If the perfect midrange is not round then we could deal with that. Simplicity is nice but expanding ones woodworking skills is always good. I, personally, am never going to have a driver not flush mounted 🙂
In my opinion, the tweeter needs to be flush mounted, and any driver next to the tweeter needs to be flush mounted. So for a vertically aligned 3-way with the tweeter on top, the tweeter and mid get flush mounted.
In a recent project, I worked up an evaluation of the SB12MNRX2-25-4. I liked this driver. The sensitivity was noticeably higher than a typical 5" midwoofer. Frequency response was nice, but for a dip centered around 2k. Distortion was lower than I expected. It is a really good driver for the money.
SB12MNRX link
In a recent project, I worked up an evaluation of the SB12MNRX2-25-4. I liked this driver. The sensitivity was noticeably higher than a typical 5" midwoofer. Frequency response was nice, but for a dip centered around 2k. Distortion was lower than I expected. It is a really good driver for the money.
SB12MNRX link
I saw that SB12MNRX last night when I was looking through drivers. What drew my attention was that I didn't know what the "M" was for and thought it was a typo. It does look very nice and should definitely stay on the "possibilities" list. I am partial to a 5" driver as kind of a sweet spot, but I'd really like to hear others input on that. I do think a 6" is probably ruled out, just due to directivity issues without a waveguide. Yes, I know it can work, but I'd like to make things simpler where we can.
That is actually a Dayton that is reasonable priced in Europe at 40-45€ so a candidte. It's specs are also not bad at all.
Hopefully someone will correct me if I am wrong here, I know there are several people contributing to the thread who know this better than I do.
I assume the SLS 12" has 90 dB/2.83v sensitivity. (In reality I'd even expect it a little lower, because speaker companies don't usually publish conservative data. Upon actual measurements, it is
Baffle step loss will be 4-6dB. I would normally assume 6dB, but since these will be a little closer to the floor and a little closer to the front wall in most cases, 4dB is probably realistic. So, our system SPL will be 84-86dB.
The midrange of a 3-way tends to have 1-2dB boost once a crossover is in place and taking into account the overlap with the woofer and tweeter output. So in a optimistic scenario, maybe the midrange would work with 82-85dB sensitivity. So I picked 86db as the minimum we should look at. I believe all of the ones on my initial list meet this. And in fact, this is not very demanding of a 5" driver so as long as we pay attention to the sensitivity, this isn't going to be a very limiting factor in our selection.
I agree with your analysis.
The mid will be operating in 2-pi space except at the very bottom of its range near the crossover (200-500 Hz). The filter Q of the woofer low pass and mid high pass can be tweaked to get a flat response even if the mid lacks a bit of sensitivity in this region... in other words, if the mid response is falling off due to baffle step, the electrical Q of the woofer-mid crossover can be increased up to 0.7 or even 0.8 and we can recover some of the lost sensitivity.
j.
The mid will be operating in 2-pi space except at the very bottom of its range near the crossover (200-500 Hz). The filter Q of the woofer low pass and mid high pass can be tweaked to get a flat response even if the mid lacks a bit of sensitivity in this region... in other words, if the mid response is falling off due to baffle step, the electrical Q of the woofer-mid crossover can be increased up to 0.7 or even 0.8 and we can recover some of the lost sensitivity.
j.
I briefly looked at BS with 37 cm width baffle around 200 Hz cut off (so I assume acoustical in the baffle step simulator), It was already progressive 4 dB loss for the woofer below that number.
I assume more a 5 dB loss from what I see in Vituix as the crossover should be nearer from 250/300 Hz acoustical in reality (that means a lower electrical cut off there).
At 300 Hz, always at looking Vituix approximation BSC (so w/o filters) the mid driver above that 300 Hz to 500/600 Hz as only 2 to 3 dB loss.
There are often a sligth plateau bump from that 500/600 hz til 1500 Hz (which is a mess btw in the case of the NE149W if you look at the IEC measurement as it has already a bump that is hard to solve w/o notch if you use that driver till 3K hz which it diserves). Aldo a little of the baffle step is mixed in the sound soap at 3/4 meters, not sure 1 to 2 dB really matters, but if there are 4 dB error in the mid in the filter sim, something is certainly wrong. After in real world at home, any way there will be peaks and deeps due to reflections...
That's how I approximate it so, 91 dB - 4 = 87 Hz, then the 2/3 dB loss for the midrange near the cut woofer cut-off, the highish 2 Pi efficienty towards the tweeter is ruled by the low pass inductance and damped shunt in case of the 12 dB low pass.
@mbrennwa is knowing a lot here as the OSMC was a 12" woofer, same cabinet size, with a highish cut off to theVolt dome mid around 450 Hz iirc. Maybe he can input ?
Edit : Jim fiered just before me 🙂
@temp25 , idealy the mid is flushed, that's smoother, but with the distance at 3 to 4 m, it is mixed in the sound soap (furnitures, surfaces diffraction, etc). which seems to be more important is to smooth the cabinet edges by chamfering > 1". I have an idea for that to simplify the design btw, but no POC of it yet.
Btw I am also very intrigued by the Dayton Sig 150, it is not so expensive at Soundimport if my memory serves me ( which doesn't most of the time !)
I assume more a 5 dB loss from what I see in Vituix as the crossover should be nearer from 250/300 Hz acoustical in reality (that means a lower electrical cut off there).
At 300 Hz, always at looking Vituix approximation BSC (so w/o filters) the mid driver above that 300 Hz to 500/600 Hz as only 2 to 3 dB loss.
There are often a sligth plateau bump from that 500/600 hz til 1500 Hz (which is a mess btw in the case of the NE149W if you look at the IEC measurement as it has already a bump that is hard to solve w/o notch if you use that driver till 3K hz which it diserves). Aldo a little of the baffle step is mixed in the sound soap at 3/4 meters, not sure 1 to 2 dB really matters, but if there are 4 dB error in the mid in the filter sim, something is certainly wrong. After in real world at home, any way there will be peaks and deeps due to reflections...
That's how I approximate it so, 91 dB - 4 = 87 Hz, then the 2/3 dB loss for the midrange near the cut woofer cut-off, the highish 2 Pi efficienty towards the tweeter is ruled by the low pass inductance and damped shunt in case of the 12 dB low pass.
@mbrennwa is knowing a lot here as the OSMC was a 12" woofer, same cabinet size, with a highish cut off to theVolt dome mid around 450 Hz iirc. Maybe he can input ?
Edit : Jim fiered just before me 🙂
@temp25 , idealy the mid is flushed, that's smoother, but with the distance at 3 to 4 m, it is mixed in the sound soap (furnitures, surfaces diffraction, etc). which seems to be more important is to smooth the cabinet edges by chamfering > 1". I have an idea for that to simplify the design btw, but no POC of it yet.
Btw I am also very intrigued by the Dayton Sig 150, it is not so expensive at Soundimport if my memory serves me ( which doesn't most of the time !)
Last edited:
@mbrennwa : according the bass woofer sensivivity in the datasheet, what should be the minimum sensivity of the midrange for an acoustical cut off around 250/300 hz. 2 DB higher, etc ? (selection phase for choosing the driver units).
I explain myself more : ideally if you don't use serie resistor in the mid for attenuation to cope with the woofer, taking into account of the DCR of an air coil for the low pass filter (we talk passive also here). We can attenuate a mid not really can boost it (but playing a little on the Qtc), so that's why I am asking a minimal conservative number from which one should not choose the mid driver below this 2.83V sensivity number according the woofer's one.
I explain myself more : ideally if you don't use serie resistor in the mid for attenuation to cope with the woofer, taking into account of the DCR of an air coil for the low pass filter (we talk passive also here). We can attenuate a mid not really can boost it (but playing a little on the Qtc), so that's why I am asking a minimal conservative number from which one should not choose the mid driver below this 2.83V sensivity number according the woofer's one.
Last edited:
I'm in favor of the SIG150. It has the sensitivity needed, and I've been told by a PE tech guy that it is a very good driver. I would for myself, prefer a non metal cone, but that's hard to find with a good price, and a round frame. And this speaker isn't going to be for me anyway.
Should we include one of these in the design? My thinking is to parallel it with a padding resistor, to allow a total SPL change of 3dB or so.
Should we include one of these in the design? My thinking is to parallel it with a padding resistor, to allow a total SPL change of 3dB or so.
Attachments
Last edited:
Here's the data page at Peerless by Tymphany's web site:
https://products.peerless-audio.com/transducer/166
Same spec for the 830668 SLS10
Why would you take a retailer's transcribed or old spec sheet over the manufacturer's?
I'm not saying PEs is right, I was just confirming that the mfg data sheet at PE does say 8.3mm.
I won't be driving the woofer to 8 or 10mm anyways so it's not very important to me 🙂
I won't be driving the woofer to 8 or 10mm anyways so it's not very important to me 🙂
You'll need to work out the exact answer by taking a good look at the baffle step, i.e., the shape of the baffle step curve. In the OSMC project we used the Tolvan Edge calculator, which seemed to work well. You could also take a look at Fig. 2 in the OSMC paper, which shows the 4pi woofer response. This was calculated from a 2pi measurement, and then the Tolvan baffle-step model was used to convert this to 4pi.
When considering the baffle step, I highly recommend to accept that the 6 dB loss is pretty darn close to reality. I often read comments that low-frequency response will be boosted by room gain, which is of course correct, but the effect is not nearly as much as one tends to hope.
I plotted the OSMC Tolvan curve (see attachment). The curve is at -0.2 dB at 250 Hz, and at +0.7 dB at 300 Hz. This means that for 4pi you'll need roughly 6 dB more from the woofer than from the mid.
Another point is to take into account the losses in the woofer x-over, which usually has a coil or two with substantial DCR in series with the woofer. I'd factor in at least 1 dB for this, maybe more if using budget coils with higher DCR.
In total you'd want the woofer sensitivity about 7 dB than that of the midrange driver. My guess is that won't happen, and you'll have to pad the mid down to match the woofer / mid outputs.
Attachments
@A4eaudio The SB12MNRX is the most bang for buck 5" cone mid that has the lowest and smoothest FR past a higher than usual LP frequency for its size. You could do a 2nd order LP @ 3 - 3.5k with this one, which vastly opens up tweeter selection. I tested several mids for consideration which are easily available and cost less than $75. The SB came out on top of all them, including the 5FE120 and 5FE125.
@temp25 That Lpad you're considering costs $20 (for the 100w model, which is the minimum I'd pick). These also vary greatly in quality and precision, plus they get intermittent and noisy after a while. I'd put that money into the right midrange.
@temp25 That Lpad you're considering costs $20 (for the 100w model, which is the minimum I'd pick). These also vary greatly in quality and precision, plus they get intermittent and noisy after a while. I'd put that money into the right midrange.
To clarify, we're interested in the 4pi for power response? Even though a mid will really only project its output forward?