Machsupport Forum

Mach Discussion => General Mach Discussion => Topic started by: Kryten on October 25, 2007, 07:19:55 PM

Title: Urgent - Threading not real time
Post by: Kryten on October 25, 2007, 07:19:55 PM
Hi, I've been threading in Mach 3 Turn ( usually with limited success ) and have finally worked out why it doesn't work properly.
 The index pulse from the spindle could be missing the odd pulse every so often and this makes the true speed display jump around in a range of about 10 RPM ( between 720 & 730 RPM ). Mach 3 grabs any number between 720 & 730 and uses it for the entire cut, even with spindle speed averaging. This causes the machine to cut a different pitch at each pass, which is usually only apparent on long threads. The feed rate should be the same for every pass of the threading cycle. The RPM is calculated before the cut starts so this feature doesn't allow for RPM changes with load ( which shouldn't occur anyway ). I tried it without " use spindle feedback in sync mode " checked, but then you don't get the index pulse to start at the same spindle angle every pass.
 There is no need to calculate the spindle speed between every pass unless you have constant surface speed ( RPM varies with diameter ).
Could you please take an average spindle speed and use it for the entire threading cycle ?
Thanks,
Glen. 
Title: Re: Urgent - Threading not real time
Post by: mick on October 26, 2007, 12:33:15 PM
hello, Glen
              i had the same problem , it was cured by making the slot in the index disc wider.
                                                                        hope this helps,
                                                                                 mick.
Title: Re: Urgent - Threading not real time
Post by: zarzul on October 29, 2007, 10:09:36 AM
Glen,

I use threading a lot.  I originally had some trouble with threading and it was caused by my spindle index pulse having a little noise.  If you can't get a rock solid spindle signal, you are going to have trouble.  After getting the noisy signal cured I have no troubles at all with threading.

Arnie
Title: Re: Urgent - Threading not real time
Post by: keithmech on February 12, 2008, 02:16:41 PM
how can you tell if it is noise?I have my true speed going all over the map.
I connected the scope to the output from the index card and the
signal looks ok?
Title: Re: Urgent - Threading not real time
Post by: Overloaded on February 12, 2008, 05:33:03 PM
Would the "Index Debounce" help here ?
RC
Title: Re: Urgent - Threading not real time
Post by: panaceabeachbum on February 12, 2008, 09:42:29 PM
adjusting the debounce value let mine settle out , I do all my threading at a lower spindle speed (usualy 2-300) also and havent had any problems and I have threaded both ends of a couple of hundred suppressor tubes recently
Title: Re: Urgent - Threading not real time
Post by: DennisCNC on February 12, 2008, 10:32:49 PM
What size is the motor on your lathe?  It might be slowing down when you start cutting and mach doesn't correct in time.  Do you have tachometer on the spindle to make sure the speed is constant?
Title: Re: Urgent - Threading not real time
Post by: keithmech on February 13, 2008, 12:01:00 AM
I have gone from 1-2000 on the debounce in intervals of 5 on the low end
50 as I got past a few hundered.and it had no effect.
Title: Re: Urgent - Threading not real time
Post by: keithmech on February 13, 2008, 12:04:10 AM
The motor is 2hp .it is connected to a vfd.All the signal wires are shielded
with the shields grounded on one end.
Title: Re: Urgent - Threading not real time
Post by: Chip on February 13, 2008, 01:59:12 AM
Hi, Glen

Even if you have a good sharp clean pulse on your scope, Mach needs enough time to see the pulse change state.

On my emco compact 5 lathe the disk is 5 inch diameter with a 3/4 to 7/8 inch slot, good pulse up-to around 3000 rpm.

Mach has allot of things going on, Scanning for this index pulse is one of many.

Last week we had a similar problem with a index pules even with a 1 inch slot it wouldn't work, Turned out that the index pulse was going through an OPTO

coupler that was to slow to see and change it's state fast enough. Wound up using an input pin on the board that went direct to the printer port.

Set your "Index De-bounce" to 500 or so, The 'De-bounce Interval" is for Limit switches and such, It should be set to 500 to 2000, Switches bounce

sometimes, Both help to eliminate/see-through noisy wiring.

What's the disk diameter and Gap ?

Hope this Helps, Chip