I have a hobby CNC machine and want to provide a few useful "system services" to improve my working environment, where such things seem to cost an arm and a leg - so why not build my own? And there comes the rabbit hole...
I have no hardware yet (pointless ordering this week!) but my first planned project is an accurate tachometer for the milling motor (spindle) to compare the actual spindle RPM against the programmed RPM. Those of you with CNCs will understand feeds and speeds, and how important it is to get them right!
IF it proves the (German-built, reliable) spindle isn't doing as it's told, I'm left with a problematic (read "dodgy") PWM Digital to Analog Converter Module, so that might become Project #2, if it's even possible with an Arduino.
The spindle is a Sorotec SFM 1000 PV-ER, which has a digital interface to control the speed of the spindle, based on the voltage output from the converter.
The machine's control board is a Duet3 MB6HC and uses RepRap firmware, The GCode (program) file provides the required spindle speed and this is converted to a voltage by the PWM board, and this output voltage sets the speed on the spindle. The history behind this issue is here, if you have the time to read it!
I was thinking optical, rather than hall effect, as the spindle speeds can reach up to 25,000rpm, although I normally use it between 10-18,000 rpm. I don't necessarily want to incorporate the tachometer within the electronics of the CNC machine, because once I get this sorted it shouldn't need 'sorting' again, but would be nice to have an actual RPM display, rather than what the Duet thinks it's doing.
How familiar am I with Arduino and coding? I'm not, but I have been reading the introductory pages ,
May I suggest the accuracy is a waste of time. Unless you have a spindle motor of at least one HP, the speed will vary quite a few percent and rapidly change from instant to instant. An average over some fraction of a second would be a much better indicator.
It seems like the controller has an idea of its speed and works to keep it constant, however i didnt see that explicitly stated in the docs of your motor.
If youre using collets like these, you can probably read the 3 black slots optically.
If youre running 16000rpms max, youd be expecting ((16000/60)*3) pulses per second
Welcome! You start by giving us a lot more information than the typical first timer, Great Job! I have an old Clausing vertical mill that has more birthdays then myself but it still works great, not that often but great. I have looked at a similar idea and please understand I am not a machinist or anything close.
Counting the collet slots sounds good until you look at the immediate area, all types of crud cuttings lubricant etc get involved. My motor is fixed but there is a set of step pulleys under a hinged cover. As it turns out the step pulleys on the spindle has the largest on top easy to get at. This is the same RPM as the spindle. I finally came to the conclusion I could use a retroreflective sensor and put some white marks on the pulley that the sensor could monitor. Probably 4 but I have not thought this through yet. There is some room in the cover.
The sensor could be monitored by an Arduino with an LCD display. There I could display RPM then RPM filtered (averaged over time) at the same time. Probably 4 different counts quite easily on one display. The code would be easy, power could come from a simple wall wart. Best to mount the electronics in a metal case if possible. In the 4" rectangular junction boxes with covers are cheep and shield well.
The motor is actually 1kW, approximately 1.34hp and is quite precise when the speed is selected manually, but less so when controlled by the PWM to Voltage converter, hence my thoughts that the converter itself may be dodgy. I'm inclined to trust the control board to provide the correct output to the converter and the spindle to respond appropriately to the correct voltage, which only leaves the converter. As everything is already built within the case it's a little tight to stick probes in to check the output voltage of the converter vs programmed speeds from the controller, but it is doable.
Here's a photo of the enclosure (I've got big, mechanic's hands!):
The collets are ER16 standard, (like these, held in a nut like this) so I could put a white mark (tape/correction fluid) on the nut for an optical sensor.
,
