HomeRoast Digest

Topic: Acoustic analysis of a roast (5 msgs / 145 lines)
1) From: Clifton Burkett
Just for kicks I set up my mike stand and mike I use for recording =
violin to
record the noise of the hottop while roasting. Very easy to see the =
when processing the resulting waveform. I think it will be interesting =
plot the distribution of snaps for various roasts. Once I figure out how =
install some sort of temperature probe in the hottop and connect it to =
computer it should make an interesting analysis of a roast. Has anyone =
this yet? Any suggestions on a temperature probe setup? 
Clif - SomeWhere in Florida

2) From: Johnny Kent
Clif, what a great idea!
So I wonder if there is a specific sound signature to an excellent roast.
Perhaps you can tell something from the frequency distribution of the
cracks. Muted cracks will have one signature and sharp cracks another.
Around 01:05 AM 1/18/2004 -0500, Clif - SomeWhere in Florida  scribed:

3) From: The Scarlet Wombat
It would be interesting to do a FFT analysis of the waveform and compare it 
with the moisture content and variety/batch of the beans.  One might gain 
some insights regarding different profiling methods.

4) From: john roberts
I have speculated some on the utility of using a microphone in roasting
control and it seems to me that we should be able to come up with a
signature for the two cracks and perhaps even bean movement (for control of
fan speed). Beyond that the "sound" of the cracks will probably be dominated
by size and perhaps other physical characteristics of the beans that may be
difficult to parse out other utility than that cracks are happening. Most
inexpensive microphones use plastic and wouldn't tolerate high temperatures
very well but I'm sure that could be managed.
I have also speculated about use of optical (color) feedback for degree of
roast, and smoke/CO/chemical sensors, for possible use in controlling the
process. Of course one would also have to know what to do with such
information once collected.
I see crack detection as a pretty obvious metric for managing hands off
the other detectors are just food for thought at this time.
I am in the middle of developing an embedded micro product design (with a
microphone connected) but haven't done more than speculate about controlling
my Gourmet (still working). The fan noise may make parsing out the cracks a
little harder but if our ear can do we should be able to get a micro to. I
haven't spent any development time on this, since I'm pretty committed to my
real new product work, but at some point in the future if somebody else
doesn't do it first.

5) From: Ed Needham
I can easily guess how it will look, but I can't wait to see it plotted
together for reference.
Great idea!
Ed Needham
To Absurdity and Beyond!
homeroaster ... d.o.t ... com

HomeRoast Digest