Sampling Rates for Tensiometer/Strain Gage 1

[gdechaine]

I have to measure the tension in the armature of a very large hydraulic piston. The tension can be as high as 250,000 lb over 1/2 sec. Does anyone know if there is a strain/tension sensor that can accurately measure the profile over the 1/2 sec. I have heard it might be difficult to find a tension/strain sensor that can handle a sampling rate in the range of a 1000 Hz. Is anyone familiar with a company that can manufacture this type of sensor or know of a different method to use.

 

[itsmoked]

Can you just use a pressure sensor in the hydraulics and infer the tension? Seem that is the most common method applied. Many pressure transducers are very fast.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[gdechaine]

We can calculate the overall tension that will be seen in the piston using the pressure measurement from the cylinder and volumetric geometry however we still need to get a precise direct measurement of tension/strain on the piston itself due to the overall applied force. The measurement will be taken at several increments along the piston.

 

[BobM3]

Load Cell or Force Transducer (e.g.

http://www.interfaceforce.com)?

The transducers are analog. I assume you are digitizing the output analog voltage? One thousand hertz for a sample rate is slow for most digital sampling hardware.

 

[gdechaine]

Well it would be nice if we could use the load cell instrumentation however there is absolutely no place to install that type of instrumentation. So we are really stuck searching for a Wheatstone bridge type strain gage for the measurement. What I am not sure about is, whether or not there's a sensor like this out there that can accurately record the profile I previously discussed.

 

[itsmoked]

Well, you keep saying [/I][/B]sensor[/I][/B]. Any strain gauge sensor build with a bridge is nearly instantaneous, many, many times faster then you require!

It is the data acquisition system that is at issue. You need to move on to it. The 'sensor' is not the issue.

Any bridge strain gauge that has the range and accuracy you need will suffice. You need to research the system that you want to capture the result with.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[8060A]

I am a scale designer and work with mechanical engineers in the design of precision load cells. What gdechaine is asking for I believe is how to attach strain gages directly to the armature of the piston. The placement of gages for accurate measurement is an art. When the structure gages must be placed on cannot be modified so that strain is linear through the gages, it is very difficult to make quality tension readings. Transducer class gages can be placed in tension and by rotating them 90deg, in compression to get more output out of a wheatstone bridge. I disagree with itsmoked regarding speed. When gages are placed on large metal objects, the frequency response is dependent on the resonant frequency of the object. I generally consider load cells with capacities above 10000 lb to have a resonant frequency no faster than 500-1kHz. As strain factors go down frequency goes up, but since the armature was probably designed for infinite life, the strain factors on any surface are probably quite low, so output from strain gages might be so low as to be in the noise floor.

Bear in mind that even properly designed strain gage based load cells have outputs in the 1-3mV/V range which will require a high quality instrumentation amp to get it up to the level most high speed A/Ds require.

Silicon gages have higher outputs, but have temperature stability problems. Might be good for a unique application.

Low speed high resolution A/Ds are the norm in the scale industry as rarely does the transducer have the speed to justify a SAR type A/D.

Bottom line, I doubt there is a simple solution to your problem. My company routinely makes high accuracy scales with capacities up to 500000 lb, but due to resonance issues, we don't bother making high speed measurements > 100 samples/s. If 50 samples over your 1/2 second interval is enough...

 

[EdDanzer]

FEA will take much of the “art” out of designing a structure to apply a strain gauge to. The sampling rate is a combination of structure and electronics. Your data collection system should have a maximum sampling rate 10X of the actual collection rate to obtain accurate results. External EMF can cause data collection problems. FEA simulation can define the shape and frequency response of the area you install the strain gauge at. The fatigue life of the strain gauge and structural part where the gauge is installed may require a separate armature for testing. If load range is wide and testing period long a heat treated alloy material may be required.

Some reference sites:

http://www.allaboutcircuits.com/vol_1/chpt_9/7.html

http://www.vishay.com/strain-gages/

Ed Danzer

http://www.danzcoinc.com

http://www.dehyds.com

 

[gdechaine]

I have to agree with 8060A. As you have said due to resonance and position issues it is very hard to make an accurate high speed measurements. I need an accurate measurement in the range of 1,000 samples/ second. The application requires high speed and high resolution which as you have said is very difficult to obtain when considering the influence from resonance and the other outside factors such as position placement, and noise. In regards to the FEA we have already completed that analysis however we still need to verify the model with actual measurements from the equipment.

 

[EdDanzer]

If you want to get the best performance out of a strain gauge you need to design an area that deforms the correct amount in the area that the gauge is glued to. This is a separate step required to make a good transducer area in your part. If you looked at the Vishay site there are gauges that can take greater strain allowing for greater sensitivity but it may shorten the gauge life.

This may require reducing the diameter of the armature or drilling a hole in the center axis. Without a cross section picture of the cylinder it is difficult to explain.

Ed Danzer

http://www.danzcoinc.com

http://www.dehyds.com

 

[stockdam]

I don't fully understand the problem.

You want to measure the tension in an armature (I can't picture this).

Any resistive strain gauge will measure at well above 1000 Hz so I don't understand why anyone says that this is a problem. Your only problem is making sure that the sensor is placed in an area where there's enough strain. If the strain is too low then you'd have to introduce a load cell somewhere into the system.

I'd need to understand your application better but at face value it seems to be quite easy.

Another way of looking at it is that if the strain is so low that it's almost unmeasurable then you don't have to worry about fatigue etc in the structure being measured. If you are worried in the strain of another component connected to the piston then you need to strain gauge it.


Send me an email to stockdam@coldmail.com - change the word cold to hot in this address - it's a hotmail.com email address.