Maybe the best approach for DIY exciters is to try and make a silk purse from a sow's ear. Stick a rare earth magnet onto the panel and bring a solenoid close to it. Add distance sensing (Hall effect, reflectance, interferometric, back emf, idk there has to be chips for that) and do bespoke closed loop control on the whole thing - magnetic non-linearities, panel dynamics, to get linear sinusoidal displacement and optimise overall response. Some class D amps already integrate DSP, and probably some of them have feedback control inputs for the PWM ?
Eric or Christian,
Would you have to have the properties and results of a run of FEM eigenmode analysis for a rectangular plate? I'm trying out Elmer, so far OK but don't have a real example to validate. Had a look in past posts and found some results but properties were in a previous post.
Length / width / thickness
Young's modulus - can only do isotropic AFAIK.
Density
Poisson ratio
Support (so far Im only sure how to set up clamped support)
Thanks, Paul
Would you have to have the properties and results of a run of FEM eigenmode analysis for a rectangular plate? I'm trying out Elmer, so far OK but don't have a real example to validate. Had a look in past posts and found some results but properties were in a previous post.
Length / width / thickness
Young's modulus - can only do isotropic AFAIK.
Density
Poisson ratio
Support (so far Im only sure how to set up clamped support)
Thanks, Paul
Hi @pway
Yes your description is right. This image from a stack exchange page may help.
If you now imagine two radial ring magnets opposing one another the field lines between the magnets will compress increasing the field lines seen by a flat coil placed between them. This is no different to a method used on Magnapans using bar magnets and linear ‘coils’. As radially magnetised ring magnets are rare And expensive intend to use a circular array of bar magnets so in a sense all I am proposing is a ‘rolled up’ Magnapan motor. As Andre’ noted above, the B in BL is constrained by Xmax but if using a DML panel from 200hz up the required Xmax is not large. DML’s are not pistonic devices so displacement is not a design requirement. Because of the possibility for relatively large diameter coils the. L in BL is less restricted than in a conventional voice coil and may compensate for any loss of B. All hand wavey theory at this time.
Burnt
Yes your description is right. This image from a stack exchange page may help.
If you now imagine two radial ring magnets opposing one another the field lines between the magnets will compress increasing the field lines seen by a flat coil placed between them. This is no different to a method used on Magnapans using bar magnets and linear ‘coils’. As radially magnetised ring magnets are rare And expensive intend to use a circular array of bar magnets so in a sense all I am proposing is a ‘rolled up’ Magnapan motor. As Andre’ noted above, the B in BL is constrained by Xmax but if using a DML panel from 200hz up the required Xmax is not large. DML’s are not pistonic devices so displacement is not a design requirement. Because of the possibility for relatively large diameter coils the. L in BL is less restricted than in a conventional voice coil and may compensate for any loss of B. All hand wavey theory at this time.
Burnt
Last edited:
A fascinating approach @lenta. It must have taken a lot of research to establish the node points. What is the bandwidth and efficiency like?
Burnt
Too low SPL, to have enough signal-to-noise ratio for accurate frequency response. Microphone should be at 2 m (or 2 yards) distance from the panel and average SPL must be at least 75 dB.
At the moment I am listening to some music, I will post a recording, I can't hear anything nasty so far.
To be honest, it's sounding very nice.
If I had spent a long time designing this whizzer I would be very pleased.
Oops just heard some splashy cymbals.
I had a listen on my headphones and it's on the recording 😃
Not bad.
Steve.
To be honest, it's sounding very nice.
If I had spent a long time designing this whizzer I would be very pleased.
Oops just heard some splashy cymbals.
I had a listen on my headphones and it's on the recording 😃
Not bad.
Steve.
Attachments
While thinking about exciter design I would like to point to this commercial dml speaker and exciter again. It has been mentioned before in this thread, but what intrigues me still is this bbx exciter they are using. It looks nothing like any of the other exciters available and it must be a very powerful exciter looking at the speaker’s specifications.
bbx - AER Loudspeakers (aer-loudspeakers.com)
bbx - AER Loudspeakers (aer-loudspeakers.com)
Good shout @twocents. From the specs the exciter is a mix of exciter acting on a panel from 2000hz down and a metal diaphragm combined mid and tweeter up to 30khz. It’s an interesting approach!
Burnt
+ @twocents
Let me play the counter-weight to balance the opinion : the DML have 2 characteristics : no crossover in the mids, wide dispersion. In addition, it is not so difficult to reach at least 10k. So why a solution with a diaphragm from 2k. But I am perhaps wrong in the understanding of the principle of this loudspeaker?
i made a few recordings but i thought this one was better ,mainly because of the microphone roll off below 100hz.
it is just the two panels with the coffee containers in place running full range.
the microphone was in-between the two panels ,the panels are about 40cm from the microphone and about 30cm below the coffee whizzers, (the rug microphone stand was too short)
steve.
it is just the two panels with the coffee containers in place running full range.
the microphone was in-between the two panels ,the panels are about 40cm from the microphone and about 30cm below the coffee whizzers, (the rug microphone stand was too short)
steve.
@homeswinghome and @twocents its not obvious that transducer bandwidth limits its DML action until you read the detailed technical description on their website and it isn't obvious even then. They do call it an exciter, and many of its design features are typical of exciters, but they have certainly chosen an unusual interpretation. They use acrylic for their panels and claim a cutoff of 40hz which is feasible with acrylic, there are other examples on YouTube. Its an interesting variation on the theme but I would like to see a similarly well engineered exciter used for a wider bandwidth myself.
Burnt
The bbx is probably a shaker?
At first I thought it was a solenoid shaker with the cone attached to the solenoid, but they seem to say they are separate drivers?
A shaker would put far more force into whatever surface it was attached to.
I would think it would be very impressive attached to a stud wall.
Although the price would probably be excessive?
I have seen one shaker that has a frequency up to 12k or 15k ?
Steve.
At first I thought it was a solenoid shaker with the cone attached to the solenoid, but they seem to say they are separate drivers?
A shaker would put far more force into whatever surface it was attached to.
I would think it would be very impressive attached to a stud wall.
Although the price would probably be excessive?
I have seen one shaker that has a frequency up to 12k or 15k ?
Steve.
I chose the recording for my coffee ☕ whizzer because of the female close microphone vocals with lots of sssss .
They made no effort to suppress this on the recording.
I compared the recording with my headphones and they were very similar with no HF ringing or excessive sibilance.
I will have to phone my friend and ask him to save the small 2cm coffee ☕ containers for me.
They already think that I am mad for taking their last supply 🤪
Steve.
They made no effort to suppress this on the recording.
I compared the recording with my headphones and they were very similar with no HF ringing or excessive sibilance.
I will have to phone my friend and ask him to save the small 2cm coffee ☕ containers for me.
They already think that I am mad for taking their last supply 🤪
Steve.
I think I might glue the coffee ☕ whizzer to an exciter and measure it.
First glue it to the exciter ring foot, then glue it to a piece of ply (or other) on the exciter foot.
It would be interesting to see the difference?
Steve.
First glue it to the exciter ring foot, then glue it to a piece of ply (or other) on the exciter foot.
It would be interesting to see the difference?
Steve.
I took the exciter off of the CD.
I thought I'd better show what I had to do to the CD to get a decent sound from it.
Plenty of bluetack similar to the rings tectonic use on their BMRs.
Obviously you'd want these around the back.
Not sure if this glue will stick to the coffee container?
Might have to use some Evo-stik.
Steve.
I thought I'd better show what I had to do to the CD to get a decent sound from it.
Plenty of bluetack similar to the rings tectonic use on their BMRs.
Obviously you'd want these around the back.
Not sure if this glue will stick to the coffee container?
Might have to use some Evo-stik.
Steve.
Attachments
That is what I call dedication🙂 I slowly continue with the Ikea canvases, yesterday I glued 160 x 270 x 4mm poplar plywood panels to them. The back side of the panels was pre-coated with 50/50 PVA/water mixture just to provide some basic protection.
I have a question about mounting the daex30hesf-4 to the spine - can I use the M6 thread/screw with a neoprene or rubber washer between the exciter and spine? I would also put on some Loctite to the thread, so that it does not get loose.
I have a question about mounting the daex30hesf-4 to the spine - can I use the M6 thread/screw with a neoprene or rubber washer between the exciter and spine? I would also put on some Loctite to the thread, so that it does not get loose.
Paul,
Yes, I have probably hundreds of them! Tell me what you want me to model and I'll do it.
Eric
Paul,
I don't know Elmer, but LISA works pretty well for me. It can handle orthotropic materials, which is important to me since the most interesting panel materials (like plywood and other sandwich composites made with carbon fiber or fiberglass skins) are all orthotropic. As far as supports, LISA can do free, simple, clamped, and any combination thereof.
There's a free version of LISA which has a pretty limited number of nodes, but actually it's enough to do some reasonable plate models. I paid for the full version which I think was $100.
Eric