Have you ever used a damped order filter? Try adding a 2 ohm resistor on either the .18mH coil, or the 10uf cap. One or the other may improve the alignment.
That's too long for typical indoor measurements. Show us the impulse response data, which is more useful to assess the gating.
Whoa, you mean to tell me the measurement conditions I stated of being right next to a highway, on a 10ft ladder isn't anechoic? Boy am I shocked lol.
Show us the impulse response data, which would be more meaningful so assess the echoes.
The sharp peaks at 3.75 ms, 4.6 ms and maybe also at 4.1 ms look a bit funky. What are they? Reflections?
Could be, who knows. Response doesn't really change with gating so I'm gonna move past whatever thing you're focusing on as I don't feel you've really been reading the information provided in the thread.
Lol sure, whatever. Thank you for leaving, you were already set to ignore and likely for a good reason. Bye bye 🙂
I don’t think the peaks should be problematic when designing a crossover. They are however to assess the system and/or components.
You ask us to rate the crossover. I’d say you don’t need the notch on the woofer coil nor the zobel (?) on the tweeter.
The proof of the pudding of course is in the eating. Build the crossover and measure the system. Then rate it yourself, I guess…
You ask us to rate the crossover. I’d say you don’t need the notch on the woofer coil nor the zobel (?) on the tweeter.
The proof of the pudding of course is in the eating. Build the crossover and measure the system. Then rate it yourself, I guess…
Regarding Filter:
I assume you are using LR4 crossovers, given the DEEP reverse null. LR4 gives a bump in the power response at the crossover but provides good phase matching. As I know you are interested in smooth power and in-room response, try 3rd order crossovers (Butterworth or Linkwitz-Riley). Simulate and save the different variations of 3rd order using xo points of 2000, 2400, 2800, and 3200. Now compare the Power and PIR and see which ones are getting you the best results, pick the one you like best and tweak from there. Note, however, you will not get a deep reverse null, but you can keep an eye on the phase of the individual drivers near the crossover point.
Measurements:
You have received a lot of feedback on your measurements because it "looks" like there may be issues. One way to check...wire up one of your crossovers, make sure all of the inductor DCR's are correct in VituixCAD, measure at 2m, and load the FRD into VituixCAD as an overlay. VituixCAD is very good IF you give it good inputs. If everything is good, you should be within about 0.5dB or less, between your measurement and your simulation, 20kHz down to 300Hz or so.
If things do NOT match well, then it is your measurements. If that ends up being the case, you should start a thread about the measurements and tell us more about the baffle dimensions, mic, REW settings, etc.
I assume you are using LR4 crossovers, given the DEEP reverse null. LR4 gives a bump in the power response at the crossover but provides good phase matching. As I know you are interested in smooth power and in-room response, try 3rd order crossovers (Butterworth or Linkwitz-Riley). Simulate and save the different variations of 3rd order using xo points of 2000, 2400, 2800, and 3200. Now compare the Power and PIR and see which ones are getting you the best results, pick the one you like best and tweak from there. Note, however, you will not get a deep reverse null, but you can keep an eye on the phase of the individual drivers near the crossover point.
Measurements:
You have received a lot of feedback on your measurements because it "looks" like there may be issues. One way to check...wire up one of your crossovers, make sure all of the inductor DCR's are correct in VituixCAD, measure at 2m, and load the FRD into VituixCAD as an overlay. VituixCAD is very good IF you give it good inputs. If everything is good, you should be within about 0.5dB or less, between your measurement and your simulation, 20kHz down to 300Hz or so.
If things do NOT match well, then it is your measurements. If that ends up being the case, you should start a thread about the measurements and tell us more about the baffle dimensions, mic, REW settings, etc.
Took a leap of faith and bought some passive xover parts. I have to say I'm very pleased with the result. The speakers sound great and measure very close to my sim. No offense but the folks saying "you don't care about your measurements" can go take a seat in the corner.
Took a quick one measurement in room, mic pointed at speaker, not the cleanest data but I won't have time to take them outside for awhile. The reverse null looks great. Subjectively they sound very nice. I will hopefully be taking these to PE's competition. Not exactly a ground breaking award winning speaker but I'm happy with the results and proud of my work. More importantly I feel a sense of confidence in my measurement and filter design that I am eager to work on more passive speakers.
Xover cost was around $20 per speaker. Cost of drivers and wood puts these at just under $200 using MDF.
Took a quick one measurement in room, mic pointed at speaker, not the cleanest data but I won't have time to take them outside for awhile. The reverse null looks great. Subjectively they sound very nice. I will hopefully be taking these to PE's competition. Not exactly a ground breaking award winning speaker but I'm happy with the results and proud of my work. More importantly I feel a sense of confidence in my measurement and filter design that I am eager to work on more passive speakers.
Xover cost was around $20 per speaker. Cost of drivers and wood puts these at just under $200 using MDF.
@wafflesomd You've gone to quite a bit of trouble in setting up a reasonably "anechoic" outdoor measurement environment. It seems reminiscent of something like this:
From: Moir, James (1980). Measuring Loudspeaker Responses. Hi-Fi News & Record Review, October, pages 71–73.
For a 10-ft (3.05 metre) ladder mounting outdoors, that should allow for the first ground reflection to be occurring at around 17ms. You could use a 15ms gate time to give you some low-frequency measurement accuracy down to around 70Hz or so.
The abovementioned sharp peaks in the impulse response could be due to cabinet diffraction effects of some kind. The delay time of the first peak is 3.75−2.926=0.824ms, which corresponds to a distance of just 0.28metres (0.93ft). Is that distance associated with any of the cabinet dimensions and driver spacings?
By "bump", do you mean that a peak is introduced into the power response? An LR4 crossover introduces a 3dB dip in the power response at the crossover frequency.
If the crossover point is at, for example, 80Hz, and the two (effectively coincident) sound radiators are assumed to be omnidirectional monopoles, won't there be a dip in the directivity response at 80Hz? Or maybe I'm getting things back to front?
Last edited:
When they add to +6dB, each radiator is at quarter power at the cross compared to their own passband.
Last edited: