I'm moving away from from my DSP processor (with it's dignity lacking sound quality) in search of better sounding, more simple, active crossover.
I will need a stereo, 4 way, 6/db oct. filters and attenuators for level matching. With such a simple, concrete configuration, it seems to me the Marchand and First Watt are way overkill.
What direction should I take with this project? Would it make sense to DIY passive line level filters followed by buffers?
Pass Labs B1 or B2 or Diamond buffer? Your thoughts?
I will need a stereo, 4 way, 6/db oct. filters and attenuators for level matching. With such a simple, concrete configuration, it seems to me the Marchand and First Watt are way overkill.
What direction should I take with this project? Would it make sense to DIY passive line level filters followed by buffers?
Pass Labs B1 or B2 or Diamond buffer? Your thoughts?
6 dB/oct is easy, just use RC filters followed by a buffer. Using opamps for this works well, although big boys tend to go the discrete route.
High pass filters may be even "easier". I have modified the DC blocking capacitor at the input of an amplifier to give the desired cut-off frequency at a 6 dB/oct slope.
But I am a bit surprised about the combination of a 4-way system with 6 dB/oct filters. Really?
(Steeper filters are also not that difficult; Sallen Key opamp filters generally work very well for this. And I've had some good experience with eBay kits here...)
Having three different DSP units over the last 10 years, I've tried a lots of configurations. 48db, 24db, 12db/octave, EQ, etc. The best long term sound on my horns -- 6db all round, nothing else. Maybe a bit of bass boost but not always.
How to determine the "R" in the RC filter? Which buffer would you recommend?
How to determine the "R" in the RC filter? Which buffer would you recommend?
Depends a bit on your setup, what is connected to your filter input and outputs.
Some useful links:
Audio Designs With Opamps -2
Filter Design and Analysis
Active Filters
Studio mixer (both BAL and SE outputs) --> line level filters (four band) --> various amps.
The selection of FOUR amps are being finalized but it's a mix of solid state and tube amps.
At some point here I will need to have attenuation of each of the channels.
The selection of FOUR amps are being finalized but it's a mix of solid state and tube amps.
At some point here I will need to have attenuation of each of the channels.
By Rod Elliott: Op amps, 4 way, 6db/octave.
Project 170
Is this a good one to build or discrete "diamond buffers" would be a better sound?
RJM Audio - The B-board Buffer Stage
>>Do you have schematics of the input sections of those amps?
I do not but maybe the two options mentioned above are good to start with?
Project 170
Is this a good one to build or discrete "diamond buffers" would be a better sound?
RJM Audio - The B-board Buffer Stage
>>Do you have schematics of the input sections of those amps?
I do not but maybe the two options mentioned above are good to start with?
The Elliot filter is a good choice. I don't think diamond buffers would give 'better sound'.
hey! I am in the early stages of building a passive 600ohm xover for my horns. I happen to like 6dB/oct slopes as well. Marchands are not constant impedance networks.
We will need custom inductors to assembly a constant impedance network and I am also considering Linkwitz Riley 12 and 24 db/oct as options and I was basically looking. We will also need PCB and a 1200/2400 ohm stepped attenuators. Once we have these basic building blocks we can assembly the thing as we like and need.
I had built 18 and 24db/oct passive filters and found out that biamping is the way to go with horns. Probably a steeper slope for the subs works better but midbass-to-mid and mid-to-high transitions sounds great with 6dB/oct once you take care of aligning the drivers physically to minimize delays.
We will need custom inductors to assembly a constant impedance network and I am also considering Linkwitz Riley 12 and 24 db/oct as options and I was basically looking. We will also need PCB and a 1200/2400 ohm stepped attenuators. Once we have these basic building blocks we can assembly the thing as we like and need.
I had built 18 and 24db/oct passive filters and found out that biamping is the way to go with horns. Probably a steeper slope for the subs works better but midbass-to-mid and mid-to-high transitions sounds great with 6dB/oct once you take care of aligning the drivers physically to minimize delays.
look here at these simulations I ran
http://www.diyaudio.com/forums/analog-line-level/300506-600ohm-passive-xover-multiamp.html
basically, the 6dB/oct filter I am planning to build is an easier load for the preamp as its impedance stays constant at 600ohm (or whatever value you size it for) while other networks do not.
http://www.diyaudio.com/forums/analog-line-level/300506-600ohm-passive-xover-multiamp.html
basically, the 6dB/oct filter I am planning to build is an easier load for the preamp as its impedance stays constant at 600ohm (or whatever value you size it for) while other networks do not.
What is the advantage of such a passive filter vs an RC filter? Is it the case that a typical pre-amp will have a hard time driving an RC filter 3-way crossover? Why not use an active crossover?
The "advantage" is that the signal does not pass any active devices, which may be considered to affect the signal in some way. The first (real!) drawback of this approach is that you will have a hard time to deal with impedance matching of the source and amplifier input. The second drawback is that it may be difficult to find suitable inductors.
Going active (i.e., buffers, opamps, etc.) removes these constraints to a large degree. And I don't believe that opamps that are suitable for the job will noticeably affect the sound quality. An active/buffered filter will certainly work and sound better than a fully passive system that is fighting with poorly matched impedances.
I do have different experience actually as I preferred the passive xover and once you have a preamp with low output (100ohm) and amps with high input (50k) you are much more than OK. I only used beringhers so maybe other active xovers sound better than my passive.
Impedance mismatch will cause insertion loss (signal will be attenuated) but as long as the filter has constant impedance over the frequency spectrum this will be a minor issue ... just cranck the volume up (I am assuming the pre can handle a 600ohm load without impairment or distortion and clearly have a constant impedance at the output).
RC networks need to be buffered because their impedance changes along the spectrum so you'll need to go active. Suitable inductors are in the 50mH to 3H range so pretty easy to source. That's why I'd like to rebuild a LC passive xover.
Impedance mismatch will cause insertion loss (signal will be attenuated) but as long as the filter has constant impedance over the frequency spectrum this will be a minor issue ... just cranck the volume up (I am assuming the pre can handle a 600ohm load without impairment or distortion and clearly have a constant impedance at the output).
RC networks need to be buffered because their impedance changes along the spectrum so you'll need to go active. Suitable inductors are in the 50mH to 3H range so pretty easy to source. That's why I'd like to rebuild a LC passive xover.
Last edited:
Just go Behringer CX3400 you can pick them up s/h for under $100
Behringer CX3400 Super-X Pro DJ/Club Crossover 2/3/4-Way XLR In & Outputs | eBay
Cheers George
Can someone build me one of these active filters?
Project 170
I ask because I can build a kit, but not a circuit like this from scratch.
Project 170
I ask because I can build a kit, but not a circuit like this from scratch.
Sure. But to do this well, it might be wise to use nicely made PCBs. Layouting the PCBs, have the PCBs made, and building the filters seems like a lot of time (in terms of money). Would you be better off to learn this kind of stuff and do it yourself?
Or maybe ask Rod Elliot if he has (or knows about) a suitable PCB?
Be careful reading fig 4
The sch as drawn gives the impression that there is a GND coming in from the input.
There is no gnd from the input, there are only two wires from the source: Signal Hot and Signal Cold.
The R2 resistor should be laid horizontally pointing towards the output.
The actual output is two wires: one from Pin1 of the opamp and the other from R2. The ground symbol at the R2 is not a ground. It is the second signal wire of the two wire signal pair.
The screen around the two wires of the balanced impedance connection goes to Chassis/enclosure, not to signal pads.
Last edited:
- Status
- Not open for further replies.