I'm starting work on an active xo for the Proac Response 2.5. My goal is to reproduce the frequency response of the passive parallel xo as closely as possible.
This has been complicated a bit by the fact that the proac xo is not exactly textbook. There may be some debate as to what is the proper response for the active xo to target.
I would submit that the simulated (or actual) voltage seen by the drivers under passive EQ is the target voltage response for the active circuit. However, it may be better to use a graph of the audible frequency response of the actual unit rather than the raw voltage--any opinions?
A good starting point for simulation is probably the following schematic:
This has been complicated a bit by the fact that the proac xo is not exactly textbook. There may be some debate as to what is the proper response for the active xo to target.
I would submit that the simulated (or actual) voltage seen by the drivers under passive EQ is the target voltage response for the active circuit. However, it may be better to use a graph of the audible frequency response of the actual unit rather than the raw voltage--any opinions?
A good starting point for simulation is probably the following schematic:
Attachments
The response is shown here:
The tweeter has a firm 18dB/octave response. The woofer starts at a 18dB slope but tapers off a bit towards the higher frequencies.
(the audible tweeter response will be higher due to higher sensitivity, but the xover point and slope are unchanged)
The tweeter has a firm 18dB/octave response. The woofer starts at a 18dB slope but tapers off a bit towards the higher frequencies.
(the audible tweeter response will be higher due to higher sensitivity, but the xover point and slope are unchanged)
Attachments
I'm planning on using the active XO that Grey was helpful enough to provide copious documentation for, probably either with a BJT or JFET. It occurred to me that it would be best to settle any question of which response to match, though, before going to a lot of work calculating the poles of the filter!
I'd greatly appreciate any responses, especially from the experts who have studied this speaker's crossover in great detail.
(the schematic below is just to provide some point of reference for where I am going with this--no need to read anything into it!)
I'd greatly appreciate any responses, especially from the experts who have studied this speaker's crossover in great detail.
(the schematic below is just to provide some point of reference for where I am going with this--no need to read anything into it!)
Attachments
No responses yet, so I've started playing with some filters. The first set will be Bessel-alignment 3rd-order xovers. There is a real compromise between matching the electrical (voltage) slopes of the proac drivers and the summed response, as you will see. I think there is still a question of whether the slopes even should be matched, or whether a crossover frequency should simply be chosen.
On to the pics! It should be pretty obvious which are the driver responses in each set and which are the filters.
Best matching: LP at 2kHz, HP at 4kHz.
On to the pics! It should be pretty obvious which are the driver responses in each set and which are the filters.
Best matching: LP at 2kHz, HP at 4kHz.
Attachments
Now for the summed responses. It is pretty clear that the closest match is probably unacceptable. While the 3kHz/3kHz filter is certainly the flattest, it may not sound as good as the 2kHz/3kHz that more closely matches the original Proac response.
Any thoughts from the gallery? I will be experimenting with a Butterworth alignment next, which will probably even out the trough in the less "correct" filters (e.g. 2kHz/3kHz), if at the expense of some other things.
Any thoughts from the gallery? I will be experimenting with a Butterworth alignment next, which will probably even out the trough in the less "correct" filters (e.g. 2kHz/3kHz), if at the expense of some other things.
Attachments
I admire your efforts and am watching this thread closely but do not have the expertise to help out. Please keep us posted as there may be more like me.
dshortt9 said:
Well, at least one more 🙂
The ProAc 2.5 (DIY probably) is definitely #1 on the list of speakers I want to hear in combination with my (still to finish) amplifier.
Have to finish my amp first though...
At least I'll get the chance to test my protoype on some Mission 702e speakers this X-mas.
Back to this thread:
My dream setup for the moment would be:
pre-amp+active x-overs+poweramps+ProAc 2.5
All DIY if possible 😉
I don't know if you already have 2.5s, but if you don't, my advice would be: build them first (or buy them if you have the $$$).
(Of course after you have listened to them first and still want it!)
Listen to them. Get to know them. Every little quirk and detail.
Then start designing, tweaking, building active crossovers.
That way you can compare the sound, since you know them already, and can decide which filter reproduces the original sound best (or surpasses it of course 😉)
Good luck!
(And please keep us posted)
Remco
I am using ProAc 2.5 DIY clones and can report that they are every bit as good as the originals, if not better with high quality crossover components and the mods on the clone website. I'm finishing up my Aleph 2's in the next couple of weeks to complete my dream system.
Renco,
I do already have a pair of DIY 2.5s. It is just a matter now of the right active XO, and the right amps. For the moment the bass amp with be a 100W P3A and the tweeter an (inverted) gainclone, but I'm toying with the idea of a class-A tube amp or the JLH design for the high end.
I do already have a pair of DIY 2.5s. It is just a matter now of the right active XO, and the right amps. For the moment the bass amp with be a 100W P3A and the tweeter an (inverted) gainclone, but I'm toying with the idea of a class-A tube amp or the JLH design for the high end.
Hi Tiroth,
I've just read through your posts. I have also simulated the passive xo response with Electronic Workbench and got the same hump at 2Khz. I think it is partially due to the .83mH in the LP filter. If you remove this inductor the response will flatten.
The other reason for the hump is the raising response of the 18W8535 driver starting from about 1,5 Khz.
The corner points of the filters approx. are 1,5Khz (LP) and 3,8Khz (HP) anyway.
Cheers,
Zoltan
I've just read through your posts. I have also simulated the passive xo response with Electronic Workbench and got the same hump at 2Khz. I think it is partially due to the .83mH in the LP filter. If you remove this inductor the response will flatten.
The other reason for the hump is the raising response of the 18W8535 driver starting from about 1,5 Khz.
The corner points of the filters approx. are 1,5Khz (LP) and 3,8Khz (HP) anyway.
Cheers,
Zoltan
If you guys have not seen this link take a look. Troels has done a lot of work with the crossover to improve it. http://members.chello.se/jpo/2_5_clone_measurements_v4.pdf
- Status
- Not open for further replies.