Automated coil spring testing

[mrrcx]

Hello,

What I am trying to accomplish is build and automated coil spring testing machine and run it with labview.

I will explain my thought and goal and what hardware I plan on using, and feel free to tell me if it looks good or not.

Linear Encoder readhead

Linear encoder shaft

DAQ chassis-cDAQ-9184 compact DAQ chassis (4 slot Ethernet)

Bridge Module-NI 9237 4-ch +/- 25mV/V, 24 bit simultanoes bridge module, D-sub

relay module- NI9481 4-channel relay, EM form A, 30 VDC (2 A), 60 VDC (1A)

DIO module-NI 9402 4 channel LVTTL/TTL module

AO Module-NI9263 4-ch, 100 kS/s, 16 bit +/-10V, AO module

fail indicator-red LED

pass indicator-Green LED

DC power supply-24VDC, 60W

37 pin D-sub-D-sub connector for bridge module

DAQ hardware enclosure-box to house DAQ hardware

DAQ hardware subpanel-subpanel aluminum fits 10x8IN

air regulator electronic pressure regulator, 0-10V input, 3-120 psig

mounting adapter-mounting adapter for displacment sensor

I also have a pneumatic ram for compressing the springs

total goals are to use a load cell and an LVDT measurement, then send it to excel so i can crunch the data and statistics

my question is, is the hardware i have listed aboev best?  or suggestions?

Does NI have anything for doing emulation and simulation?

since I do not actually have the hardware i would like to simulate some sort of input acting as the load cell.

I had asked that question on here yesterday and recieved a reply, but i need a more detailed reply since i am far from an expert at labview.

i have most of the GUI setup how i want it, now i just need to connect wires and simulate the generation of data so it can be collected to a csv where it can be worked in excel.

Any hepl is greatly appreciated.

Thank you

-Mrrcx

[RavensFan]

What kind of testing are you doing on the spring?

Is it a fast acting cyclic loading where you would be testing for fatigue?  Or a slow acting single cycle to test for something like an axial stiffness?  What kinds of loads are  you trying to exert on these springs?

How tolerant is your test in the event a load or position exceeds some target value?

I ask because this can help determine whether a cDAQ on a PC is the correct choice for running the test.  If something needs to happen quickly or react precisely to some target load or position, I wouldn't rely on a PC to control it.  I would start looking at hardware such as a compact Fieldpoint or compactRIO with a real-time operating system to give you tighter time control over the process.

Something slow enough, a PC might be just fine.  But I would still look at some hardware safety backups.  What is the risk and damage that could be caused if the test didn't stop when it is supposed to?  I wouldn't rely on a PC to stop a process if the risk of something going to far means a chance of injury or death.  I'd look at having some sort of backup system such as a limit switch properly set to be able to cut off a motion process in the event the PC fails to react when it should.

For simulating the load cell, like you were talking about in your other post, since you are dealing with a linear stiffness, I would just read the position, multiply by a stiffness factor, and output a load in the subVI that simulates the load cell.  You could even had some "noise" to the system by using a random number generator to add or subtract from the value by a few percent so the "load cell" output doesn't look too perfect.

[mrrcx]

Excellent thoughts

the test is relativley slow and the springs are small and only have a few hundred pounds of force.

I need to know:

nominal height

distance to coil bind

spring rate.

so each test will cycle about 5 times max per spring and probably do it at a rate of 1 cycle per second if that

i figured the load cell could be programmed to cut off at a max rate held for a certain time as well as the operating pressure can be limited

[RavensFan]

One cycle per second seems pretty fast to me.  How many data points do you need to collect during that second?  I would think 10's at the minimum, but more like 100's.  The data acquisition would be no problem with that.   You could easily set up a fast enough DAQ rate to collect all of your data from start to finish of the 5 cycles.

My concern would be the start and stop of a cycle if the PC is controlling it.  If your test involves the PC commanding a valve to open to start moving the pneumatic cylinder, and it is going to arrive at its stop point in about a second, how are you going to detect that quickly enough and command the valve close to stop the cylinder?  I think relying on a PC to be guaranteed to respond that quickly (I'd say you'd be looking at a control loop that needs to be 1 Hz or even faster) is asking too much of a PC.  Limiting operating pressure should help, at least it should stop advancing beyond that pressure even it it takes a few seconds for the PC to detect it and respond.  (Remember that a Windows PC is just too likely to decide to go off and run a virus check and steal CPU cycles right at the very second you need it to react to your test.)

Assuming you are able to work through and control issues.  The data you collected ought to be able to easily analyzed once the test is complete to get the stiffness value, and detect when the stiffness appeared to go non-linear when the spring binds up.