Design considerations:
1. Will be using a fully active crossover.
2. Relatively high efficiency desired so that I have head room to perform various types of convolution/equalization and not be limited and get the SPL that I want out of them i.e. a little over reference level listening.
3. I desire 'fast' dynamics and transient response to steal an amorphous audiophile term. I've noticed that in my speakers an MTM design usually gives that immediate midbass impact and great transient respond.
4. I have a pair of 0.56 cubic foot parts express cabs laying around that I can leverage. These just happen to have the perfect internal volume per winisd to accommodate the two PA165 midbass drivers in a sealed configuration which is my preference to make sub integration easier.
5. I will be using these with dual subs so low bass is not needed, I just need a lot of impactful midbass down to 100hz to avoid directionality issues.
Initial thoughts:
1. Midbass driver: Dayton PA165: https://www.parts-express.com/Dayton-Audio-PA165-8-6-PA-Driver-Speaker-295-015
1. Will be using a fully active crossover.
2. Relatively high efficiency desired so that I have head room to perform various types of convolution/equalization and not be limited and get the SPL that I want out of them i.e. a little over reference level listening.
3. I desire 'fast' dynamics and transient response to steal an amorphous audiophile term. I've noticed that in my speakers an MTM design usually gives that immediate midbass impact and great transient respond.
4. I have a pair of 0.56 cubic foot parts express cabs laying around that I can leverage. These just happen to have the perfect internal volume per winisd to accommodate the two PA165 midbass drivers in a sealed configuration which is my preference to make sub integration easier.
5. I will be using these with dual subs so low bass is not needed, I just need a lot of impactful midbass down to 100hz to avoid directionality issues.
Initial thoughts:
1. Midbass driver: Dayton PA165: https://www.parts-express.com/Dayton-Audio-PA165-8-6-PA-Driver-Speaker-295-015
a. Very sensitive at 92db and change.
b. Listed xmax is 0 - how is that possible? That has to be in error right?
c. Low cost, good power handling, I have friends in the Knoxville Audio Society that state that they love the sound of this mid although we all know sound is subjective.
d. WinISD shows that they are at -5db at 100 hz and have an F3 of about 120hz with two of these drivers per 0.56 cubic foot enclosure which is the optimal volume for sealed with these drivers.
e. Since these drivers are sensitive and I will have two of them for cabinet giving me a +3db gain that puts them at around -2db at 100hz. I will leverage equalization to just drop the overall response to compensate for this (since I will have the head room) and make them relatively flat.
f. Cone break-up over 4khz, will need a tweeter that can be crossed over as low as ~4k with a steep slope or lower.
g. Since these are PA speakers that go relatively high in frequency response there might be directivity concerns. This is where things get fuzzy for me especially with an MTM configuration.
2. Tweeter OptionsPricey option if I want to play with ribbons. On sale at Madisound for 70 dollars and change right now. One of the more durable ribbon designs that goes relatively low. Pretty damn sensitive well at 95.5db.
b. Tectonic TEBM46C20N-4B BMR 3" Full-Range Speaker 4 Ohm
I honestly just wanted to play with a BMR or a FAST. Open to other options/opinions. These are the most sensitive of the BMR's at around 86db.
c. https://www.parts-express.com/Peerless-BC25SC08-04-1-Silk-Dome-Neodymium-Tweeter-with-Wav-264-1390
I have these on hand, they can be crossed low enough and they are very sensitive at 98.8db. These are giant killers and I like their sound.
d. Maybe play with vifa TC9FD-08 since they are cheap and seem to be some of the darlings of the FAST world.
d. Anything else I should be considering?
I'm thinking that I may cut the baffle in the center between both of the midbass drivers and go 'modular' i.e. create several pieces of wood of the same size as the 'cut out' center piece so that I can sub in different tweeters. Given that this is an active crossover I can level match or change my profile between tweeters at will and use rope caulk to keep it sealed between experiments.
Things I'm probably missing that I'm aware of but am fuzzy on and likely need help with:
1. MTM spacing.
2. Baffle width vs beaming width of the drivers.
3. What you don't know you don't know so there are likely other concerns.
Open to any feedback.
Thanks!
John
Nope, 0 Xmax is perfectly possible - and the clue is in your second statement: if it's designed as a mid (ie not a mid-bass) it typically won't need much excursion, and it's easier to design for high efficiency if high excursion is not needed. The description on the specsheet of "evenhung" voicecoil confirms this isn't a mistake. So, don't expect it to stay particularly clean if you want to use it as a midbass. In reality, as the magnetic field doesn't stop dead the instant the coil passes the edge of the top plate, there will probably be a mm or so of excursion available, but that won't necessarily be enough depending on your expected listening SPL.
These 2 statements contradict each other. If you already modelled both drivers (as in statement d) then you won't get an extra 3dB from using 2 drivers as that's already accounted for in the model.
Also, WinISD doesn't account for bafflestep, so expect up to 6dB more rolloff as frequency falls from the mids into the bass. That means you'd need to do a lot of EQing, which, given the extremely limited Xmax, suggests this is not the right driver for your application.
Thanks for the clarification David, I am a combat veteran as far as building documented designs, not so much when trying to design things myself but it's time to start experimenting. I don't THINK that excursion is needed, from my understanding (again, correct me here) excursion is needed to compensate for positive boost based equalization or as another way to get desired response without increasing cone size. What does the term 'evenhung' voice coil mean? I've seen the term but was not aware of it's implications so this is definitely good info. So with these drivers I won't get any 100-200 hz impact is that what this means to me in this case?
As to the statements contradicting each other I get what you're saying, the 2 x drivers will increase my overall sensitivity but it doesn't effect the actual shape of the response i.e. they'll still be -5db (of the increased sensitivity median baseline) at 100hz and have an F3 of about 120hz just +3db higher. I guess it's just another way to state it and I didn't articulate it well.
Okay, THAT is useful information regarding bafflestep. I had no idea. So are you saying that the crossover slope will increase by about 6db because of the bafflestep?
All of my EQ'ing will be negative not boosts so I'm not looking for increased excursion to perform positive eq'ing.
Looking at the frequency graph in order to make my speakers flat at 100hz I'd have to drop the entire rest of the frequency range by 5db to match the fact that it's -5db in the bass. This is the 'headroom' I was speaking of earlier, the fact that all of the drivers are pretty sensitive enables me to do cuts instead of boosts to keep myself from increasing excursion or going into distortion. Given that the woofers are 92db sensitivity this now puts me at around 87db overall efficiency if I want to use cuts to make a flat response. And with the information you're giving me regarding baffle step I might lose even more but likely not enough to jeopardize my design goal of a little over reference level, let's say 90ish db without having any significant distortion.
Am I thinking sanely or missing other variables.?
As to the statements contradicting each other I get what you're saying, the 2 x drivers will increase my overall sensitivity but it doesn't effect the actual shape of the response i.e. they'll still be -5db (of the increased sensitivity median baseline) at 100hz and have an F3 of about 120hz just +3db higher. I guess it's just another way to state it and I didn't articulate it well.
Okay, THAT is useful information regarding bafflestep. I had no idea. So are you saying that the crossover slope will increase by about 6db because of the bafflestep?
All of my EQ'ing will be negative not boosts so I'm not looking for increased excursion to perform positive eq'ing.
Looking at the frequency graph in order to make my speakers flat at 100hz I'd have to drop the entire rest of the frequency range by 5db to match the fact that it's -5db in the bass. This is the 'headroom' I was speaking of earlier, the fact that all of the drivers are pretty sensitive enables me to do cuts instead of boosts to keep myself from increasing excursion or going into distortion. Given that the woofers are 92db sensitivity this now puts me at around 87db overall efficiency if I want to use cuts to make a flat response. And with the information you're giving me regarding baffle step I might lose even more but likely not enough to jeopardize my design goal of a little over reference level, let's say 90ish db without having any significant distortion.
Am I thinking sanely or missing other variables.?
Hi, sorry for the late reply, work got a bit busy.
So, more about excursion.
All (traditional) drivers need some amount of excursion, and it increases exponentially as frequency goes down, hence mids and tweeters not needing much at all.
Getting down to 100-ish Hz or so however will start to need a bit, especially if you have to account for bafflestep. You can check the needed excursion (before bafflestep) in WinISD, there's already a graph for that.
"Evenhung" refers to the relationship between the height of the magnetic gap (AKA top plate or front pole of the motor assembly) and the height of the voice coil.
Here's a cross section of a typical driver.
In this case, the voice coil is taller than the top plate, so the voice coil can move some distance in or out and still keep the same amount of turns of wire fully in the magnetic gap, so the force generated by the voice coil stays nice and linear with respect to changing input signals. That would be called overhung.
An Even hung coil would be exactly the same height as the gap, so as soon as the coil starts moving, it no longer has the same number of turns of wire in the gap, so won't trach the input signal in as linear a manner.
This is mostly going to result in increased distortion, but also would mean that the driver just doesn't continue to get louder as the input signal is turned up.
Bafflestep: Rod Elliott has a good article on it here: https://sound-au.com/bafflestep.htm. It isn't so much an extra 6dB/Octave applied to the rolloff of the driver but rather a shelf downwards of all frequencies below a certain point.
Even if you do your compensation by cutting frequencies above the bafflestep point instead of boosting below it, the net effect in terms of the frequency response of the signal getting sent to the driver will be the same for a given SPL - it'll be having to work harder at the lower end of its range, exactly where it needs the most excursion already.
HTH,
David.
So, more about excursion.
All (traditional) drivers need some amount of excursion, and it increases exponentially as frequency goes down, hence mids and tweeters not needing much at all.
Getting down to 100-ish Hz or so however will start to need a bit, especially if you have to account for bafflestep. You can check the needed excursion (before bafflestep) in WinISD, there's already a graph for that.
"Evenhung" refers to the relationship between the height of the magnetic gap (AKA top plate or front pole of the motor assembly) and the height of the voice coil.
Here's a cross section of a typical driver.
In this case, the voice coil is taller than the top plate, so the voice coil can move some distance in or out and still keep the same amount of turns of wire fully in the magnetic gap, so the force generated by the voice coil stays nice and linear with respect to changing input signals. That would be called overhung.
An Even hung coil would be exactly the same height as the gap, so as soon as the coil starts moving, it no longer has the same number of turns of wire in the gap, so won't trach the input signal in as linear a manner.
This is mostly going to result in increased distortion, but also would mean that the driver just doesn't continue to get louder as the input signal is turned up.
Bafflestep: Rod Elliott has a good article on it here: https://sound-au.com/bafflestep.htm. It isn't so much an extra 6dB/Octave applied to the rolloff of the driver but rather a shelf downwards of all frequencies below a certain point.
Even if you do your compensation by cutting frequencies above the bafflestep point instead of boosting below it, the net effect in terms of the frequency response of the signal getting sent to the driver will be the same for a given SPL - it'll be having to work harder at the lower end of its range, exactly where it needs the most excursion already.
HTH,
David.
Thank you so much for the education! I'll dive back into WinISD, I am proficient with the basics but still have a lot to learn. That diagram was a great context that clearly showed the relationships, thank you for that as well.
Your bafflestep explanation was valid as well, so it shifts the entire response down at a certain point like a shelf filter would. Totally makes sense. So it's a hard limiter on the SPL you can get below a certain frequency range. Your explanations made total sense to me. More variables to consider.
Your bafflestep explanation was valid as well, so it shifts the entire response down at a certain point like a shelf filter would. Totally makes sense. So it's a hard limiter on the SPL you can get below a certain frequency range. Your explanations made total sense to me. More variables to consider.
98.8 at 2.83V into 4 ohms...so kind more like 95.8 depending how you look at it. Still extremely sensitive. Since you have them I can't imagine why you wouldn't try these first and let us know how it goes.
Right, my apologies, I noted the abnormal (it kills me when a vendor uses 2.83vrms instead of 1 watt at 1 meter for their measurements) but yes, extremely sensitive. I absolutely would try those first what I'm missing is the midwoofer to use along with them and my other options. David helped me out with my last prospective set of drivers by letting me know that they would have midrange output but be lacking any slam or dynamics so I excluded them from my list of options.
Well at least then you know what you are getting...doesn't Stereophile do the same? What kills ME is brands randomly boosting the sensitivity for "room gain" or "summed up as a pair" or whatever bull-oney reason: Klipsch, Zu, all other liars please line up for the whipping, to be followed by tar and feathers.
