Hi:
 
I have had some requests for information about our bi-directional, bee
counters.  To date, we have deployed 21 of these units at three sites.
 
At the moment, we are designing and building these from scratch.  We
intend to publish circuit diagrams and software code this fall.  Perhaps
the Bee-L subscribers can recommend an appropriate, peer-reviewed journal.
 
Several authors have published papers concerning a variety of counters
ranging from fairly simple to very sophisticated.  The state-of-the-art
seems to be a commercial bee counter called BeeScan that is sold in
Europe by Lowland Electronics bvba.  We havenot seen or used one; mainly
because it if far too expensive for our research needs. We could afford
one or two, but not 21 counters.
 
The Lowland Counter has 32 entrance portals, each equiped with
bi-directional, micro-gap, infra-red detectors.  Each unit has its own
self-contained computer processor. It is designed for mounting to the front of
a full-size beehive.
 
Based on a published paper and advertising, the Lowland system
appears to be a very nice piece of equipment.  If anyone has used one, we
would be interested in your comments.
 
 
Our units are quite a bit different.  We use inexpensive infra-red
emitters and detectors available at our local electonic parts stores.  Our
counters mount to a nucleus hive - a two story unit 1/2 the size of a single
Dadant deep hive body.
 
Each of our counters has 14 entrance portals.  The electonics are
mounted in a polyethylene porch with an overhanging roof.  Poly is relatively
easy to machine with simple tools and remains dimensionally stable in wet and
dry weather.  However, you need to build some special jigs to meet rather
critical machining and alignment tolerances.
 
Like the Lowland counter, our counters are bi-directional (counting bees
going in and coming out of the hive).
 
Our software uses a very rigorous algorithm.  Because bees do not always go
all the way through an entrance (balking or backing out), our software
only scores a bee if it registers on both of the detectors mounted in the
portal.  That is a very important consideration for count accuracy.
 
Where we differ from all other counting systems is our use of PC
computers and a commercially available digital interface so that we can
monitor several colonies with a single computer processor.
 
Because seven hives at a single location yield 10 MgB of data per day, we need
a fairly good computer, minimally a 75 MgHz 486 with 500 MgB of storage and 8
MgB of memory.  A Pentium machine is even better.
 
The computer has to have an enhanced, bi-directional parallel port.
Serial ports are just too slow, as are most of the commercially available
digital interfaces.  (And watch the manufacturer specifications - many
make exaggerated claims - we bought and sent back several systems).
 
We sample at rates up to 200/second.  Multiple 2 detectors per entrance
portal, 14 portals per counter, and seven to ten counters per site, and
you get an idea of the data flow.
 
The good news is that the electronic components for the counters are
readily available and inexpensive, and many beekeepers already have a
PC.  So, if you are handy, have access to woodworking or metalworking
tools, can solder, and read a circuit diagram - you can build one of
our units.
 
I don't have a ready answer for an inexpensive substitute for the high
speed parallel interface.  It costs as much as the computer.  Because the
interface is modular, it is easily expanded to monitor more hives.  The
cost of monitoring ten hives is only about 25% more than for monitoring
one hive.  The major costs of our system are the computer
and interface, not the counter electronics.
 
Most Bee-L subscribers already have one part of the system - a PC.
 
If you are willing to accept general activity patterns rather than
real-time counts and only need to monitor one or two hives, you can
purchase a simpler A-bus interface.  (The A-bus system can be obtained
for 1/4 the cost of our parallel port interface).
 
Our system is designed to plug into IBM-compatible PCs.  Bruce King, a UM
graduate student, who designed the counter electronics is a MacIntosh fan.
But he couldn't find a digital interface for the Mac that would meet the data
flow requirements.
 
In the near future, I will publish some of our counter results to our Web
site.  In the meantime, I hope this answer some questions.
 
At present, we are hard pressed to make enough to keep up with our
research needs.  We haven't seriously considered manufacturing these
systems.  Common sense suggests that these units are still too expensive
for the average beekeeper or bee club and that even if we could make an
inexpensive system, there wouldn't be much demand.
 
Cheers
 
Jerry J. Bromenshenk
[log in to unmask]
The University of Montana-Missoula
 
http://grizzly.umt.edu/biology/bees