Gluing an accelerometer to a cone or speaker cabinet

[Amphiprion]

OK, so apparently these things need to be solidly affixed. And they should be perpendicular to the axis of movement.

1) I would like to stick this on the backside of the cone at an angle to the direction of movement. Could I just adjust the voltage output of the accelerometer with a gain stage to account for the fact the acceleration isn't normal to the plane of the accelerometer?

2) I would just use epoxy to stick it to the backside of the cone, but I want it to be undoable (ACH-01's are $30 a pop from Digikey, and I'd like to use the woofer for something else later too). Any adhesive or mounting suggestions? Hank (aka Mr. 3M) I am looking in your direction!!!

3) Sort of like #2 - what's the best way to stick this on the side of a speaker cabinet a la Stereophile reviews?

Thanks for any help,



Amphiprion

[Hank]

Hey, I don't work in the adhesives division, El Nino. I think any adhesive you use will be difficult, if not impossible to completely remove. You could certainly slice through an adhesive with a razor blade to save the accelerometer, but getting every bit of the adhesive off the cone would be the problem. Try hot melt glue. Epoxy would be difficult to slice through. I'm interesed in cabinet resonance measurement and would like to measure a solid wall cabinet versus a constrained layer wall cabinet. Whaddya say? 8)

[Amphiprion]

Hey, this measurement is your measurement system, too. I've mooched enough saw usage off you just lemme know when you wanna use it. I don't think I'll have the accelerometer setup ready until I spend the cash on Praxis (might be within the next month). Depends on how my Intel stock gamble does. I'm up $300 on it in the last 3 weeks, so maybe not too much longer

ETA: Just checked the close, make that $350 :B

[Biff]

Re #1

Mark, would it be easier to set a wedge in between so the direction of motion IS perpendicular to the body of the accelerometer?

[Amphiprion]

Ooh, that could work too. Thanks for the idea!

[soho54]

What are you trying to accomplish with the accelerometer on the driver? Just curious.

[Amphiprion]

Get an acceleration signal from the driver

[soho54]

I mean for what purpose? You wanted to be able to remove it and mentioned Praxis, so I thought it might be for some type of measurement. So this is just for a DIY servo/feedback trial run?

EDIT: Asking about the cone mount only. I get the cabinet resonance part.

[joecarrow]

Are you working on motion feedback? I know some people made good progress on that a while ago, but I seem to remember reading that the latest subwoofers provide enough performance at a low enough price to make active feedback much less attractive.

Some quick links I found:

http://www.analog.com/en/content/0,2886,759__7004,00.html

http://danmarx.org/audioinnovation/servosub.html

[soho54]

I was thinking:
Experimental modal analysis of vibration modes on the cone, or a force balance accelerometer displacement sensor, maybe? Something crazy technical, you know. If only I could get into a house with a garage, and out of this duplex. I could stop tinkering vicariously through others. :B

[Amphiprion]

It's mainly for playing with motional feedback. I started thinking about it a few months ago and posted some questions here, finally getting around to playing with it. But yeah, I'm not ruling out going all mad-scientist with them

[kingpin]

Is this gonna start a:

"My woofer is faster than your woofer" thread. :B :B

Mike

[capslock]

You might want to stick it to the VC former to preserve phase margin, then.

[Amphiprion]

Would there really be a measurable phase difference in placing the accelerometer on the cone as close to the former as possible vs. on the former itself (like inside where the dustcap is)?

[mmoeller]

Would there really be a measurable phase difference in placing the accelerometer on the cone as close to the former as possible vs. on the former itself (like inside where the dustcap is)?

Unless you assumed an infintly stiff cone there would be some dampening and some time difference. A very small difference, but with the accuracy of the accelerometer I doubt that you would see it. Put one on the former and one on the cone. You could track the difference.

Would the accelerometer effect the cone vibration modes? How much does it weigh?

[JonP]

Unless you assumed an infintly stiff cone there would be some dampening and some time difference. A very small difference, but with the accuracy of the accelerometer I doubt that you would see it. Put one on the former and one on the cone. You could track the difference.

Would the accelerometer effect the cone vibration modes? How much does it weigh?

Interesting thread... I've wanted to explore box vibrations vs different bracing and dampening methods. Someday...

I'd think that putting it on most cones would alter modes a fair amount, even if it only weighs a few grams. Now, on a RSS 15", not so much, but on smaller ones it probably would change everything and not give valid data.

Putting it on the box would be easy and give accurate data, though, due to the relative mass between the box panel and the sensor, as well as the high stiffness.

There are many kinds of thin double sided tape that would work well. I've found a "carpet tape" at hardware stores that is a thin film a few mils thick, sticks very well, and removes without leaving much trace. I use it for tacking down things for router work, etc. Being so thin, it won't compress much with vibration, unlike a thicker or foam tape which would dampen the vibration you're trying to measure.

Good luck, and post some graphs when you get it going!

[Arneson]

In experiment mode you could find the the dampening effect of the attachment to be a good thing.
Why not prototype on a crap speaker to monitor the device and work out kinks.

[///M5]

Beeswax can work for attaching but the frequency response will start to roll off typically in the 2kHz range. You will have some mass loading as well since the mass of the accel is going to be significant compared to the mass of the cone itself.

As to your off axis question, using a cosine correction is fine.

The preferred way to measure cone velocity is to use a laser vibrometer like those available from Polytec; however, an entry level one is in the $25k range. I have one, but am located in Minneapolis.