Carbon Fiber open baffle dipole panels

Status
Not open for further replies.
Jump to Latest
T

thadman

Member
Joined 2005
#1
I spoke with one of the Materials Science professors at my University and apparently they have some surplus Carbon Fiber I might be able to use. The material is currently being used by his Grad students for fatigue testing for military aircraft. I am not sure of its exact parameters, however I believe it is very expensive. The splinters are not very fun to pull out of your hands.😀

Would carbon fiber be a good material for an open baffle? Intuitively, I would not assume it to be optimal since a resonance will probably occur within the bandwidth of the loudspeaker. However, I'm not sure what would be more optimal other than maybe Corian.

How many plys would you recommend? 8?

0, +45, -45, 0, 0, -45, +45, 0

He also has a large amount of honeycomb Zylon I might be able to use for a core. However, I'm concerned the panel may become excessively thick (1/4")...or maybe I'm overthinking things.

Would you use a Zylon core?

How would you mount the drivers to the Carbon Fiber? I'd like to use the SS 15Ms I have laying around.

I've got a pair of Peerless HDS tweeters I could use for the HF. However, they wouldn't be dipole. Alternatively, I could build a small DIY open back ribbon and bond the CF directly to the C1010 steel used in the magnetic circuit. He has some VERY strong epoxy (can withstand 20,000lbs of force).

If I used the SS 15Ms in a MTM format, how long would you make the ribbon? I'm thinking a 1500hz crossover would be a good place to start.

Should I let the baffle "swing" or use steel bracing for floor mounting? If braced, the resonance may be directly excited.

Thanks,
Thadman
 
Last edited:
#2
The most interesting application I see for composites in a dipole is shaping H and, to a lesser extent, U baffles---if you think about it, as usually made they're basically poorly built horns and should benefit from reduced rectilinearity. Can you clarify your design, the design goals, and link the 15M datasheet? From the above it sounds like you're thinking a flat baffle two way but desired SPL levels and so on aren't specified. If you check this forum there are several of active or recently active dipole build threads that cover baffles extensively.

Most carbon fiber's neither particularly cheap or particularly expensive, though if the research in question is using milspec weaves it'll be expen$ive. But for an audio baffle I wouldn't necessarily expect any great advantage of fiberglass, kevlar, or other composite layups. I'm curious, however, of your motivation for keeping baffle thickness under 6mm.
 
#3
I'm working on open baffles made out of kerfed mdf as a core and several layers of fiberglass as a skin. Your carbon fiber should work fine, but its main advantages (high stiffness to weight ratio) may not be used to full advantage in your application.

The choice of core--material, thickness, etc.--will have as much impact on your baffle's behavior as will the fiber skin. I think Rockport is really going about this the right way so you might check out their site: highly damped core with thick fiberglass or carbon fiber skin. Of course if you're of the "stiff and light baffle" school then you might take a different approach. I just don't happen to be a member of that school.

I see twest820 posted while I was typing. I second his question regarding the thickness of your baffle. Why so thin? Stiffness scales very nonlinearly with thickness so a little extra thickness goes a long way.

Few
 
Last edited:
#4
Few said:
I see twest820 posted while I was typing. I second his question regarding the thickness of your baffle. Why so thin? Stiffness scales very nonlinearly with thickness so a little extra thickness goes a long way.

Few
Click to expand...

Airflow.

At high frequencies, the magnetic circuit will begin to look like an H-frame. As a result, standing waves and diffraction will result if the baffle is too thick. For 20khz reproduction, the full depth should not exceed 4.5mm.

With midrange / midbass drivers, the baffle is often chamfered behind the frame. This is done to minimize turbulence and maximize airflow. If the baffle is made thin, chamfering will not be required.
 
#5
Few said:
I
The choice of core--material, thickness, etc.--will have as much impact on your baffle's behavior as will the fiber skin. I think Rockport is really going about this the right way so you might check out their site: highly damped core with thick fiberglass or carbon fiber skin. Of course if you're of the "stiff and light baffle" school then you might take a different approach. I just don't happen to be a member of that school.
Click to expand...

I hold a similar opinion. Resonances are unavoidable above 500Hz. It's simply unrealistic to assume a baffle could achieve such things. We might minimize the amplitude of the resonance by damping as you suggest.

At low frequencies, however, light and stiff would be the way to go.

I understand these materials are not optimal per say, but you can't beat free.
 
#7
twest820 said:
Not exactly clear what the previous post is about; taking a tweeter apart to better understand how they're built before proceeding further might be in order. If you're referring to the ribbon case I would suggest taking a closer look at the Neo3.
Click to expand...

I'm familiar with how tweeters work. The ribbon is perhaps the simplest mechanical transducer in terms of topology. We might think of the magnetic circuit as a "duct" in which the ribbon is located.

http://www.linkwitzlab.com/images/graphics/hframe1.gif

At high frequencies, the thickness of the baffle approaches and often exceeds the physical wavelength of sound waves. Since the maximum depth for the duct is already defined (<4.5mm for 20KHz), the depth of the magnetic circuit will decrease as the depth of the baffle increases. Obviously this is undesirable since we are losing linear Bl.

The idea is to place the h-frame resonance at least an octave outside the bandwidth of the transducer, since it cannot be equalized.
 
Last edited:
#8
Can you post a figure describing this design and explain the requirements to baffle the ribbon and why the ribbon magnet structure determines the thickness of the baffle everywhere rather than just near the ribbon? Seems to me this approach makes things unnecessarily difficult.
 
#9
twest820 said:
Can you post a figure describing this design and explain the requirements to baffle the ribbon and why the ribbon magnet structure determines the thickness of the baffle everywhere rather than just near the ribbon? Seems to me this approach makes things unnecessarily difficult.
Click to expand...

My last post may have been a little unclear. I was talking about distribution of variables. If the baffle depth (as a design variable) increases, then the magnet depth (as a design variable) must decrease to satisfy our total depth requirement (4.5mm).

baffle --> magnet --> baffle

A little more clear?

Here's the Scanspeak pdf

http://www.madisound.com/catalog/PDF/scanspeak/15w_8530k01e.pdf

I've also got a pair of Beyma TPL150 AMTs that I've been wanting to use. They've been up for sale for a while, but no buyers, so I might as well use them if they'll work. However, they're probably overkill. I only need 100dB peak as I live in a small apartment. I doubt the other renters would appreciate any more than that.

I see you've got a Neo3 in your avatar. Do you have a lot of experience with that particular driver? How do you think it would handle a 1500Hz crossover point dipole?
 
Last edited:
#11
thadman said:
A little more clear?
Click to expand...
Not really; still looking for that explanation of why you need the entire baffle to be the same thickness as it is near the ribbon. I'm guessing by increased baffle depth causing a reduced magnet thickness you're actually referring to increased carbon fiber layup thickness within a fixed baffle thickness.

thadman said:
How do you think it would handle a 1500Hz crossover point dipole?
Click to expand...
I seem to end up answering that in just about every dipole thread. 😛 With that and the latest round of questions about midranges I'd also suggest Saurav's dipole midrange suggestion in HTGuide's Mission Possible forum.

From the low crossover orders and need for baffle I take it you're planning passive crosses and only one amp channel per speaker. I originally read your remarks about ribbon magnets forming an H baffle as being about as being in reference to the back wave of a small driver, so I'm sure you can see the back wave symmetry problems the use of small drivers creates. From the choice of small drivers it seems you may have bookshelves in mind; in addition to greater specificity on SPLs some clarity on the baffle sizing and low frequency targets seems in order.
 
#12
twest820 said:
Not really; still looking for that explanation of why you need the entire baffle to be the same thickness as it is near the ribbon. I'm guessing by increased baffle depth causing a reduced magnet thickness you're actually referring to increased carbon fiber layup thickness within a fixed baffle thickness.

I seem to end up answering that in just about every dipole thread. 😛 With that and the latest round of questions about midranges I'd also suggest Saurav's dipole midrange suggestion in HTGuide's Mission Possible forum.

From the low crossover orders and need for baffle I take it you're planning passive crosses and only one amp channel per speaker. I originally read your remarks about ribbon magnets forming an H baffle as being about as being in reference to the back wave of a small driver, so I'm sure you can see the back wave symmetry problems the use of small drivers creates. From the choice of small drivers it seems you may have bookshelves in mind; in addition to greater specificity on SPLs some clarity on the baffle sizing and low frequency targets seems in order.
Click to expand...

I have by far and away enough displacement for low frequency. 4x NS15 and 2x TD15M. However, I don't need very high SPL. I live in a small apartment close to campus and the other renters aren't very kind to the idea of >100dB.

I will be listening to the speakers in the nearfield (<1m). As far as baffle width goes, as narrow as possible to preserve the dipole radiation pattern.

I'm planning on purchasing an E-MU 1616m and GTX470 fermi for my PC, so PC XOs w/ CUDA processing are a possibility in the future. In addition, I've got a DCX2496 so active crossovers wouldn't be a problem. However, I like the idea of a minimalist passive crossover (1st or 2nd order crossovers) so I'd prefer to go that route as it affords me more channels from my PC (1616m has 6 balanced outputs).
 
Last edited:
#14
twest820 said:
Why not skip the baffle entirely and go with a nude swinging three way? You've the EQ capability for it.
Click to expand...

The drivers will be basically nude anyways with the CF baffle. I think it would look a lot nicer than having each individual driver swinging.

I currently have 3 Crown XLS602D amplifiers. They have enough firepower for whatever I need, but aren't really noted for their midrange or high frequency performance.

I'd like to buy a high quality ~50 watt Class A or Class AB amplifier for the midrange section. If I went active, I'd have to buy 3 amps vs 1. I believe they might benefit from superior amplification. I might even settle for 10 watts if I could build the amp myself. If I built the amp myself, active may look more attractive.

Any thoughts on chip amps vs tube amps?

The Soviet GU-50 tubes are inexpensive and supposedly offer high performance. I could operate the tubes without any negative feedback and connect the preamp to the grid for simple triode operation. However, this will probably require an output transformer.
 
Last edited:
#16
Variac said:
You might also consider some class A S.S. amps. The Pass ones are about as simple as a tube amp, sound fantastic, and no output transformer
Click to expand...

Great idea😀

Are there any specific ones you'd recommend?

Any ideas on the total cost for 3x 10w Pass Class A SS amps?

Any thoughts on running minimalist passive crossovers between the SB17NRXC35-4, SB12NRXF25-4, and Neo3 and using FIR filters via CUDA or IIR filters via the DCX2496 for phase linearization / response shaping and implementation of an active HPF for the 6.5"?

The E-MU 1616m only has 6 balanced outputs. I'm gonna need at least 1 output for the NS15s. If I used the DCX2496 for bass management, I believe I could use 1 input for 4 outputs with different levels of equalization on each output.
 
Last edited:
#17
It looks like matlab has a plug-in for CUDA acceleration already😀

http://developer.nvidia.com/object/matlab_cuda.html

I'd assume writing the code for the FIR filters in matlab would be 10x easier than writing them in CUDA.

If I used CUDA acceleration, I might be able to save some cash on PC components. Do you think the system would take advantage of 12GB of RAM or would 6GB DDR3-1600 be fine? I'd be using an i7 930 w/ the Gigabyte X58A-UD3R motherboard.
 
Last edited:
#18
I've got 4x Aurasound NS15 and 2x Dayton RSS390HF.

I'd like to use the Daytons for LF reinforcement (20-80Hz)

There are two possibilities for the Aurasounds.

2x dipole + 2x monopole or 4x dipole

Should I mount the Aurasounds on the same "swinging" baffle as the SB17 / SB12 / Neo3 or use a separate baffle?
 
#20
Keriwena said:
I don't mean to be a wet blanket, but all I can think of is they use carbon fiber sheets to make guitars because it resonates so well. While lacking in exotic appeal, you may be better off with a Masonite/Homasote open sandwich.
Click to expand...

What about sandwiching a layer of 3M 112P05 viscoelastic dampling polymer between the carbon fiber panels for constrained layer damping? Alternatively, 3M VHB tape could also be used.

What about the CF --> Zylon honeycomb core --> CF combination? Would you still consider that prone to resonance?

Could anybody recommend any other damping materials?
 
Last edited:
Status
Not open for further replies.
Top Bottom