings. Speaker is made in cheap MDF now, only mono. Next step is to build the speakers in 2x18mm Birch multiplex. Also the wings will be extended a bit for somewhat higher bass efficiency. I will make the final filters in this configuration, but will continue with experiments with the test baffle for now...
I have not mounted the diagonal wings yet, only the lower wings. Speaker is made in cheap MDF now, only mono. Next step is to build the speakers in 2x18mm Birch multiplex. Also the wings will be extended a bit for somewhat higher bass efficiency. I will make the final filters in this configuration, but will continue with experiments with the test baffle for now...
The woofer better near the floor, you will gain some bass .
need to be at last 6db more than mid for OB
I considered to put the woofer lower, but this also has some downsides. I wanted to put the woofer close to mid for better integration and phase alignment (and easier measurements). Also, higher up are less obstructions, e.g. the couch. Finally, visually I prefer the current config.
But yes I may lose a few dB's...
Still, the woofer has high power capability and a very high xmax so I can boost bass a lot.
But yes I may lose a few dB's...
Still, the woofer has high power capability and a very high xmax so I can boost bass a lot.
I certainly hope so. I have spent quite some time to try to find the (in my eyes) right components for this build. They should be in the higher end of what is possible. I attached a photo of the backside of the speaker (in the meantime I have improved the wiring...
):
It would indeed be nice to be able to simulate with measured driver frequency response and phase data. Do you have any reference to the optimal procedure to follow here? Would a Umik1 with REW do or do I need a loopback?
I plan to spend some more time tonight to debug the phase issues I have. This week I measured a very deep dip at 260-280 Hz @ 1m distance, although I am quite sure that the drivers individually have the right phase setup. It looks like that the filters or phase shift of low/mid drivers results in almost exactly 180 degree phase shift...
Adding roughly 600 us delay to mid and tweeter solves it largely, though I did not search the optimal value yet.
To illustrate my point, hereby some in room measurements at 1m distance, all at AMT height. Harsch XO is done at 300 Hz and 1600 Hz, delays are only set to Harsch XO values (no driver delay compensations).
Non-inverted:
Clearly a phase mismatch is around 300 Hz. Step response is not a nice traingle.
Inverted:
Better phase alignment at 300 Hz. New dip at 500 Hz. Nice step triangle, but strange start of step response, probably due to HF in wrong direction.
I will do some more experiments with changed delays...
Non-inverted:
Clearly a phase mismatch is around 300 Hz. Step response is not a nice traingle.
Inverted:
Better phase alignment at 300 Hz. New dip at 500 Hz. Nice step triangle, but strange start of step response, probably due to HF in wrong direction.
I will do some more experiments with changed delays...
I just tried what mid delay would result in the best crossover at 300 Hz (measured at height between mid and woofer, roughly 1m distance). The best value is around 3200 us (1664 theoretical for Harsch XO). Next, I measured system again at 1m, AMT height.
The step response is no improvement...
(the additional delay would also result in roughly 180 degree phase shift at 300 Hz, T = 3.33ms)
The step response is no improvement...
(the additional delay would also result in roughly 180 degree phase shift at 300 Hz, T = 3.33ms)
Last edited:
After an evening filled with delay related measuring experiments, it looks like I got my test open baffle system in a better shape. I made a LR4 preset with exactly the same driver compensation EQs as my Harsch XO and searched for deepest dip around 300 Hz using inverted woofer. The mid delay needed to be 405 us compared to woofer. Applying this delay to Harsch XO (added to mid+tweeter) gave satisfactory results. Next, I spent some effort optimizing the frequency response curve.
Response is now:
Mid response is still up for improvement, but no extreme things anymore.
Nice sharkey curve eh?
Response is now:
Mid response is still up for improvement, but no extreme things anymore.
Nice sharkey curve eh?
Thanks for the feedback XRK. I am glad I sorted this out. I added cushions on the floor during measurements to be less confused by reflections.
Still, I need to get used to the new sound a bit. On first (low level) listen it became more soft, perhaps due to missing peaks or improved timing. Though I need to listen longer for final conclusions.
What would you (or anybody reading this) recommend for woofer placement? I did not CNC the new baffles yet, so in principle I can still change the design a bit. In current placement, the woofer and mid centers are roughly at quarter wavelength of 300 Hz XO, helping proper integration. Lower placement, near the floor, would probably raise efficiency with 2-3 dB, also nice (as nicoch also commented). But I am afraid that it will be audible that lower freq comes from the floor...
Still, I need to get used to the new sound a bit. On first (low level) listen it became more soft, perhaps due to missing peaks or improved timing. Though I need to listen longer for final conclusions.
What would you (or anybody reading this) recommend for woofer placement? I did not CNC the new baffles yet, so in principle I can still change the design a bit. In current placement, the woofer and mid centers are roughly at quarter wavelength of 300 Hz XO, helping proper integration. Lower placement, near the floor, would probably raise efficiency with 2-3 dB, also nice (as nicoch also commented). But I am afraid that it will be audible that lower freq comes from the floor...
Last edited:
I think it is not so practical to put an ESS AMT on a different spot as on top. Delay is no problem with DSP anyhow. So I would like to keep WMT, only question is the height of W. My current feeling is to keep locations as is and only add longer wings to raise sensitivity. Perhaps a few years later I will buy a 2nd pair of woofers to get 6 dB more. But not required now, I listen mostly low volume...
Flexunits 15
I cannot follow you. Do you propose to lower woofer a bit to find a compromise on sensivity, integration and looks?
Nice to see this thread is still going strong.
I often return to Harsch for topology between midwoofer and treble. Somehow it results in less fatiqued in my ears than ie. LR12.
Harsch makes me want to listen to the music, letting me forget to adjust things forever. Harsch is in my experience less "in your face", and may be a tad less detailed than LR12. For the lack of better words, it reminds me a little bit about a analog glow in the music, that I have yet to experience with LR12.
I have a theory that with a slanted baffle, Harsch may disperse less sound to the ceiling, hence the sound from this 1st reflection is less prominent.
Can someone point to a diagram showing the lobe of Harsch-filter. May have been posted in this thread?
I often return to Harsch for topology between midwoofer and treble. Somehow it results in less fatiqued in my ears than ie. LR12.
Harsch makes me want to listen to the music, letting me forget to adjust things forever. Harsch is in my experience less "in your face", and may be a tad less detailed than LR12. For the lack of better words, it reminds me a little bit about a analog glow in the music, that I have yet to experience with LR12.
I have a theory that with a slanted baffle, Harsch may disperse less sound to the ceiling, hence the sound from this 1st reflection is less prominent.
Can someone point to a diagram showing the lobe of Harsch-filter. May have been posted in this thread?
Last edited:
I think that Harsch XO sounds less "Hifi" than LR24. Hifi in the less positive meaning as somewhat artificial and flat...
I have the impression that Harsch XO makes music more interesting and fun to listen to...
But as I did not do any scientifically proven AB experiments, I may be deluding myself...
I have the impression that Harsch XO makes music more interesting and fun to listen to...
But as I did not do any scientifically proven AB experiments, I may be deluding myself...
with an incident from one of my 2x4HD got bricked, i have set my previous 4way into 2way, additional beston ribbon filtered with 3.3uf to add more sparkle. xo is set to 240Hz and being 2way is still good sounding setup and less complicated
i made measurement on prv 5" and gave several adjustment to make it flat, i find that without eQ is better.
I want to utilise my audionirvana classic 8" with some of my class a amplifier (prv 5" is powered by myreff gainclone) but not using dsp, so my idea is to make additional fullrange OB pllxo bessel to match existing dsp filter. i think i have 2 options:
1. AN 8" is placed to cover 12"excess hole but no delay, any advice whether to keep Harch xo or maybe change to another type?
2. AN 8" on separate baffle and placed above my subwoofer, there is about 60cm difference behind woofer which i think equivalent to my dsp delay about 2ms. but is it OK to place them separate baffle?
thanks
i made measurement on prv 5" and gave several adjustment to make it flat, i find that without eQ is better.
I want to utilise my audionirvana classic 8" with some of my class a amplifier (prv 5" is powered by myreff gainclone) but not using dsp, so my idea is to make additional fullrange OB pllxo bessel to match existing dsp filter. i think i have 2 options:
1. AN 8" is placed to cover 12"excess hole but no delay, any advice whether to keep Harch xo or maybe change to another type?
2. AN 8" on separate baffle and placed above my subwoofer, there is about 60cm difference behind woofer which i think equivalent to my dsp delay about 2ms. but is it OK to place them separate baffle?
thanks
