Good work, that's an awesome result down to 20Hz. Hiow much input signal dd that take to get up there? I mean your input data for boost in dB and signal input in W
2V, but that's near field. That's not at listening distance. It's noted in the heading, but just wanted to be clear. As for how I got the results, I cut away signal with the DSP and then raised the total signal level.
Very best,
The data sheet for the GRS 12SW-4HE shows a small bump in the impedance curve at about 470Hz or so. This corresponds to a peak in the SPL response between 400Hz and 500Hz. It appears that there is a strong cone vibration mode occurring that is also showing up in the impedance curve due to electrodynamic coupling effects.
It looks like there was a lot of boost at 20 Hz — by 20dB or so relative to 100 Hz, judging by the curve shown below. That's to be expected, as the natural response of the compact closed-box loudspeaker system in question starts to roll off relatively early.
Motional FeedBack i.e. correcting the nonlinearities of the driver by actively monitoring the cone movement and modifying the signal accordingly. The GRS are quite poor in terms of IMD so should be a huge improvement https://www.diyaudio.com/community/...on-and-15-dollars.360855/page-10#post-6583177
So the Fb is ~36 Hz with one driver used as a PR.
The PR would be ~90 out of phase (lagging by 270 degrees) with the driver at Fb and more out of phase at your 30Hz vibration test frequency.
If you were to sweep upwards from 36Hz, you would find a frequency where the driver and PR excursion would be equal and in phase (lagging by 360 degrees), at that frequency the cabinet vibration would be the same (very little) as both drivers driven with the same polarity and excursion.
Art
Thank you for taking the time to write this up in such detail. FWIW I built a similar test box using two LAB12 type drivers, but didn't measure it in such depth as yourself, so this is all very good to read. What conclusions I reached, though, broadly correlate with your findings, which is encouraging.
One other thing i did test was a brace between the magnets, which is a common approach. It took a little effort to get a good tight fit, but after all that I didn't actually find any discernible difference with or without it, so I probably wouln't bother again. Perhaps it depends on the rigidity of the basket as to how well the magnet is restrained from the front flange. The LAB12 has a pretty solid cast frame, so a thick (small or braced) baffle may be all that is needed, though I'd guess even pressed baskets could be similar, if well shaped and thick enough.
The choice of impedance is difficult; 4, 6, 8 or 16ohms might be best depending on the number of drivers and how they're arranged for different amps. FWIW my personal choice has been 4ohms per driver because a pair can be either driven by two channels of most stereo amps, or some pro class-D power amps are rated to 2ohms per channel for quite a lot of power into a parallel pair. Unfortunately many of the better plate amps that are so popular can be quite expensive outside the US (power isn't necessarily very cheap, as is often blindly asserted). But on the up-side, I think it is probably healthier for the amp to be in a separate vented case, and I like not having to run both signal and mains-power cables to the sub when a single (thick) low voltage bit of flex will work very well over living-room distances.
Kev
One other thing i did test was a brace between the magnets, which is a common approach. It took a little effort to get a good tight fit, but after all that I didn't actually find any discernible difference with or without it, so I probably wouln't bother again. Perhaps it depends on the rigidity of the basket as to how well the magnet is restrained from the front flange. The LAB12 has a pretty solid cast frame, so a thick (small or braced) baffle may be all that is needed, though I'd guess even pressed baskets could be similar, if well shaped and thick enough.
The choice of impedance is difficult; 4, 6, 8 or 16ohms might be best depending on the number of drivers and how they're arranged for different amps. FWIW my personal choice has been 4ohms per driver because a pair can be either driven by two channels of most stereo amps, or some pro class-D power amps are rated to 2ohms per channel for quite a lot of power into a parallel pair. Unfortunately many of the better plate amps that are so popular can be quite expensive outside the US (power isn't necessarily very cheap, as is often blindly asserted). But on the up-side, I think it is probably healthier for the amp to be in a separate vented case, and I like not having to run both signal and mains-power cables to the sub when a single (thick) low voltage bit of flex will work very well over living-room distances.
Kev
@weltersys
would that frequency be as long as the distance between the driver and passive radiator? Like the 4 x 1/4 wavelength freq in a folded TL at the cancelation notch?
would that frequency be as long as the distance between the driver and passive radiator? Like the 4 x 1/4 wavelength freq in a folded TL at the cancelation notch?
No, a passive radiator works like a bass reflex port, a phase inverting Helmholtz resonator.
PRs are generally used in place of ports in small boxes where a port of adequate cross sectional area would result in an impractical length and port volume for a given Fb (box tuning frequency).
The moving mass (Mms) of the PR (237g for the GRS 12SW-4HE) is equivalent to the mass of the air in a port of equivalent cross sectional area as the PR's Sd (effective piston area, 507.3cm² for the GRS 12SW-4HE) that would tune the 2.6 Ft^3 net internal box to the same frequency (Fb) ~36Hz.
That port would be over 4 foot long, and since it's volume would be greater than the box (impractical...), it would no longer function as Helmholtz resonator, and would have pipe resonances related to length...
In the present compact enclosure, the circular cutouts on the front and rear baffles are in very close proximity to the side walls of the cabinet. As a result, there is not a lot of scope for much flexing of the baffle that might lead to any associated motion of the driver basket on each baffle. The forces the driver generates transfer directly into the four enclosure walls, and equal and opposite forces cancel them from the opposing driver. It also helps that that baffle is quite thick, being comprised of two layers of plywood. That makes it quite stiff, as the baffle's bending stiffness is proportional to the cube of the baffle thickness.
I just did a quick near field measurement with 1 driver active, 1 driver passive.
Red is active driver near field.
Blue is passive driver near field.
This measurement implies a tuning near 35hz. So, the measurement is close to this too, the near field active driver shows it immediately rolls off and shelves in the mid-30's hz range. Then, the blue line, the passive driver is humped up in the upper 30's hz range. Pretty neat to see it.
Note, I do have a physical low pass filter on that is dragging everything down above about 70~80hz range on the amp I'm using. But I don't care about the output up there, just to note it was active in this measurement. But no EQ otherwise.
Very best,
Last edited:
Thanks,
I think bracing more and bracing the magnets and them being closer likely matters more with much larger higher power drivers that have a lot more force being applied to the enclosure, so larger cones with more excursion and power handling. I think with average 12" drivers like this with 12mm xmax and 250w power handling, there's not enough energy to really show a major issue. But certainly something that handles 2kw of power and has nearly 30mm xmax and a larger cone, back to back uncoupled and not braced well enough would exhibit problems I imagine (example driver being the Stereo Integrity HST-18 with 2kw of power on each one).
Impedance wise, I think it only really matters depending on how you want to amplify it. These are budget drivers and I don't really want to put a costly amp on it, so 4ohm in series presenting an 8ohm load is inefficient, no acoustic power increase, and just divides power and really drops output of most inexpensive amplifiers. I think if I re-did this, with budget drivers still, I would use something with 8ohm per driver so I could wire total load to be in parallel at 4ohm and benefit acoustic power increase and keep the output of typical inexpensive amps, and ultimately would have more output for the same cost. The Skar SDR-12's are $20 more expensive and can come with dual 4ohm voice coils and would satisfy this. I would do this instead if starting over. But I already had 1 GRS and just had to buy a 2nd, so this cost me $59 to do this.
Very best,
It can be done with any number of pairs.
I've seen a tower with 6 total, 3 front 3 rear in dual opposed. It was awesome.
Very best,
Hrm, not sure I follow what you're saying. Height position in a room is just going to show the room response. Ground level is about measuring a near anechoic response to see what the design is doing without room influence.
I have no idea what you're saying about internal or screw positions or whatever MBL or whatever tis, let alone how that screen shot is relevant. Sorry.
Very best,
I bought 8 pcs. I would make 2 towers with 4 pcs each (in dual opposed). Crossover frequency 40Hz. I wonder if it is better to go with sealed or ported. I have 8.5 ft3 available for each tower. The amplification is made with a Dyaton SA1000. It has an internal parametric equalizer in order to compensate the freq response.
As far as I see you tested both configurations (sealed vs ported).
What did you preferred? It will be used for music only.
You can do sealed or bass reflex. The sealed volume of air is not why the vibration is canceled, its the alignment of the rear of the drivers to each other so that the forces they generate oppose each other. At least that I'm aware of.
8 ft^3 is plenty per tower for sealed.
For bass reflex, it would depend on the tuning. The curve would likely be a slope unless you tuned high with so many drivers in such a small volume relative to what a bass reflex would require for most drivers. These drivers are ok in about 1.2~2 ft^3 each, but their measured Qts is much higher so they do better in more volume in reality and 8 ft^3 would be small (for these GRS drivers).
If you want the most output possible to 40hz, then ported will get you there.
If you want the most bandwidth, sealed gets you there.
I went sealed for music, because I can shape it however I want with EQ.
For example, here's my current house curve on this dual opposed 12 that I run. It's near field, plenty of output, and I focused on 30hz output. That's the beauty of multiple drivers, shared xmax, more power handling. You can EQ them with more latitude. I prefer this for music. I can change it without worry. If I want to get 20hz output, I can. If I did bass reflex and tuned higher like 30~40hz, then it wouldn't have 20hz output. So I like having the options and I didn't need max SPL output. If you want max SPL output to 40hz though, yes, bass reflex.
Here's simulation of GRS 12 SW4HE x 4 in a 8.5 ft^3 tuned at 40hz bass reflex with 1000w of power. See how the peak is before 40hz? It would be better if we port tuned at 30hz or 35hz, but this will result in a larger, longer port.
But the response curve at 35hz is so much better I think:
Excursion is good to 25hz without a high pass filter needed. So for music, this is probably fine.
A slot port will do this easiest to get enough surface area and its length is short. I like how 12" diameter looks for air velocity here and the length. This is impractical as a pipe but as a slot port it would be easy.
Let's convert to a slot with end correction 2.227:
So maintain 12" area, which is 113.1 square inches in any slot dimensions you want, with a 6.4cm or 2.5 inch length and you get 35hz tuning in that volume. Easy port for an enclosure like this, low air velocity, clean for music, no chuffing or noise at all at max power. And it doesn't eat up a huge volume to make this.
Two of these would have huge output.
++++++++++++++++++++++++++++++++++++++++++++++++++++
That said, look at this:
The GRS 15SW4HE is on sale for $96 each currently. Four of them in a sealed tower in dual opposed configuration (2 front, 2 rear) in about 12 ft^3 volume with 1000w of power at 4ohm gives a nice response with a lot of output:
GRS 15SW4HE x 4
12 ft^3 sealed (dual opposed configuration)
1000w at 4 ohm
1 meter
I would build this over a bass reflex for music. Yea, its bigger. But full bandwidth and no port to fuss with. EQ to your heart's whim.
Very best,