January 2004
In editorials you walk a fine line. We're not trying to be mean or belittle anyone. We're just trying to pass along more of our experiences to keep people from getting ripped off.
Most people would rightly surmise, "How in the world could anyone spend two editorials talking about sensors?" After all, you mount them, put power on them, and see if their lights blink. What more is there to discuss?
The latest marketing technique of taking something complex and calling it a "sensor" gives us plenty to talk about.
A customer bought a new "vision
sensor" and wanted us to give them a quote on installing it. The
customer expected about a day's work because the salesman told them that
"it does not require any programming". You mount it, put 24 VDC
on it, configure it and you're done. So why were we trying to rip them
off and charge for programming?
Over the years we have specified, selected,
wired, installed, or programmed probably hundreds of thousands of
sensors. We cannot think of one that did not require programming. Before
PLCs and computers you didn't have to program them -- because there were no
computers to program. But you still had to wire them into some relay
network -- hence, program them.
Funny thing about sensors. You can put
power on them to make their little lights turn on and off. You can even
mechanically mount them so that they "sense" what they are suppose to
"sense" when they are suppose to "sense" it. But
until you program them into the rest of the system -- what good are they?
They're just sitting out there with their little lights flapping on and off.
How do you tactfully point this out to a
customer, that is already convinced that sensors do not require programming,
has dozens of sensors already on their machine (all of which are programmed), and
not sound like a cocky, arrogant, SOB?
We don't like "intelligent
sensors". Usually what someone did was take an intelligent system,
rip its brains out -- to dumb it down (i.e. limit it's capabilities), and then
slap a "sensor" label on it. Kind of like finding out right
before your operation that your neurosurgeon is really just a taxidermist.
For simple operations, we mean applications, -- they're great. We highly
recommend them. But we can't remember the last time someone gave us an
easy application.
For example, we wouldn't mind a taxidermist
removing a splinter from my finger. But if there is any serious cutting
going on -- we'll stick with a professional.
Our experience has been that
"intelligent sensors" are so limited that it winds up taking longer
to find work-arounds and overcoming the limitations, that if they had bought a
real system, you could have done more for less. Most of these
"intelligent sensors" severely limit your options.
If you are currently sensing some conditions
with multiple sensors -- then these simple "vision sensors" should be
able to replace them. Anything more complex -- we would hesitate to use
them. For example, if you wanted to know how many bolts were installed on
one side of an assembly and all the heads of the bolts were clearly visible in
a side view. This is something you could put a sensor on each bolt head
OR you could use on of the vision sensors for this. Problem is -- normally
an application is more like one bolt on this side, one bolt on the other side,
etc. which would require multiple vision sensors at $5,000 to $10,000
each. In which case regular sensors or a single vision system with
several cameras makes more sense.
Then again, regular sensors are only about
$100. So you would need to sense about 50 to 100 of them to break
even.
Another good application for "vision
sensors" is if a number is put on a part that is easily read (OCR) by
vision. If there are just a few clearly different character codes (such
as E, T, Z, and A) and a high contrast difference between the character and the
background then a vision sensor should be able to handle the job.
There are huge risks involved with
automation projects. Rule #1 of successful systems integration is NEVER
limit your options. You never know what your problems are going to be in
a systems integration project. As soon as you cut a corner -- it will
come back to haunt you -- 100 times over. As a systems integrator, you
might have to throw away a $7,500 "vision sensor" and go buy a real
$20,000 vision system. So you're starting a project $20,000 in the
hole. It is hard to stay in this business if you cut even one
corner. If you aren't careful, a single mistake can cost you your entire
business.
We try to offer a fair and balanced opinion
on every page of our website. We would appreciate more information from
other users to express their opinions, which we will then incorporate. If
you have questions or comments please post them on our message board (see
button in left hand column) so that others can read and benefit.
