I was hoping to accomplish a few things here.
First I would like to rebuild the crossover so that I could active tri-amp the speaker and keep the HF network intact. (Would I be correct to assume I could just build the section circled in red?)
Also at some later date just for fun I would like to try and duplicate the HF network in a DSP. What driver information would I need to understand what is being done to each set of drivers in the HF section?
First I would like to rebuild the crossover so that I could active tri-amp the speaker and keep the HF network intact. (Would I be correct to assume I could just build the section circled in red?)
Also at some later date just for fun I would like to try and duplicate the HF network in a DSP. What driver information would I need to understand what is being done to each set of drivers in the HF section?
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
With all three sections you'd need to work out what effect the current crossover is having, unless it is no good and you want to design it yourself. There are a bunch of ways to find out. Also yes, you can build part like you suggested.
The crossover is plenty good, the crossover points are LF - 350hz - mid - 1500hz tweeters (although I have no clue as to the order, slope or type)
The issue is I prefer to use active crossovers and there is only a single set of binding post, in addition to that I know that the tweeter section is separated in to 4 sections and I know for a fact that there is attenuation but not sure of any additional frequency filters.
At the time the crossover was designed it was absolutely bleeding edge technology but I am kind of curious as to what if any difference a modern day DIY person could do with modern DSP, pc design software and masurment kit available to the average joe these days.
I fully intend to active tri-amp (quad if counting my subs) but for ease and simplicity want to keep the original HF network in place. (Will probably send full frequency to HF network but will play around with different HP filters and see what I can do to mitigate any I'll effects of cascade filters and phase alignment. )
Doubt I will ever actually run 14 channels of dsp amd amplification in my system but I might try it out just to see so to speak.
The issue is I prefer to use active crossovers and there is only a single set of binding post, in addition to that I know that the tweeter section is separated in to 4 sections and I know for a fact that there is attenuation but not sure of any additional frequency filters.
At the time the crossover was designed it was absolutely bleeding edge technology but I am kind of curious as to what if any difference a modern day DIY person could do with modern DSP, pc design software and masurment kit available to the average joe these days.
I fully intend to active tri-amp (quad if counting my subs) but for ease and simplicity want to keep the original HF network in place. (Will probably send full frequency to HF network but will play around with different HP filters and see what I can do to mitigate any I'll effects of cascade filters and phase alignment. )
Doubt I will ever actually run 14 channels of dsp amd amplification in my system but I might try it out just to see so to speak.
So how do you plan to get the transfer functions.. are you going to reverse engineer in a simulation, measure the electrical signal, something else? Does it fit with your programming options?
So far my plan is to using PCD or a similar program and getting my frequency response data from REW but am unsure what tools to use to get impedance data.
I'm hoping to find some of it in online databases but doubt it.
I sent an email to the original speaker designer and will see what he comes up with.
I'm hoping to find some of it in online databases but doubt it.
I sent an email to the original speaker designer and will see what he comes up with.
Well I have the actual original crossover Networks.
What do you think would be the easiest way to get the data out of them?
What do you think would be the easiest way to get the data out of them?
You can use REW if you have it. Sweep the inputs to the speakers like you were going to measure them. Measure the response directly at each speaker terminal in turn. You don't need a mic, and take care with your soundcard inputs in the process.
The result of each measurement will be what you have to copy/reproduce with your filters.
The result of each measurement will be what you have to copy/reproduce with your filters.
Connect the Woofer and Mid directly to the output of the amps that will be driving them.
Are you going to put the amps inside the speaker cabs or have them separate.?
If you put them inside you will need connections for signals and power.
If the amps are separate then you will need three sets of connections to the speakers .
Either way you can't get away with just using the existing set of binding posts.
If you have two stereo amplifiers you could hook up a mono four way set up to try it out.
As a starting point set the DSP crossover frequencies to suit the response of your sub to woofer, 350Hz, and 1500Hz.
Then you can make any adjustments depending on the results of the measurements that AllenB suggested.
Sid.
Thank you sid.
That's kind of my plan at this point. I will be using external amps, LF crossed over with sub's, LF crossed over to mid at 350hz and a low pass for the mids at 1500hz
As for the HF I will start with full range and let the passive HF network do its thing and depending on what dsp I I end up with either leaving it full range, setting a much lower conventional high pass or using a FIR high pass in order to covid a hole in my frequency response due to cascaded filtering. The issue and need to leave the high side intact is due to the HF consisting of 16 drivers per side and they are separated in to groups of 4 as can be seen in the 2ed diagram in my original post. I do know for sure that the 4 groups have different amounts of power to each group but do not know if they have different HP and or LP filters for each group until I learn more about what the original crossover does.
That's kind of my plan at this point. I will be using external amps, LF crossed over with sub's, LF crossed over to mid at 350hz and a low pass for the mids at 1500hz
As for the HF I will start with full range and let the passive HF network do its thing and depending on what dsp I I end up with either leaving it full range, setting a much lower conventional high pass or using a FIR high pass in order to covid a hole in my frequency response due to cascaded filtering. The issue and need to leave the high side intact is due to the HF consisting of 16 drivers per side and they are separated in to groups of 4 as can be seen in the 2ed diagram in my original post. I do know for sure that the 4 groups have different amounts of power to each group but do not know if they have different HP and or LP filters for each group until I learn more about what the original crossover does.