Some Interesting Drivers, a New 3-way Project

I see you have the same 5db dip in the response of the SB12MNRX2-25-4 that I have.
Yes, I am still thinking about and processing all this data.
What was heard i would call "wooly", but reminded me of 300Hz cabinet resonances. Like I said, it could have been the box, driver, or implementation.
I appreciate your thoughts.

For the price, the SB12MNRX2 is a lot of driver. Yes there is that dip from 1.5k to 2.4k, but there it rolls off smoothly in the high end, so no notch filter is needed. The SB15CRC has a smoother response, but it definitely requires a notch filter at 6.5k.

As I said, I am reviewing all the data, and I will summarize my thoughts in a future post.
 
  • Like
Reactions: AlmaAtaKZ
This is my assessment of the two drivers. I would certainly appreciate any opinions that anyone has,

Frequency Response

The SB15CRC has a nice even response over the intended bandwidth from 400 – 3k, except for a narrow -2 dB dip at 2.4k. That dip is present all the way out to 90 degrees, so it effects the power response also. The cone breakup peak at 6.5k will have to be dealt with, and I have two different notch filters in sim that would work.

The SB12MNRX2 has a -4 dB dip centered on 2k, and the dip is more than a half-octave wide, so this will be audible. A dip is much less harmful than a peak, so the audibility may be low. On the plus side, the driver has nice extension all the way to 10k, without a big breakup peak. This would allow the designer to use a simple low pass crossover.

Spectral Decay / Burst Decay

Looking at the burst decay plots, the SB15CRC has a cleaner decay in the 500-4k region of interest. It has a very clean, even decay, and a very rapid decay. This performance is very similar to what I observed in the SB15CAC and SB15NBAC, except the breakup peak is a little lower with the composite cone.

The SB12MNRX2 has a slower decay over the whole region, and there is some energy storage in the 1.5k – 3k range. It is obvious that the engineers at SB tried to dampen this resonance as much as possible within their cost constraints, because it dies out pretty quickly. The only evidence in the on-axis curve is the -4 dB dip centered on 2k, and as we all know, dips are much much preferable to peaks.

Distortion

The distortion testing was interesting. I expected the SB15CRC to have an advantage here, and it did, but not as much as I was expecting. The SB15CRC distortion was at or below the limit of my test equipment resolution over the whole intended bandwidth (400 – 3 kHz). H3 stayed below 0.2% over the whole range.

The SB12MNRX2 was higher, with H3 having a broad peak of 0.3% range from 1k to 3k, and H2 reaching above 0.5% in the same range. Overall, this is low distortion for a $60 driver. The distortion of both drivers is low enough that I would not use the distortion results as a deciding factor between the drivers.

Assessment

So, looking at all the data, the SB15CRC is the better driver for my purposes. The frequency response curve is more manageable, the spectral decay is slightly better, and the distortion is slightly better. However, if my budget for this project required that I use the lower cost SB12MNRX2, I feel it would be a good driver that could work well.

j.
 
No, I had not considered that driver. My local dealer has the 15W/4434G and the 15M/4624G midrange, but not the one you mentioned, so I was unaware of it. You are correct, it looks like a very nice driver. I wonder if I can easily get one in the US ? I will have to look beyond Madisound I suppose.
 
@b_force - - you mentioned wavecor earlier in this thread... Do you have thoughts on a Wavecor driver that I might consider? Looking at the published specs, it seems the WF120BD04 (8 ohm round) would be suitable, but perhaps the WF152BD06 would be suitable as well... Thoughts?
 
interesting @b_force

@hifijim

The Audix TM1/Plus is a fine microphone for frequency response measurements. I can see you've taken care to measure the distortion within it's limits.

IIRC, from the specifications, the THD is 1% (-40dB) @130dB. So it's self-distortion of 0.01%, (-80dB), dominated by H2, is at 130-40 = 90dB.

Suppose your amplifier, pre-amp and microphone all have self distortion of -80dB, to measure a real H2 of -60dB (0.1%) you need a delta of 20dB. ie:
1718254238038.png


Reference:
https://sengpielaudio.com/calculator-thd.htm

So the CRC is 0.1% across most of it's bandwidth. The SB12 gets up to almost 0.6% in the middle frequencies. Whether you like an extra H2 is preference, but the CRC is the more "honest" driver.

As for the drip, it's a little fly in the ointment. But depending on your baffle dimensions you may be able to work around it...
 
Last edited:
  • Like
Reactions: hifijim
@tktran303
I think we're starting to split hairs at this point, lol

Main advantage over a 5 inch is just maxSPL incl the (obvious) better distortion at higher levels.

Penalty is slightly higher directivity.

So the main question to @hifijim is what he's looking for?

I personally would pick a 5 inch always over a 4 inch, unless it doesn't aesthetically look right or when someone really wants a wide dispersion.
 
  • Like
Reactions: tktran303
IIRC, @DcibeL has some experience with a driver with a narrow 2Khz dip (not drip) - are you able to chime in?
Which driver? Eton 5-880 comes to mind. In-cabinet measurements:
1718284671832.png


Discussion here is correct, a dip is favourable to a peak, you won't be offended by what you don't hear. But, one day you will experience a speaker that is absent of the response anomalies, and hear what you've been missing this whole time. The hole in your speaker response is akin to the hole in your life, you've got to make some big changes to to be fulfilled across the entire spectrum.

Main consideration is that a response like this will "force your hand" on the crossover point, which may limit somewhat on overall power & di optimization. But, don't think that it is an unusable response, a filter with a bump at the knee is easily accomplished passively, so a crossover frequency 2k to 2.5kHz is not out of the question. The above Eton was used with a 2kHz filter as a midrange in a 3-way, and although it's not the lowest distortion driver in the world, the final result is well balanced and very enjoyable to listen to, nothing stands out as problematic.
1718285888464.png

1718285932245.png

1718286045919.png


Here's a visual aid of another mystery driver. A filter response like this is not difficult to achieve passively. Even though this particular driver has a dip at 2kHz, response can be managed to a smooth rolloff without too much effort.
1718286065682.png

1718286080887.png



I've also used both WF152 and WF120 being mentioned here. Both are excellent, I'd recommend either. Difference in directivity I would not describe as a "penalty", just different. I tend to prefer a more directive speaker, but overall balance of all aspects of power response, sensitivity, power handling, distortion capabilities of the drivers all needs to be balanced, so it's not as cut and dry as smaller midrange = better. Keen internet readers would find my driver testing at HTGuide.
 
  • Thank You
Reactions: tktran303
a dip is favourable to a peak, you won't be offended by what you don't hear. But, one day you will experience a speaker that is absent of the response anomalies, and hear what you've been missing this whole time.
Well, if we take a bit of psycho-acoustics with this, I am willing to argue that in some cases (most) people are physically not capable to be able to hear small dips.

There are absolutely tons of articles about this and is by now just common part of any psycho acoustics book and other literature.

In general I personally find "missing" way to much of a strong word to describe this.
If anything I would describe it as a slight difference in nuance

But obviously, it all depends

Changing room acoustics, seats/position and other stuff can have a much bigger impact.

Difference in directivity I would not describe as a "penalty", just different. I tend to prefer a more directive speaker, but overall balance of all aspects of power response, sensitivity, power handling, distortion capabilities of the drivers all needs to be balanced, so it's not as cut and dry as smaller midrange = better.
I agree and that is a more nuanced version of what I meant basically 🙂

Fact is that with a 4 inch you quickly run out of SPL
Based on my own experience, that is also the first complaint most people and clients have.
 
  • Like
Reactions: DcibeL
No, I had not considered that driver. My local dealer has the 15W/4434G and the 15M/4624G midrange, but not the one you mentioned, so I was unaware of it. You are correct, it looks like a very nice driver. I wonder if I can easily get one in the US ? I will have to look beyond Madisound I suppose.
Jim, you can try contacting Madisound to see if they can get 15M/4624G for you. They ordered the SB12MNRX25-4 for me when they were first released, and offered the same service for a Mudorf AMT I was interested in.
 
  • Like
Reactions: hifijim
So the main question to @hifijim is what he's looking for?
My simulations of baffle layout show that I get an acceptable response using either a midrange Sd of 50 cm^2 (4" driver) or 80 cm^2 when paired with the chosen tweeter. There ARE small differences in directivity, but those differences are within the margin of error of the simulation, given the sim is preliminary.

I have not built a cabinet yet, and I have not even defined the exact driver spacing on the baffle, although I have a rough idea. So all measurements, simulations, and calculations are preliminary at this point. As @DcibeL said ...
Main consideration is that a response like this will "force your hand" on the crossover point, which may limit somewhat on overall power & di optimization.
Every decision I make restricts my future choices and options. So I simulate at every stage to make sure I am not locking myself onto a path to nowhere. As I get new information and data, the sim becomes more refined and the margin of error goes down.

At this point of the design process, I am trying to manage the power and directivity performance. The primary drivers of directivity for this type of speaker are (1) the directivity of the tweeter, which has a small waveguide, (2) the directivity of the midrange driver from 1k - 3k, (3) the size, shape, and edge treatment of the baffle, (4) The vertical spacing of the drivers, (5) the crossover frequency and slopes.

Once the cabinet is built and I start to take real measurements, 4 of the 5 directivity influences are fixed, and the crossover is the only variable left open. So I want to manage the power and directivity performance to the extent possible.

There are two styles of directivity I consider desirable. With a large waveguide tweeter and a low crossover, this is what I aim for
1718292591041.png


With direct radiator tweeters and highly profiled cabinets, I can get a directivity performance like this,
1718292674856.png

Either style sounds quite acceptable in my room.

j.
 
  • Like
Reactions: AlmaAtaKZ
There ARE small differences in directivity, but those differences are within the margin of error of the simulation
Either style sounds quite acceptable in my room.
In that case, my pick would be a 5 inch.
They are not that much more expensive (in some cases even cheaper), there is an awful lot more choice, gives you more headroom, meaning more wiggle room as well.

I personally rarely use or advice a 4 inch for exactly these reasons.
 
Well, if we take a bit of psycho-acoustics with this, I am willing to argue that in some cases (most) people are physically not capable to be able to hear small dips.

There are absolutely tons of articles about this and is by now just common part of any psycho acoustics book and other literature.

In general I personally find "missing" way to much of a strong word to describe this.
If anything I would describe it as a slight difference in nuance

But obviously, it all depends

Changing room acoustics, seats/position and other stuff can have a much bigger impact.
Hi, everyone who have not tried Harman how to listen test should do it now to realize this stuff themselves! 🙂 Do not shy away, nobody is ranking you, it is for you to know your own listening skill better, for your own reference!

Having tried the software I noticed it is easy to detect peaks, while really hard to detect dips, and the difference was very apparent from statistics in the end. Statistics show which tests you leveled most with least tries, and which tests you had to do multiple takes before getting really stuck. It's not just peaks/dips, but there are many other usefull tests.

Playing with the software one gets some kind of a reference to own listening skill, for example, if you can plow through all levels on some test, you can be quite sure you can trust what you hear in that regard. But if you cannot, like the dip listening test might show, it is very likely many dips go unnoticed. Very very helpful to have this kind of personal reference to build personal listening skills, and trust yourself, or not 🙂 we are all alone with our perception, so it's good to have some kind of a reference to reflect against.

I think completing level 8 in the tests would qualify as "trained listener" in Harman listening tests, if you ever take part. If you cannot get to level 8 on some tests, consider spending some time to exercise to get confidence to evaluate your own systems better.
 
Last edited: