These slopes are far from proper LR2 or LR4 curves, at least for the specified xo points. Look at that tweeter response for example, the -6dB point is around 1400 Hz instead of 2000 Hz.
As predicted, the natural driver responses are far from flat in the specified crossover regions. You need to use custom Q crossover filters or response flattening (before applying xo filters) if your goal is proper LR slopes.
As predicted, the natural driver responses are far from flat in the specified crossover regions. You need to use custom Q crossover filters or response flattening (before applying xo filters) if your goal is proper LR slopes.
Before you apply the desired filters, EQ the drivers flat in the stop-band, say 2 octaves from the intended Xo-frequency. Then, apply the filters. Beware of the tweeter and use low levels - start your sweeps at say 500 Hz - this is configurable in the Measure window in REW.
//
//
This is a good idea, a passive crossover for one component. Make your measurements with the capacitor connected. Pretend it is part of the amp.
What if you built your own amps, wouldn't you custom band limit them?TNT said:
I see now, I will retry these parameters.
I have done earlier sweeps of the drivers without any filter and run sweeps that may be done for the drivers. I'll have another go at this and then flatten them throughout the stop band.
However, I don't necessarily need LR filters, I just want to make my system sound good!
So if there are any tips, feel free to share.
Anyway, I'll start over completely
Last edited:
Stick with your initial actual x-over points. The suggestion for 1400 was more I think what came out of yur first trial - not a suggestion to change them...
//
//
LR filters are developed for loudspeaker crossovers, these are in-phase type filters (LR2 one-driver-reversed polarity is in-phase either), maximizing the reinforcement of the drivers. This way we can minimize the driver distortion, because each driver work as narrow bandwidth as possible for flat frequency response (on-axis).
Of course there are more ways to "good" sound.
Not sure what you mean by this, but LR produces the minimum required for a flat response, producing a power response dip. And since you will not get a flat response without lining up the phase, you won't compensate for lobing.
Allright, I will continue this journey in two weeks (going abroad). Final question. Dus a lowe or higher q shift the xo point or just loadness.
In other words, should I also try and adjust the freq in Hypex to realize the goal of LR4 (or 2) at the right frequencies?
In other words, should I also try and adjust the freq in Hypex to realize the goal of LR4 (or 2) at the right frequencies?
You can give it a try yourself (because it's your loudspeakers and your ears), and with DSP, it's the simplest. And remember, every crossover is a compromise in some ways.
However, based on the driver sizes (bass 5 inch, mid 5 inch with tight kevlar, tweet 1 inch), can you recommend a logical crossover point?
I mean you can get a flat frequency response without the drivers are being in phase, but this requires other slopes than LR. And if the drivers are not in phase, they are not reinforcing their sound pressure maximally each other.
Tweeter to mid point: as low as possible where the tweeter is not strained yet. 2 kHz sounds reasonably.
I wouldn't go higher than 3 kHz.
Mid to woofers: here we have more freedom because of larger wavelength, no breakup region but I wouldn't go lower than the original 350Hz, especially if you want to use LR2. This kevlar driver is a pure mid (with low xmax), not a midwoofer.
This is all IMO.
I wouldn't go higher than 3 kHz.
Mid to woofers: here we have more freedom because of larger wavelength, no breakup region but I wouldn't go lower than the original 350Hz, especially if you want to use LR2. This kevlar driver is a pure mid (with low xmax), not a midwoofer.
This is all IMO.
Last edited:
Problem or not, it depends.
If the drivers are not in phase, then we need more bandwidth from the drivers to get flat frequency response and thus more possible non-linear distortions. At least that's what I thought.
We're only talking a dB or two, if your drivers are this close to audible non-linear distortions then you have a problem either way. On the other hand, there are linear distortions to think about which might be audible.
Thanks, but given the new approach I will take, I may use an LR4 instead. I just haven't looked at the right details when designing the filters. I know know what to now.
So I'll start with an LR4 at 350 and 2000 Hz respectively and I will turn this into a proper LR4.
The way to check in REW is checking the crossover point and the slope (24 db / oct) right?
By the way, I'll upload new photos of the original filters in a sec.
Original MF/HF filter (front and back).
The exiting black/red cables go to the tweeter.
Photo front
DSC-0069 — imgbb.com
Photo back
DSC-0070 — imgbb.com
Just to let you all know, but I don't think this information is really necessary anymore.
The exiting black/red cables go to the tweeter.
Photo front
DSC-0069 — imgbb.com
Photo back
DSC-0070 — imgbb.com
Just to let you all know, but I don't think this information is really necessary anymore.
Last edited:
Just to make sure what this suggestion means: if I were to design an LR 4 with XO at 2000 Hz.
Would I flatten the frequency response for the tweeter from 500 Hz and upwards? Because the natural response of the tweeter starts descending from 1500 K and onwards. Do I boost the lower frequencies to flatten the tweeter response curve from 500 Hz up to 22,000 and afterwards apply a filter? That seems unlogical.
If you look at the natural tweeter response: IMG-20191217-WA0003 — imgbb.com of a sweep from 300 Hz to 22,000 Hz.
What steps would I take before applying crossover filters.
Put in a shelf with boost from 1 Khz and downwards, flatten it completely and then apply a 4th order LR filter?