Using Mach as a camshaft measuring device

[jonny quest]

I'm thinking of building a camshaft measuring test stand, needs to be accurate to .00005" or at worst case .0005".

My original plan was to use a linear digital indicator with a SPC cable going into excel for graphing. But... I do need a rotation means of measurement also for my formula's.

I don't know how to let the rotation record simultaneously... or in sync with the linear.

Any thoughts?

[BR549]

I can tell you how I do it, but it is NOT the best way it can be done. MACH does NOT have good recording capability. It was NOT designed to record data but to create machine movement.

I set up the cam in the 4th axis as A. Next take a hi resolution probe with a flat faced tip the same diameter as the lifter face. Next rotate the axis in steps and do a probe routine and gether both the A axis rotational and Z axis linear height of the lobe. THe end results is a high resoluton porofile of the lobe based on points which in turn can be translated into lines and arcs through additional software steps. Crude but it does work.

IT really is NOT the preferred method to profile a cam.  Best is to use a single rotary encoder for the rotational axis AND a linear encoder for the linear axis then record both into a data base and then process the data from there.

NOW you could bring in the linear encoder thru the encoder inputs and then you would have a continous value stream from the A axis DRO and the Z axis encoder BUT no easy way to do the actuall recording in a precise manner. VB is not that precise in syncing with mach to have a smooth controlled recording stream.


Hope that helps, (;-) TP

[jonny quest]

Thanks, I knew about the probe... but just not accurate enough for my work. Thank you very much for explanation on how you do it!

There's those fancy 2 chanel data aquisition boxes (have to look up names again) that can be had for under $100 with software. I guess that will be my best option... unless there are other options?

[BR549]

EVEN at best you will only get a collection of points not a profile(;-) IF you are serous to the .00005" you need professional equipment. AND it is not cheap.

NOTE: You cannot get repeat .00005 accuracy without a climate controlled room.

BEEN there Done that, (;-) TP

[Chestermarine]

My first question is: Why is the 50 millionths degree of precison necessary, if you are looking at the cam lift/rotation parameters in Excell?  Perhaps you are making this more difficult than necessary.

I have made all manner of cams using 4-axis CNC machines, beginning in 1981. Once the cam is digitized, in a series of angle/displacement readings, you can import the csv file into Excell, and see the cam profile.

When taking/recording the readings, you must be sure to sync the angle value, and the displacement. It may seem simple, but unless you have built a fully automated machine, it is easy to get the number pairs out of sync.  For a simple setup, you can use a dividing head to hold one end of the camshaft, and set up a linear probe with spc output. Usually, a spc output has a two number output line, comma separated. The first one is a line count, and the other is the digital value. If you make note of the the start angle cam position relative to the starting line count, and advance the rotation exactly one degree at a time, you can edit the file when you are finished, and put in the correct angle values in place of the line count. One degree steps in an Excell file will give a very good picture of the cam. You will also have a table of offsets to whatever degree of precision your linear indicator reads to. Doing this by hand, you can digitize a cam lobe in about 45 minutes.

Making a dedicated machine would not be very difficult, and using Mach3 to do it is also not difficult. Mach could turn the rotary table in whatever increments you would choose, and operate a relay to clock the digital probe. The recording would not be done by Mach3, as it would be just stepping the rotary table, pause, clock the linear probe, step again etc.  The linear probe could be sending it's readings to your laptop computer, i.e. a terminal program.


Regards, John
 







   









 

 

[jonny quest]

It's not so much the physical profile that i'm interested in.

Its the inches per degree that gives you the acceleration, and the finer measurement of jerk.

These are very important in making a cam that is as fast as possible... But also easy on valvetrain.

.001" isn't tight enough linear resolution.1 degree may be fine, .5 would be good i think.

[BR549]

IF you are not using a lifter face model to do the measusuring THEN you WILL have to translate the points into a value based on the lifter face not just a indicator point on the lobe.

Hum somethings amiss  you will except 1 deg accuraccy for rotation but need .00005 for the displacement??  your jerk data will be all over the place.

NOT knocking your idea, it is a noble thought to build your own stuff. BUT reality has to be factored in somewhere and a reasonable expectation brought into play. 

Just a thought, (;-) TP