A Study of DMLs as a Full Range Speaker

Can you please share how low your 60x40 cm panels play before they start to roll off (from the average level above say 100 Hz. or so)? Thank you.
This is the best I can do I'm afraid. The crossover was set to a 4th order Butterworth high pass filter @ 200hz. In theory the response should only have been 3db down @ 200hz. However, as you can see, it's much lower than this. My method of attaching the panel to the frame may be partly to blame.
 

Attachments

  • DML with 200hz XO.jpg
    DML with 200hz XO.jpg
    109.2 KB · Views: 21
Can you please share how low your 60x40 cm panels play before they start to roll off (from the average level above say 100 Hz. or so)? Thank you.
Since my very beginning with DML, the bass roll off has been a mystery. We are getting now interesting information from the simulation. The first output of the simulation is the FR of the panel as it is mounted in a window of a giant wall that prevents the rear wave to interact with the front (half space). In such conditions, the FR extend to the lowest mode. In practice we see a roll off higher in frequency. The hypothesis is the front and rear waves cancellation like in an open baffle so in first approach linked to the dimensions of the panel compared to the sound speed in the air. A DML is much complex than a classical open baffle as the location of the source change with the frequency (anti-node location due to the mode shapes).
Christian
 
  • Thank You
Reactions: moray james
@homeswinghome
Hi Christian.
I've given the panels with the exciter not attached to a spine a more serious workout. It's definitely an improvement, to the point that I'm probably not going to bother with EPS. Response curve follows RTINGS to with 1db from 20hz to 20khz.
Simon.
In my current understanding, a rigid spine interacts in 3 ways with the panel and its exciter:
  • as a vertical support to avoid the long term deformation of the exciter spider due to its weight. I have no experience with exciters from Xcite but Dayton Audio like a DAEX25FHE it is mandatory.
  • it stops the axial movement of the magnet and so the low frequency magnet resonance with 2 sub consequences:
  • a change in the response at low frequency. Not fully clear for me but I think Eric shown the LF extension goes a bit lower with a free magnet. Maybe the free/fixed option of PETTaLS allows to get a better idea.
  • a reduction of the distortion in fixed magnet. This is probably true only when the panel is acting without high pass filter. I can be check easily with REW.
  • by an acoustic mechanism I not able to describe, the surfaces parallel at the panel on the rear side creates a resonance in the midrange. This already exist with the exciter alone but can be increased with a spider. It is the case in my plywood panel where the exciter is supported by a 10x10cm plywood part.
In addition, the spine is a mechanical problem as there is a risk to modify the exciter magnet position, either in the axis direction or by rotation in the other directions putting the voice coil a bit out the position of minimum distortion.
So the common position is a spider is mandatory for the long term and we can add design in a way not to add constraints on the exciter and of reduce dimensions in the panel plan.
Christian
 
Since my very beginning with DML, the bass roll off has been a mystery.
I tried playing around with a Tectonic Elements BMR driver in a 46 x 50 cm sheet of wood. The bass falloff was very similar to that of my DML. This leads me to suspect something weird with the room. That said, DSP corrections in my miniDSP Flex sorts it all out very nicely. BTW, in my previous response curve, the frequency response follows the RTINGS curve from 30hz not 20hz. However, that's good enough as the lowest note on a bass is 41hz, and a piano 32hz. Only organ enthusiasts my feel the lack of another octave of bass extension!

I was playing some Philip Glass yesterday evening. Either the electricity was extra clean, or my ears were, but the system sounded very good indeed, with more of those leading and trailing edges apparent. The only place left to go now is either an amp upgrade or another DIY mains cable to my mains block based on the Russ Andrews Evolution 300 but at a fraction of the cost, and even that may not be an improvement on my Lapp Kabel DIY job.
 
@spedge
Steve, have you had any more thoughts on using Proplex or similar fluted polypropylene? Wickes do a 2400x1200 sheet of the 2mm thick stuff for £5.60.
My current panel is a 60x40cm bit of 2mm ceiba plywood semi-rigidly attached to a wood frame with spaced bits of 6mm thick foam glued to panel and frame. I'm using Xcite exciters which are a cut above what I was using before.

Bass duties are taken care of by H frame units with 15" drivers. Crossover & DSP is done via the miniDSP FLEX, with overall tuning to the RTINGS curve.
Christian is correct in that the round flueted 6mm panel is my preferred material.
The square flutes have very flimsy walls which buckle and bend, the round flueted proplex panels are designed to withstand a forklift trucks weight.
It would be interesting to see how a 3mm or 4mm round flute would sound ?
I also tend to have a raised response in the lower frequencies, it just sounds more natural to me.
It also matches the sound from my headphones, giving me a more "being their " sound.
A lot of my panels tend to have a natural rising response in the lower midrange anyway, so no need for any EQ.
I just match the low frequency drivers response to match the panels response levels.
With my usual overlap of frequencies in the 100hz to 300hz or so depending on panel type.
It works for me 😃
Steve.
 
I just used random pink noise and ran it for about 100-150 samples. I used to use the ,moving mic method', but found a central position gave me the same results.
I also tried the Harman curve (same source as the one you mention), but found it a little lacking in the lower midrange.

I'm thinking of trying something other than ceiba ply. The sound is very 'safe', but lacks some transparency. Perhaps the ply is 'slower' than say EPS or Proplex leading to masked leading and trailing edges. It seems frequency response isn't everything!

Whatever panel material I decide on, I may try the late, great Zygadr's cloth tape mounting method.

I've lost track of Steve's (@spedge ) preferred damping method for EPS. I think it's sand both sides and apply 50/50 PVA/water to both sides. I just need to decide on EPS grade. Zygadr used to promote ultra high density stuff, whilst I think Steve prefers lower grades. I may hedge my bets and go for EPS100 which is one rung above standard EPS70.
Zygadr used to have vary large panels , as I did in the early days.
he used 10mm HD eps , I used 25mm lower density eps
A 10mm 8ft tall low density EPS panel would not have been able to support itself and would have just flopped over, similar to the large 1mm ply panel I made the recording of.
He did not even get a chance to try my sanding technique.

I don't believe he got around to trying lower density materials before he passed away.

If you like the sound of ply , I would recommend using thin flexible ply.
I thought it was very good.


Steve.
 
I tried playing around with a Tectonic Elements BMR driver in a 46 x 50 cm sheet of wood. The bass falloff was very similar to that of my DML. This leads me to suspect something weird with the room.
With a peripheral suspension, the anti node of the first mode of the panel is a the center of the panel which gives it a behavior similar to an open baffle of about the same size with the driver not too far from the center. I guess this explain the similar bass roll off. Above the comparison is not possible as the anti nodes move, the driver of an open baffle has a fixed position.
 
The hypothesis that the DML's bass response is affected by the back wave, similar to an OB, has me wondering about putting "wings" around the panel, similar to what I've seen for some OB designs. That might be an interesting experiment.

Based on some experimenting I've done with placing sound-absorbing panels behind my DMLs, I'd expect some changes in the apparent depth of instruments and vocalists as a result. Interestingly, FR measurements didn't reveal significant differences with and without the panels.
 
The hypothesis that the DML's bass response is affected by the back wave, similar to an OB, has me wondering about putting "wings" around the panel, similar to what I've seen for some OB designs. That might be an interesting experiment.
My bass drivers (15" Monacor) have been in asymmetrical semi-topless U frames and are now currently in H frames. Even with a Qts of 0.75 and an Fs of 25hz, bass response drops heavily below 30hz. The Fs of some exciters can be in the 200s, so perhaps bass falloff is to be expected. However the Xcite datasheet doesn't reflect what I'm measuring.
 
The hypothesis that the DML's bass response is affected by the back wave, similar to an OB, has me wondering about putting "wings" around the panel, similar to what I've seen for some OB designs. That might be an interesting experiment.
I have the same question... Note that as for an OB, it could happen the bass roll off is lowered but also some effect on the directivity. One additional possibility is to use some fabric as wing. It seems it is something used in ribbon microphones.
 
My understanding is that one has an oversized frame e.g. with internal measurements of 70x50cm for a 60x60cm panel. One then uses self-adhesive cloth tape (like gaffer tape) - so for 50mm wide tape 20mm is stuck to the panel material, 5mm gap then 25mm stuck to the frame. The tape is put on the front. The tape is on each complete side i.e. no gaps.
I think that this can be a pretty good mounting method. I've been calling this the "bridge" suspension, because the suspension bridges across the gap between the panel and the frame. I know I posted these images before but I don't think I shared the frequency response. Far from perfect but decent output down below 50 Hz. And about the best looking spectrogram I've generated. This was an outdoor measurement so really minimal reflections compared to most.
Eric

1751306324062.png 1751306463007.png

1751307123240.png

1751306917738.png
 
However the Xcite datasheet doesn't reflect what I'm measuring.
Good remark.
Below is an extract of the XT25 datasheet.
The panel is 0.3m² with a ratio of 2 so about 0.39 by 0.78m. Something unexplained or simply the panel is said open back because without rear load like a closed back but each side radiating in its own space?
1751307345037.png
 
  • Like
Reactions: lordtarquin
However the Xcite datasheet doesn't reflect what I'm measuring.
Can you clarify exactly what you mean by this? I'm not sure that Fs of an exciter can be interpreted the same as you would for a regular driver, especially not the Fs coil. Typically, the Fs magnet correlates better with the the exciter's low frequency cut-off in my experience, although I can't say for sure that that is a reliable rule either. Xcite seems to be the only supplier reporting Fs magnet.
Eric


1751308873989.png
 
I think that this can be a pretty good mounting method. I've been calling this the "bridge" suspension, because the suspension bridges across the gap between the panel and the frame. I know I posted these images before but I don't think I shared the frequency response. Far from perfect but decent output down below 50 Hz. And about the best looking spectrogram I've generated. This was an outdoor measurement so really minimal reflections compared to most.
Eric
@Veleric This was the first method I tried a few years ago but with the self-adhesive foam strip on the back rather than the front. Currently, my panel is attached to the frame with 12mm wide foam glued to the frame and the panel. Each bit of foam is about 40mm long and the gaps are similar. I had toyed with the idea of going back to my previous method but with front mounted foam i.e. your method, but I'd need some new frames - something like 70x50cm external.