Large Lathe Retrofit

for improving our cutting processes is to mount an accelerometer and a vibration sensor on the tool post so that I may measure chatter in order to identify situations that cause vibration or chatter and improve our cutting parameter selection. It would be interesting if I could for instance identify the vibration signature of when an insert wears out or breaks and then possibly output an error message on the user interface, retract the tool post or even stop the job immediately if the case warrants it.

I am interested too in doing this, because i am planning to use the lathe in a production environment (not hobby).
Right now i am testing to measure the motor load in order to know if i can improve the feed or speed in some parts of the program.
I am measuring it througth modbus, because the servo drive have an Rs-485 port (rs232 too) and export it in registers.

Some time ago i was searching some advanced products but they are expensive
www.toolmonitoring.com/produkte.html
www.techna-tool.com/index.php?id=36



A first step could be implemented in LinuxCNC attaching an acelerometer to the A / D input card and showing the values.

A more sophisticated product may be a FPGA card that convert the input in mm/sec (or mm/sec2?) and a program like the showed in youtube (ATAMsystems) running in userspace and connected to HAL.

May be somebody that read this post could help us

Excuseme for my english
VNR

I don't see the need for an fpga, I think a simple setup with an MCU, a fast ADC+AMP, and a accelerometer. The accelerometer could be a cantilever beam with weight at the end and a strain gauge attached to it since then the frequency response would only be limited then to the speed of the ADC. The MCU would poll the ADC through I2C or SPI and then submit the data stream with a timing stream via USB to be analyzed or displayed on the PC motion controller. Since this data wouldn't be used for any type of real time adjustment, a serial over USB interface would work just fine and it would make it very simple to write a little python scrypt to visualize or analyze it. Said script could capture what lines of the g-code are being run and present the vibration profile as a timeline with a reference to the g-codes being run.

Since its all AC coupled data from the accelerometer/microphone, a sound card input would probably be fine...