I want to push the limits of big sound from tiny sealed boxes.
I have access to custom EQ (I am designing my own electronics for the task and have tons of experience in DSP).
I am thinking of 5 - 15L or so. Single 6.5 inch to 8 inch drivers. Long throw.
I have already purchased TB W6-1139SIF and plan on using cheap class D amps.
I am interested in trying a few different designs and have just bought a router and table so boxes are not such a big deal any more.
Just looking at a few different drivers .. some car audio subs are suited to this type of thing. eg. JL Audio 8W3v3-4, Hertz ES 200.5 and more regular
SB Acoustics SB23MFCL45-4. Plus the Daytons of course.
From a design perspective it makes sense to have the guts hanging out of the bottom, with the box on 4 legs. Almost all of the power is lost in the motor. If the motor is exposed to air it will cool down better. The advantages of the car subs are twofold:
- low Vas so suited for smaller enclosures
- the motor is generally quite stylish and wires can be hooked up directly via bananas and not look awful and avoids internal box wiring.
Just wondered if anyone has tried this approach. Are there any disadvantages to having the driver backwards? Saves on box volume too. I must say I am leaning toward one of the car subs. I will use Linkwitz TF and whatever else I see fit for EQ.
Any recommendations on drivers? Budget not so much of an issue within reason ... eg. up to USD200 or so but cheaper is better if quality is the same.
PS. I can think of one good reason not to.
Motor noise ...
If motor is inside the sub any chuffing from the motor will be reduced compared to outside at higher volumes.
That is probably good enough reason not to do it.
Efficiency is another minor one due to decreased cone area at the back but may only be 1 dB or so.
I have pretty much decided on down firing BTW.
I have access to custom EQ (I am designing my own electronics for the task and have tons of experience in DSP).
I am thinking of 5 - 15L or so. Single 6.5 inch to 8 inch drivers. Long throw.
I have already purchased TB W6-1139SIF and plan on using cheap class D amps.
I am interested in trying a few different designs and have just bought a router and table so boxes are not such a big deal any more.
Just looking at a few different drivers .. some car audio subs are suited to this type of thing. eg. JL Audio 8W3v3-4, Hertz ES 200.5 and more regular
SB Acoustics SB23MFCL45-4. Plus the Daytons of course.
From a design perspective it makes sense to have the guts hanging out of the bottom, with the box on 4 legs. Almost all of the power is lost in the motor. If the motor is exposed to air it will cool down better. The advantages of the car subs are twofold:
- low Vas so suited for smaller enclosures
- the motor is generally quite stylish and wires can be hooked up directly via bananas and not look awful and avoids internal box wiring.
Just wondered if anyone has tried this approach. Are there any disadvantages to having the driver backwards? Saves on box volume too. I must say I am leaning toward one of the car subs. I will use Linkwitz TF and whatever else I see fit for EQ.
Any recommendations on drivers? Budget not so much of an issue within reason ... eg. up to USD200 or so but cheaper is better if quality is the same.
PS. I can think of one good reason not to.
Motor noise ...
If motor is inside the sub any chuffing from the motor will be reduced compared to outside at higher volumes.
That is probably good enough reason not to do it.
Efficiency is another minor one due to decreased cone area at the back but may only be 1 dB or so.
I have pretty much decided on down firing BTW.
Last edited:
Mounting a driver on the outside of a custom built cabinet is almost always pointless.
Consider: a 20x20x20cm cube would provide 8 litres volume (gross). But also consider you'd have to raise the cabinet by 10cm to allow the driver to hang out of the bottom. Your cabinet now occupies a space 20x20x30cm, If you'd simply built the cabinet that size the gross volume would be 12 litres. Rule-of-thumb: bigger box, bigger bass.
Your entire design concept seems to go against the basic rules of speaker design.
(1) Around the tuning frequency a ported sub will always be superior to a sealed sub by 3dB. (2) High Xmax drivers are less efficient than their counterparts. (3) Ported subs offer superior heat dissipation to their sealed alternatives.
From experience: you'll end up throwing 200 watts at a sealed 6.5" driver to produce the bass a good ported 8" can provide with less that 50 watts.
Consider: a 20x20x20cm cube would provide 8 litres volume (gross). But also consider you'd have to raise the cabinet by 10cm to allow the driver to hang out of the bottom. Your cabinet now occupies a space 20x20x30cm, If you'd simply built the cabinet that size the gross volume would be 12 litres. Rule-of-thumb: bigger box, bigger bass.
Your entire design concept seems to go against the basic rules of speaker design.
(1) Around the tuning frequency a ported sub will always be superior to a sealed sub by 3dB. (2) High Xmax drivers are less efficient than their counterparts. (3) Ported subs offer superior heat dissipation to their sealed alternatives.
From experience: you'll end up throwing 200 watts at a sealed 6.5" driver to produce the bass a good ported 8" can provide with less that 50 watts.
Playing with small high excursion woofers is fun. It is however not the most sensible path to a well performing system from a costs perspective.
High power + small box excludes ported boxes, as the port will not fit. Passive radiator is an option. Check the simulations whether the passive radiator will actually increase efficiency and not just voltage sensitivity. Thermal power handling of the motor might limit sound output before the cone excursion limit is reached in really small boxes.
I would advise against down firing. A small high power subwoofer will have both a high moving mass and cone excursion. The subwoofer will rattle on the floor as it jumps up and down 40 times per second. Furthermore, it costs precious cabinet volume as the space below the cabinet could have been used within the cabinet.
A better measure for efficiency at low (below resonance) frequencies -for a given cone area- is a large BL / sqrt(Re), instead of a small Vas.
The equalizer must have a dynamic bass boost to increase the subjective bass experience. Maximum bass extension at low SPL, while keeping the amplifier and driver within their limits when the volume is turned up.
Agree. Pick the largest box you can get away with and then search for a suitable woofer. Unless absolute performance is not the goal, but performance-per-liter.
High power + small box excludes ported boxes, as the port will not fit. Passive radiator is an option. Check the simulations whether the passive radiator will actually increase efficiency and not just voltage sensitivity. Thermal power handling of the motor might limit sound output before the cone excursion limit is reached in really small boxes.
I would advise against down firing. A small high power subwoofer will have both a high moving mass and cone excursion. The subwoofer will rattle on the floor as it jumps up and down 40 times per second. Furthermore, it costs precious cabinet volume as the space below the cabinet could have been used within the cabinet.
A better measure for efficiency at low (below resonance) frequencies -for a given cone area- is a large BL / sqrt(Re), instead of a small Vas.
The equalizer must have a dynamic bass boost to increase the subjective bass experience. Maximum bass extension at low SPL, while keeping the amplifier and driver within their limits when the volume is turned up.
Front and back have the same effective cone area.
Last edited:
It sounds like an interesting project, instead of a single 6.5", what about two opposing 4" long throw drivers? - although even with two long throw 4"s I guess you'd be struggling to get enough displacement. I've use drivers the wrong way around a few times ( see " a cheep omni... " in the full range section ) and envision no problems. I wonder it anyone's recorded the temp inside a closed box sub, not really anyway for heat to escape.
Thanks, good points all.
Comments:
I am open to suggestions on drivers.
On the 8" front I have narrowed the field down to 3:
SB23MFCL45-4, UM8-22 (2 ohms nice can drive that with class D and massive throw), RSS210HF-4
On 6.5 inch (maybe PRs would help with this size in particular)
DCS165-4, TB W6-1139SIF
I am in Australia with limited choice of drivers as well. Thanks for your suggestions.
Comments:
- I am a big fan of ported subs and all I have built so far have been ported for the reasons Surtsey says. However, friends who live in small places object to big subs (and lets face it a ported sub must be much bigger than 10l just for the port). Of more concern than total volume is floor space. Hence looking at maximum 8 inch and preferably 6.5 inch. Height is less of a concern. You can put a 40 cm tabletop on it and turn it into a coffee table or a big pot plant for better SAF. Computer audio is another area of interest. I have been playing around with 4 inch FR drivers in small boxes and they are great. A small sub is a natural companion to these rather than gigantic 50l boxes. Of course I understand it will never be as loud or as good as a big sub but there are many instances where small is desirable. I want to push the limits to see what you can get from a tiny sub. And that means sealed box, long throw, for better or worse. 4th order BP is another option but does not really seem to offer much advantage.
- Class D amps are cheap now. I would be using the tda7498e base design. I will end up designing my own PCBs. Watts are cheap so amp power/expense/size is less of an issue than it was 10 years ago.
- I will be putting my own biquads into a microcontroller with dynamic EQ based on volume level. I have toyed around with the MAXIM MAX98395 evaluation kit designed for tiny 1 inch speakers and what it achieves is incredible but it does not really do anything you cannot do with a microcontroller. I was thinking of stealing the dynamic EQ aspect and trying it out with a bigger sub.
Attention Required! | Cloudflare
- I decided against putting the guts outside almost as soon as I posted as noted in my PS.
- Side firing is an option as opposed to bottom firing. I will try both but I have measured both before and found there is little difference.
- A practical aspect is the jumping around thing. Big subs of course weigh more and are less prone to this. I may need to mass load the box to reduce it. Driver orientation may also have an effect.
- I started off with the idea of 4 inch but the driver piston area is just too small for decent volume in a small box from models. 6.5 inch seems about the minimum to get acceptable bass extension and volume. 4 inch drivers are also not as long throw as 6 inch so 1x8>>1x6.5 >> 2x4.
- Passive radiators are an option but again modelling shows the PR area needs to be bigger than the driver area for it to be viable, thus increasing the box volume. I will take another look at this though. Maybe 2 PRs on the sides but I am not convinced this will be better than throwing watts at a single driver. I will model it and see.
I am open to suggestions on drivers.
On the 8" front I have narrowed the field down to 3:
SB23MFCL45-4, UM8-22 (2 ohms nice can drive that with class D and massive throw), RSS210HF-4
On 6.5 inch (maybe PRs would help with this size in particular)
DCS165-4, TB W6-1139SIF
I am in Australia with limited choice of drivers as well. Thanks for your suggestions.
I have decided to stick with sealed box and dynamic EQ for this reason.
Ports and PRs are limited in extension to what you tune the port/PR to. This is fixed as it is mechanical. The frequency chosen is what you can get with expected listening volume.
With a Linkwitz transform it is quite feasible to adjust it based off volume as it is electronic. At low listening levels you can feasibly go as low as you like by extending the LF poles. As volume increases, excursion increases and you can adjust the filter so that excursion is limited by decreasing the filter time constants. You can theoretically do this with a microcontroller.
Of course you need experience with microcontrollers, of which I have plenty. In practice it will be harder than this simple explanation but you only know by trying.
EDIT - This would require storing the waveform for some amount of time (eg. 100msec) in order to see how loud the music will be coming up. Remember in LT, f0 and Q0 are fixed. fp and Qp are chosen based off expected listening volume. I am suggesting fp and Qp are time varying filters based off the volume in the next 100 msec (or whatever interval is chosen). I would expect possible problems in audio changing filters on the fly like this. Of course this is a digital solution but that is all I listen to now these days as most others do as well.
Another option is to just look at average RMS volume over the last few seconds and more slowly change the filters assuming the volume remains relatively constant. This might work better in practice. Plenty of options. Does anyone know of any commercial systems that do this?
Ports and PRs are limited in extension to what you tune the port/PR to. This is fixed as it is mechanical. The frequency chosen is what you can get with expected listening volume.
With a Linkwitz transform it is quite feasible to adjust it based off volume as it is electronic. At low listening levels you can feasibly go as low as you like by extending the LF poles. As volume increases, excursion increases and you can adjust the filter so that excursion is limited by decreasing the filter time constants. You can theoretically do this with a microcontroller.
Of course you need experience with microcontrollers, of which I have plenty. In practice it will be harder than this simple explanation but you only know by trying.
EDIT - This would require storing the waveform for some amount of time (eg. 100msec) in order to see how loud the music will be coming up. Remember in LT, f0 and Q0 are fixed. fp and Qp are chosen based off expected listening volume. I am suggesting fp and Qp are time varying filters based off the volume in the next 100 msec (or whatever interval is chosen). I would expect possible problems in audio changing filters on the fly like this. Of course this is a digital solution but that is all I listen to now these days as most others do as well.
Another option is to just look at average RMS volume over the last few seconds and more slowly change the filters assuming the volume remains relatively constant. This might work better in practice. Plenty of options. Does anyone know of any commercial systems that do this?
Last edited:
You seem to be working from a base misconception. I have been down your rabbit hole and it's not pretty. Commercially available options show a 6.5" driver needs about 12 litres in a ported cabinet, and 8" driver needs about 20 litres. There is an unloaded Tannoy SFX box next to me as I type. It has an internal volume of 19 litres.
I once owned a Bose Companion 5 series II sub. The 7.5 litre unit uses a 6" dual-coil driver and attempts to throw a couple of hundred watts at it. The results are unimpressive.
My best results in building a micro-sub used a 210h x 210w x 250d mm (8.5 litres) cabinet and a 6.5" front-loaded driver with a port in the side elevation. It used a 50w Class D amp, and kicked the Bose's ***. However, it could never compete with the 20 litre subs I use in my living room. The microsub survives because it's the only sub that fits under my bed.
My rabbit hole is dynamic LT alteration based off volume. Have you tried that? If not you have not been down the same hole. I always challenge accepted beliefs and practice - that is my nature. I am often wrong but not always. You cannot possibly expect any fixed mechanical tuning to behave in the same way as dynamic electronic tuning.
I like the idea of two drivers in force cancellation mode. This reduces required xmax, meaning cheaper drivers can be used and reduced shaking. Initial modelling is showing -3dB at 33Hz and 97dB with peak amp power of 140VA, 6mm max excursion (2 6.5 inch drivers) in an 8.5l box. At lower listening levels than this bass could be extended further.
I like the idea of two drivers in force cancellation mode. This reduces required xmax, meaning cheaper drivers can be used and reduced shaking. Initial modelling is showing -3dB at 33Hz and 97dB with peak amp power of 140VA, 6mm max excursion (2 6.5 inch drivers) in an 8.5l box. At lower listening levels than this bass could be extended further.
Anyone can simulate what I am looking at with WINISD.
Choose 2 x Dayton DCS165-4 drivers in a closed 8.5 l box. Their xmax is listed at 6mm but this is a fairly typical driver. There are better ones (and worse ones).
Under filters add a Linkwitz transform, f0= 67 Hz, Q = 0.7, fp = 34 Hz, Q = 1.
Add an HPF (to reduce xmax) n = 4 fc = 30 Hz. Signal level 8W.
The result is average level 96 dB, fc ~ 33Hz (93dB). Excursion 6mm at about 33Hz. Amplifier apparent load power 60W.
Now you can increase loudness, keeping the same xmax and increasing fp in the LT to compensate. You can also reduce listening level to say 90 dB, change the LT to 25 Hz and the HPF to 15Hz, keep the xmax the same and get an fc of about 23 Hz.
It is completely impossible to do this with mechanical tuning. If people keep replying about bigger is better we are talking about different things entirely. Please keep that in mind. I know bigger is better FFS but many times people want small.
Now I know you cannot do this dynamic EQ with MiniDSP or similar commercial products. I have vast experience with DSP and microcontroller programming (30 years+) and am talking about designing a new product which as far as I know is unavailable. I don't see any reason why dynamic EQ such as this cannot be done though with a cheap microcontroller. I guess I will find out.
Choose 2 x Dayton DCS165-4 drivers in a closed 8.5 l box. Their xmax is listed at 6mm but this is a fairly typical driver. There are better ones (and worse ones).
Under filters add a Linkwitz transform, f0= 67 Hz, Q = 0.7, fp = 34 Hz, Q = 1.
Add an HPF (to reduce xmax) n = 4 fc = 30 Hz. Signal level 8W.
The result is average level 96 dB, fc ~ 33Hz (93dB). Excursion 6mm at about 33Hz. Amplifier apparent load power 60W.
Now you can increase loudness, keeping the same xmax and increasing fp in the LT to compensate. You can also reduce listening level to say 90 dB, change the LT to 25 Hz and the HPF to 15Hz, keep the xmax the same and get an fc of about 23 Hz.
It is completely impossible to do this with mechanical tuning. If people keep replying about bigger is better we are talking about different things entirely. Please keep that in mind. I know bigger is better FFS but many times people want small.
Now I know you cannot do this dynamic EQ with MiniDSP or similar commercial products. I have vast experience with DSP and microcontroller programming (30 years+) and am talking about designing a new product which as far as I know is unavailable. I don't see any reason why dynamic EQ such as this cannot be done though with a cheap microcontroller. I guess I will find out.
The MCMs work quite well too in simulations of this configuration and are dirt cheap
MCM 6.5 inch drivers
Many of the drivers I try out seem to end up with similar Bode plots but each with their own filter configurations. ie. 96dB/35Hz at xmax for driver. Amp power varies a bit though. Small changes in fc have large effects on power. This is just for 6.5 inch drivers. Bigger drivers/enclosures would yield further improvements.
I think this is the future of EQ. Amp power has been wasted so far but it can be used to extend bass response at lower listening levels (which is probably most of the time for most of us).
EDIT: AND needing smaller box size for most listening except when you want to utilise maximum party output.
Some say the glass is half empty. Some say it is half full.
Engineers say the glass is twice as big as it needs to be.
MCM 6.5 inch drivers
Many of the drivers I try out seem to end up with similar Bode plots but each with their own filter configurations. ie. 96dB/35Hz at xmax for driver. Amp power varies a bit though. Small changes in fc have large effects on power. This is just for 6.5 inch drivers. Bigger drivers/enclosures would yield further improvements.
I think this is the future of EQ. Amp power has been wasted so far but it can be used to extend bass response at lower listening levels (which is probably most of the time for most of us).
EDIT: AND needing smaller box size for most listening except when you want to utilise maximum party output.
Some say the glass is half empty. Some say it is half full.
Engineers say the glass is twice as big as it needs to be.
Last edited:
Tsardoz,
I like where you're going here. A lot of commercial products do stuff like this, including small bluetooth speakers which have plenty of bass at low volumes, but sound thin when pushed.
In my experience, these mini-subs need quite a lot of power to really get going. I've hit those Tang Band 6.5" mini-subs with around 500w peaks before, but harmonic distortion was getting high at that point.
If you do come up with a reasonably-user-friendly way to implement excursion limiting, I'd be really interested to hear about it.
Chris
I like where you're going here. A lot of commercial products do stuff like this, including small bluetooth speakers which have plenty of bass at low volumes, but sound thin when pushed.
In my experience, these mini-subs need quite a lot of power to really get going. I've hit those Tang Band 6.5" mini-subs with around 500w peaks before, but harmonic distortion was getting high at that point.
If you do come up with a reasonably-user-friendly way to implement excursion limiting, I'd be really interested to hear about it.
Chris
Bag End speaker introduced the concept of dynamic LF EQ back in the early 1990s with their analog ELF-M2 module. The future is already in the pasture
."The Continuous Variable Recovery limiter circuit provides a fast attack time with a variable recovery time dependent upon the amount of limiting originally called for. The ELF (Extended Low Frequency) concealment performs dynamic control of reducing the low frequency extension. It will allow the system to play with the upper bass notes unaffected while protecting the system from overload caused by playing the lower notes louder than the amplifier and/or loudspeaker can reproduce safely and undistorted.".
https://bagend.com/wp-content/uploads/2014/09/ELF-GUIDE-COMPLETE-V-1.3.pdf
Interesting discussion on parameter variation here including BL ...
Is High BL or Low BL Better? The Question is Wrong! - YouTube
The key parameters in my proposed design after trial modelling many speakers seem to just be Xmax then Re for a given speaker diameter. The filters seem to be able to fix any other irregularities but some have quite high power requirements that might be unrealistic. I would think the less the EQ filter needs to change the basic response the more successful it will be.
I don't believe for a second that is all there is to it. Drivers will respond to the EQ in different ways.
What is really concerning is how old all those guys are in the video link.
I have added another driver the super cheap Peerless SDS-160F25PR01-08 which also has similar curves but needs a bit more amp power. It will be interesting to see how these drivers differ.
Last edited:
Bag End speaker introduced the concept of dynamic LF EQ back in the early 1990s with their analog ELF-M2 module. The future is already in the pasture
"The Continuous Variable Recovery limiter circuit provides a fast attack time with a variable recovery time dependent upon the amount of limiting originally called for. The ELF (Extended Low Frequency) concealment performs dynamic control of reducing the low frequency extension. It will allow the system to play with the upper bass notes unaffected while protecting the system from overload caused by playing the lower notes louder than the amplifier and/or loudspeaker can reproduce safely and undistorted.".
https://bagend.com/wp-content/uploads/2014/09/ELF-GUIDE-COMPLETE-V-1.3.pdf
Yah I just found this google patent too
US Patent for Dynamically extending loudspeaker capabilities Patent (Patent # 10,200,003 issued February 5, 2019) - Justia Patents Search
so the Bag End speaker makes their patent invalid. I will fight Google if I need to. I could not possibly lose to Google in court ...
EDIT
I got onto this field because I was designing a 1 inch BT speaker and found this MAXIM amp chip
MAX98390
It does dynamic EQ with amazing results. A Maxim engineer told me they use Linkwitz TF with power adjusted HPF so they dont blow up the speaker. This gave me the idea.
I tried it with hifi speakers and results were crap but it was designed for tiny 30c speakers not hifi.
When I looked at it I realised it did nothing you could not do with a microcontroller.
Except the amp cost $2 cf 11c for a Chinese class D 5V amp chip. Except it gives 3dB extra gain for the voltage boosting.
Anyway we will see how I go. I am not claiming to make the best sub ever but maybe we can make better use of cabinet volume for some situations like PC speakers etc.
Last edited:
Just having a look at 4 inch drivers there is a big performance hit
eg. SB Acoustics SB12NRX25-4 5mm xmax
5L box, 2 drivers, 91.5 dB f-3dB = 50Hz amp power needed though is only 25W
This might be OK for PC speakers but I think the sweet spot is more like 6.5 inch
(and of course this will still not fill a big room)
eg. SB Acoustics SB12NRX25-4 5mm xmax
5L box, 2 drivers, 91.5 dB f-3dB = 50Hz amp power needed though is only 25W
This might be OK for PC speakers but I think the sweet spot is more like 6.5 inch
(and of course this will still not fill a big room)
I take that back.
There seems to be more improvement the smaller the driver.
Of course 2x4 is not as good as 2x6.5 but it seems you can get away with a 5 litre box and drivers that cost half as much for a loss of about 6dB and some extension but much less power.
This size would be good for PC or other very close applications.
I will make one of these too (SB12PFC25-8).
Models showing when the cabinet volume is pushed too small the amp power rapidly escalates near fp ... if this happens for any given driver pair increasing enclosure volume has the greatest improvement.
An HPF seems essential to limit cone excursion below fp, unless it is very long throw (in which case it needs more amp power anyway).
There seems to be more improvement the smaller the driver.
Of course 2x4 is not as good as 2x6.5 but it seems you can get away with a 5 litre box and drivers that cost half as much for a loss of about 6dB and some extension but much less power.
This size would be good for PC or other very close applications.
I will make one of these too (SB12PFC25-8).
Models showing when the cabinet volume is pushed too small the amp power rapidly escalates near fp ... if this happens for any given driver pair increasing enclosure volume has the greatest improvement.
An HPF seems essential to limit cone excursion below fp, unless it is very long throw (in which case it needs more amp power anyway).
Small speakers are great, but there are limits to what is useful output.
For a given excursion, 2x6.5” have 8.4dB more Sound Pressure Level (SPL) potential than 2x4”.
For example, with 6mm excursion at 31.5Hz, a pair of sealed 6.5” can do around 93.4 dB SPL, the 2x4” 85dB. Finding a 6mm Xmax on a 4” would be unusual…
At 1kHz, an increase of around 9-10dB sounds about “twice as loud”.
At low frequencies, a louder level is required to sound equal in level, that difference being called the Equal Loudness Contour (ELC) , which changes with SPL.
Although the SPL difference between the 4” and 6.5” drivers is 8.4dB at 31.5Hz, the ELC drops 15dB, from 50 Phon (SPL @1kHz) down to only 35 Phon.
50 dBA is pretty quiet, less than most conversations at one meter, 35 dBA drops into the whisper level, below the ambient noise levels common in many rooms.
For a given excursion, 2x6.5” have 8.4dB more Sound Pressure Level (SPL) potential than 2x4”.
For example, with 6mm excursion at 31.5Hz, a pair of sealed 6.5” can do around 93.4 dB SPL, the 2x4” 85dB. Finding a 6mm Xmax on a 4” would be unusual…
At 1kHz, an increase of around 9-10dB sounds about “twice as loud”.
At low frequencies, a louder level is required to sound equal in level, that difference being called the Equal Loudness Contour (ELC) , which changes with SPL.
Although the SPL difference between the 4” and 6.5” drivers is 8.4dB at 31.5Hz, the ELC drops 15dB, from 50 Phon (SPL @1kHz) down to only 35 Phon.
50 dBA is pretty quiet, less than most conversations at one meter, 35 dBA drops into the whisper level, below the ambient noise levels common in many rooms.
Attachments
OK preliminary results are in with the TB & Dayton 6.5 inch subs.
The maximum level with acceptable levels of distortion (to my ear - no obvious motor noise) is really much quieter than the models would indicate and really too low even for PC speakers. I was EQing to about 40 Hz.
I think the main reason is Xmax is not really linear at all - it is defined by geometry and not distortion levels.
I know 2x6.5" can give more than acceptable output. I have a commercial sub with similar drivers (much bigger bass reflex box) which is plenty loud for this type of listening.
The ways to increase loudness:
1) increase box size so less amplification is needed
2) add a port or passive radiator
3) increase driver size
4) less bass extension
or combination of all 4
The box needs to be MUCH bigger for a decent port area that won't chuff (too big for what I want).
I can add a passive radiator and make a bigger box (double the volume) for little extra cost.
I can tune the PR to (say) 30 Hz and apply some sort of LF shelf (like a Linkwitz transform) to better drive the PR than from the drivers alone. This only needs around 4-5 dB gain. However the concept of variable bass extension with volume becomes void once a mechanical resonator is installed.
Bottom line - you cannot design subwoofer systems just based off the specs, particularly Xmax which vastly overestimates the travel before there is obvious audible distortion.
I will continue to experiment - I am sure I will come up with something but I was definitely too optimistic.
The maximum level with acceptable levels of distortion (to my ear - no obvious motor noise) is really much quieter than the models would indicate and really too low even for PC speakers. I was EQing to about 40 Hz.
I think the main reason is Xmax is not really linear at all - it is defined by geometry and not distortion levels.
I know 2x6.5" can give more than acceptable output. I have a commercial sub with similar drivers (much bigger bass reflex box) which is plenty loud for this type of listening.
The ways to increase loudness:
1) increase box size so less amplification is needed
2) add a port or passive radiator
3) increase driver size
4) less bass extension
or combination of all 4
The box needs to be MUCH bigger for a decent port area that won't chuff (too big for what I want).
I can add a passive radiator and make a bigger box (double the volume) for little extra cost.
I can tune the PR to (say) 30 Hz and apply some sort of LF shelf (like a Linkwitz transform) to better drive the PR than from the drivers alone. This only needs around 4-5 dB gain. However the concept of variable bass extension with volume becomes void once a mechanical resonator is installed.
Bottom line - you cannot design subwoofer systems just based off the specs, particularly Xmax which vastly overestimates the travel before there is obvious audible distortion.
I will continue to experiment - I am sure I will come up with something but I was definitely too optimistic.
Last edited:
Tsardoz,
By definition, Xmax is maximum linear excursion, typically the point at which the length of the voice coil in the magnetic gap begins to decrease and distortion starts to increase. Xmax should be either measured as the greater of either voice coil height minus top plate thickness, divided by 2, or the excursion where THD reaches 10%.
Most people find THD of 10% acceptable below 100 Hz.
If you happened to measure using the dB "A" scale, it would read -35dB at 40 Hz, "C" scale about -3dB at 40 Hz.
Room modes can cause large peaks and dips (+/- 12 dB is not uncommon) depending on sub and listening locations.
What was the SPL and excursion you measure at 40Hz which you found distortion unacceptable?
Art