I bought a pair of ATC SCM20 Pro PSL Mk2 speakers about a year ago. They are a 20 litre 2 way studio monitor - passive.
My plan is to make them active using a dsp crossover on PC, through my RME Babyface Pro interface and out to a couple amps I already own.
The frequency I wanted to cross to the tweeter was 1.2khz, but I learned in a discussion here that the 10 inch baffle would cause a suck out starting at about 1.4khz if the tweeter was used for the upper midrange frequencies
The way I see it I have two options: raise the crossover to 1.65khz 24db/oct, or make a 15 inch wide box that's 2/3 the depth to move the suck out down to 1khz and then cross just a bit higher than planned at 1.25khz.
The closer I come to having to make the decision, the more I think the tweeter might sound better crossed at 1.65khz. It does have the excursion capability for a 1.2khz crossover point at the levels I plan to use, I'm just concerned that having it work so close to the ~935hz resonant frequency might impact sound quality. There's no Qms Qes or Qts listed anywhere for it. The tweeter is really great, but unfortunately it has almost no documentation... it's an underhung design with a 2.0B or BI, 3mm gap/2mm coil. The manual says the crossover is 2.1khz, but it's actually 1.975khz...
I took a bunch of measurements to find out how tall and wide the impedance rise is at resonance. Can this be used to find the Q? I also know that the -6db point is about 875hz.
Resistance rises to ~24 ohms at 935hz, is about 18 ohms at 1200hz, 12 ohms at 1500hz, 8 ohms at 2000hz. It seems like a wide peak - is this a good sign?
My head hurts from taking all the measurements this afternoon lol. What else can I do to determine where the lowest crossover point is which still sounds good?
My plan is to make them active using a dsp crossover on PC, through my RME Babyface Pro interface and out to a couple amps I already own.
The frequency I wanted to cross to the tweeter was 1.2khz, but I learned in a discussion here that the 10 inch baffle would cause a suck out starting at about 1.4khz if the tweeter was used for the upper midrange frequencies
The way I see it I have two options: raise the crossover to 1.65khz 24db/oct, or make a 15 inch wide box that's 2/3 the depth to move the suck out down to 1khz and then cross just a bit higher than planned at 1.25khz.
The closer I come to having to make the decision, the more I think the tweeter might sound better crossed at 1.65khz. It does have the excursion capability for a 1.2khz crossover point at the levels I plan to use, I'm just concerned that having it work so close to the ~935hz resonant frequency might impact sound quality. There's no Qms Qes or Qts listed anywhere for it. The tweeter is really great, but unfortunately it has almost no documentation... it's an underhung design with a 2.0B or BI, 3mm gap/2mm coil. The manual says the crossover is 2.1khz, but it's actually 1.975khz...
I took a bunch of measurements to find out how tall and wide the impedance rise is at resonance. Can this be used to find the Q? I also know that the -6db point is about 875hz.
Resistance rises to ~24 ohms at 935hz, is about 18 ohms at 1200hz, 12 ohms at 1500hz, 8 ohms at 2000hz. It seems like a wide peak - is this a good sign?
My head hurts from taking all the measurements this afternoon lol. What else can I do to determine where the lowest crossover point is which still sounds good?
What slope do You plan to use at 1.2 kHz? I don't know this tweeter but very few 1" domes can be crossed that low, and usually at the cost of high-ish 2nd and 3rd order distortion. 1,65 looks safer but it really depends on the tweeter.
Planned slope is 4th order
I'm more concerned with how it will sound near resonance - q doesn't matter much an octave+ above Fs. Resonance won't be attenuated much if I cross at 1.2khz.
Edit: I plan to do distortion and detailed off axis measurements of the drivers individually and as a system through the included passive crossover edit2: to tweak exactly where the crossover should go and if I need to make a new, wider box
I'm more concerned with how it will sound near resonance - q doesn't matter much an octave+ above Fs. Resonance won't be attenuated much if I cross at 1.2khz.
Edit: I plan to do distortion and detailed off axis measurements of the drivers individually and as a system through the included passive crossover edit2: to tweak exactly where the crossover should go and if I need to make a new, wider box
Last edited:
It doesn't have to be a problem if you can navigate the interactions to make sure the rolloff you have works together with the natural rolloff.
One way to do this is to put your impedance and response in a simulator, and create a target to aim for. Then you don't need to know the Q.
One way to do this is to put your impedance and response in a simulator, and create a target to aim for. Then you don't need to know the Q.
I don't have a measurement mic yet, but I did a frequency sweep with the tweeter on one side, speaker on the other - it's how I found 875hz is the approximate -6db point. The tweeter is 6db more efficient than the speaker - when the levels matched was 875hz.
To use the tweeter's natural slope with the crossover, I estimate from what I heard that I'd need to lower the crossover a bit more to 1150 or 1125hz, which might or might not be a problem with excursion with some material (electric guitar I'm thinking).
Anyway, those are things I need a microphone to find out for sure, happening soon. Harmonic distortion at frequency and amplitude will be extra helpful.
About the extremely wide impedance rise around resonance - a return to 8 ohms only at 400hz and 2khz - do you know if this is this a good thing re audibility of the resonant mode? It's a 6 ohm tweeter, too, I think it only finally falls to that by over 3khz
To use the tweeter's natural slope with the crossover, I estimate from what I heard that I'd need to lower the crossover a bit more to 1150 or 1125hz, which might or might not be a problem with excursion with some material (electric guitar I'm thinking).
Anyway, those are things I need a microphone to find out for sure, happening soon. Harmonic distortion at frequency and amplitude will be extra helpful.
About the extremely wide impedance rise around resonance - a return to 8 ohms only at 400hz and 2khz - do you know if this is this a good thing re audibility of the resonant mode? It's a 6 ohm tweeter, too, I think it only finally falls to that by over 3khz
Last edited:
Firstly, you can get the ball rolling by using manufacturer data (your measured impedance would be good), until you get your mic. This is because at these lower frequencies the tweeter plays wide and the baffle step has limited effect so to some degree this data will be similar to reality.
The impedance peak won't affect sound as long as you get the response equalised to the proper rolloff. Of course if you ignore the peak it will interfere with your attempts. One way is to design an RLC to take down the peak, another is to adjust your crossover to work with it.
Create an ideal curve in your simulator, and use whatever crossover is necessary to give you that result.
The impedance peak won't affect sound as long as you get the response equalised to the proper rolloff. Of course if you ignore the peak it will interfere with your attempts. One way is to design an RLC to take down the peak, another is to adjust your crossover to work with it.
Create an ideal curve in your simulator, and use whatever crossover is necessary to give you that result.
I don't have a simulator (I'm new to designing) - is there one you'd recommend? Hopefully free but if there's something amazing that doesn't cost a tonne that's cool too.
Everything is going to have to be measured for this driver unfortunately lol. The datasheet just has 3mm gap 2mm coil, 20,000 gauss. I guess it comes from using OEM parts
Everything is going to have to be measured for this driver unfortunately lol. The datasheet just has 3mm gap 2mm coil, 20,000 gauss. I guess it comes from using OEM parts
You do also need to consider distortion. There are very few 1" domes that will cross over below 1.8Khz cleanly.
Considering that you will use DSP, at 1.2 kHz you could try a higher order xo. Generally its a bit pointless until You get a mic to see how high is distortion at this point and what group delay You would be getting from xo. So hold Your horses and spend some time with xo simulator 🙂 You can also check XSim to use with measured response, once You get a mic.