I will call this reply the blind leading the one that can see.
The only spindle stuff i have done is using steppers on my lathe............
Just something i found on the Yahoo group, FWIW, as it's almost a year old.....
RICH
--- In mach1mach2cnc@yahoogroups.com, Steve Blackmore <steve@...> wrote:
>
> On Mon, 09 Mar 2009 00:41:36 -0000, you wrote:
>
> >I am using the latest dev version, downloaded it a few days back. Its very
broken as far as I can see.
>
> Just a thought - There are a few tuning things that you need to do to
> get anywhere like an "as designed" speed response anyway.
>
> The following is taken from many emails testing with Art and assuming
> Mach is set for 25kHz (I've edited out some of the settings peculiar to
> my setup and the cursing
>
> Before doing anything else, delete the linearity.dat file, Mach will
> create a new one.
>
> >Set 1000 steps / unit for the spindle, set the velocity slider 99% of the
> >way up, slide the accel slider so it reads .1 seconds when the first ramp is
> >straightening out on top. Save that. ( it shoudl be about 1000steps/unit,
> >vel 1450, accel 220 , .022G's.. )
> >
> > Set Pulley 1 for a max speed of 20,000 RPM, min speed of zero.
> >
> >Now command a M3S20000 .
> >
> >Let the spindle speed up and measure its actual RPM, ( from the screen or a
> >tach..whatever..
)
> >
> >Now enter that number in the pulley cfg as the max RPM, so it may be 1300RPM
> >or whatever.
> >
> >Now, when you command 1300, youll be getting close to 25Khz,
> >( this is (velocity/60)*accel = freq )
>
> Try the above - it won't do any harm and may improve things.
>
> BTW - restart Mach before testing.
>
> >The difference may be quite large between vfd's in how they react to
> >frequency and ramping.
>
> > Analysing it though, any vfd shoudl be tuned by sending it max frequency,
> >( hence the S20,000 test which will put out 25Khz) and then telling the
> >pulley max what the max frequency actually gives you.. hence the tach or rpm
> >reading at max. ). The code then takes a command like S600, and if the max
> >freq gives 1200RPM, then it seems reasonable that 12.5khz should give 600RPM
> >and thats what mach will do. Now spindle linearity is meant to correct for
> >small discrepancies between a 1:1 relationship and some non-linear responce,
> >but Im wondering if that 8Khz limitation is really making the vfd work too
> >hard at regulating.. Is there any literature on the VFD to say what
> >frequency range it expects or is it justa 0-5 volt analogue in the end?
>
> Explanation here - my VFD seems to max out at 8-10kHz or so!!
>
> > Now run a series of speed tests, say 100,200, 400, 600, 800..ect..
>
> From this we did find a problem peculiar to step/dir spindles that has
> been fixed in a pre-release version but we also know from this and other
> tests that the calibrator is definitely screwed, but the above should
> get you as close as you can get for now.
>
> What I have found so far is the calibrator does not wait (or wait long
> enough) for the spindle speed to settle before commanding the next speed
> step in the calibration sequence and it is on the list for a fix. As you
> can guess, this completely screws up the linearity file. It's
> particularly bad on some high inertia spindles that are slow to respond.
> The object of all of this is to get as linear response as possible,
> without calibration, on a wide a selection of VFD/spindles, then to fix
> calibration, finally to fine tweak PID so it doesn't have to work too
> hard.
>
> Steve Blackmore
> --
>
Guest
