pick and filter frequency from microphone

[Kris_worm]

Hi all,

I'm currently working on a project picking up stone or steel rolling with timber logs in a huge drum (the drum is made of steel), through a laptop with a microphone attached, using Labview.

All I need to do is getting sound from the mic, do some analysis and then determine if there is any stone or steel inside the drum, in which case may damage the drum. (we believe when stone or steel knock on the drum, they create higher frequency than the one created when timber hit the drum, maybe we are wrong???).

I'm not an acoustic specialist and I'm good at sound waveform processing at all, could anyone please give me any idea of how I filter the sound and pick up what I want? Or any recommendation of any website that I can go learn acoustic principles?

Appreciation in advance.

[mikeporter]

This certainly sounds do-able however unless someone has done this before, it will take a lot of experimentation to figure out how to make it work. For example, you will need to  determine the significant bandwidth, and then learn what to look for in the spectral data.

What you will want to do is read the output of some sort of vibration transducer (a microphone probibly won't be the best choice due to the possibility of picking up outside sounds) into LV via a DAQ card and perform a FFT on the data so you can examine the spectral contents. You will also need to learn what the running gear of the drum itself looks like in the frequency domain. Motors, gears, drive chains, bearings and a host of other stuff will all make their own "contributions" to the data.

Mike...

[Kris_worm]

Hi Mike, I really appreciate for your quick reply as I'm having no idea at all of how I should start, at least now I can tell my manager that its not something that we can get done in a week time.

Your post is really really helpful to me, the worst thing is that I haven't even had access to any of these drums, I can only do development upon guess the moment. now it sounds like I have to go do a degree in acoustic somewhere first before I can understand what I'm doing.

Would there be any chance that we can buy a good vibration transducer you mentioned from somewhere, which might do a lot work for us and leave the simple one? otherwise we have to give up then since I'm the only developer working on that this moment.

Thank you so much again.

Kris

[mikeporter]

Sure, there are lots of people making these sorts of sensors. Essentially what you will probibly end up with is a sensor that is attached to the drum in some way and an amplifier. The output of this amplifier is what will get sampled by the DAQ hardware.

However there are also some hardware issues that will need to be addressed - like for instance the drum is rotating. Visualize twisting wires... You will need to either bring the data out through an arrangement of slip rings after amplification or do the digitizing remotely on the drum and access it wirelessly.

How big are these drums in terms of dimensions and weight? How fast do they turn? What is the overall operating environment like (temperature, moisture, dirt, etc)?

Mike...

[Kris_worm]

Mmmm... as I said I haven't even got access to any of these drums, I have no idea of how fast they rotate, but won't be very fast, up to 6 rpm I guess.

Diameter: 7'-16'

Lengths: up to 120'

The overall operating environment can be outdoor, raining, very rough condition. That adds up the difficulty, doesn't it?!

Cheers,

Kris

[mikeporter]

Not necessarily more difficult - just different.

So everything needs to be in a weather-tight enclosures. What will the outdoor ambient temps be summer to winter? Where will these systems be located? Will there be standard power available?

Oh yes, something that I meant to ask earlier, but forgot. Are you looking for loose metal or metal embedded in the tree (i.e. spiked)?

Mike...

[Kris_worm]

Yes, Mike, temp ranges from -15 C to 40 C I guess, the system will be pretty simple as long as it picks up and determine there is something (either stone or metal) inside the drum, it would fire a digital output to client's control to stop the drum and get an operator to find the stone or metal and throw out.

No worry on power source. We were thinking of placing couple mics outside the drum colse to the surface, so any of these mics picks up a stone or metal, we will send signal to stop the drum.

We are not looking for embeded metal, stone or metal will only roll with timber and tap inside the drum which the client doesn't want that to happen. Maybe there is some alternative way of avoiding that.

Mmmm.. anyway, I appreciate for your help.

Kris