Welcome! Need support, you got it. Or share your ideas and experiences.
Return to Planar Speaker Asylum
85.188.73.4
In Reply to: RE: "When Andyr fits those Neo8's then I will be able to pop around and hear them!" ... posted by contiguous1 on June 02, 2016 at 22:38:50
A rewiring kit from Magnepan is $ 35+ shipping. Four rolls of wire and glue. It is not that difficult to do it by yourself.Sure, the IV is brighter sounding than the IVa. The BG Neo-8-drivers are not without faults, removing the front plate with its magnets improves the respone and lowers the distorion. Output level will sink but is still okey. There is a resonance caused by the trapped air between the front and rear plates/magnets. A single sided diaphragm radiates more freely.
A row of six Neo-8 makes a rather small diaphragm. You will need about 11 to get the same size area as for a Tympani IVa mid. Neither the Neo-8 or the Tympani IV or IVa mids can move very much air.
Edits: 06/02/16Follow Ups:
The stronger motor on the Neo8 can move more air so can cover a lower frequency once the line is long enough and what is restricting output below 800 hz is dipole cancellation, which you can eliminate by using more baffle. The main source of distortion is the cavity resonance and it cancels out. The lower freq distortion can be eliminated if you are willing to use a back cup and lose the dipole pattern, but I saw the THD measurements for the neo 8 drivers and the pro version PA Neo based long drivers and they are low..
Roger kudos to your technical ribbon skills!
I will pass on that amount of effort until I have a few spare panels (just case).
Maggies Tympani IV's
Hi Roger,
So the Neo 8s have magnets both sides ... which causes a resonance/distortion??
So how come there is not the same resonance/distortion on the 20.1/20.7 mids (which have magnets both sides)?
Andy
The 20.7 is also likely to have that cavity resonance as well, but it is probably around the 10khz mark so it is probably there at close mic measurements, however there is a cancellation effect of this resonance as you increase the length of the driver or array of drivers the FR bump from the cavity resonance reduces in size and disappears entirely in a 5 ft driver or array at a distance.
I would not imagine that the tradeoff of less sensitivity and reintroduction of single ended driver compression is worth the gain in resolution and the absence of the cavity resonance. you get by removing the front magnet board. particularly since the cavity resonance peak is cancelled out at listening distance.
So by having 6x Neo 8s in row (in a T-IV or T-IVa), you won't experience the effect that only 1x Neo 8 delivers?
Andy
Yes, each doubling of the number of drivers in the line reduces the resonance peak by about 3+ db. That is because the resonance freq are only produced from within the cavity space between the holes so do not have an on axis component. The cavity resonance output are at random phase from each other so cancel out as you increase the length of the line. A similar resonance exists in the larger RD units that becomes smaller as you go from the 22" unit up to the 75" unit.
The normal model of a line array of point sources does not really hold for a planar line array because it is a real cylindrical pressure waveform constrained by a lack of vertical dispersion due to the confines of the floor and ceiling thus not allowing a vertical pressure differential or gradient In the case of the Beveridge waveguide they make sure nearly no possible vertical component output can leave the speaker above a certain freq, which is part of the reason for its amazing clarity, it produces a true cylindrical waveform and pretty much nothing else. The cavity resonance output has a 0-180 deg phase offset to it and delay of at least one cycle relative to on axis. So when you are sitting at the midpoint axis of the line the cavity resonance output from the bottom has a counterphase output from the top and will cancel out as random noise does when you double the number of devices in an amp's output stage.
The ET planar speakers don't have a cavity resonance since their magnets are held far apart from each other without any excess material between the vertical lines of magnets.
I also tried a CD horn on the line array as I added pairs of drivers to it and saw that in addition to the CD horn EQ effect I got a progressive cancellation of the cavity resonance as the CD horn was lengthened (more so with the 90 deg horn than 120deg). In this case, like in the Beveridge, the waveguides were causing a horizontal axis cancellation of the cavity resonance by reflecting the sideways output back on itself from the horn walls..
ESLs have a similar cavity resonance but you never see it on far field FR measurements since it is cancelled out by then. And you probably noticed that ESLs are nearly all tall line arrays occupying 1/2 or more of the typical floor to ceiling height. with the exception of the Quad 57 and the later point source ESL63 and onwards..
Can you demonstrate (with measurements) that multi-driver cavity resonance reduction without including the inherent line-array propagation aspect within those measurements? This seems a catch-22, to me. Do you see what I mean?
Dave.
Can;t separate it out of course, What I did do was leave the mic at the midpoint height at the listening seat and block the bottom half with a board at a couple of distances on the mic speaker line and the closer one ~3ft did not have as much of a reduction in the cavity resonance, but the 6-7 ft distance did not show any different result than the listening position without the board. Just shows crudely that the cancellation effect takes place in proximity to the drivers within a distance the length of the driver line rather than further out.
I did close mic FR sweeps on the 6 driver line but don't recall the results exactly. I think it showed some reduction from the peak on the single driver FR but it was no different than measurements at a distance on a 2 driver line so would have been a 3-4db reduction in the peak.. I do remember setting the distance at 8" to correspond to the driver height.
So, you are saying the cavity resonance does not exhibit a minimum-phase characteristic and each driver is different than its neighbor?Dave.
Edits: 06/05/16
The cavity resonance is a trapped air resonance so has to build over time so would have random phase relative to the signal and a time delay of at least one cycle. It is exactly a random noise with amplitude related to freq, not a freq response variation due to any other cause like a driver's breakup and self absorption modes related to non pistonic behavior. The diaphragm itself is not producing the resonance. It is more like a cabinet resonance.So should not be thought of as a part of the freq response of the driver but a noise with its own separate behavior.
What is different among drivers above and below the midline listening and measurement axis is opposing directions, nothing more.
When I started trying out the Neo8 drivers my plan was to use a notch filter if my line did not squelch the resonance as some others had shown in their Neo8 line array measurements. But as I increased the number of drivers in he array the resonance peak progressively declined in magnitude and was pretty much vestigial with 6 drivers per side.
Can you post your measurement of a single (or two) driver operating in the array, and then a measurement with all 6 drivers operating?
(Same microphone position obviously and equivalent electrical drive voltages to the transducers.)Dave.
Edits: 06/05/16
I don't have the old measurements anymore. But I posted other's and some of my own before. The old laptop they were on died suddenly some 5 years ago I didn't back up since that was the backup pc for work.
I have this left from 1/3 octave warbles from about 8ft. It should be on axis from the equidistant arc setup. The leftover peak is smaller at the listening seat @11ft by about 2 db. I will look for those. There should also be one channel sweeps somewhere.I can look through the spreadsheets to see if anything else interesting is left. The old REW sweeps I don't have any longer.
Edits: 06/06/16
The area of interest has very low resolution on that plot, and the primary peak (around 8khz) is not the driver cavity resonance.
Regardless, I still think I would apply the appropriate notch filter to all the drivers. The cavity resonance is inherent in the transducer construction and it can't be "averaged" away by acoustic "coupling." If you move your microphone closer you will start to see it. I know they allude to this coupling "alleviation" characteristic in the driver datasheet, but I'm not buying that's the most preferable implementation.
Anyways, extending on your hypothesis......if you were to apply the 12khz notch filter to the array configuration, you would create a dip/anomaly in the free-field response, yes/no? I don't believe that to be the case. I think the measured acoustic response would look even better/flatter. And if you moved your microphone close it would look even better/flatter too. Win/win.
Cheers,
Dave.
I posted a Neo8pdr 9 segment line array from a DIY forum that was done with better equipment and there is nothing remaining of the cavity resonance. It should still be on the asylum. The forum is defunct so I can't post the notes from the projects Yes, the use of a notch filter produces a dip in the line array. The 12khz resonance is entirely gone but the residual portion at the onset of the peak requires more distance to dampen.
The resonance peak does get cancelled at a distance, most of it is gone by 1 m. I already knew to expect it from other's line array projects so never used a notch filter. As you can see in the FR plot the combined drivers are not showing anything. The plot I posted is the LR4 symmetrical at 5khz. I got very similar results with 1st order at 6-8 khz though a tiny bit of the 8khz peak is visible in the combined speaker I will see if I can find a trace, but the 1st order work was done 5 years ago. The LR4 and B3 crossovers were from nearly 10 years ago. So I can redo measurements to show you how it works but it very obviously does.
The FR at the listening seat is not showing any need for a notch filter. You should not view the resonance as an inherent part of the driver's FR but as a noise with its particular behavior. At normal listening distances of 2-3 m there is no need to pay any attention to it. Definitely no notch filter.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: