I've been into audio for as long as I can remember - building simple amps when I was a kid and more recently more involved projects, like a passive preamp project, some DAC stuff and a small music player for my kids. I've always wanted to build speakers and I just took the first step - I'm currently building a pair of compact bookshelves (Soundblab's Dynamites). I love the idea of designing/building something like a pair of line transmission speakers, but I'll build up to that!
So, now I want to design my own, and I thought I start with something simple - I have a spare pair of spare Dayton Audio DSA90-8 full-range drivers that I thought would be a good place to start. These will be used for near-field, probably without a sub. I used Dayton's recommendations for enclosure side and modeled three options, a sealed enclosure, a ported enclosure and a passive radiator option with Dayton's matching DSA90-PR.
I've done a lot of reading (/lurking), and it seems that a good option with this driver is a single driver in a small cabinet, or a larger cabinet that also has a woofer to add some bass. I wanted to keep things simple to start with, so I'm looking at the single-driver option. Maybe I'll have a stab at crossover design next time!
The enclosure sizes are 0.04 cuft for sealed, 0.06 cuft for ported and passive radiator.
I'm using WinISD so I added the specifications for the DSA90-8 and DSA90-PR and modeled a few options. Orange is sealed, green is ported and purple is passive radiator.
Transfer function magnitute. The passive radiator seems like the most linear response.
Cone excursion. This one I'm less certain on what to do with this information. I think it means that a high-pass filter would be needed with the ported enclosure? Would is be recommended on all of these? I've seen recommendations for high-pass filters to protect the drivers on full-range designs.
Cone excursion (PR). It seems that a single passive radiator is sufficient in this design
Air velocity. It seems that 25m/s and below are considered safe, from my research, so there shouldn't be any chuffing on the port. This is a 1" diameter port.
From this, my takeaway is that the passive radiator design seems to give the smoothest response and have the least compromises, although the ported design will go slightly lower and have a bass lift before rolloff. Is that accurate? I would assume the bass in the ported design could be tamed with a high-pass filter?
I know that placement of the driver on the baffle, and baffle size make a difference, but I'm not sure how to model this. Does anyone have any recommendations on where to start with that?
I'm really new to this but I'm trying to research as best I can so please be gentle! Any recommended reading would also be appreciated! I know that's a lot of questions, but I'm not quite sure how to get answers to them all.
So, now I want to design my own, and I thought I start with something simple - I have a spare pair of spare Dayton Audio DSA90-8 full-range drivers that I thought would be a good place to start. These will be used for near-field, probably without a sub. I used Dayton's recommendations for enclosure side and modeled three options, a sealed enclosure, a ported enclosure and a passive radiator option with Dayton's matching DSA90-PR.
I've done a lot of reading (/lurking), and it seems that a good option with this driver is a single driver in a small cabinet, or a larger cabinet that also has a woofer to add some bass. I wanted to keep things simple to start with, so I'm looking at the single-driver option. Maybe I'll have a stab at crossover design next time!
The enclosure sizes are 0.04 cuft for sealed, 0.06 cuft for ported and passive radiator.
I'm using WinISD so I added the specifications for the DSA90-8 and DSA90-PR and modeled a few options. Orange is sealed, green is ported and purple is passive radiator.
Transfer function magnitute. The passive radiator seems like the most linear response.
Cone excursion. This one I'm less certain on what to do with this information. I think it means that a high-pass filter would be needed with the ported enclosure? Would is be recommended on all of these? I've seen recommendations for high-pass filters to protect the drivers on full-range designs.
Cone excursion (PR). It seems that a single passive radiator is sufficient in this design
Air velocity. It seems that 25m/s and below are considered safe, from my research, so there shouldn't be any chuffing on the port. This is a 1" diameter port.
From this, my takeaway is that the passive radiator design seems to give the smoothest response and have the least compromises, although the ported design will go slightly lower and have a bass lift before rolloff. Is that accurate? I would assume the bass in the ported design could be tamed with a high-pass filter?
I know that placement of the driver on the baffle, and baffle size make a difference, but I'm not sure how to model this. Does anyone have any recommendations on where to start with that?
I'm really new to this but I'm trying to research as best I can so please be gentle! Any recommended reading would also be appreciated! I know that's a lot of questions, but I'm not quite sure how to get answers to them all.
Last edited:
Going down the rabbit hole, eh? You'll have to decide how far down you want to go. There are far, far, more knowledgeable people here, than I, so hopefully they'll chime in. Ported, with the slight peak, will give a bit more perception of bass as this driver isn't going very low. You are correct that excursion could be an issue but what wattage level was that graph for? You said this would be near field so it's doubtful you'd driving them with much power, so this might not be an issue. For driver placement on your baffle, try (assuming you have access to Excel) "Baffle Diffraction and Boundary Simulator" from audio.claub.net. You might also look at Basta! from Tolvan Data. I haven't used it, but it's free, and well regarded.
Should be fun. As with any 3" driver dont expect much bass below 80 to 100 Hz.
Dayton packs a rather powerful magnet for the price driver Qts .37
Friendly to small boxes
It is a nice aluminum cone for a Mid as well, same thing rather friendly size mid chamber
Basically works well in 1.6 liters Vent would be tuned slightly above or below Fs for 3rd or 4th order.
As with many small drivers tune at Fs and call it a day.
Normal alignments show a smooth roll off to me. In advanced properties in WinIsd
Default leakage ( QL) is set to 10 which is typically impossible in ported.
For a small box like this leakage be about 8 average is 7 and larger boxes move toward 6
With more realistic leakage models wont show such a high Q peak in the roll off.
This driver isnt very High Q .37 Qts regardless so shouldnt have any Q at cutoff unless you tuned to high.
When you model higher Q drivers in the future Qts = .4 to .6 or higher drivers, good to use average leakage of 7
Usually shooting for lowest tuning when possible, since you noticed yes everything unloads below that.
Alignments rather easy with .3 Qts drivers
would just use standard 4th order alignment BB4 or C4 and likely add weight to the passive to match 66 Hz tune.
For low distortion and higher power usually 2x more area for passive, so typically 2 passive drivers. Then weight is needed.
I have noticed with one passive, Dayton did a nice job with matching passives usually no weight is needed. Unless your picky about true alignment.
It is a 3" so this case no bass anyways below 100 Hz in real life with real world looses. Fun project for desktop , low listening.
Dayton packs a rather powerful magnet for the price driver Qts .37
Friendly to small boxes
It is a nice aluminum cone for a Mid as well, same thing rather friendly size mid chamber
Basically works well in 1.6 liters Vent would be tuned slightly above or below Fs for 3rd or 4th order.
As with many small drivers tune at Fs and call it a day.
Normal alignments show a smooth roll off to me. In advanced properties in WinIsd
Default leakage ( QL) is set to 10 which is typically impossible in ported.
For a small box like this leakage be about 8 average is 7 and larger boxes move toward 6
With more realistic leakage models wont show such a high Q peak in the roll off.
This driver isnt very High Q .37 Qts regardless so shouldnt have any Q at cutoff unless you tuned to high.
When you model higher Q drivers in the future Qts = .4 to .6 or higher drivers, good to use average leakage of 7
Usually shooting for lowest tuning when possible, since you noticed yes everything unloads below that.
Alignments rather easy with .3 Qts drivers
would just use standard 4th order alignment BB4 or C4 and likely add weight to the passive to match 66 Hz tune.
For low distortion and higher power usually 2x more area for passive, so typically 2 passive drivers. Then weight is needed.
I have noticed with one passive, Dayton did a nice job with matching passives usually no weight is needed. Unless your picky about true alignment.
It is a 3" so this case no bass anyways below 100 Hz in real life with real world looses. Fun project for desktop , low listening.
Thanks everyone for the input!
@WhiteDragon
Yes, I have an idea of what I'm getting into. I built a 3D-printed NFC-card speaker system for my kids with slightly smaller ND65 2-1/2 drivers from Dayton Audio using a ball-park estimated waveguide design. They went much lower than I expected, so while I know the DSA90 won't go low, I'm hoping they'll be a fun first project that won't disappoint!
Thanks for all the info, especially the leakage comments - I haven't seen that anywhere. I did briefly consider these as a mid, but I decided starting with a 3-way was probably too much and a pair of compact bookshelves would be very handy for one of my spaces anyway. My next attempt will probably be a 2-way, when I'll tackle a crossover.
I can't work out where to put QL in WinISD, any pointers? I also realized I input the size of the box wrong in WinISD for the vented design (2.2l instead of 1.6l). The larger cabinet provided a lower range, but came at the trade-off of a hump-and-dip, where the correct size is much smoother.
I read somewhere that double the surface area was good for passive radiators, but when I modeled that in WinISD, bad things happened. I thought it was just a bad idea, but I must have got something wrong. Trying it again now, and adding 4g of mass to the passive radiators, it seems like it may be the better of the models. I'm not 100% sure about the trade-off of sharper roll-off with more power around 90-100Hz vs a smoother roll-off that starts much higher.
Green is vented, purple is 2 x PR and red is 1 x PR
Haha, probably to the bottom!
Good point. I think I actually had this set at 20W, more than I would need for nearfield. I have spare pair of Schiit Rekkrs that I will probably use in mono, from memory I think they're good for 4W, so excursion probably isn't an issue for any of these.
I have! I'm actually just reading it a second time, it's full of great info!
@WhiteDragon
Yes, I have an idea of what I'm getting into. I built a 3D-printed NFC-card speaker system for my kids with slightly smaller ND65 2-1/2 drivers from Dayton Audio using a ball-park estimated waveguide design. They went much lower than I expected, so while I know the DSA90 won't go low, I'm hoping they'll be a fun first project that won't disappoint!
Thanks for all the info, especially the leakage comments - I haven't seen that anywhere. I did briefly consider these as a mid, but I decided starting with a 3-way was probably too much and a pair of compact bookshelves would be very handy for one of my spaces anyway. My next attempt will probably be a 2-way, when I'll tackle a crossover.
I can't work out where to put QL in WinISD, any pointers? I also realized I input the size of the box wrong in WinISD for the vented design (2.2l instead of 1.6l). The larger cabinet provided a lower range, but came at the trade-off of a hump-and-dip, where the correct size is much smoother.
I read somewhere that double the surface area was good for passive radiators, but when I modeled that in WinISD, bad things happened. I thought it was just a bad idea, but I must have got something wrong. Trying it again now, and adding 4g of mass to the passive radiators, it seems like it may be the better of the models. I'm not 100% sure about the trade-off of sharper roll-off with more power around 90-100Hz vs a smoother roll-off that starts much higher.
Green is vented, purple is 2 x PR and red is 1 x PR
It is in advanced parameters
These drivers are dipping into .3 Qts so Q is not high. Wont be as visible change.
Reference for the future if you do sim high Q drivers .4 Qts or higher.
Standard leakage models from Thiele/ Small was 7 so WinIsd set at 10 assumes no leakage.
It is much more apparent with High Q drivers. In fact Thiele / Smalls theory had a lot of errors, their models did not match real world as well.
That is why they added extended models which accounted for Leakage, Absorption and port Efficiency
WinIsd includes extended models Ql Qa and Qp leakage being the most important for transfer function Q or roll off at cut off
Passive has a sharper cutoff, that is known behavior.
Better not to stare at transfer functions and imagine sound. Real life they typically have same bass content.
Passive is just a alternative to ports. It is also a option for low tuning when a enclosure doesn't accept long ports.
The passive port sound is slightly different. Not being a air velocity through a tube, but a piston
Most the turn off in the past was at high volumes the passive had audible distortion. That was from not enough passive area.
So 2x passive solves the problem often 3x with higher power. This system wont be expecting much more than 5 to 12 watts with a high tune 60 Hz
Ball Game changes at 100 plus Watts with 30 Hz systems.
It is really more just for fun, Real world listening is heard or the sound. It is the drawback I guess with modern spreadsheets being visual.
In the old days you just did the math, then built the system. Again most negative opinions of real world listening of passive.
Is experience from 70's and 80's systems where there wasn't enough passive area and you hear mechanical distortion.
As opposed to standard ports you hear the chuff or ring of the air velocity. So neither system is perfect.
At least with computer design your aware of the passive excursion and more aware of suspension and weighting.
Best way to find out, is comically find out in real world.
Last edited:
I use older WinIsd so advanced parameters easy to find in project window.
That " New" version is utterly annoying for finding anything or navigation.
Believe it came out for junk windows Vista
Now With Windows 10 old software runs no problem.
Back in the day older software crashed on vista.
For some reason WinIsd changed the navigation flow and graphics.
Instead of just fixing the 64 bit errors from junk Vista. Hate it.
With a FR box you can always add woofers later. If that happens they are best sealed or aperiodic.
A PR is a vent substitute, you should be able to get a PR & vented alignments that are the same. I’m not a fan of PRs.
Go for it. Build anything, learn stuff. It is a small box. You’ll have lots odf material from the sheet or half sheet of plywood you get to rebuild them if the first try doesn’t work out.
dave
Thanks again, that all makes sense.
@planet10 I did think about trying to leave space, but in all honesty I would probably just rebuild them if I find the low-end too disappointing. This is more of a learning experience.
You're both right, I think I'm probably spending way too much time looking at charts - probably because I'm new at this. With most projects I usually run the numbers and go with gut instinct on what I want to build and take it from there, so that's what I'm going to do with this too! I'm going to build a pair with 2xPR for fun and see what I think. I may even build a pair of both cabinets (PR + vented) and compare - they're a small build so it's not like I'm going to waste a ton of material. With no crossover, I'm not going to building extra parts I don't need either.
Last question - I did read in some places about protecting the drivers with a high-pass filter. Nobody has mentioned it, so is that unnecessary in this case? I'm just wondering, will the fact that the driver is trying to reproduce low frequencies create distortion in higher frequencies? I know it's a "full range" driver, but it's not really full-range!
@planet10 I did think about trying to leave space, but in all honesty I would probably just rebuild them if I find the low-end too disappointing. This is more of a learning experience.
You're both right, I think I'm probably spending way too much time looking at charts - probably because I'm new at this. With most projects I usually run the numbers and go with gut instinct on what I want to build and take it from there, so that's what I'm going to do with this too! I'm going to build a pair with 2xPR for fun and see what I think. I may even build a pair of both cabinets (PR + vented) and compare - they're a small build so it's not like I'm going to waste a ton of material. With no crossover, I'm not going to building extra parts I don't need either.
Last question - I did read in some places about protecting the drivers with a high-pass filter. Nobody has mentioned it, so is that unnecessary in this case? I'm just wondering, will the fact that the driver is trying to reproduce low frequencies create distortion in higher frequencies? I know it's a "full range" driver, but it's not really full-range!
Just finished my Dynamites build, so now I'm going to start buying parts for my compact nearfield design above. I'll be building them while I enjoy the Dynamites
