As a long time lurker and stupid question poser I thought I'd better put my £ where my mouth is.
So here are my Satorels.
SB SATori WO24P-4, mOREL Em1308 and ET448, xovers are with a Minidsp Nanodigi.
There is a load of info in the linked pictures but basically the aim was to have a xover scheme (active) that was as simple as possible with limited EQ so that I could try and replicate in passive components in the future.
Caveat - measurements are in a small 9ft x 9ft room that has horrible reflections (below 300hz) so I did my best. Measurements taken with REW and Behringer UMC2002HD and ECM8000 (generic calibration file) at 85cm from baffle on centre line and at Tweeter height. Note dB scale is only indicative as I did not calibrate this but it was at the loud end of normal listening. I did try a spinorama for each driver but they were not good due to the room. A quick check showed good correlation with manufacturer data.
Xovers are - Woofer LR24 > Mid BW18 at approx 700hz and Mid LR24 > Tweeter LR24 at approx 3200hz, small notch in the woofer to cover baffle step and tweeter to save very high end. And that's it (plus delays and some gain adjustment).
The result is a pretty flat response with fairly low distortion at loud listening levels and I am getting a small amount of reverse nulls (probably room issues), not sure about the step response though.
I also knocked up some shaping curves to see which I likes most and will listen to these over the coming few weeks and decide which I like. I would also like to do a nominally 2nd order crossover to see what the difference is (with similar crossover points and flat response).
So, will now don my flack jacket and tin hat as I am very happy to have comments made as to what's good (hopefully) and what can be improved.
Cabinet Layout - repurposed Akai SW125 (approx 40litres sealed)
Raw driver measurements - Note protection capacitors on Mid 39uF and tweeter 12uF (also BW48 filters at 200hz and 700hz for protection)
Step response to develop delay - no idea if this is the correct methoid but seemed to work - result approx 0.3-0.4ms
Filters used in Mindsp and frequency response, distortion and full system Step response (not sure this look particularly good, comments please).
Shaping Filters and response
Manufacturer data checks (noting baffle step and room on Woofoer, a buit of diffraction and protection capacitor on mid, and I may have left a filter on for the tweeter, Doh ???)
So here are my Satorels.
SB SATori WO24P-4, mOREL Em1308 and ET448, xovers are with a Minidsp Nanodigi.
There is a load of info in the linked pictures but basically the aim was to have a xover scheme (active) that was as simple as possible with limited EQ so that I could try and replicate in passive components in the future.
Caveat - measurements are in a small 9ft x 9ft room that has horrible reflections (below 300hz) so I did my best. Measurements taken with REW and Behringer UMC2002HD and ECM8000 (generic calibration file) at 85cm from baffle on centre line and at Tweeter height. Note dB scale is only indicative as I did not calibrate this but it was at the loud end of normal listening. I did try a spinorama for each driver but they were not good due to the room. A quick check showed good correlation with manufacturer data.
Xovers are - Woofer LR24 > Mid BW18 at approx 700hz and Mid LR24 > Tweeter LR24 at approx 3200hz, small notch in the woofer to cover baffle step and tweeter to save very high end. And that's it (plus delays and some gain adjustment).
The result is a pretty flat response with fairly low distortion at loud listening levels and I am getting a small amount of reverse nulls (probably room issues), not sure about the step response though.
I also knocked up some shaping curves to see which I likes most and will listen to these over the coming few weeks and decide which I like. I would also like to do a nominally 2nd order crossover to see what the difference is (with similar crossover points and flat response).
So, will now don my flack jacket and tin hat as I am very happy to have comments made as to what's good (hopefully) and what can be improved.
Cabinet Layout - repurposed Akai SW125 (approx 40litres sealed)
Raw driver measurements - Note protection capacitors on Mid 39uF and tweeter 12uF (also BW48 filters at 200hz and 700hz for protection)
Step response to develop delay - no idea if this is the correct methoid but seemed to work - result approx 0.3-0.4ms
Filters used in Mindsp and frequency response, distortion and full system Step response (not sure this look particularly good, comments please).
Shaping Filters and response
Manufacturer data checks (noting baffle step and room on Woofoer, a buit of diffraction and protection capacitor on mid, and I may have left a filter on for the tweeter, Doh ???)
@Rallyfinnen thanks for the comment, can you explain a bit more and what I need to do to reduce it. I am new to this game and willing to learn.
I guess it could be something in the filters(?) or actual physical problem with the tweeter, installation of tweeter causing reflections, or measurement setup including reflections. Could be reflections from something near the speaker/microphone, maybe even something holding the microphone giving reflections? There should be no reflecting stuff nearby when measuring. Alu-domes can have ringing above the audible band, and then it's not a major problem. It should in some way be visible in frequency response or waterfall plot though.
Normally you would delay a unit in respect to the other one, according to the distance between voice coils. Delayed by 0.3 ms=10.32 cm; 0.4ms=13.76cm. By the manufacturer sketches, these should be shorter. Deepness of the nulls depends on filter type and delay. Did you try LR all the way?
Last edited:
@Lojzek Thanks for the info - I did do the calculation for offset using good old Pythagoras for different listening distances but used the tip of the dome the mid and tweeter and the middle of the dust cap on the woofer (physically measured). At 80cm the delay including the vertical distance was 0.25ms approx so 0.3ms is not too far away from that, the values used were those that gave the deepest nulls.
I should have some time on Tuesday for more investigation and will try LR24 but the BW18 plus the protection cap gives you a good approximation to LR24 and the Morel rolls off nicely at the low end. I'll also try other delays, also with a lower delay the woofer/mid only step response did not look very good.
I should have some time on Tuesday for more investigation and will try LR24 but the BW18 plus the protection cap gives you a good approximation to LR24 and the Morel rolls off nicely at the low end. I'll also try other delays, also with a lower delay the woofer/mid only step response did not look very good.