The free air resonance frequency of the woofer is Fs=21.5Hz. Placing the woofer in a 42-litre closed-box enclosure increases the woofer's resonance frequency to Fb=33.5Hz. That means that the stiffness of the combined enclosure plus driver system is about 2.42 times that of the driver alone. That is, the stiffness of the air in the enclosure is 1.42 times that of the driver alone. This indicates that for an enclosure volume of Vb=42 litres the "air spring effect" contributes quite substantially to the enclosure's tuning.
In vintage "acoustic suspension" loudspeakers such as the Acoustic Research AR-7, the aim was to improve low-frequency linearity and output by making the air spring dominate over the mechanical spring in the woofer. The 8-inch woofer used in that system had a free-air resonance frequency of Fs=25Hz, with a closed-box resonance frequency of Fb=68Hz. This indicates that the stiffness of the combined enclosure plus driver system was about 7.40 times that of the AR-7's woofer alone.
Today's woofers have suspension systems that are linear up to high values of Xmax, particularly for drivers that are suitable for use as subwoofers. Hence, relying on the linearity of the air spring effect is not necessary to the same extent that it was in the AR-7 design.
Last edited:
Unfortunately, that's not possible. But you definitely opened up a new rabbit hole for me... So thank you 🙂
Thanks again @witwald for your very interesting and helpful responses. What you write corresponds quite well with what I (subjectively) hear from the current config. I'm actually quite happy with it, and can probably live with it. I will test out your EQ idea though. I do feel the mid bass is a bit too loud compared to the very lowest frequencies, and your L/R filter could maybe help. I believe the SPL should still be enough for my listening, I'm not doing any crazy loud stuff.
... and still, I have this idea that it could be interesting to test out how a vented config would compare. I guess it wouldn't be a big deal to remove the vent and plug the hole if it turns out it doesn't work out. But for now, I'll try the sealed EQ option and see what it can add.
... and still, I have this idea that it could be interesting to test out how a vented config would compare. I guess it wouldn't be a big deal to remove the vent and plug the hole if it turns out it doesn't work out. But for now, I'll try the sealed EQ option and see what it can add.
An update on this little project... I used the sealed box for a few weeks and really enjoyed it. For music it worked very well, with higher output and deeper extension than my old XTZ 10.17. I also liked the sound better, even though I can't really explain why. More "clean" sound is a word that comes to mind, but I'm not sure if that's a scientific term... For movies it also worked well, but I knew there was potentially more to get from this driver. I tested different EQ settings, but was of course always limited by xmax at the lower frequencies. So a couple of days ago, I put in a vent. I used 100 mm pipe with flared ports, three 90 degree bends to fit it in the box, total length 110 cm. With a net volume of 34 liters, WinISD gave me a 24 Hz tuning. First port resonance should be 152 Hz which I thought was acceptable.
When I tested the fininshed box, it turned out that the tuning frequency is actually 21 Hz. Not sure why it's considerably lower than WinISD gave me with this port length? The box seems quite well sealed so I don't think there's any considerable air leakage or anything like that, Maybe the bends add some air resistance and changes the tuning? I used a bit of stuffing (not a lot), so that might be the reason. Anyway, I'm actually very happy with the end result. I use a 10 db PEQ boost at 21 Hz with Q=2, and a 20Hz 4th order Butterworth high pass filter. This is the only EQ I've used so far. According to WinISD, it will keep xmax well within acceptable limits with 700 Watts of power.
Today I set up my UMIK and did some in-room measurements. Measured at listening position, with EQ as mentioned above, it gives a nice slightly rising curve until around 25 Hz, less than 3 db down at 20 Hz, and still meaningful response at 17Hz. After that, it drops off very quickly. I haven't tested max SPL, but it seems to be quite enough for my use.
I was a bit suprised that as far as I can notice, it didn't lose any of the muscial qualities from the sealed version. So I can basically only see positives from going to vented. No mechanical noises, no port chuffing, no timing issues, no audible port resonance (I let the AVR cross over at 80 Hz). Maybe I'll discover problems later when I test more extensively, but so far I'm very happy with the results. It's also quite fun that it's in a sofa chaiselounge, so for all intents and purposes it's invisible. I'm very happy with this Dayton driver, looks like it's behaving very well in small enclosures. Not that I have much to compare with, since this is my first build...
One question though, I have a pretty bad peak at around 34 Hz. I had the exact same issue with the sealed version, so I'm pretty sure it's a room problem. How would you go about fixing this? My old Pioneer AVR doesn't seem able to take care of it by automatic room correction. I'm thinking to try some narrow band PEQ in the Crown's DSP, I guess it's resonable to start with a PEQ at 34 Hz with narrow Q? Will probably also add a second subwoofer later (in the sofa armrest...), which might help?
... and a big thank you! to the ones in this thread who spent so much time modeling and educating me on different issues.
When I tested the fininshed box, it turned out that the tuning frequency is actually 21 Hz. Not sure why it's considerably lower than WinISD gave me with this port length? The box seems quite well sealed so I don't think there's any considerable air leakage or anything like that, Maybe the bends add some air resistance and changes the tuning? I used a bit of stuffing (not a lot), so that might be the reason. Anyway, I'm actually very happy with the end result. I use a 10 db PEQ boost at 21 Hz with Q=2, and a 20Hz 4th order Butterworth high pass filter. This is the only EQ I've used so far. According to WinISD, it will keep xmax well within acceptable limits with 700 Watts of power.
Today I set up my UMIK and did some in-room measurements. Measured at listening position, with EQ as mentioned above, it gives a nice slightly rising curve until around 25 Hz, less than 3 db down at 20 Hz, and still meaningful response at 17Hz. After that, it drops off very quickly. I haven't tested max SPL, but it seems to be quite enough for my use.
I was a bit suprised that as far as I can notice, it didn't lose any of the muscial qualities from the sealed version. So I can basically only see positives from going to vented. No mechanical noises, no port chuffing, no timing issues, no audible port resonance (I let the AVR cross over at 80 Hz). Maybe I'll discover problems later when I test more extensively, but so far I'm very happy with the results. It's also quite fun that it's in a sofa chaiselounge, so for all intents and purposes it's invisible. I'm very happy with this Dayton driver, looks like it's behaving very well in small enclosures. Not that I have much to compare with, since this is my first build...
One question though, I have a pretty bad peak at around 34 Hz. I had the exact same issue with the sealed version, so I'm pretty sure it's a room problem. How would you go about fixing this? My old Pioneer AVR doesn't seem able to take care of it by automatic room correction. I'm thinking to try some narrow band PEQ in the Crown's DSP, I guess it's resonable to start with a PEQ at 34 Hz with narrow Q? Will probably also add a second subwoofer later (in the sofa armrest...), which might help?
... and a big thank you! to the ones in this thread who spent so much time modeling and educating me on different issues.
Wow. I have never boosted a ported enclosure this much. Really surprised its needed.
That's a lot of bends, less than I deal.
No surprise here, there are a lot of variables winisd is not considering. Port adjustment is often needed to get the tuning your after. I would alter the port to get your desired tuning and hopefully minimize some of the boost required.
A high boost at tuning freq with a high pass right below is an advice I got from a forum member, and I'd say it worked very well. The net effect is not 10db. I wouldn't say it's absolutely needed, it's more about preferences I think. I'm sacrificing some SPL above 30 Hz. I have found that I really like the low end extension and the "house curve" that slopes up to 25 Hz. The difference is easy to hear, and also to see in WinISD. But I'm definitely a novice so it could very well be that there are better ways to get there.
I'm sure you're right, but for my purposes it has worked quite well.
Yeah, I was prepared to do this, but I'm pretty happy with the results I'm getting now, so I'll probably leave it the way it is. This is a "stealth" subwoofer in a living room, and it can of course not compete with bigger vented boxes. Everything is a compromise. My goal was to get meaningful low end extension from a small box, without too many drawbacks.
An example to illustrate what I mean. On my old XTZ 10.17 (non Edge), a 25 Hz test tone was weak, and a 20 Hz tone resulted in mostly just mechanical noise and port chuffing. On my sealed DIY sub, the 25 Hz was louder, and 20 Hz was audible but weak. On my vented version, the 20 Hz tone is loud and clear, without chuffing or mechanical noise.
Next up is getting rid of the room mode.
Nice! that you have tested all of em and you like your last DIY version best
do you have a NapkinNPen sketch ?
so i can draw a 3D of it 🙂
i would like to make one for my HT
( well just a VIZIO bar ) but it have sub out
and my current 8" sub is just MEH
The narrow Q PEQ should take care of the peak at the listening or measurement microphone position.
A second sub in the sofa probably won't change the narrow Q room peak the first sub is exciting, but will add some headroom and may smooth room response a bit.
Since you already have enough LF SPL for your use without excessive excursion, nothing to be gained adding the second sub other than a more even bass distribution around the room.
I don't have a sketch, but it's just a rectangular box. I removed a brace in my couch's footstool part of the chaiselounge. That left room for a rectangular box with outside dimensions l=650 w=510 h=190 (millimetres). I used 18mm MDF, made sure the box was a snug fit under the footstool, and glued the whole thing in place. The Dayton driver is mounted down-firing, off center. I used 100 mm vent pipes, bends and flares, I think the brand is Jantzen Audio, but I guess you could use any PVC pipe with decent rigidity. (I had originally planned to use HVAC pipe, but the walls seemed a bit too thin). I put polyester mat on the inside walls and wrapped a little of it around the vent a few places. The port fires down beside the driver. I use the original legs on the footstool, and the bottom of the box is 7 cm above the floor. The footstool was quite sturdy, and with the added weight of the box, driver etc it's pretty heavy. I can hear no rattling.
That's quite a low tuning frequency for such a small enclosure. It could be described as impressively low...
The formula that WinISD uses is likely based on a simplified/idealized formulation that doesn't apply as accurately when the port length is 1.1 meters. Having three sets of 90-degree turns could have created a bit of a blockage, which may show up as an added-mass effect. The extra mass drops the port resonance frequency.
The stuffing increases the effective volume of the enclosure, which in turn reduces the port resonance frequency for a given set of port dimensions.
Below is a simulation of the filter-assisted vented-box 34-litre enclosure with a tuning frequency of 21Hz.
If we drop the PEQ boost at 20Hz to be only +6.5dB instead of +10dB, we get the following flatter response in the passband of the subwoofer.
If we drop the PEQ boost at 20Hz to be only +6.5dB instead of +10dB, we get the following flatter response in the passband of the subwoofer.
I notice that you've labelled this with RSS315HO-4. Note that my driver is the DVC version, 315HO-44. Not sure how much it means, but some of the parameters are different. For instance the Fs is lower for the 44. I'm running it in 2 ohm config now, with 700 W available from one channel on the Crown.
I didn't get a curve as steep as yours when I modelled in WinISD, but I'm sure your model is more accurate. I do like the upwards sloping curve. Still, you're probably right that it might be too much. Will do some tests with reduced boost.
What you describe as "the upwards sloping curve" is a downwards slope with rising frequency as preferred by most listeners.
https://www.avsforum.com/threads/jb...54077/page-214?post_id=57291428#post-57291428
As Floyd Toole wrote:
"Almost 50 years of double-blind listening tests have shown persuasively that listeners like loudspeakers with flat, smooth, anechoic on-axis and listening-window frequency responses... When such loudspeakers are measured in typically reflective listening rooms the resulting steady-state room curves exhibit a smooth downward tilt.
Peaks can be attenuated by EQ, but narrow dips should be left alone... Once the curve is smoothed there is the decision of what the bass target should be. Experience has shown that one size does not fit all. Music recordings can vary enormously in bass level..The upshot is that we need a bass tone control and the final setting may vary with what is being listened to, and perhaps also personal preference. In general too much bass is a "forgivable sin" but too little is not pleasant 🙂"
Whether you prefer a low frequency target slope of "flat", +3dB, +6dB or more is up to you.
In general, the lower your listening volume, the steeper the downward slope, or "too much bass" is preferred.
Art
Apologies for my mistake. I've changed over the driver, and the simulations for it with the two voice coils connected in series are shown below. In this case, the PEQ at 20Hz can be reduced by 3dB to be set at +7dB.
The Thiele–Small parameters are certainly quite different, with the Fs value being noticeably lower. However, as you are not using an empirical alignment, having detuned the enclosure and applied compensating filter assistance, the response is an amalgam of many factors. This seems to be borne out by the simulation results presented above.
A 2-ohm load will draw a lot of current from the amplifier, and for long-term use, I'm not sure how robust that configuration will be. If your Crown amplifier is a stereo unit, I think you'd be better off driving the two coils independently, one coil per channel of the Crown. You will still get to use the full power range of the amplifier, but it might be less prone to overloading.
Last edited:
It's pleasing that the overall response that you've managed to achieve is in line with your preference.
It's a very compact box, that's for sure. Capable of excellent bass extension in your configuration, but a bit limited in terms of maximum output. But that's a good compromise overall, I think.
When using the system tuned for maximum passband flatness, if you dropped the cut-off frequency of the 4th-order high-pass Butterworth filter from 20Hz down to 18Hz, you would be able to eke out an extra 1–2Hz of bass extension. This wouldn't significantly change the maximum excursion of that configuration.
I think that means that the Qes of the driver in this voice coil configuration is 1/2 that which was provided in the datasheet. On that basis, I've performed a simulation of the vented-box enclosure with the 21Hz tuning but without any filter assistance. The dashed cyan curve is the response of the driver in the series voice coil configuration. The very droopy response when the voice coils are connected in parallel is quite evident. Note that the DC resistance is 1.63ohms, which is almost 20% lower than 2 ohms.
After adding the 4th-order Linkwitz–Riley low-pass filter set to 80Hz, together with a 4th-order Butterworth high-pass filter set to 20Hz, the following simulated response is obtained. This produces an upper −6dB cut-off frequency that much higher than the desired 80Hz.
We can adjust the L–R filter's setting downwards to 63Hz to achieve an acoustic −6dB point at approximately 80Hz. It is also possible to apply −3.0dB of PEQ at 20Hz to tame the slight peakiness in the low-frequency response. The result is shown below.
The system can be tweaked to another configuration, this time using a 20Hz 2nd-order Butterworth high-pass filter. This keeps the driver displacement at low frequencies nicely under control, but has the added benefit that the group delay will be a little less than with a 4th-order high-pass filter in the signal chain. Setting the PEQ at 20Hz to be −1.5dB instead of −3.0dB, the simulation yields the following response curve.