Strain gauges or load cells cheap/homebrew

[Jim Weir]

I have a one-off application where I need to measure weights in the 1500-3000
pound (600-1500 kg) range. There are four points of support on the item being
weighed (an automobile, actually) and all support more or less an equal part of
the weight.

I am weighing inside a closed shelter (garage) but it is unheated. The local
environment can be from 30 to 100°F and I'm looking to measure with an accuracy
of 10 pounds or half a percent. I can calibrate the scale nearly to the weight
of my measurement with a known good weight immediately prior to the weighing.

It would seem that strain gauges/load cells would be relatively inexpensive,
especially considering that I can go down to "Q-Mart" and buy a digital bathroom
scale accurate to a pound in 200 for less than $10. It doesn't seem to be so.
Lots of folks sell strain gauges. Lots of folks get an arm and a leg for them.

Any thoughts on how to do a "cheap" strain gauge or load cell? The supporet
metalwork and bending arm don't seem to be a problem. That is a fairly
straightforward calculation. How to get weight into an electrical signal of
some sort is the problem.

The electronics is rather trivial. The sensor is the sticky wicket.

Jim

 

[Spehro Pefhany]

On Tue, 23 Mar 2004 08:04:39 -0800, the renowned Jim Weir
<jim@rst-engr.com> wrote:

I have a one-off application where I need to measure weights in the 1500-3000
pound (600-1500 kg) range. There are four points of support on the item being
weighed (an automobile, actually) and all support more or less an equal part of
the weight.

I am weighing inside a closed shelter (garage) but it is unheated. The local
environment can be from 30 to 100°F and I'm looking to measure with an accuracy
of 10 pounds or half a percent. I can calibrate the scale nearly to the weight
of my measurement with a known good weight immediately prior to the weighing.

It would seem that strain gauges/load cells would be relatively inexpensive,
especially considering that I can go down to "Q-Mart" and buy a digital bathroom
scale accurate to a pound in 200 for less than $10. It doesn't seem to be so.
Lots of folks sell strain gauges. Lots of folks get an arm and a leg for them.

Well, strain gauges are cheap (maybe 5 dollars each). You can always
have at it and try machining your own bending arms (don't forget
mechanical overload protection) and attach the strain guages to them.
The output is pretty low level as well for such a wide ambient range
(you're looking for 50 microvolts accuracy, give or take). Good luck
with it.

Best regards,
Spehro Pefhany
--
"it's the network..." "The Journey is the reward"
speff@interlog.com Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog Info for designers: http://www.speff.com

 

[John Popelish]

Jim Weir wrote:

Any thoughts on how to do a "cheap" strain gauge or load cell? The supporet
metalwork and bending arm don't seem to be a problem. That is a fairly
straightforward calculation. How to get weight into an electrical signal of
some sort is the problem.

The electronics is rather trivial. The sensor is the sticky wicket.

The problem is that there has to be distortion of a spring member that
is measured. Measuring that distortion is hard enough, but having the
spring member distort repeatably with no memory of previous
distortions is the hard part. It takes a very well designed spring to
distort enough to be measured and do it in a better than 0.5%
repeatable way, especially with that temperature range.

--
John Popelish

 

[John Fields]

On Tue, 23 Mar 2004 08:04:39 -0800, Jim Weir <jim@rst-engr.com> wrote:

I have a one-off application where I need to measure weights in the 1500-3000
pound (600-1500 kg) range. There are four points of support on the item being
weighed (an automobile, actually) and all support more or less an equal part of
the weight.

I am weighing inside a closed shelter (garage) but it is unheated. The local
environment can be from 30 to 100°F and I'm looking to measure with an accuracy
of 10 pounds or half a percent. I can calibrate the scale nearly to the weight
of my measurement with a known good weight immediately prior to the weighing.

It would seem that strain gauges/load cells would be relatively inexpensive,
especially considering that I can go down to "Q-Mart" and buy a digital bathroom
scale accurate to a pound in 200 for less than $10. It doesn't seem to be so.
Lots of folks sell strain gauges. Lots of folks get an arm and a leg for them.

Any thoughts on how to do a "cheap" strain gauge or load cell? The supporet
metalwork and bending arm don't seem to be a problem. That is a fairly
straightforward calculation. How to get weight into an electrical signal of
some sort is the problem.

The electronics is rather trivial. The sensor is the sticky wicket.

---
Omega's got some for about $5 each...

http://www.omega.com/toc_asp/frameset.html?book=Pressure&file=STRAIN_GAGES

--
John Fields

 

[MikeM]

Jim Weir wrote:

I have a one-off application where I need to measure weights in the 1500-3000
pound (600-1500 kg) range. There are four points of support on the item being
weighed (an automobile, actually) and all support more or less an equal part of
the weight.

I am weighing inside a closed shelter (garage) but it is unheated. The local
environment can be from 30 to 100°F and I'm looking to measure with an accuracy
of 10 pounds or half a percent. I can calibrate the scale nearly to the weight
of my measurement with a known good weight immediately prior to the weighing.

It would seem that strain gauges/load cells would be relatively inexpensive,
especially considering that I can go down to "Q-Mart" and buy a digital bathroom
scale accurate to a pound in 200 for less than $10. It doesn't seem to be so.
Lots of folks sell strain gauges. Lots of folks get an arm and a leg for them.

Any thoughts on how to do a "cheap" strain gauge or load cell? The supporet
metalwork and bending arm don't seem to be a problem. That is a fairly
straightforward calculation. How to get weight into an electrical signal of
some sort is the problem.

The electronics is rather trivial. The sensor is the sticky wicket.

Jim

Hi Jim,

Stain Gauges are relatively cheap, hard to apply and take special
low noise, low drift, high gain electronics.

Load cells are expensive!

What makes load cells expensive is building mechanical structures which
concentrates the load induced strain along predictable axes, without stressing
the structure so that it takes a permenant set (hysteresis).

look here http://www.entran.com/pl/pllusa.htm


MikeM
(Skylane '1mm)

 

[Tim Wescott]

"Jim Weir" <jim@rst-engr.com> wrote in message
news:ugn060de7aibl3b9et005i9srlle7mdpul@4ax.com...

I have a one-off application where I need to measure weights in the
1500-3000
pound (600-1500 kg) range. There are four points of support on the item
being
weighed (an automobile, actually) and all support more or less an equal
part of
the weight.

I am weighing inside a closed shelter (garage) but it is unheated. The
local
environment can be from 30 to 100°F and I'm looking to measure with an
accuracy
of 10 pounds or half a percent. I can calibrate the scale nearly to the
weight
of my measurement with a known good weight immediately prior to the
weighing.

It would seem that strain gauges/load cells would be relatively
inexpensive,
especially considering that I can go down to "Q-Mart" and buy a digital
bathroom
scale accurate to a pound in 200 for less than $10. It doesn't seem to be
so.
Lots of folks sell strain gauges. Lots of folks get an arm and a leg for
them.

Any thoughts on how to do a "cheap" strain gauge or load cell? The
supporet
metalwork and bending arm don't seem to be a problem. That is a fairly
straightforward calculation. How to get weight into an electrical signal
of
some sort is the problem.

The electronics is rather trivial. The sensor is the sticky wicket.

Jim

4x cheapo bathroom scales.

4x 10:1 lever arms, so 10lb on the pad = 1lb on the bathroom scale.

A calculator.

 

[Roger Hamlett]

"nospam" <nospam@nospam.invalid> wrote in message
news:fhf360hhtde9dc8e5428blpamgb9nchlu4@4ax.com...

Jim Weir <jim@rst-engr.com> wrote:

Any thoughts on how to do a "cheap" strain gauge or load cell? The
supporet
metalwork and bending arm don't seem to be a problem.

Accurate load cells are not easy, that is why they are expensive.

Using load cells is not that easy, they have limited (typically 150%)
overload capability and with their tiny deflection protection with simple
mechanical end stops is not usually an option.

I pulled apart my (more that $10) bathroom scales (something I have been
meaning to do for years). They have two very robust hard steel bending
bars
(////////) with a single element strain gauge (GGG) stuck in the middle of
each.

(fixed pitch font required)

|=================================================================|
V V
/////////////////////////////////////////////////////////////////
A GGG A

=======================================================================


They appear to be robust enough not to bother with overload protection.
They also appear not to bother with temperature compensation unless they
do
a bit with a diode or something on the PCB.

The construction obviously had price in mind.

The display resolution is a little better that 0.5% but no accuracy is
specified and what temperature range do you expect in a bathroom?

Stick on strain gauge elements are cheap enough. I guess if you can zero
and calibrate before you measure you could do something as crude and get
the accuracy you require.
I suspect Jim, might be wanting to weigh an aeroplane. I looked a while ago

at building a simple system based on the designs in the home scales to do
this. At heart it is not difficult, provided you return a simple 'number'
from the strain gauge, and then calibrate with a known load, at the same
temperature. It is possible the home units do 'tweak' their values based on
a simple temperature reading, in the software. The actual gauge elements are
very cheap, and it seemed quite practical. However I then discovered that a
company just down the road from me, made the systems used to weigh racing
cars, and were prepared to lend my one, so the project 'died'. If the demand
is infrequent, hire is not expensive, and they then come with a calibration
certificate, making it much easier to satisfy the aircraft inspector...

Best Wishes

 

[Rene Tschaggelar]

Jim Weir wrote:

I have a one-off application where I need to measure weights in the 1500-3000
pound (600-1500 kg) range. There are four points of support on the item being
weighed (an automobile, actually) and all support more or less an equal part of
the weight.

No. Your 4 point's won't have the same force.

Think again. The 4 wheels won't have the same force either.

I am weighing inside a closed shelter (garage) but it is unheated. The local
environment can be from 30 to 100°F and I'm looking to measure with an accuracy
of 10 pounds or half a percent. I can calibrate the scale nearly to the weight
of my measurement with a known good weight immediately prior to the weighing.

[snip the underestimation]

You need some frame where the car drives on. Then underneath the 4
corners wave your loadcell. The loadcell is some flexing piece of
steel. It shouldn't permanently deform under standard load, including
driving on it as well as off it. There you glue the strain gauge on.
The signal is in the uV range with substantial temperature dependent
offsets. A chopper OpAmp is used to bring the signal up to a decent
range. Zero balance frequently and you're done.
Good luck.

Rene
--
Ing.Buero R.Tschaggelar - http://www.ibrtses.com
& commercial newsgroups - http://www.talkto.net

 

[The Real Andy]

"Rene Tschaggelar" <none@none.net> wrote in message
news:4060b6f5$0$701$5402220f@news.sunrise.ch...

Jim Weir wrote:
I have a one-off application where I need to measure weights in the
1500-3000
pound (600-1500 kg) range. There are four points of support on the item
being
weighed (an automobile, actually) and all support more or less an equal
part of
the weight.

No. Your 4 point's won't have the same force.
Think again. The 4 wheels won't have the same force either.

The 4 points may well be different, but sum them up and you will get a
single weight of the item you are weighing. No platform is required. In
fact, using 3-4 single 'weigh pads' is fairly typical when weighing aircraft
and vehicle.


<snip>

 

[The Real Andy]

<snipped all>

You buy $10 bathroom scales, you get what you pay for, crap. These things
vary from using mechanical levers and optinterruped discs to capacitive load
cells (bad in the bathroom) to genuine load cells. In AU you pay $300-400
for a decent set of personal scales that can be recalibrated and are
accurate, as opposed to the $20 cheapies. I bet if you throw a calibrated
mass onto your bathroom scales they would probably be within a few kg or so,
until they get to the common weight ranges where they are probably 6-8kg+
out.

If you wan't accurate weighing, use load cells. If you are using it for a
backyard app, scratch up on you mech eng maths and eng material science. If
you want cheapish loadcells, loacate you local CAS distributer. I am sure as
shit that CAS would have to be the cheapest L/C manufacturer in the world!!

 

[Roger Hamlett]

"Jim Weir" <jim@rst-engr.com> wrote in message
news:ugn060de7aibl3b9et005i9srlle7mdpul@4ax.com...

I have a one-off application where I need to measure weights in the
1500-3000
pound (600-1500 kg) range. There are four points of support on the item
being
weighed (an automobile, actually) and all support more or less an equal
part of
the weight.

I am weighing inside a closed shelter (garage) but it is unheated. The
local
environment can be from 30 to 100°F and I'm looking to measure with an
accuracy
of 10 pounds or half a percent. I can calibrate the scale nearly to the
weight
of my measurement with a known good weight immediately prior to the
weighing.

It would seem that strain gauges/load cells would be relatively
inexpensive,
especially considering that I can go down to "Q-Mart" and buy a digital
bathroom
scale accurate to a pound in 200 for less than $10. It doesn't seem to be
so.
Lots of folks sell strain gauges. Lots of folks get an arm and a leg for
them.

Any thoughts on how to do a "cheap" strain gauge or load cell? The
supporet
metalwork and bending arm don't seem to be a problem. That is a fairly
straightforward calculation. How to get weight into an electrical signal
of
some sort is the problem.

The electronics is rather trivial. The sensor is the sticky wicket.

Jim
I see I was wrong in thinking about an aircraft, since you say in the

original post that it is an automobile.
Anyway, I'd make some comments. In the UK at least, 'strain gauges' are very
cheap (about $5, in one off quantities). These are the little 'sensor'
elements, designed to be bonded to the surface of the material where you
want to measure strain. Generally, in most designs, you have to use these in
pairs, effectively adjusting a bridge. This immediately doubles the cost.
Then you have to chose the right type for the substrate (aluminium/steel
etc), and special adhesives/preparation are required to get the gauge to
work right. Load cells, are quite 'neat'. They are usually designed so that
they have mechanical stops to limit the maximum strain on the test element,
to prevent complete destruction, but also need a fairly complex
'protection' system, if they are going to survive more normal overload
conditions. They often are directional, with the structure designed so
that they read the applied force largely in one direction only. The cost
of these depends on the design, and maximum load (you would probably
need something like a 500Kg module). These are a lot more expensive
(perhaps $300!...). Have a look at the 'Entran' ELA series though as
one of the cheapest reasonable cells.
It is perhaps worth remembering, that though four sensors is 'nice',
allowing you to take the total weight in one measurement, there is nothing
really to stop you using a single load cell, and making four platforms, all
the same height, with provision to attach the cell, and then just taking the
readings seperately from the four wheels. A bit more work, but cheaper.
There are some much cheaper load sensors, in the piezo-resistive sensors
(though these are generally designed for low forces). These are cheap (<$10
each), and have the advantage of a relatively simple voltage output. Built a
a bridge (so that temperature compensation is provided), and used with a
mechanical beam design, could potentially provide a budget system of good
repeatability (this is more important than absolute 'accuracy', since you
are able to use a calibration close to the measured value).

Best Wishes

 

[Gary Peek]

Jim Weir wrote:

Any thoughts on how to do a "cheap" strain gauge or load cell? The supporet
metalwork and bending arm don't seem to be a problem. That is a fairly
straightforward calculation. How to get weight into an electrical signal of
some sort is the problem.

Jim, Assuming you have more time than money, you may want
to make your own load link, which means gluing strain gauges
onto some metal piece than stretches.

See our resource page:
Collecting Parachute Test Drop Data
at http://www.industrologic.com/cptdd.htm
The link: Constructing your own load links
shows how it can be done.

Our SGAMP-2 may be what you need to amplify it too.


--
Gary Peek mailto:mylastname@mycompanyname.com
Industrologic, Inc. http://www.industrologic.com
Phone: (636) 723-4000 Fax: (636) 724-2288

 

[Fred Bloggs]

Roger Hamlett wrote:

Anyway, I'd make some comments. In the UK at least, 'strain gauges' are very
cheap (about $5, in one off quantities). These are the little 'sensor'
elements, designed to be bonded to the surface of the material where you
want to measure strain. Generally, in most designs, you have to use these in
pairs, effectively adjusting a bridge. This immediately doubles the cost.
Then you have to chose the right type for the substrate (aluminium/steel
etc), and special adhesives/preparation are required to get the gauge to
work right. Load cells, are quite 'neat'. They are usually designed so that
they have mechanical stops to limit the maximum strain on the test element,
to prevent complete destruction, but also need a fairly complex
'protection' system, if they are going to survive more normal overload
conditions. They often are directional, with the structure designed so
that they read the applied force largely in one direction only. The cost
of these depends on the design, and maximum load (you would probably
need something like a 500Kg module). These are a lot more expensive
(perhaps $300!...). Have a look at the 'Entran' ELA series though as
one of the cheapest reasonable cells.
It is perhaps worth remembering, that though four sensors is 'nice',
allowing you to take the total weight in one measurement, there is nothing
really to stop you using a single load cell, and making four platforms, all
the same height, with provision to attach the cell, and then just taking the
readings seperately from the four wheels. A bit more work, but cheaper.
There are some much cheaper load sensors, in the piezo-resistive sensors
(though these are generally designed for low forces). These are cheap (<$10
each), and have the advantage of a relatively simple voltage output. Built a
a bridge (so that temperature compensation is provided), and used with a
mechanical beam design, could potentially provide a budget system of good
repeatability (this is more important than absolute 'accuracy', since you
are able to use a calibration close to the measured value).

The dependence of ultrasonic longitudinal wave velocity upon material
stress looks to be a far more interesting, durable, and accurate
methodology:
http://content.aip.org/RSINAK/v68/i1/102_1.html

 

[Pat Ford]

"The Real Andy" <i_luv_beer_especially_pilsners@yahoo.com.au> wrote in
message news:40616162@dnews.tpgi.com.au...

"Rene Tschaggelar" <none@none.net> wrote in message
news:4060b6f5$0$701$5402220f@news.sunrise.ch...
Jim Weir wrote:
I have a one-off application where I need to measure weights in the
1500-3000
pound (600-1500 kg) range. There are four points of support on the
item
being
weighed (an automobile, actually) and all support more or less an
equal
part of
the weight.

No. Your 4 point's won't have the same force.
Think again. The 4 wheels won't have the same force either.

The 4 points may well be different, but sum them up and you will get a
single weight of the item you are weighing. No platform is required. In
fact, using 3-4 single 'weigh pads' is fairly typical when weighing
aircraft
and vehicle.


snip

The 4 points will be different and that is probly why he wants to weigh the
car. Changing the cg is a common method of tuning a chasis, and knowing
where the cg is requires the 4 weights at the wheels. Brake tunning is
affected by cg height as well so what I'd do is have the 4 scales on a
platform so you can lock the suspension, and tilt the car. The cg height is
then figured out using trig, after you see where the apparent cg moved.

Pat

ps I used 4 bathroom scales and reworked the springs, I didn't care about
accuracy as much as repeatability

 

[Dan Major]

Jim Weir <jim@rst-engr.com> wrote in
news:ugn060de7aibl3b9et005i9srlle7mdpul@4ax.com:

I have a one-off application where I need to measure weights in the
1500-3000 pound (600-1500 kg) range. There are four points of support
on the item being weighed (an automobile, actually) and all support
more or less an equal part of the weight.

...snip...
Any thoughts on how to do a "cheap" strain gauge or load cell? The
supporet metalwork and bending arm don't seem to be a problem. That
is a fairly straightforward calculation. How to get weight into an
electrical signal of some sort is the problem.

The electronics is rather trivial. The sensor is the sticky wicket.

I've made a lot of load cells, or more correctly, I've instrumented

strained elements to measure load. The easiest load cell is one made up of
a single-ended cantilever. Next easiest is a double-ended cantilever. The
math to calculate the strain give the load (stress) and physical dimensions
and material of the cantilever is straightforward. Get ahold of either the
folks at Omega (www.omega.com) or Vishay Micromeasurements
(http://www.vishay.com/brands/measurements_group/guide/guide.htm). Both
sell strain gages for making up your own load cells. Both will provide you
with written references on how to calculate the physical properties (sizes)
of your load cells. Another reference book I would recomend is Beckwith
and Buck "Mechanical Measurements". If you're still having problems after
contacting thtese folks, post back.



--
Sooner dot boomer at gbronline dot com

 

[Jim Weir]

The question was not how to machine the bending arms or how low a level the
signal was. The question was how to make or buy strain gauges for little money.
RTFQ.


Jim




Spehro Pefhany <speffSNIP@interlogDOTyou.knowwhat>
shared these priceless pearls of wisdom:

->
->Well, strain gauges are cheap (maybe 5 dollars each). You can always
->have at it and try machining your own bending arms (don't forget
->mechanical overload protection) and attach the strain guages to them.
->The output is pretty low level as well for such a wide ambient range
->(you're looking for 50 microvolts accuracy, give or take). Good luck
->with it.
->
->Best regards,
->Spehro Pefhany

 

[Jim Weir]

Did you really read the question?

Jim



John Popelish <jpopelish@rica.net>
shared these priceless pearls of wisdom:


->The problem is that there has to be distortion of a spring member that
->is measured. Measuring that distortion is hard enough, but having the
->spring member distort repeatably with no memory of previous
->distortions is the hard part.

 

[Jim Weir]

Thanks, John. FOr whatever reason all my searching didn't turn up these folks.
That's exactly what I asked for and was looking for.

Jim


..
->
->---
->Omega's got some for about $5 each...
->
->http://www.omega.com/toc_asp/frameset.html?book=Pressure&file=STRAIN_GAGES

 

[Jim Weir]

And having done special low-noise, low-drift, and high gain electronics damn
near all my 45 year professional career, I think I can handle the electronics.
The question was how to make or buy a cheap strain gauge.

Jim





MikeM <trashcan@yahoo.com>
shared these priceless pearls of wisdom:

->
->Stain Gauges are relatively cheap, hard to apply and take special
->low noise, low drift, high gain electronics.

 

[Jim Weir]

Now THAT'S clever. THAT's CLEVER. I appreciate the thought, Tim. I'll see if
that isn't a solution.

Jim



"Tim Wescott" <tim@wescottnospamdesign.com>
shared these priceless pearls of wisdom:


->>
->
->4x cheapo bathroom scales.
->
->4x 10:1 lever arms, so 10 lb on the pad = 1 lb on the bathroom scale.
->
->A calculator.
->