Hey Folks!
I want to build a small and lightweight speaker that I can take with me while traveling.
Yes, I know there are tons of products out there already, but I really enjoy building my own gear instead of just buying something off the shelf.
I'm also hoping to get better performance than a typical JBL Xtreme-style speaker.
I'm already quite experienced with electronics, so that's not a problem. However, I’d really appreciate some feedback from the diyAudio hive mind on my current plan.
The enclosure will be around 18×18×8–10 cm and made from carbon fiber to keep it lightweight. All components were chosen based on size and weight.
So far, WinISD suggests I should be able to get around 99 dB SPL @ 30 W without exceeding Xmax (for both the woofers and PR), and around 49 Hz @ -3 dB.
However, distortion looks decent up to 2 kHz, so I’m wondering if I can push the woofer that high and just tame the response peak with the DSP?
Even so, the tweeter would still need to go quite deep. I'm also unsure whether the ND20F-4 is the right choice from a distortion standpoint.
HiFiCompass did some measurements: LINK
Looking at the HD response at 315 mm, -30 dB H2 at 2 kHz doesn’t inspire much confidence.
I’m wondering if I should go for a 1" tweeter instead – but which one could still fit, offer lower distortion, and cross low enough?
I’d really prefer a neodymium-magnet tweeter to keep weight down. Any recommendations?
I've also considered using a compact full-range driver as a mid+tweeter, but couldn’t find anything that fits the size constraints, has enough sensitivity, and uses a neodymium motor.
Some preliminary renderings of the planed enclosure:
Thanks Folks!
Have a nice day
Farbe
EDIT:
I just found the SB Acoustics SB14ST-C000-4, its smaller with just 0,6" but seems to show better distortion (again HifiCompass). However sensitivity is also 3db down, not sure about this one.
EDIT2:
I might also be able to fit a Dayton Audio ND25FA-4 1" by milling two screw holes off. Not sure how this would change the response. I also found no distortion measurments.
I want to build a small and lightweight speaker that I can take with me while traveling.
Yes, I know there are tons of products out there already, but I really enjoy building my own gear instead of just buying something off the shelf.
I'm also hoping to get better performance than a typical JBL Xtreme-style speaker.
I'm already quite experienced with electronics, so that's not a problem. However, I’d really appreciate some feedback from the diyAudio hive mind on my current plan.
My goals:
- Small – max. 3 L internal volume
- Loud – ~100 dB SPL
- Lightweight – <2 kg
- Low-end extension – 50 Hz @ -3 dB
How do I plan to achieve this?
- 3× Tang Band W3-2052S – 3" woofers
- 2× Dayton ND20F-4 – ¾" tweeters
- 1× Dayton DSA175-PR – 6.5" passive radiator
- Custom PCBwith:
- Custom 4s 21700 battery pack
The enclosure will be around 18×18×8–10 cm and made from carbon fiber to keep it lightweight. All components were chosen based on size and weight.
So far, WinISD suggests I should be able to get around 99 dB SPL @ 30 W without exceeding Xmax (for both the woofers and PR), and around 49 Hz @ -3 dB.
The real issue: Woofer/tweeter crossover
The W3-2052S shows some signs of ??cone breakup?? around 750 Hz, according to Voice Coil Magazine measurements (available on Tang Band’s site).However, distortion looks decent up to 2 kHz, so I’m wondering if I can push the woofer that high and just tame the response peak with the DSP?
Even so, the tweeter would still need to go quite deep. I'm also unsure whether the ND20F-4 is the right choice from a distortion standpoint.
HiFiCompass did some measurements: LINK
Looking at the HD response at 315 mm, -30 dB H2 at 2 kHz doesn’t inspire much confidence.
Mounting constraints and alternatives
I’m planning on an 18×18 cm front baffle, with the three woofers arranged in a triangle. That leaves me with only around 50 mm of space for a tweeter.I’m wondering if I should go for a 1" tweeter instead – but which one could still fit, offer lower distortion, and cross low enough?
I’d really prefer a neodymium-magnet tweeter to keep weight down. Any recommendations?
I've also considered using a compact full-range driver as a mid+tweeter, but couldn’t find anything that fits the size constraints, has enough sensitivity, and uses a neodymium motor.
Some preliminary renderings of the planed enclosure:
So… what do you think?
- Will such a low crossover work?
- Will my ears bleed from the tweeter distortion?
- Is this whole setup a bad idea?
- Any better tweeter or full-range suggestions?
- Any issues with the three woofers? (Cancelation?)
Thanks Folks!
Have a nice day
Farbe
EDIT:
I just found the SB Acoustics SB14ST-C000-4, its smaller with just 0,6" but seems to show better distortion (again HifiCompass). However sensitivity is also 3db down, not sure about this one.
EDIT2:
I might also be able to fit a Dayton Audio ND25FA-4 1" by milling two screw holes off. Not sure how this would change the response. I also found no distortion measurments.
Last edited:
I also found the SB26STCN-C000-4 which would also be able to fit if i cut its mounting flange down.
Its distortion figures look much better.
Does anybody have experience with cutting the flat part of the waveguide off? I would imagine that it shouldnt matter much as the tapered part wouldt be touched. However the flat plane would be missing as it would be mounted directly adjecent to the Tangband woofer and therefore experience some sort of baffle step on the surround of the woofer (when mounted completely flush with it).
Its distortion figures look much better.
Does anybody have experience with cutting the flat part of the waveguide off? I would imagine that it shouldnt matter much as the tapered part wouldt be touched. However the flat plane would be missing as it would be mounted directly adjecent to the Tangband woofer and therefore experience some sort of baffle step on the surround of the woofer (when mounted completely flush with it).
Just some thoughts:
It's an interesting project, keep it up!
- Two tweeters nearby will exhibit strong lobing and resulting comb filtering. Use one tweeter instead.
- Three midwoofers driven up to the tweeter may also show some lobing. Maybe use two of them as 2.5 way woofers?
- are you sure you included baffle step in your max spl calculation? Or do you plan to use the speaker near-wall?
- your enclosure walls look very thin. Pressure in a resonator enclosure can get quite high.
It's an interesting project, keep it up!
By the way: another 100 dB down to 50 Hz project, just for comparison:
I am not sure how bad the comb filtering will be in real life. Most of the protable speakers have two tweeters side by side.
But i am also not sure how much the second tweeter actually offers in terms of performace.
I imagined it would enchance the stereo image -- granted, there isnt much of an image anyways with all the sources beeing that close together.
Plan is to use the speaker on the ground, thats also what WinISD is simulating e.g. Half space. So ~100db @ 50Hz should work outdoors on the ground.
I would take advantage of a wall / corner if i would find one outdoors 🙂
The enclosure is very thin indeed, I intended to laminate 3x 200g/m^2 carbon fiber with a 2mm PET foam core, which should give me around 3mm wall thickness with an equivalent strength of roughly 1,5mm of mild steel. That is hopefully enough to keep it stiff / damped enough (due to the foam core) while still being incredable light. Braces are planed, i might end up connecting the inner walls with a cross brace depending on the actual performance of the housing.
I have seen the build, its incredible! However its way to big and heavy for me. Less performance but more portablility is what i want to arcive.
But i am also not sure how much the second tweeter actually offers in terms of performace.
I imagined it would enchance the stereo image -- granted, there isnt much of an image anyways with all the sources beeing that close together.
Plan is to use the speaker on the ground, thats also what WinISD is simulating e.g. Half space. So ~100db @ 50Hz should work outdoors on the ground.
I would take advantage of a wall / corner if i would find one outdoors 🙂
The enclosure is very thin indeed, I intended to laminate 3x 200g/m^2 carbon fiber with a 2mm PET foam core, which should give me around 3mm wall thickness with an equivalent strength of roughly 1,5mm of mild steel. That is hopefully enough to keep it stiff / damped enough (due to the foam core) while still being incredable light. Braces are planed, i might end up connecting the inner walls with a cross brace depending on the actual performance of the housing.
I have seen the build, its incredible! However its way to big and heavy for me. Less performance but more portablility is what i want to arcive.
I'm a fan of early prototypes being fast, ugly, and cheap. If I were undertaking a similar project, I would test the woofer system in a plain MDF box as early as possible to determine if it has the desired output.
I've used the TB W3-1876S in desk speakers with passive radiators at around 1 liter each, and while they are one of the better performers in small enclosures, I also wouldn't classify their output at 50 Hz as "loud." Their specs look pretty similar to the W3-2052S.
I actually 3D printed a enclosure for tests and i am so far not very staisfied with the results. The woofers are quite good but i am not quite sure what to think about the tweeters and/or the crossover.
I cant really measure indoors (not enough space), so I lack good data to programm the DSP correctly.
I wanted to do that when i get the chance to measure outdoors.
It somehow sounds very harsh. It might be the comb filtering, the distortion or that the response isnt corrected enough.
I want to try some FIR filters (to keep the phase correct) but so far i only did some rough 4ms gated close to walls measurments and played with multiple EQs to get is somewhat flat.
The 3D printed housing also isnt stiff or airtight enough, however the bass extension looks to be like simulated. I dont have enough amplifier power to test at full power (and the housing might explode then 😀 ).
I am currently waiting on the Bluetooth modules and working on the DSP + AMP pcbs.
The W3-1876S puts out a tiny bit more but runs into X-max. (Red: 2052S / Green:1876S)
Both are very similar, the 1876S might even be a better choice depending on how high it goes..
However i already have the 2052s 🙄
Running three, with 3x the power, gives them +9,7db compared to running a single one. So they should go close to 100db, maybe even over if i just care about x-max and exceed the power rating by a little.
I cant really measure indoors (not enough space), so I lack good data to programm the DSP correctly.
I wanted to do that when i get the chance to measure outdoors.
It somehow sounds very harsh. It might be the comb filtering, the distortion or that the response isnt corrected enough.
I want to try some FIR filters (to keep the phase correct) but so far i only did some rough 4ms gated close to walls measurments and played with multiple EQs to get is somewhat flat.
The 3D printed housing also isnt stiff or airtight enough, however the bass extension looks to be like simulated. I dont have enough amplifier power to test at full power (and the housing might explode then 😀 ).
I am currently waiting on the Bluetooth modules and working on the DSP + AMP pcbs.
The W3-1876S puts out a tiny bit more but runs into X-max. (Red: 2052S / Green:1876S)
Both are very similar, the 1876S might even be a better choice depending on how high it goes..
However i already have the 2052s 🙄
Running three, with 3x the power, gives them +9,7db compared to running a single one. So they should go close to 100db, maybe even over if i just care about x-max and exceed the power rating by a little.
It's pretty common to start with quasi-anechoic measurements indoors to get the basics of the crossover worked out. 4 ms isn't that bad given your cross point (if that's reflection free).
Later, in-room response can also be used to make adjustments. If you set your window to something like 100-200 milliseconds, you'll have much more resolution and you'll get a better feel for what the bass is doing. Of course room reflections and modes will be heavily influencing this measurement. And keep in mind that most speakers will have an in-room response that gradually slopes down with increasing frequency. How much is a matter of personal preference. As you implied, your EQ for use outside will likely be significantly different though, since you won't have any bass reinforcement from the room.
It's hard to say where you are at this point. Your design is definitely stretching things a bit with that tweeter, woofer, and cross point, but whether that's the issue or a more basic frequency imbalance is to blame is hard to say without seeing more data.
I doubt comb filtering is the explanation for the harshness though.
Later, in-room response can also be used to make adjustments. If you set your window to something like 100-200 milliseconds, you'll have much more resolution and you'll get a better feel for what the bass is doing. Of course room reflections and modes will be heavily influencing this measurement. And keep in mind that most speakers will have an in-room response that gradually slopes down with increasing frequency. How much is a matter of personal preference. As you implied, your EQ for use outside will likely be significantly different though, since you won't have any bass reinforcement from the room.
It's hard to say where you are at this point. Your design is definitely stretching things a bit with that tweeter, woofer, and cross point, but whether that's the issue or a more basic frequency imbalance is to blame is hard to say without seeing more data.
I doubt comb filtering is the explanation for the harshness though.
I played with it a little bit and it might really just be too much treble.
I will measure it again in the upcoming days.
Thats my housing and prototype setup so far:
Just some very cheapo class D boards and an Wondon AMP2 as the DSP.
The amps are crap, not sure what they did wrong of the the chip is just a copy and not a real TPA3116, it has a crap ton of noise / high noisefloor.
So maybe the reconstruction filters are also crap and therefore they introduce quite a bit of distortion. I didnt measure that.
Maybe i should go with a single bigger tweeter to get the lower cross and eliminate comb filtering. I could also eliminate a complete amplifier channle and save space/weight/increase battery runtime.
Any opinions on stereo image for these small setups? Worthless anyways? I played some music while switching between mono sum and stereo and honestly couldnt hear much of a difference.. However it feels wastefull not to include two channels. Some phones are able to create good stereo seperation with two very close speakers. Maybe i can do some stereo widening with the DPS.
Electronics:
Thats my current plan.
The TPS61288 will boost the 4S battery to a steady 18V for the amps.
It should be able to deliver ~125W continous with just 2W of losses.
This should give the TPA3126 enough voltage to deliver steady 34W @ 1%THD / 42W @ 10%THD for the woofers.
I dont think i will use the full 42W, the W3s are rated for just 15W / 30W. They should in theory stay inside their Xmax even at 42W but they might overheat or experience quite a bit of power compression.
The TPS25751 will be the USB-C PD controller, it supports a dual role port so charging (100W) and discharging with 60W (in my implementation).
Battery will be 4S 21700 with 5800mAh so a total of ~80Wh should be way more than enough for 8h playtime on resonable volumes.
And if the battery is empty, the 100W PD input power should be enough to keep it running indefinitly with typical music.
So charging a laptop from the speaker should be very doable.
I will measure it again in the upcoming days.
Thats my housing and prototype setup so far:
Just some very cheapo class D boards and an Wondon AMP2 as the DSP.
The amps are crap, not sure what they did wrong of the the chip is just a copy and not a real TPA3116, it has a crap ton of noise / high noisefloor.
So maybe the reconstruction filters are also crap and therefore they introduce quite a bit of distortion. I didnt measure that.
Maybe i should go with a single bigger tweeter to get the lower cross and eliminate comb filtering. I could also eliminate a complete amplifier channle and save space/weight/increase battery runtime.
Any opinions on stereo image for these small setups? Worthless anyways? I played some music while switching between mono sum and stereo and honestly couldnt hear much of a difference.. However it feels wastefull not to include two channels. Some phones are able to create good stereo seperation with two very close speakers. Maybe i can do some stereo widening with the DPS.
Electronics:
Thats my current plan.
The TPS61288 will boost the 4S battery to a steady 18V for the amps.
It should be able to deliver ~125W continous with just 2W of losses.
This should give the TPA3126 enough voltage to deliver steady 34W @ 1%THD / 42W @ 10%THD for the woofers.
I dont think i will use the full 42W, the W3s are rated for just 15W / 30W. They should in theory stay inside their Xmax even at 42W but they might overheat or experience quite a bit of power compression.
The TPS25751 will be the USB-C PD controller, it supports a dual role port so charging (100W) and discharging with 60W (in my implementation).
Battery will be 4S 21700 with 5800mAh so a total of ~80Wh should be way more than enough for 8h playtime on resonable volumes.
And if the battery is empty, the 100W PD input power should be enough to keep it running indefinitly with typical music.
So charging a laptop from the speaker should be very doable.
Last edited:
Looks like i will order two SB Acoustics SB26STCN-C000-4
They should be able to play much lower, have less distortion across the board and show better decay charactaristics.
They should be able to play much lower, have less distortion across the board and show better decay charactaristics.
I ordered two and they surely measure and sound better. 😀
It doenst sound as harsh anymore I attribute this to lower distortion.
Hopefully the sound difference isnt just my subconscious telling me i payed more so they must be better.
Here are some measurements with REW
(nearfield directly on tweeter and woofer after crossing and EQing - ignoring PR response)
I moved to a linkwitz riley 48db crossover at 1khz to better control the cone breakup of the woofer and safe the tweeter from too much low end.
The measurements in REW are still with LR 36db crossover.
The measurement at the end is at just 30cm distance with reflections. I seem to still have phasing issues, but i think i will work that out with the real housing.
I have the tweeter just sitting on top of the 3D printed housing.
I also need to check how much the baffle will influence this in the far field.
I included the raw measurements of the woofer and tweeter before EQing and crossing.
It doenst sound as harsh anymore I attribute this to lower distortion.
Hopefully the sound difference isnt just my subconscious telling me i payed more so they must be better.
Here are some measurements with REW
(nearfield directly on tweeter and woofer after crossing and EQing - ignoring PR response)
I moved to a linkwitz riley 48db crossover at 1khz to better control the cone breakup of the woofer and safe the tweeter from too much low end.
The measurements in REW are still with LR 36db crossover.
The measurement at the end is at just 30cm distance with reflections. I seem to still have phasing issues, but i think i will work that out with the real housing.
I have the tweeter just sitting on top of the 3D printed housing.
I also need to check how much the baffle will influence this in the far field.
I included the raw measurements of the woofer and tweeter before EQing and crossing.
Hi, a few things there. I think 100db/50hz at 3l is not achievable with your system. You simply wont move enough air.
I've been building such speakers for around 10 years, including custom amps with nearly 1kW.
Several effects: the provided xmax is mostly at an already high distortion, the coil is moving outwards of the usable magnet field.
Your PR is already compressing alot, same as for the woofers. (Also, coil heating reduces the effective power)
Your amp's THD is very high already.
You have to account for the electronics and battery as well as their cooling within the enclosure, as well as speaker actual self-volume. To achieve 3l internal but include all other components, you will be around 6l external.
In your orientation, the speaker will "rubble", all the mass is moving in one direction. There is a reason JBL and most others place Passive radiators adjacent of each other. Or on the same plane as the active driver.
For tweeters, i would use a single but bigger one. Something with inbuilt backside volume or horn to get a low coupling frequency. You do avoid comb filtering with you will get in your config and profit from the improved output of the bigger cone.
I think you are developing something very cool, but you have to manage the expectations a little 😉
As reference, my current project is based on the dayton epique 150 Sub in 6.5l internal volume with 3 6" PRs (SB Acoustics 16 Types)... all hooked up to a 400W Custom Amp. With all that, i barely get 50hz@100db without too much distortion 😅
I've been building such speakers for around 10 years, including custom amps with nearly 1kW.
Several effects: the provided xmax is mostly at an already high distortion, the coil is moving outwards of the usable magnet field.
Your PR is already compressing alot, same as for the woofers. (Also, coil heating reduces the effective power)
Your amp's THD is very high already.
You have to account for the electronics and battery as well as their cooling within the enclosure, as well as speaker actual self-volume. To achieve 3l internal but include all other components, you will be around 6l external.
In your orientation, the speaker will "rubble", all the mass is moving in one direction. There is a reason JBL and most others place Passive radiators adjacent of each other. Or on the same plane as the active driver.
For tweeters, i would use a single but bigger one. Something with inbuilt backside volume or horn to get a low coupling frequency. You do avoid comb filtering with you will get in your config and profit from the improved output of the bigger cone.
I think you are developing something very cool, but you have to manage the expectations a little 😉
As reference, my current project is based on the dayton epique 150 Sub in 6.5l internal volume with 3 6" PRs (SB Acoustics 16 Types)... all hooked up to a 400W Custom Amp. With all that, i barely get 50hz@100db without too much distortion 😅
You seem to imply that WinISDs simulation can’t be trusted?
It’s telling me I would get 97,4db@50hz = 100db with -3db @ 50Hz
At just about 30W with just 1,75mm excursion on all 3 drivers.
Xmax for the active driver is rated at 4mm - it measured 4,2mm and 4,8mm for 20% distortion based on Bl (moving out of the magnet field) and compliance.
So i dont think it will be the issue here.
PR is right at the limit of 8mm, that might show some more distortion.
Even with 50°C temp rise on the voice coil i still get 97db.
However that's all debatable if that really matters. First i dont expect super clean and audiophile sound out of a tiny portable speaker and secondly i will add harmonic distortion ON PURPOSE anyways!
You might now ask why i would do that: Answer is simple, Psychoacustics! You can use that to make your brain believe that there is more low end extension by adding harmonics. (Search for missing fundamental)
The amps do 34W@1%THD which isnt that high compared to acceptable THD values in that frequency range. (It seems to be that even above >5% isnt really detectable in that frequency range).
As for the internal vs external volume: I intend to do a full 3D model including battery's and electronics to calculate real internal volume. Quick back of the envelope calculations show me that i need roughly 200-300ml of extra volume to account for 80Wh of LiPo battery capacity and the electronics. So certainly not 3L.
I will surely blow my weight budget though. It will be closer to 3kg in the end.
I am still not sure on the 2 vs 1 tweeter thing.
Two have the advantage of needing less power - being more efficient and reducing distortion where it matters, it could also help with stereo imagine even though it will be a stretch as they are quite close.
While one saves weight and shows no comb filtering.
So far i could not really hear the comb filtering in practice. I can measure it but in direct comparison by ear with stereo music: No difference - cant spot anything.
However I also dont get much / any stereo separation, i might be able to improve that with some DSP algorithms but i doubt it.
So for me in practice it seems like its only a weight/efficiency/distortion issue.
It’s telling me I would get 97,4db@50hz = 100db with -3db @ 50Hz
At just about 30W with just 1,75mm excursion on all 3 drivers.
Xmax for the active driver is rated at 4mm - it measured 4,2mm and 4,8mm for 20% distortion based on Bl (moving out of the magnet field) and compliance.
So i dont think it will be the issue here.
PR is right at the limit of 8mm, that might show some more distortion.
Even with 50°C temp rise on the voice coil i still get 97db.
However that's all debatable if that really matters. First i dont expect super clean and audiophile sound out of a tiny portable speaker and secondly i will add harmonic distortion ON PURPOSE anyways!
You might now ask why i would do that: Answer is simple, Psychoacustics! You can use that to make your brain believe that there is more low end extension by adding harmonics. (Search for missing fundamental)
The amps do 34W@1%THD which isnt that high compared to acceptable THD values in that frequency range. (It seems to be that even above >5% isnt really detectable in that frequency range).
As for the internal vs external volume: I intend to do a full 3D model including battery's and electronics to calculate real internal volume. Quick back of the envelope calculations show me that i need roughly 200-300ml of extra volume to account for 80Wh of LiPo battery capacity and the electronics. So certainly not 3L.
I will surely blow my weight budget though. It will be closer to 3kg in the end.
I am still not sure on the 2 vs 1 tweeter thing.
Two have the advantage of needing less power - being more efficient and reducing distortion where it matters, it could also help with stereo imagine even though it will be a stretch as they are quite close.
While one saves weight and shows no comb filtering.
So far i could not really hear the comb filtering in practice. I can measure it but in direct comparison by ear with stereo music: No difference - cant spot anything.
However I also dont get much / any stereo separation, i might be able to improve that with some DSP algorithms but i doubt it.
So for me in practice it seems like its only a weight/efficiency/distortion issue.
Just to check a different way: (as it really sounds much from such a small enclosure)
Svante posted the formula to calculate SPL based on volume displacement: Link
If we plug in the values from the PR (103cm^3) and the W3s (16,8cm^3) running at 30W we get:
101db SPL in 1/2 Space = standing on a ground plane not suspended in the air.
At least in theory it should work. I just added the displacements of the PR and the drivers, I assume this to be false.
If i just use the PR I get 99db.
If i subtract the drivers displacement, I get 98db.
So i am not 100% sure why there is some sort of loss, maybe its radiation efficiency or some out of phase behaviour of the pr and the drivers.
At least it looks like the math/simulation cant be far off - the amount of air moved by the driver + PR is enough to produce the SPL I am after.
100db at 50Hz isnt much though the dbC scale calls for a -30db weighing - so you would need +30db to have the same apparent volume as 100dbs @ 5000hz.
I will also implement some sort of house curve on the final speaker, its intended to be used outdoors so i dont get any room gain which means i would likely need >+12db at 50Hz to make it not lack bass. I would therefore just get <90db in the mids/highs. (some kind of dynamic loudness correction will also be needed to still get loud and pull bass at higher volumes)
Your force cancellation point still stands. The PR is throwing around ~ 90g of mass (with tuning weigth), some of it is offset by the drivers on the front, but as they move much less and have less moving mass, its not a force cancelling design. There will be some amount of movement.
As far as i know it seems to be generally accepted to have roughly 1% moving mass compared to enclosure mass to limit vibrations to acceptable levels.
With roughly 2-3kg max weight I will look at close to 3-4,5% of mass ratio that also directly translates to movement of the enclosure.
So if the PR is moving 8mm, the housing will move 4,5% = 0,36mm assuming it is floating in free space.
My gut feel tells me that <0,4mm isnt ideal but acceptable considering size and weight constraints.
I will certainly invest in some very rubbery feet to help couple the enclosure to the ground and limit the movement.
Svante posted the formula to calculate SPL based on volume displacement: Link
If we plug in the values from the PR (103cm^3) and the W3s (16,8cm^3) running at 30W we get:
101db SPL in 1/2 Space = standing on a ground plane not suspended in the air.
At least in theory it should work. I just added the displacements of the PR and the drivers, I assume this to be false.
If i just use the PR I get 99db.
If i subtract the drivers displacement, I get 98db.
So i am not 100% sure why there is some sort of loss, maybe its radiation efficiency or some out of phase behaviour of the pr and the drivers.
At least it looks like the math/simulation cant be far off - the amount of air moved by the driver + PR is enough to produce the SPL I am after.
100db at 50Hz isnt much though the dbC scale calls for a -30db weighing - so you would need +30db to have the same apparent volume as 100dbs @ 5000hz.
I will also implement some sort of house curve on the final speaker, its intended to be used outdoors so i dont get any room gain which means i would likely need >+12db at 50Hz to make it not lack bass. I would therefore just get <90db in the mids/highs. (some kind of dynamic loudness correction will also be needed to still get loud and pull bass at higher volumes)
Your force cancellation point still stands. The PR is throwing around ~ 90g of mass (with tuning weigth), some of it is offset by the drivers on the front, but as they move much less and have less moving mass, its not a force cancelling design. There will be some amount of movement.
As far as i know it seems to be generally accepted to have roughly 1% moving mass compared to enclosure mass to limit vibrations to acceptable levels.
With roughly 2-3kg max weight I will look at close to 3-4,5% of mass ratio that also directly translates to movement of the enclosure.
So if the PR is moving 8mm, the housing will move 4,5% = 0,36mm assuming it is floating in free space.
My gut feel tells me that <0,4mm isnt ideal but acceptable considering size and weight constraints.
I will certainly invest in some very rubbery feet to help couple the enclosure to the ground and limit the movement.
Too late to Edit my post:
I found whats wrong: I ignored the RMS part.
If we use the RMS of the displacement like Svante said:
We get 98db halve space with the PR and woofer added. So much closer to WinISD almost spot on.
I found whats wrong: I ignored the RMS part.
If we use the RMS of the displacement like Svante said:
We get 98db halve space with the PR and woofer added. So much closer to WinISD almost spot on.
Hey,
for the calculation, you must only take the PR at the tuning point. For BR Enclosures, its either the driver moving or the PR, the relationship is shifting depending on frequency. You can see that quite well with WinISD. So for your calculation, i would take only the PR.
WinISD overall doesnt calculate the powercompression. Try to push your PR to 8mm by "hand", you will feel an increasing resistance above a certain threshold. I never encountered an afforderable PR which is fully linear to Xmax aside from Purifiy and Accuton Types.
Also vastly different: the actual SPL of the driver at 50hz (see the manufacturer measurements) and the SPL in WinISD. This one is a little tricky due to how the drivers are measured.
From experience, you can subtract around 6db from the winISD max. SPL Prediction (WinISD doesnt simulate free Air btw.)
The Harmonics addition is a well known practise, but the THD is a controlled one and cannot be compared to the THD coming from a driver out of Spec. General guideline for non-hifi would be <10% THD up to 200hz and <1% THD above.
I think you will achieve around 95db@70hz. By the way you dont happen to be german, judging from your name? 😀
for the calculation, you must only take the PR at the tuning point. For BR Enclosures, its either the driver moving or the PR, the relationship is shifting depending on frequency. You can see that quite well with WinISD. So for your calculation, i would take only the PR.
WinISD overall doesnt calculate the powercompression. Try to push your PR to 8mm by "hand", you will feel an increasing resistance above a certain threshold. I never encountered an afforderable PR which is fully linear to Xmax aside from Purifiy and Accuton Types.
Also vastly different: the actual SPL of the driver at 50hz (see the manufacturer measurements) and the SPL in WinISD. This one is a little tricky due to how the drivers are measured.
From experience, you can subtract around 6db from the winISD max. SPL Prediction (WinISD doesnt simulate free Air btw.)
The Harmonics addition is a well known practise, but the THD is a controlled one and cannot be compared to the THD coming from a driver out of Spec. General guideline for non-hifi would be <10% THD up to 200hz and <1% THD above.
I think you will achieve around 95db@70hz. By the way you dont happen to be german, judging from your name? 😀
Hallo fellow German 😉
WinISD shows exact excursion for both the PR and the active driver. There is no sense in not including the drivers excursion. However even if we do that its still in the ballpark as the active drivers only move very little like you said.
Power compression: I think you got this wrong, Power compression only happens on active drivers- The voice coil heats up and increases in resistance resulting in less current at the same voltage so you end up providing lower power at the same voltage.
WinISD can calculate its input, you just need to input the voice coil temp rise - which i estimated to be 50K above ambient = 70°C.
This still didnt result in significant power compression.
95db at 70Hz is way off. I will certainly measure that when i get a suitable housing.
I think you are very much overestimating the impact of the distortion and "power compression"
In my experience, WinISD is close to spot on.
Do you have any real measurements / comparisons with WinISD on known power with a calibrated mic?
WinISD shows exact excursion for both the PR and the active driver. There is no sense in not including the drivers excursion. However even if we do that its still in the ballpark as the active drivers only move very little like you said.
Power compression: I think you got this wrong, Power compression only happens on active drivers- The voice coil heats up and increases in resistance resulting in less current at the same voltage so you end up providing lower power at the same voltage.
WinISD can calculate its input, you just need to input the voice coil temp rise - which i estimated to be 50K above ambient = 70°C.
This still didnt result in significant power compression.
95db at 70Hz is way off. I will certainly measure that when i get a suitable housing.
I think you are very much overestimating the impact of the distortion and "power compression"
In my experience, WinISD is close to spot on.
Do you have any real measurements / comparisons with WinISD on known power with a calibrated mic?
HA I KNEW IT FROM THE NAME!
Power Compression maybe was the wrong term for it, more like mechanical compression.
I do actually, with the minidsp Umik during actual use. Thats why my opinion stems from😅 winlsd works 100% linear while reality isnt. If you manage to get good measurements id love to see them. If you live in the munich regions, i would even help if you want to.
Power Compression maybe was the wrong term for it, more like mechanical compression.
I do actually, with the minidsp Umik during actual use. Thats why my opinion stems from😅 winlsd works 100% linear while reality isnt. If you manage to get good measurements id love to see them. If you live in the munich regions, i would even help if you want to.
A Geneva speaker uses the tweeters in the middle of the enclosure right next to each other, at the top edge. They are no flange type so the C-C distance makes comb filtering only in very high frequencies where it matters less.
Does Dayton nd90 or 91 fit?
They would have less issue at 1kHz.
There's also Epique PR which may have more excursion?
Does Dayton nd90 or 91 fit?
They would have less issue at 1kHz.
There's also Epique PR which may have more excursion?
https://www.baudline.com/erik/bass/xmaxer.html
NB: at the PR tuning frequency, the main drivers won't be moving very much. Almost all of the SPL is produced by the PR.
Above the PR tuning frequency, the main drivers are on their own. At 70Hz, 3x 3" drivers will each require 8.35mm Xmax.
This is a cool design, and I like what you're doing, but I think the LF output is still a little bit under what you're looking for.
Finally, close-up measurements don't account for the baffle-step effect on the frequency response, which may explain the "harsh" sound previously observed.
Chris
NB: at the PR tuning frequency, the main drivers won't be moving very much. Almost all of the SPL is produced by the PR.
Above the PR tuning frequency, the main drivers are on their own. At 70Hz, 3x 3" drivers will each require 8.35mm Xmax.
This is a cool design, and I like what you're doing, but I think the LF output is still a little bit under what you're looking for.
Finally, close-up measurements don't account for the baffle-step effect on the frequency response, which may explain the "harsh" sound previously observed.
Chris