Re: Strain gauge

[nobody@xxxxxxxxxxxx (Dave Houston)]

About 25 years ago I researched using piezo load cells for another
application but they might have an application here, mounted as the OP
suggested.

I ran the US operations of a French company that made machine tools and I
was looking for a way to sense the load on the machine (hydraulic press that
came in 30-440 ton capacity) for precise process control and to sense a hand
or finger between the dies for safety purposes. I don't recall the name of
the company (Swiss) that made the particular load cells I was interested in
but I made a trip to upstate NY to visit their US office. Their engineers
thought it quite feasible and, to my surprise, were not concerned with the
potential liability issues. I also talked to engineers at Battelle Labs and
some people from DARPA, all of whom thought it feasible. I passed the idea
along to the engineers with my company on a trip to Paris, thinking they
could follow up with the Swiss company from there. They never did, but a
couple of years later, after the US operation was sold to a former
competitor and I was consulting for them, I met with a group of engineers
(with whom they were working) at the University of Minnesota. They also saw
no difficulty in implementing it other than the need for auto-ranging
electronics to deal with the relatively small loads presented by a finger
vs. the potentially large loads presented by the workpiece.

I don't think anything came of it as the parent company (KKR) shut down the
Minnesota plant soon after that. My health had deteriorated to the point
where I had to quit traveling at about that time and had to stop working
altogether a little later.

The sensors were not terribly expensive although, in those days, the
electronics were more costly than they are now. They are certainly sensitive
enough to act as occupancy sensors although noise would be a problem as
would discriminating between a child and a pet.

ddl@danlan.*com (Dan Lanciani) wrote:

>In article <pr2iv11p9jqtkhq18oe8htehuardbp2n3l@xxxxxxx>, sales@xxxxxxxxxxxxxxxxxxxxx (Robert L Bass) writes:
>
>| Thanks, Dan, you're right.  They don't generate a current as I
>| erroneously stated.
>|
>| > It might be possible to do something over a longer period if you
>| > could factor out various non-occupancy-related noise, though
>| > thermal changes in the structure to which the Pulsor is attached
>| > could be a big problem...
>|
>| There's more to it than that.  The resistance change is momentary
>| and occurs when the joist flexes.  It returns to its previous
>| state once the joist stops moving.  In short, they detect a
>| moving load on the floor above.  They do not report the existence
>| of a static load.
>
>No, that is not how they work.  The resistance changes when the
>Pulsor is stressed.  It remains changed as long as the stressing
>force is applied.  The detection of (only) a moving load is entirely
>a function of the processor electronics.
>
>You can see this for yourself by simply connecting a Pulsor to an
>ohm meter.  Apply stress by pushing the middle and pulling on the
>ends.  (This can be done easily by placing your thumb on the middle
>of one side and your first two fingers on either end of the other
>side.  The two "sides" are the one with the potting compound visible
>and the one opposite.)  Observe that the resistance changes and remains
>changed until you release your grip.  Now apply stress in the opposite
>direction by switching sides.  Observe that the resistance changes in
>the opposite direction and again remains changed until you release
>your grip.
>
>| > Given the phenomenal signal processing power that you can
>| > throw at a problem for relatively low cost these days (even
>| > compared to a few years ago) it may be worth looking at.
>|
>| I wouldn't discourage anyone from researching possible solutions
>| to the problem of occupancy sensing.  However, before starting it
>| helps to understand the limitations of the sensors themselves.
>| These devices detect joist movement.
>
>No, Pulsors really are static strain transducers.  It is the processors
>which are in effect "AC coupled" (really they probably use some sort of
>self-balancing bridge, though I may be abusing the terminology slightly).
>
>| A small kerf is cut into
>| the joist and the detector is cemented in place under that cut.
>| When the joist flexes the detector is stretched, causing a
>| momentary electrical change.  It's not like a pressure switch
>| that closes when you stand on it and opens when you leave.
>
>Actually, it is a lot like a pressure switch--at least an analog
>version.  The real problem is that live loads typically produce
>much more joist deflection than static ones.  That may be the
>cause of confusion about how the sensors work.  I can't say whether
>it is possible to pull useful static stress information from the
>inevitable noise, but this would be true for any stress-based
>detection approach.  It is not a limitation of the Pulsor sensors.
>
>| Also note that the sensors can't tell in which direction a load
>| is moving -- into or out of an area.  They only know that a load
>| has been placed upon the floor somewhere within a certain radius.
>| TTBOMK, the transducers don't differentiate between "load placed"
>| and "load removed."
>
>The resistance changes one way when stress is applied and the other
>when it is removed.  The sensors can definitely tell the difference
>between load placed and load removed.  Again, do not be fooled by
>the fact that the processors discard most of the available information
>and give you a momentary motion detect output.  They are merely trying
>to emulate an alarm-style motion detector to make life easy--and they
>do a very good job.
>
>				Dan Lanciani
>				ddl@danlan.*com