Designing an open back midrange speaker

[a4eaudio]

I am considering attempting to design an open-back midrange speaker. This would have a tunnel through the cabinet that is open on the back, not just an open-baffle.

I have searched the Internet quite a bit before posting, but if there are any good resources that you know of, feel free to point me in that direction. I have found very little on designing an open-back midrange speaker, most of what comes up are opinions on open vs sealed. I am aware of the Finalists and Statements that are open-back speakers, but there are no detailed descriptions of how the cabinet and crossover decisions were determined.

So to start:

1. Are there certain T/S parameters that suggest a driver is a good or bad candidate for an open-back design?
2. How is the optimal tunnel width determined (assuming a circular pipe) - use something like 1" larger than the driver diameter or rely on trial and error? (For example, I know the Statement IIs use a 5" tunnel for a 4" driver and the Finalists use a 6" tunnel for a 5" driver.)
3. How is the optimal tunnel depth determined? This would be tough to do based on trial and error as I need to build the cabinets before measuring. The Statement IIs are 16.5" deep and the Finalists are 15.5". If I just used 16" and figure out step 4 below, will the crossover take care of the details or does this need to be figured out more precisely beforehand?
4. Is there a specific way to measure the frequency response and then use the FRD files in simulation software (PCD) in the normal fashion? Or (i) is it going to be much more trial and error in a real room environment to pick up the rear reflections; or (2) some kind of mathematical transformation of the measured FRD? (I read John K's method for measuring for dipole speaker design, but that doesn't seem to apply exactly to an open-back midrange design. Although admittedly it could apply, as I didn't understand all of the details.)

[Chris Roemer]

Ideally to run a driver w/out a box behind it, I think you pick one w/a Qts near 0.70 (maybe in the .6-.8 range).

[hitsware2]

Same driver requirements as O.B. ( high Qts (as noted) )
If the length of the pipe becomes ~> than the diameter ,
or the proportion deeper than a square ( for a box )
then pipe resonances start to manifest.
The diameter + the depth becomes the 1/2 wave cancellation
dimension .

[PWR RYD]

For midrange duty there is no need for high Qts or even 0.6 - 0.8 as suggested.

[hitsware2]

Same driver requirements as O.B. ( high Qts (as noted) )
If the length of the pipe becomes ~> than the diameter ,
or the proportion deeper than a square ( for a box )
then pipe resonances start to manifest.
The diameter + the depth becomes the 1/2 wave cancellation
dimension .

Depends on the baffle size .... and range of ' midrange '

[jhollander]

The tunnel configuration add the complication of the tunnel resonance. You can design one with Horn response or if you like math google U-frame sub-woofer design. Mart King gives a formula for the resonance. ABC Dipole will also do a U-frame.

The other option simply ignore the issues and focus on the boxless mid range. Measure your results and remember that you need to tweak by taking real time measurements. The wrap around response will not follow the frequency response of crossover simulation programs

[a4eaudio]

I have seen some of the rules of thumb about high QTS, but...I think that is for open baffle bass. I think PWR RYD is right when it comes to midrange, although I have no idea behind why any of this is the case. (And yes, hitsware2 your point is taken that it depends on how low I try to go with the midrange, but I don't think that will be an issue in this design.)

Not to dwell on Jim Holtz and curt_c 's speakers, but they are the only open-back mids that I have much information on, and obviously perform well. The midranges are:

• NE123W-08 - QTS 0.35, xmax 3.1 mm
• NE149W-04 - QTS 0.42, xmax 5mm
• Tang Band W4-1337SD - QTS 0.37, xmax 3mm

I am planning on using the Tang Band W5-2143 - QTS 0.38. xmax 2.5 mm. I already have these drivers, so if they don't work I'll end up just building a sealed design.

My current plan would be to make sure I have enough room for a sealed midrange compartment and if I can't get the open-back concept to work I can just take out the tunnel and seal it up.

[a4eaudio]

The other option simply ignore the issues and focus on the boxless mid range. Measure your results and remember that you need to tweak by taking real time measurements. The wrap around response will not follow the frequency response of crossover simulation programs

Thank jhollander , what do you mean by "real time" measurements? If I have a USB mic (UMIK-1) and REW, can I do this?

[jhollander]

Yes, measure and design as you normally would then measure the resulting FR and adjust components to get the mid response you want. With the rear response combining with the front response the summation is not going to match PCD or XSim. Only other watch out is if your tunnel resonance is in the pass band then that is difficult to deal with without stuffing.

[PWR RYD]

[/LIST]I am planning on using the Tang Band W5-2143 - QTS 0.38. xmax 2.5 mm. I already have these drivers, so if they don't work I'll end up just building a sealed design.

My current plan would be to make sure I have enough room for a sealed midrange compartment and if I can't get the open-back concept to work I can just take out the tunnel and seal it up.

Great driver choice! I am using that Tang Band W5-2143 as an OB mid for my 20-20 Indy project.

[a4eaudio]

Great driver choice! I am using that Tang Band W5-2143 as an OB mid for my 20-20 Indy project.

Awesome, that makes me feel better! I have all the drivers from 2 years ago to make a sealed speaker but then the InDIYana 2020 project came up and I thought I'd try something for the competition. Then the cold, hard truth hit me that there is not a lot of information out there on designing open-back mids, and I'm sure not ready to do some of the more advanced options. Given your's and John's comments above, I'll get started on my cabinets and try to make it work. Thanks everyone!

[curt_c]

Some thoughts on tunnel mid designs.

My whole purpose of the ‘tunnel mid’ design is to eliminate or attenuate the resonance that plagues most all sealed box designs. Standing wave propagation as well as simple reflections of the back wave can cause aberrations and smearing of the sound. A cylindrical ‘enclosure’ of a slightly larger diameter than the driver eliminates any standing wave propagation with the exception of the reflection along its length, which is due to the abrupt change in acoustic impedance at its terminus whether open or closed. Lining the tunnel with 1” foam and adding a foam plug somewhere near the rear of the tunnel provides sufficient length of material to attenuate most audible resonance due to standing waves. While the cylindrical enclosure can be utilized with the terminus closed, any practical length is still plagued with some acoustic energy reflected back through the cone. My experience suggests this excess energy is less intrusive if allowed to leak out the rear of the enclosure where it blends with the reverberant sound field and generally not perceived as first arrival sound.

I have not found driver Qts to be a significant factor in tunnel mid designs when the passband is limited to the driver’s mass controlled region. That is, where the cone excursion is predominantly limited by Mms, rather than Cms. While this can be easily modeled for any specific driver by comparing excursion vs. frequency at various excitation levels, generally choosing the high pass crossover frequency roughly 3 octaves above fs would be a good rule of thumb.

I recommend the tunnel length results in the ½ wave cancellation appearing in the stop band of the crossover. An easy method to determine this is by using the formula:

Tunnel length = (speed of sound / crossover HP frequency) / 2

or: (13560/400)/2 = 16.95” for a 400 Hz Xo.

While this does not accurately model the actual path length or ½ wave cancellation frequency, one can easily calculate this by measuring the actual linear distance incorporating half the front baffle width and half the rear baffle width. The easy method errs on the side of a lower ½ wave cancellation frequency being buried further down in the stop band, so it’s generally good enough for my purposes. Putting the ½ wave cancellation frequency near the woofer to mid crossover makes it easy to compensate for in the woofer LP and mid HP networks. Also note that due to the attenuation of the sonic energy coming from the tunnel terminus, it will not result in a true cancellation, but merely a dip in response.

As far as measurements go, I generally measure as if it were a sealed design a 1 meter and a 7 mS gate. I model the crossover for the best on and off axis response average, with the knowledge I may need to adjust the woofer and mid networks a bit in the final voicing. Since my favorite part of speaker design is the voicing, I don’t mind having to do the extra work.

– It’s extra fun for me.

C

[a4eaudio]

Some thoughts on tunnel mid designs...

Curt - thank you very much for sharing your thoughts and experience!

[rpb]

Some thoughts on tunnel mid designs.

My whole purpose of the ‘tunnel mid’ design is to eliminate or attenuate the resonance that plagues most all sealed box designs. Standing wave propagation as well as simple reflections of the back wave can cause aberrations and smearing of the sound. A cylindrical ‘enclosure’ of a slightly larger diameter than the driver eliminates any standing wave propagation with the exception of the reflection along its length, which is due to the abrupt change in acoustic impedance at its terminus whether open or closed. Lining the tunnel with 1” foam and adding a foam plug somewhere near the rear of the tunnel provides sufficient length of material to attenuate most audible resonance due to standing waves. While the cylindrical enclosure can be utilized with the terminus closed, any practical length is still plagued with some acoustic energy reflected back through the cone. My experience suggests this excess energy is less intrusive if allowed to leak out the rear of the enclosure where it blends with the reverberant sound field and generally not perceived as first arrival sound.

I have not found driver Qts to be a significant factor in tunnel mid designs when the passband is limited to the driver’s mass controlled region. That is, where the cone excursion is predominantly limited by Mms, rather than Cms. While this can be easily modeled for any specific driver by comparing excursion vs. frequency at various excitation levels, generally choosing the high pass crossover frequency roughly 3 octaves above fs would be a good rule of thumb.

I recommend the tunnel length results in the ½ wave cancellation appearing in the stop band of the crossover. An easy method to determine this is by using the formula:

Tunnel length = (speed of sound / crossover HP frequency) / 2

or: (13560/400)/2 = 16.95” for a 400 Hz Xo.

While this does not accurately model the actual path length or ½ wave cancellation frequency, one can easily calculate this by measuring the actual linear distance incorporating half the front baffle width and half the rear baffle width. The easy method errs on the side of a lower ½ wave cancellation frequency being buried further down in the stop band, so it’s generally good enough for my purposes. Putting the ½ wave cancellation frequency near the woofer to mid crossover makes it easy to compensate for in the woofer LP and mid HP networks. Also note that due to the attenuation of the sonic energy coming from the tunnel terminus, it will not result in a true cancellation, but merely a dip in response.

As far as measurements go, I generally measure as if it were a sealed design a 1 meter and a 7 mS gate. I model the crossover for the best on and off axis response average, with the knowledge I may need to adjust the woofer and mid networks a bit in the final voicing. Since my favorite part of speaker design is the voicing, I don’t mind having to do the extra work.

– It’s extra fun for me.

C

Do you think that there is much perceived quality to be gained OB compared to a relatively large, and well damped mid chamber? By well damped, I mean roughly a .2 cu-ft chamber that is chocked full, or close to full, with dense fiberglass like that used in ceiling tiles?

I have a mid-bass that I like, but it seems to have a thin / transparent cone. I damp excessively, but if OB will be even better, I might give it a try again. (Last OB I built was close to 40 years ago, and the x-over was a WAG.)

[ChrisLudz]

Some thoughts on tunnel mid designs.

My whole purpose of the ‘tunnel mid’ design is to eliminate or attenuate the resonance that plagues most all sealed box designs. Standing wave propagation as well as simple reflections of the back wave can cause aberrations and smearing of the sound. A cylindrical ‘enclosure’ of a slightly larger diameter than the driver eliminates any standing wave propagation with the exception of the reflection along its length, which is due to the abrupt change in acoustic impedance at its terminus whether open or closed. Lining the tunnel with 1” foam and adding a foam plug somewhere near the rear of the tunnel provides sufficient length of material to attenuate most audible resonance due to standing waves. While the cylindrical enclosure can be utilized with the terminus closed, any practical length is still plagued with some acoustic energy reflected back through the cone. My experience suggests this excess energy is less intrusive if allowed to leak out the rear of the enclosure where it blends with the reverberant sound field and generally not perceived as first arrival sound.

I have not found driver Qts to be a significant factor in tunnel mid designs when the passband is limited to the driver’s mass controlled region. That is, where the cone excursion is predominantly limited by Mms, rather than Cms. While this can be easily modeled for any specific driver by comparing excursion vs. frequency at various excitation levels, generally choosing the high pass crossover frequency roughly 3 octaves above fs would be a good rule of thumb.

I recommend the tunnel length results in the ½ wave cancellation appearing in the stop band of the crossover. An easy method to determine this is by using the formula:

Tunnel length = (speed of sound / crossover HP frequency) / 2

or: (13560/400)/2 = 16.95” for a 400 Hz Xo.

While this does not accurately model the actual path length or ½ wave cancellation frequency, one can easily calculate this by measuring the actual linear distance incorporating half the front baffle width and half the rear baffle width. The easy method errs on the side of a lower ½ wave cancellation frequency being buried further down in the stop band, so it’s generally good enough for my purposes. Putting the ½ wave cancellation frequency near the woofer to mid crossover makes it easy to compensate for in the woofer LP and mid HP networks. Also note that due to the attenuation of the sonic energy coming from the tunnel terminus, it will not result in a true cancellation, but merely a dip in response.

As far as measurements go, I generally measure as if it were a sealed design a 1 meter and a 7 mS gate. I model the crossover for the best on and off axis response average, with the knowledge I may need to adjust the woofer and mid networks a bit in the final voicing. Since my favorite part of speaker design is the voicing, I don’t mind having to do the extra work.

– It’s extra fun for me.

C

Curt - Your post is much appreciated.

I am planning to build the Statement center channel. If I use the unshielded version of RS180-08 and reduce the depth of the enclosure by 2 inches, the length of the open transmission tunnel also will be reduced by 2 inches. Will this have a negative effect on the mid range acoustics?

CL