OK for my tiny sub project I am trialing 2 8 inch opposing SB Acoustics subs
8″ SB23MFCL45-4 / Polypropylene - Sbacoustics
They are beautiful drivers.
I am trialing this method of coupling with 4mm threaded rod, 4 mm bolts and hex rod coupler (coupling nut). See pic. Yes I know they are too close - mybad did not measure depth. I might need to increase box size a little.
I tightened it so much the MDF started to bow in the middle.
At the very least I think it needs some sort of extra support, even just a bit of pine in the middle where it bows.
But the main problem is that it needs top and bottom off to be able to install the rods. This is almost impossible to do when you consider finishing (I was going to wood veneer).
Is all this really worth it? Are just normal wood screws really that bad?
It is such a small box I would think the drivers are pretty well coupled anyway.
8″ SB23MFCL45-4 / Polypropylene - Sbacoustics
They are beautiful drivers.
I am trialing this method of coupling with 4mm threaded rod, 4 mm bolts and hex rod coupler (coupling nut). See pic. Yes I know they are too close - mybad did not measure depth. I might need to increase box size a little.
I tightened it so much the MDF started to bow in the middle.
At the very least I think it needs some sort of extra support, even just a bit of pine in the middle where it bows.
But the main problem is that it needs top and bottom off to be able to install the rods. This is almost impossible to do when you consider finishing (I was going to wood veneer).
Is all this really worth it? Are just normal wood screws really that bad?
It is such a small box I would think the drivers are pretty well coupled anyway.
Attachments
That's an unusually small enclosure for what I normally build, so I typically have a fair bit of bracing as well, but you don't need to get out the long torque wrench and max it out either.
MDF does not have great screw retention, using machine screws and "Hurricane" or "T" nuts are a good idea for supporting heavy drivers and maintaining even screw tension.
Metal rods don't serve any useful purpose with such small baffles, and may "ring". With larger baffles, plywood or wood bracing works better, and does not require removal of a side to install the drivers.
Member
Joined 2009
Paid Member
What I did was install some additional cross-bracing and then the rods will not be able to bow the box at all. I added some stuffing into the box after installing the rods. I hope the photos explain it.
I also used 8” drivers, which provide plenty of ooomph for my home theatre and have been in use for a decade.
Push-Push or Push-Pull
I also used 8” drivers, which provide plenty of ooomph for my home theatre and have been in use for a decade.
Push-Push or Push-Pull
Attachments
Member
Joined 2009
Paid Member
Never was measured, but i do now own a measurement mic, haven’t opened the box yet since i got it last year but this winter i will give it a go.
4 drivers would need a bigger box, maybe a problem dealing with all those threaded rods, plus I have fitted an integral plate amp to my speaker which would be a tight squeeze if using 4 drivers. I’d rather have two boxes with 2 drivers each, giving more placement options.
4 drivers would need a bigger box, maybe a problem dealing with all those threaded rods, plus I have fitted an integral plate amp to my speaker which would be a tight squeeze if using 4 drivers. I’d rather have two boxes with 2 drivers each, giving more placement options.
Thanks for replies. I have a B&W PV1D which is even smaller (Here is teardown of its predecessor PV1 Hilltop Workshop | Subwoofer teardown and compare. B&W PV1, Velodyne SPL10, and a couple of ferals. - YouTube). The drivers are long throw and I figure I will need about 10dB of gain for EQ. I will be using hundreds of watts to drive it which will be under control of a microcontroller for dynamic bass EQ and clipping detection/prevention. If it were even slightly bigger I could fit my hand down the side.
I just found that B&W teardown. I note they had metal rods coupling via the cabinet directly and not via the speaker mounting holes. I think I will start with a little wood bracing to prevent the warp. I think there is just enough room to get my hand down the side if I make the bottom removable. My wood working skills are not the best but my electronic skills are. It is a work in progress. Thanks for the hornresp ideas.
My wallet hurts from just 2 of those drivers, so 4 is out of the question. That would also invalidate the idea of a tiny enclosure. Nice pics BTW too.
I just found that B&W teardown. I note they had metal rods coupling via the cabinet directly and not via the speaker mounting holes. I think I will start with a little wood bracing to prevent the warp. I think there is just enough room to get my hand down the side if I make the bottom removable. My wood working skills are not the best but my electronic skills are. It is a work in progress. Thanks for the hornresp ideas.
My wallet hurts from just 2 of those drivers, so 4 is out of the question. That would also invalidate the idea of a tiny enclosure. Nice pics BTW too.
Internal dimensions
w = 26 cm
d = 22.4 cm
h = 28.2 cm
I think 1cm more all around would make it a bit more practical. As I said a work in progress. I will make another box if this one does not work out.
w = 26 cm
d = 22.4 cm
h = 28.2 cm
I think 1cm more all around would make it a bit more practical. As I said a work in progress. I will make another box if this one does not work out.
I should add that my plan is to use the following filters:
- variable HPF to limit excursion at low frequencies and high power
- dynamic EQ so that at lower powers bass extension is better
- at higher powers the boost will be less at low frequencies again to limit excursion.
- a variable Linkwitz transform (based off power)
A microcontroller will be monitoring power levels on the fly ... FFTs etc and adjusting the filters accordingly. Not as hard as it sounds.
So the raw SPL curves might not look like much but each driver has 12 mm of Xmax to play with. It will just need a lot of amp power to get there. I have been using WinISD pro alpha which has pretty good flexibility with the filters you can use. I will probably end up using stereo amp one for each driver.
- variable HPF to limit excursion at low frequencies and high power
- dynamic EQ so that at lower powers bass extension is better
- at higher powers the boost will be less at low frequencies again to limit excursion.
- a variable Linkwitz transform (based off power)
A microcontroller will be monitoring power levels on the fly ... FFTs etc and adjusting the filters accordingly. Not as hard as it sounds.
So the raw SPL curves might not look like much but each driver has 12 mm of Xmax to play with. It will just need a lot of amp power to get there. I have been using WinISD pro alpha which has pretty good flexibility with the filters you can use. I will probably end up using stereo amp one for each driver.
Here are some results.
Currently pissy little 50W amp x2
The gain needed is about 20 dB and I was expecting 10dB so more than I thought. I was working off a 14.4L box volume.
The distortion is also weird. What is that at 50Hz? This is unboosted distortion too.
I am wondering if it needs a bit more cabinet depth to give the rear of the drivers some room. Even 1cm or so. Modelling does not show volume improves performance much until you have more than double the volume which is not going to happen.
It sounds pretty good - better than one of my other commercial subs (not the B&W which is the gold standard).
I am waiting for my new amps to arrive which should improve things a lot. Volume is still a bit limited due to all the EQ and small amp.
PS. It passes the beer on the sub test easily. I cannot see any disturbance at all so the opposing drivers plus bracing is working.
Currently pissy little 50W amp x2
The gain needed is about 20 dB and I was expecting 10dB so more than I thought. I was working off a 14.4L box volume.
The distortion is also weird. What is that at 50Hz? This is unboosted distortion too.
I am wondering if it needs a bit more cabinet depth to give the rear of the drivers some room. Even 1cm or so. Modelling does not show volume improves performance much until you have more than double the volume which is not going to happen.
It sounds pretty good - better than one of my other commercial subs (not the B&W which is the gold standard).
I am waiting for my new amps to arrive which should improve things a lot. Volume is still a bit limited due to all the EQ and small amp.
PS. It passes the beer on the sub test easily. I cannot see any disturbance at all so the opposing drivers plus bracing is working.
Attachments
Last edited:
too close? Magnets touching would give best coupling.
I am not a fan of MDF for speaker boxes, a particularily poor choice for woofers.
dave
I am not a fan of MDF for speaker boxes, a particularily poor choice for woofers.
dave
Dave,
The SB23MFCL45-4 has a vented back plate, magnets touching would reduce voice coil cooling and affect compliance.
Last edited by a moderator:
I’m not a big fan of polepiece vents, but that means some sort of vented spacer to allow them to breath.
dave
dave
I would recommend using much more substantial bars to link the drives together. The 50hz distortion might be due to your current bars resonating. You have them under high tension. They will very likely act like the strings of a guitar. Maybe use something like 12mm aluminium bar. The ends can be drilled and tapped to accept the bolts that mount the drivers. This will allow you to mount the drivers after the cabinet has been fully assembled and veneered. Make the bolt holes in the cabinet large enough so that the aluminium bars can easily slide through them, the drivers will be tightened against the ends of the bars. Mounting the drivers this way will mean that the bars are not pre-tensioned and will be much less likely to resonate.
Decoupling the drivers from the cabinet will further reduce energy transfer from the drivers to the cabinet. Use thick compliant gaskets and make the aluminium bars just long enough so that when the mounting bolts are fully tightened the gaskets are only slightly compressed. The holes that the aluminium bars pass through should be only slightly larger than the diameter of the bar as this will prevent the drivers from sagging. The bars do need to be able easily slide through the holes as friction in the holes will result in energy transfer to the cabinet.
Hope this helps.
Niffy
Decoupling the drivers from the cabinet will further reduce energy transfer from the drivers to the cabinet. Use thick compliant gaskets and make the aluminium bars just long enough so that when the mounting bolts are fully tightened the gaskets are only slightly compressed. The holes that the aluminium bars pass through should be only slightly larger than the diameter of the bar as this will prevent the drivers from sagging. The bars do need to be able easily slide through the holes as friction in the holes will result in energy transfer to the cabinet.
Hope this helps.
Niffy
I continue to be unimpressed by the unbridled hubris of the male ego, attempts to defy proven science to become a 'hero'.
The cabinet is too small.
To circumvent the design flaws, the first Neanderthal thought is to throw more power at it. (Remember all those American cars with 5 litre engines producing the same output as European cars could with 2?) Ford Maverick: 4.1 litres - 90Bhp. VW Golf: 1.6 litres - 112Bhp.
But we're in a different age we can throw EQ and DSP at the problem. But the problem remains, you have two drivers competing for the same displacement.
The cabinet is too small.
I'll bet my mortgage that a single 8" driver, or two 6.5" opposed will out-perform your design because:
The cabinet is too small.
The cabinet is too small.
To circumvent the design flaws, the first Neanderthal thought is to throw more power at it. (Remember all those American cars with 5 litre engines producing the same output as European cars could with 2?) Ford Maverick: 4.1 litres - 90Bhp. VW Golf: 1.6 litres - 112Bhp.
But we're in a different age we can throw EQ and DSP at the problem. But the problem remains, you have two drivers competing for the same displacement.
The cabinet is too small.
I'll bet my mortgage that a single 8" driver, or two 6.5" opposed will out-perform your design because:
The cabinet is too small.
I have PhD in electronic engineering and specialise in DSP FWIW. I design by science. Probably better qualified for this build than anyone bagging me.
It is not a matter of ego - just trying to see what is possible with modern drivers, amps and DSP. Tech improves with time people ...
Anyway, it has taken me a little while to sort all these combinations out given RL commitments.
8 litres was indeed too little when using two drivers. I settled on 14 litres and 20 litres depending on drivers.
8 litres would be OK for 1 driver at lower listening levels than I used for some of the drivers but then you lose the vibration cancelling effect.
I settled on the JAB5 power amp with their recommended 36V LRS-350-36 power supply. 24VDC supplies are insufficient even with 4 ohm speakers to compensate for the necessary boost as they clip at moderate volumes with much bass input. This is a great amplifier.
Anyway, the following drivers were tested from approximate worst to best...
My satellites are Scanspeak 10F/8414G10 3.5 inch in 3 litre boxes used for near field listening.
All below are two drivers without passive radiators or ports unless otherwise stated.
I kept the 4mm steel rods as ringing did not seem to be an issue but appreciate the comments about this. Plucking them showed quite a high resonance well outside the bass region and was inaudible in practice with music.
Additional bracing (pine) was also used to prevent warping of the panels when the rods were tightened.
Peerless SDS-160F25PR01-08 6-1/2" these were eventually installed in a IKEA/SONOS subwoofer (using IKEA guts to integratewith a SONOS system). The excursion was insufficient for the gain needed. This speaker needed quite a lot of EQ with the Linkwitz transform (not designed for a small enclosure). Given to a friend with a SONOS soundbar. 14 litres. Calculated f3dB was about 45 Hz. PSU 24V 6A. Not recommended even for a cheap system but better than soundbar alone.
Dayton DCS165-4 x2 - installed in a IKEA/SONOS subwoofer (using IKEA guts to integratewith a SONOS system). The excursion was insufficient for the gain needed. In the end they worked quite well with a SONOS soundbar and 2 passive radiators (8 inch) but f3dB was calculated at about 50Hz so not really what I wanted. Given to friend as gift. No EQ was needed or used. 14 litres. PSU 24V 6A. 24V is fine without any extra EQ.
Element 14 MCM 55-5665 - will also be installed in an IKEA/SONOS system. Yes, 3 friends with SONOS soundbars! These drivers look the best quality of these 3 but I have not yet finished building the system. Modelling suggests it will be the best of the 3 lower quality units.
I should add an internal class D amp was mounted inside each enclosure (not the JAB5) along with analog SP from a custom PCB I designed for the above 3.
Tang Band W6-1139SIF in 14 litre enclosure. Also given to a friend studying electronic engineering in her final year so she could learn about DSP in practice. JAB3+ amp used with 24V supplies (she does not listen loud). At louder volumes these did not sound as good (some distortion evident) as the next two (with JAB5 and 36V supplies). Although x-max is listed at 11.5mm I suspect this is a "fake" x-max measured from voice coil and gap heights and does not represent real linear range. OK for moderate levels though.
See https://www.klippel.de/fileadmin/_migrated/content_uploads/AN_04_Peak_Displacement_Xmax.pdf for an explanation.
SB Acoustics SB23MFCL45-8. Wow these are beautiful drivers. I ended up putting them in a 20 litre box with a 10 inch SB passive radiator just to see how well Linkwitz transforms work in this configuration. JAB5 amp used. They have 12mm xmax and played the loudest and sounded the best but I put them second only because I broke the rules a little (8 inch WITH PR in a bigger box). The EQ worked just fine but you need to tweak the design of the LT parameters to get a flat response. Basically the target f3dB must be higher than the calculated resonance of the PR. I also do not think it is so wise to do dynamic EQ of such a system with a PR. But it sounds great with target f3dB of 35 Hz.
Finally Dayton Epique E180HE-44 7" in 14 litre sealed box, tuned to 30 Hz with LT and JAB5 amp. I have attached the JAB5 project file for this build if anyone wants to try it. This also plays plenty loud enough for near field listening. I measured it at 93 dB at my listening position. The great thing about the Analog Devices SigmaStudio is that it can also show amp headroom in real time which was measured at about 11-14 dB playing bass heavy tracks (eg. Murder Weapon 16 BL). SigmaStudio also have dynamic bass EQ module but I have not used this yet. It is deep enough for me tuned to 30 Hz but there is some distortion at these loud volumes.
The best drivers are also quite expensive, especially if you use 2 in one box.
The idea of dynamic bass EQ would certainly work with single drivers in smaller boxes at lower listening levels. I will try this next. I typically listen more at 80 to 85 dB so in theory a single Epique 7" in a 7 litre box with a JAB5 amp would be fine. This can be done using SCIENCE and not exceeding design parameters.
WINISD Pro Alpha was used which also enables checking design parameters of the drivers are not exceeded. I am keeping the last two subs for myself ...
Just a gentle reminder I am talking about near field 2.1 or FAST type systems... not designed for nightclub use.
It is not a matter of ego - just trying to see what is possible with modern drivers, amps and DSP. Tech improves with time people ...
Anyway, it has taken me a little while to sort all these combinations out given RL commitments.
8 litres was indeed too little when using two drivers. I settled on 14 litres and 20 litres depending on drivers.
8 litres would be OK for 1 driver at lower listening levels than I used for some of the drivers but then you lose the vibration cancelling effect.
I settled on the JAB5 power amp with their recommended 36V LRS-350-36 power supply. 24VDC supplies are insufficient even with 4 ohm speakers to compensate for the necessary boost as they clip at moderate volumes with much bass input. This is a great amplifier.
Anyway, the following drivers were tested from approximate worst to best...
My satellites are Scanspeak 10F/8414G10 3.5 inch in 3 litre boxes used for near field listening.
All below are two drivers without passive radiators or ports unless otherwise stated.
I kept the 4mm steel rods as ringing did not seem to be an issue but appreciate the comments about this. Plucking them showed quite a high resonance well outside the bass region and was inaudible in practice with music.
Additional bracing (pine) was also used to prevent warping of the panels when the rods were tightened.
Peerless SDS-160F25PR01-08 6-1/2" these were eventually installed in a IKEA/SONOS subwoofer (using IKEA guts to integratewith a SONOS system). The excursion was insufficient for the gain needed. This speaker needed quite a lot of EQ with the Linkwitz transform (not designed for a small enclosure). Given to a friend with a SONOS soundbar. 14 litres. Calculated f3dB was about 45 Hz. PSU 24V 6A. Not recommended even for a cheap system but better than soundbar alone.
Dayton DCS165-4 x2 - installed in a IKEA/SONOS subwoofer (using IKEA guts to integratewith a SONOS system). The excursion was insufficient for the gain needed. In the end they worked quite well with a SONOS soundbar and 2 passive radiators (8 inch) but f3dB was calculated at about 50Hz so not really what I wanted. Given to friend as gift. No EQ was needed or used. 14 litres. PSU 24V 6A. 24V is fine without any extra EQ.
Element 14 MCM 55-5665 - will also be installed in an IKEA/SONOS system. Yes, 3 friends with SONOS soundbars! These drivers look the best quality of these 3 but I have not yet finished building the system. Modelling suggests it will be the best of the 3 lower quality units.
I should add an internal class D amp was mounted inside each enclosure (not the JAB5) along with analog SP from a custom PCB I designed for the above 3.
Tang Band W6-1139SIF in 14 litre enclosure. Also given to a friend studying electronic engineering in her final year so she could learn about DSP in practice. JAB3+ amp used with 24V supplies (she does not listen loud). At louder volumes these did not sound as good (some distortion evident) as the next two (with JAB5 and 36V supplies). Although x-max is listed at 11.5mm I suspect this is a "fake" x-max measured from voice coil and gap heights and does not represent real linear range. OK for moderate levels though.
See https://www.klippel.de/fileadmin/_migrated/content_uploads/AN_04_Peak_Displacement_Xmax.pdf for an explanation.
SB Acoustics SB23MFCL45-8. Wow these are beautiful drivers. I ended up putting them in a 20 litre box with a 10 inch SB passive radiator just to see how well Linkwitz transforms work in this configuration. JAB5 amp used. They have 12mm xmax and played the loudest and sounded the best but I put them second only because I broke the rules a little (8 inch WITH PR in a bigger box). The EQ worked just fine but you need to tweak the design of the LT parameters to get a flat response. Basically the target f3dB must be higher than the calculated resonance of the PR. I also do not think it is so wise to do dynamic EQ of such a system with a PR. But it sounds great with target f3dB of 35 Hz.
Finally Dayton Epique E180HE-44 7" in 14 litre sealed box, tuned to 30 Hz with LT and JAB5 amp. I have attached the JAB5 project file for this build if anyone wants to try it. This also plays plenty loud enough for near field listening. I measured it at 93 dB at my listening position. The great thing about the Analog Devices SigmaStudio is that it can also show amp headroom in real time which was measured at about 11-14 dB playing bass heavy tracks (eg. Murder Weapon 16 BL). SigmaStudio also have dynamic bass EQ module but I have not used this yet. It is deep enough for me tuned to 30 Hz but there is some distortion at these loud volumes.
The best drivers are also quite expensive, especially if you use 2 in one box.
The idea of dynamic bass EQ would certainly work with single drivers in smaller boxes at lower listening levels. I will try this next. I typically listen more at 80 to 85 dB so in theory a single Epique 7" in a 7 litre box with a JAB5 amp would be fine. This can be done using SCIENCE and not exceeding design parameters.
WINISD Pro Alpha was used which also enables checking design parameters of the drivers are not exceeded. I am keeping the last two subs for myself ...
Just a gentle reminder I am talking about near field 2.1 or FAST type systems... not designed for nightclub use.