What's possible with a modern 3" Midrange ... 3way Prototype

I'm a fan of midrange domes since I heared big PMC monitors in a big room. Did a lot of designs with different drivers over the years and it seems like I settle for a design after all these years and trials.

I made a reference speaker for microphone measurements from 10Hz to 100kHz and a little striped down version of that you can see here:
Studio1 Proto.JPG

Unpacked my photography gear to get you a nicer few :cool: - but it's still not the final version as I plan it.
2x SB34NRXL75-8
1x M74A with custom frontplate
1x T25B
heavy and well made cabinet, closed volume for the 12"

Active amplification Hypex FA503 and a VERY quick DSP setup. Limiting is set to 10W for the 1" and 100W for the 3" - I need these drivers for the project :geek:
Measured in my listening/measurement room, close to the backwall. There are some room influences under 400Hz - the driver itself measures flat and goes down at about 40Hz, within the room level goes down 2dB but reaches deeper. Earthworks M50, Audio Precision ... measurement stuff is good. SPL level in 1m, measured with Sine Sweeps.

Frequency response at different levels:
Studio1 close wall - SPL.png

Here you see how positioning and room size influences and expands LF response. Closed speakers often go lower in the room and need less EQing as you think (5dB of boost for these measurements). Tweeter protection get's active at about 100dBSpl, the tweter itself should be good for maybe 105dBSpl. But these are SINEWAVES for measurements - it will be able to do significant higher peaks with music!
On the low frequency side you see that both 12" woofers go into limitations at low frequencies. This is impressive to lock at to be honest, there is a LOT of movement involved at these levels! And the woofer can take that missuse with no issue.
Only that single 3" is here to still play without any sign of power compression. Crossover was 500Hz and 2000Hz btw.
No resonances over 20kHz, piston behaviour up to 35kHz at least. The Hypex module is going down in level here - the tweeter would still be linear.

How is THD at these levels?
Studio1 close wall - THD.PNG

It more or less stays under 1% :oops:
About 0,2% is the lower measurement limit of the selfnoise from the measurement mic. You reach this limit at lower leves (not shown here).

You stay <0,1% from 60Hz to 5kHz at 90dBSpl - that's already LOUD in a room!
THD rises a little faster with the 3" from 500-800Hz but still is lower as most drivers available. Crossover at 500Hz is no problem at all, if you are not crazy with SPL level you can cross lower. And crossover 2kHz is also no problem for the 1" driver!

Of course there are many measurements missing here, not sure if I can find the time to do useful off axis measurements (cause I have to do them outdoors, needs setup time and a lot of measurements). But it should give a hint what a 3" dome midrange is able to do.
  • Like
  • Thank You
Reactions: 13 users
I'm thinking about an ATMOS concept with 7 of these as floor speakers. Combined with Dirac ART for chancelling standing waves and reverb in the room at low frequencies this would be an absolute reference setup and should make a double bass array obsolete.

Dirac ART works best with distributed woofers. Here we have 14 long stroke 12" drivers on 2 heights in the room to chancell the first vertical room resonances.
When all drivers are used for the LFE channel ... you can produce 118dBSpl at 30Hz - with 2% THD! About 124dBSpl with the common <10% rule.
Main speakers with T34B, rear speakers with the smaller tweeter to get acoustic centers closer together.

Overall costs are for sure not cheap but in the context of a home cinema built it simplifies things ... looks like great value for me!
  • Like
Reactions: 2 users
Nice concept!
Some questions:
  • did you consider to place the domes off centre to "spread" baffle diffraction?
  • though the tweeter is capable of a 2 kHz crossover, this is right in the middle of the presence band (1 to ~ 5 kHz) where our hearing is most sensitive to phase irregularities.. Isn't one of the major benefits of a 3.5 inch dome mid that it covers the presence band so well? Furthermore, IME tweeters just sound better when crossed over not lower than 3.5 kHz.
I have to generally agree with the higher tweeter xover from the mid being a benefit for those needing more clean SPL. I typically prefer higher xovers on tweeters than most people use, mainly due to the uneven rise in THD associated with lower xovers and small diameter domes.

My guidelines for crossing HF domes when used in higher SPL setups is the following -

19-25mm >4.5kHz LR2,>4kHz BW3
25-30mm >3.0kHz LR2,>2.7kHz BW3
30-35mm >2.5kHz LR2,>2.3kHz BW3

The exact frequencies depend on the type of music, specifically the crest factor, compression and bandwidth energy concentration. Obviously WGs also affect this.

The other factor is the perceived loudness curve around the 2 to 5kHz bandwidth, being smaller amounts of THD are more noticeable in this range to begin with. Phase jumps are really audible here. This is why I usually try to avoid crossing in this range, preferring the midrange driver to play uninterrupted.

Most of the time mid domes used for high output I will HP at -

38mm >1.8kHz LR2, 1.6kHz BW3
50mm >1kHz LR2, 800Hz BW3
65mm >800Hz LR2, 700Hz BW3
75mm >700Hz LR2, 600Hz BW3

Exact numbers vary with specific model drivers, VC xmax and overall power compression/ handling figures.

The D7608 can handle about 80W long term if xmax is regarded and the VC has a place to shed its heat.

On very high output monitors, I have used aluminum plates stacked behind the motor and extended out past the chamber. There isn't much metal mass on the driver to remove heat effectively, so the driver needs external cooling measures.

Ferrofluid works well with this dome. I've used 1100 mg of the Ferrotec APG312 FF in the gap of the D7608. It increases power handling by at least 25% if the motor is cooled externally. I've seen people use peltier coolers on mid and HF drivers mounted in sealed compartments managed with CPU style controller hardware.

The other thing to watch with most mid domes is dust accumulation. The Morel domes are VERY STICKY! If you have forced air heating, its very advisable to use some form of a grille. If you damage the dome trying to clean it, there's no VC replacement available and you're buying a new driver. Dust can degrade the sound if there's a lot of it.
  • Like
Reactions: 1 user
@daanve @profiguy
Ähm - as the measurements show there is no THD from the tweeter. It produces less harmonics at 2kHz and 100dBSpl as 2 12" at 300Hz! And all Tweeter from ScanSpeak and other modern ones behave the same - you don't gain better THD with a higher crossover.
Our ear is more sensitive at 3kHz as at 2kHz. You would need to cross higher as 5kHz ... that would behave horrible.
I'm useing a digital crossover to get perfect addition and delay between the drivers. With the low crossover you have fitting (wide) off axis response too. It works very well ;)

When you need more SPL from the tweeter it's a better solution to get a louder tweeter. T34 fits the concept.

Metal midranges are not sticky btw.

Depth is 30cm, space behind a screen in a home cinema build is normally more as that. But the final version will be probably 22-25cm deep cause it also needs to fit in the side and backwall covers. And the height 12" should go a little higher when possible to get better behaviour in the room.

Centered position was needed for this project of the measurement speaker. There is also the 2nd midrange used but I prefer the sound with only one midrange - gives more "point source" sound, 2 midranges give some vertical "dimension" of the source. And one gives enough SPL to compete with the 2 12".
With 2 midrange you could cross even lower ...

I will make bigger roundovers between the 12" in the MHT area to avoid corner effects - then the middle position should be fine. That's a point of further investigations, if it doesn't give the results I want to see I will put them off center. (which only strays the effects, not avoid them. But still helps)
  • Like
Reactions: 1 user
Really great project, @IamJF ... I can see that the SB34 woofer performs well from a frequency response and harmonic distortion standpoint.

In my application, I noticed that my SB34NRX75-6 seems to lose a bit of clarity and definition if I ask it to play above 300 Hz. It is a small decrease in subjective performance, but with a good midrange (i.e. PTT6.5M), the system works better with a 200 Hz crossover than a 300 Hz crossover. I am not sure if it is the woofer or some cabinet resonance that is the cause, and for me this is no issue at all because of the larger midranges which I tend to use.

My application has a single woofer per side, so I do not have the SPL capability that you have, and my woofer is working twice as hard as your pair of woofers to get the same SPL. You are also using the SB34NRXL driver which has a superior motor with shorting rings. So I expect differences.

Subjectively, can you comment on the clarity, detail and definition in the 200 - 500 Hz range?

  • Like
Reactions: 1 user
Do you know my measutements of the 10" NRX drivers? They all are not suited for HighEnd 3way speakers, to much THD above 100Hz. I guess the 12" NRX will be similar and NRXL is a different beast.
I'm just now setting up the speaker to compare it to my normal reference setup. Let's see if the quick & dirty crossover plays good enough ...

I did teste before and often had tighter low frequencies from the SB drivers as with my ScanSpeak 8". But position in the room was not the same ... maybe it was just the higher H3 which gives that impression!
  • Thank You
Reactions: 1 user
Yes, now I remember that thread.

When I first measured my SB34NRX75-6 drivers they were part of a 3-way system. The mid driver was an SB17CAC35 and the tweeter was an SB26CDC. Crossovers were at 200 and 1.8k LR4. This is the harmonic distortion of the system I measured at 1 m, at about 95 dB SPL. ... circa June 2020.


It seems the SB34NRX75-6 performs better than the SB29NRX series. Retesting this driver is on my to-do list, full range without crossover, using STEPS and my new microphone.
  • Like
Reactions: 1 user
Something like that:
View attachment 1270588

View attachment 1270589

Normally you can switch through harmonics in a different sheet but not working for my saved measurement ... will have a look what's going on there.
Such peaking in the 3rd harmonic distortion trace can be characteristic of current dependent non-linearity due to induced eddys reflected in the driver impedance. This can be verified (or not) by repeating the test with a high impedance source where the driver impedance is relatively small. If my hunch is right, then current drive will deliver a significant measurable and audible benefit in this aspect beyond the already good results.

IMHE switching from voltage driver to current drive reduces a noise-like graininess on transients, such as when listening to a grand piano, and reduces what I can best describe as an "electronic haze" that I find reminiscent of crossover distortion in not-so-well designed class B amplifiers. Once heard, it's presence is instantly noticeable despite its low-level.
  • Like
Reactions: 1 user