Profile cover photo
Profile photo
Darren Dawes
3 followers -
Product Manager at Adcole Corporation
Product Manager at Adcole Corporation

3 followers
About
Posts

Post has attachment
In case you haven't heard, I have left my employer of 27 years and started something new with my partners, at CGK Gage Group (CGKGage.com). CGKGage offers Crankshaft and Camshaft measurement and reverse engineering. Using a precision cylindrical coordinate measuring machine, we can make sub-micron measurements to ensure your parts are fit for your engine build. Or, we can reverse engineer that mystery cam that you've had lying around the shop. We also sell used Adcole Gages, as well as new and used crankshaft and camshaft measuring machines from other manufacturers. Please stop by CGKGage.com, or contact me on LinkedIn:

https://www.linkedin.com/in/darrendawes

In an effort to be respectful to my former employer, I will no longer be posting about Adcole camshaft or crankshaft gages.

Post has attachment
Adcole’s Chatter Analysis package is a valuable tool, used to diagnose repetitive patterns left on a cylindrical part by the grinding process.  This software uses a Fast Fourier Transform (FFT) to provide amplitude and frequency information about the undulation pattern.  Amplitude results are often in the sub-micron range, with 50nm to 300nm being typical.  Because the undulation pattern occurs over a round profile, its interval is referred to by the term “Undulations Per Revolution” (UPR).  This pattern occurs independent of time, hence the “UPR” label, rather than a “frequency” label.

Determining chatter amplitude accurately requires sensitive measurement equipment.  Unfortunately, equipment that is sensitive enough to detect these chatter patterns can also detect spurious signals that are not present on the part.  Outside vibration sources, or sound pressure, can induce vibration in the gaging system or in the part being measured, which is then detected by the measurement probe.  If this vibration occurs over a tight enough UPR band, it will appear at a specific UPR and look very similar to an actual chatter condition.  Without the proper training, the inspector might assume that these signals are actually present on the part.  Possible noise and vibration sources include (but are not limited to) air conditioning units, pumps, stamping machines, electric motors (including those in the grinder), and even the gage system itself.  These signals may induce vibration directly into the measurement probe, or they may induce an oscillation in the part, at the part’s natural frequency.  On a tactile gage, it is even possible for the probe itself to excite the part to resonance.

Fortunately, there is an easy way to discern the real signals from false signals, by inspecting the part over different time frames.  On an Adcole gage, this means changing the speed of rotation.  The goal is to collect the data set over two different time windows; in order to separate the oscillation of the part or gage system from the pattern that is ground into the part.  Here is a theoretical example for explanation, followed by a real world example:

Assume that we have performed a radial measurement on a shaft and through use of FFT, determined that there is a 150 UPR signal in the measurement results. If this signal was induced by the grinder and actually present in the part, then we should see a signal at 150 UPR regardless of the inspection time frame.

In order to determine if the signal is from real part chatter or some other source, we will vary the time window (headstock speed, on an Adcole gage) during which the data is collected.  For this test we will inspect the part at the speed that initially gave us 150 UPR and then measure it a second time, doubling the time required to collect the data.  This means reducing the speed of rotation of the part in the Adcole gage, by half.  In our test, we might now see that the signal moved to 300 UPR.  As previously mentioned, we know that anything that is actually ground into the part is going to give a result at the same UPR, regardless of inspection speed.  However, after inspecting the part at two different speeds, we found that the UPR moved.  This means that the 150 UPR signal was not part chatter.

We can also determine the frequency of the oscillation caused by the external source.  Here, we need to know the total time required to collect the data.  Continuing our example, let us say that the first measurement window (time for one rotation, in this case) was one second.  If we see a signal at 150 UPR, we can easily translate this to 150 Hz, since 150 oscillations were recorded in one second.  If the longer measurement interval was two seconds, and we detected 300 oscillations, then we still have 150 oscillations per second, or 150 Hz.  Thus, our source of signal pollution is 150 Hz, keeping in mind that we may be detecting the actual external noise frequency, or a sympathetic vibration in the part or gaging system.  If we can locate a source of vibration or sound pressure at 150 Hz, we may find our cause.  But it’s not usually that simple.  Often, the source of vibration causes the subject of measurement (in this case, a shaft) to vibrate at its natural frequency.  Many automotive camshafts oscillate at frequencies that are close to those of AC current (50-60 Hz) causing the mistaken conclusion that an electrical source is the cause of the spurious signal.

Here’s an actual real-world example of part resonance polluting the measurement result, taken from  cam lobe measurement...

Read the rest, on the Adcole.com news page, below.

Post has attachment

Post has attachment

Post has attachment

Post has attachment
Just made another post on LinkedIn.  This one is fun to plug your business name (or your own name) into and see what people are saying about you.

5 Things you should know about Adcole gage correlation for chatter measurements

When comparing results for FFT Chatter Analysis across Adcole gage models, or even gages from different manufacturers, there are many factors that can affect correlation of the results.  Here are the most common things you should look at.

1) Angular Range.  Are you checking enough of the angular range and is the range similar, between gages?  This refers to the angular range of the data being processed, rather than the UPR range set up for the output report.
2) Resolution.  Is the gage resolution similar?  Lower gage resolutions of 360 or 1440 points per revolution can affect results, especially in higher UPR regions.  As of Jan 2015, 3600 pts/rev should be the minimum requirement for your gauging system.
3) Filltering and Analysis method.  Are the same filtering and processing options being used?  Although Chatter Analysis on an Adcole gage disregards all standard filters, there are still options in the chatter analysis routines, that can affect the results.
4) Follower Tip Size.  Is the same follower size being used?  Depending on the UPR of the chatter, it is possible to get different results for significantly different follower sizes.
5) Maintenance of Equipment.  Is the follower tip worn?  A worn spot on the follower tip appears as a much larger radius, or even a flat, to the part.  Typically, this will reduce the amplitude of the signal.

Do you have your own ideas on what affects correlation?  Please post them in the comments.

Post has attachment
One of the more attractive factories I've had a chance to visit.  The Kohler entrance looks more like a building on the Cornell campus, than a factory.
Photo

Post has attachment
3 Things your Inline Gage Company Wants You to Know

1) Temperature correction is a great idea in theory, but difficult to make work in practice.  Where inline gaging is concerned, the part has likely just come from a machining process, then a washer, possibly a blow-off, then into the gage.  The surface temperature is usually not the same as the core temperature.  So, a more sophisticated cool-down model must be fit to the data (a.k.a.  A Fudge Factor).  If the speed of the line, the size of the part buffer, or the temperature of the wash fluid changes, the correction goes out the window.  You can attempt it if you want, but you would be just as well off saving the money and dialing in the nominal for the expected part temperature.

2) The best way to increase your reject rate is to "extend the life of the wash fluid" beyond the manufacturer's recommendation.  The build up of foreign matter in the fluid means that more grinder swarf will end up on the part and be carried to the gage.  This results not only in a higher reject rate, but the need to clean the probe tips more often.

3) Executing a GR&R for size features, in a factory environment without temperature control, is nearly impossible.  A lot depends on the time frame that the GR&R takes place in (and the tolerance), but in general if it takes more than an hour, there must be a temperature stable environment.  Older factories without temperature control may also have fans or air ducts in place that aggravate the problem.  If you DO decide to attempt a GR&R for size, in an environment without temperature control, then winter is your best bet.  Heating tends to keep the plant at a more constant temperature than passive summer cooling.
Photo

Post has attachment
Wait while more posts are being loaded