I needed to balance the spindle of my CNC retro fit lathe..
It had particularly bad vibration at 580 rpm
I stuck an accelerometer (same as in a mobile phone) to the side of the head casting
The signal from the accelerometer was fed into one trace on a dual beam scope..
The signal from the spindle synch pulse was fed into the other trace and used as the trigger..
By judging how far along the pulse the peaks and troughs were, I was able to add and remove weight to the spindle to balance the shaft..
By the time I finished, the accelerometer was showing less than 1/100 of the movement when I started.. and it was getting hard to see what was gpoing on as the apindle vibration was getting swamped by vibration from the 3ph motor and interference from the VFD
but all in all very succesful...
video below
sorry about the quality of my commentary, I was getting tired by this time and I wanted to get the machine back together
http://www.youtube.com/watch?v=hLsVMxmxQys]Balancing Lathe Spindle with Accelerometer and 'Scope - YouTube
I love the ingenuity between the application of the accelerometer and the balancing through the application of 'wheel weights', and the positioning based on using the spindle index as a reference.
Just curious - what specific accelerometer did you use? It must be one that gives an analog output. What G range does it handle?
Cheers.. just wish the video was better.. but I had had enough by then.. just wanted to record for posterity !
It was an STMicroelectronics LIS344AL
RS part No. 716-7030 at £2.37 ea.
It goes up to +- 3.5g
I got down to where the vibration was causing less than 0.5mV fluctuation
That is about the limit of the scope i was using, and also the level the backgound noise (mechanical and electrical) was swamping the signal
It is supposed to be sensitive to 0.100*Vdd (3V) per g.
so it was showing less than 0.0015 g at 580 Hz
I don't know how that compares in real terms to a professional machine,
Adding an op-amp, using a vibrationless source to drive the spindle would allow greater sensitivity..
or using a more sensitive accelerometer eg an RS 719-0999 Freescale MMA7361LCT which is 800mV / g almost 3 times more sensitive
ps I found it easier to overcompensate with the lead weights, then shave tiny amounts off with a blade..
its easier to cut a bit off than stick a bit on...