RE: Homing problem in Mar 15 build

["Tim Goldstein" <tgoldste-at-mlh-services.com>]

Jon,

You are correct about the process. it has been that way for the last 3 or 4
builds. Previous to that it was just hit the switch and stop. But, as I
mention on this one it is acting quite abnormal and seems to have lost the
repeatability I had in homing since last summer (while the old method was
theoretically not as accurate it was pretty darn close).

While I agree that using the index on an encoder is more accurate, the micro
switch I am using seems to give me +-.0005 which is good enough for me.

Tim
[Denver, CO]

> -----Original Message-----
> From: emc-at-nist.gov [emc-at-nist.gov]On Behalf Of Jon Elson
> Sent: Monday, March 27, 2000 10:20 PM
> To: Multiple recipients of list
> Subject: Re: Homing problem in Mar 15 build
>
>
>
>
>
> Tim Goldstein wrote:
>
> > I just tried actually homing with the Mar 15 build on RedHat
> 5.2 and I am
> > having a problem. once I contact the home switch the axis is
> continuing to
> > travel about .1 " then it reverses direction and backs up a few
> thousands
> > then it starts back in the original direction of the homing at
> a rate of a
> > few steps per second and keeps going for 5 or so seconds.
> >
> > This is unlike homing has been working. Previously the direction would
> > reverse as soon as the homing switch was contacted and then it
> would come at
> > the switch again very slowly and stop the instant the contact closed. No
> > extra movement past the switch closing at all.
> >
> > Anyone else try to home under freqmod with this build? I have
> not tried it
> > with steppermod.
>
> The old homing procedure was to locate the switch, period.
>
> I believe the new procedure (should only be used with encoders with a
> Z (index) channel), is to use the switch to get close to one specific
> pulse from the Z channel, and then locate the first Z pulse after the
> switch closes.  If you do this right, it is possible to set up a home
> position that will precisely locate from either direction, and be accurate
> down to 1 or 2 encoder counts.
>
> Jon
>