Machsupport Forum

Mach Discussion => General Mach Discussion => Topic started by: tubular2000 on January 14, 2009, 01:27:14 PM

Title: thread problem
Post by: tubular2000 on January 14, 2009, 01:27:14 PM
hi all

iM trying to cut a thread on the lathe with not much success. the situation is the following: I need to cut an acme thread with 12 tpi and 6 inch long on a 1.5 inch ss304 shaft. my lathe is equiped with a vfd on the 7.5 hp spindle motor and ac servo motor on x and z axis. i first tried to cut the thread with the 1 pulse/revolution index and the thread wizard. the problem is that the tool doesn't seem to hold its accuracy over a long thread. the thread is good for an inch and gets thinner as the z distance gets larger, after 6 inche there is no more thread on the shaft! the first thing i did was to change from index to timing with a 6 pulses/rev, and do spindle tunning. I then increased the speed from 400 to 700 rpm with lighter cuts. while the thread is cutting i keep an eye on the actual speed dro and it is on the spot and very stable. Any idea what cause the z axis to drift with distance? The starting point of the thread is always perfect. the infeed angle is set to 14 degrees.

one strange thing i've noticed is that on the first pass the tool is cutting air and the x axis is moving slowly in the plus direction over the 6 inchs distance!!
I'm using mach 3 ver. 3.042. it is installed on a newly formated computer with nothing else on it.
Title: Re: thread problem
Post by: Hood on January 14, 2009, 03:53:52 PM
Enable the turn diagnostics plugin then restart Mach and open the plugin from the Plugin Control menu. Do a threading op (just in air) and see if there are any problems reported. If not try again actually cutting and see if its still ok.
 The threading was worked on a few months ago and it now seems perfect for nearly everyone.

Hood
Title: Re: thread problem
Post by: tubular2000 on January 14, 2009, 07:19:16 PM
thanks i will try this tomorow, but did you actually cut any long thread? because i would not have noticed this problem if the lenght of my thread had been shorter then 2 inches.
Title: Re: thread problem
Post by: Hood on January 14, 2009, 07:47:01 PM
Yep, done plenty, below is a pic of four threads, think 2 were  with the parallel port and two with the SmoothStepper but it was a few months back when I was testing things out during the threading upgrades. They are 2mm pitch and about 150mm long if I remember correctly.
Hood
Title: Re: thread problem
Post by: RICH on January 14, 2009, 08:53:09 PM
TUBULAR2000 & HOOD,
Haven't fooled wiith threading for a while and don't remember at what version of 3.042 it was really fixed.
I think 3.042.12 and above.
It was 3.042.012 to the now lockdown of 3.042.20
Hood, do remember?

Tubular2000 if you installed 3.042.20 you should be good to go.
RICH

MODIFIED: STRIKE THRU'S ABOVE
Title: Re: thread problem
Post by: tubular2000 on January 14, 2009, 10:00:51 PM
this threads sure looks better than mine. what spindle speed and first cut depth do you use? my threads cut in about 20 pass.
Title: Re: thread problem
Post by: Hood on January 15, 2009, 03:04:34 AM
Think the spindle speed was 1000rpm for 3 of them and 750 for the other. I had to slow down the spindle speed as I had lowered the kernel speed due to testing and with that lower kernel speed my axis would not have been able to keep up. Not certain on the first pass depth but  it was done in 10 passes(including 1 spring)
Hood

Title: Re: thread problem
Post by: zarzul on January 15, 2009, 10:31:52 AM
Tubular,

You said you watch the X axis moving slowly positive over the 6 inch length.  Is your X axis DRO actually showing it moving out during the threading pass?   It should be rock solid during the pass unless you are doing some sort of a tapered thread.

Arnie
Title: Re: thread problem
Post by: fisherjim on January 15, 2009, 10:38:27 AM
Go back to 1 pulse per rev on the index and add a 2 second delay to the program before starting the threading cycle. This worked for me even though mach3 has a delay built in to the threading. You could also try using a G32 thread as you can then see each cut of the thread in the program.

Jim.
Title: Re: thread problem
Post by: tubular2000 on January 15, 2009, 11:01:05 AM
here is a picture of the finish acme thread. it is not very clear but you can see the tickness of the thread decreasing as the z axis gets further from z zero. in the mean time, i tried with an older version of mach 3 that used to work with no better success. for the x axis this is very strange, on the first pass the x axis is moving out about .03" over the six inch lenght, i can see the motor turning and the x dro moving on the screen. this proble realy sucks, hope someone is going to come up with a good idea
Title: Re: thread problem
Post by: Hood on January 15, 2009, 11:12:07 AM
can you attach your code?
Hood
Title: Re: thread problem
Post by: zarzul on January 15, 2009, 11:25:37 AM
Tubular,

Looks like you may have 2 problems,  maybe they are related.    Are you using a wizard for the Gcode?

The X moving looks like a taper or chamfer command in the threading Gcode command,  don't know how else the x would move during the cut passes.

The z pitch change is very puzzling.  I think I would maybe write a short gcode series to move the z 6 " at a slow feed rate towards the chuck then rapid back to the start and repeat 20 times.  Put a dial indicator on the chuck end
and watch to see if it comes to the right place each time.

Good luck!!

Arnie
Title: Re: thread problem
Post by: Hood on January 15, 2009, 11:43:06 AM
Is the pitch actually changing, kind of looks even in the pics?
Hood
Title: Re: thread problem
Post by: Overloaded on January 15, 2009, 12:51:48 PM
I'd say the pitch MUST be changing. If the X was moving + ONLY, the thread top would be getting THICKER...
Seems so anyway.
RC
Title: Re: thread problem
Post by: Hood on January 15, 2009, 01:26:42 PM
Yes, you are right, brain dead today (even more than normal ) with rewiring the lathe and doing the PLC Ladder.
Hood
Title: Re: thread problem
Post by: Overloaded on January 15, 2009, 01:28:39 PM
You need to get a smaller PLC...one that you don't need a ladder to work on.
RC
Title: Re: thread problem
Post by: Hood on January 15, 2009, 01:30:41 PM
LOL, I would still need a ladder, I am a short arse :)
Hood
Title: Re: thread problem
Post by: RICH on January 15, 2009, 03:04:22 PM
Just some thoughts:
- in config>ports& pins> spindle setup ? have you checked use spindle feedback in Sync Mode for using the index

I would try air cutting and also do just a g32 thread and see if all is good as suggested. If ok,  then try the g76 threading. At least this way it you can eliminate something.
RICH
Title: Re: thread problem
Post by: tubular2000 on January 15, 2009, 10:50:46 PM
ok still have this problem , here is wath i did today

- remove mach3 with all related files, folder and driver from pc
- download and install a fresh copy of ver 3.042.020 plus config
- ran driver test at 35 kHz, everything fine here
- removed the ball screw from the lathe and clean. i tough a chip in the ball nut might give an intermitent slow down on the z axis and damage the thread
- change the electronic gearing on the z axis drive from 10 to 4. this gives me 10 000 ppr, this limits the rapids to 210 ipm but thats fine for wath i do
- ran turn diag. cut a thread in the air, everything fine.
- when everything done, i cut a thread on a piece of scrap, and got a perfect thread

I then loaded a good shaft in the lathe... guess wath same problem!!
after cuting a few more on scrap shaft, i found that the x axis moving problem dissapear. from wath i can see when the thread is cutting, the first 4 or 5 cuts are ok, then on the next one it starts fine then somewhere along the cut, it looks like mach3 stop sending the pulse to the z axis for a breif moment, the axis gets behind a few thousand for the rest of the cut, then the process repeats over the 16 toal cuts!!
i've tried suggestion posted here and still at the same spot!!!

Title: Re: thread problem
Post by: Hood on January 16, 2009, 02:28:20 AM
Have you run the Turn Diagnostics plugin?
Hood
Title: Re: thread problem
Post by: tubular2000 on January 16, 2009, 10:44:55 AM
yes a ran the turn diagnostic, dont see anything wrong there.

now to thread with G32, do i have to write manualy the code or is there another whizard to do it? if i have to do it manualy, how do i tell the program to wait for the index pulse to start the cut?
Title: Re: thread problem
Post by: Hood on January 16, 2009, 10:55:04 AM
If you open the M1076 macro (in your C:\Mach3 folder you cant change the line where it says Test =false  to Test=True an the wizard will write G32 code.
Hood
Title: Re: thread problem
Post by: tubular2000 on January 16, 2009, 08:27:08 PM
the final chapter

i think i get it working for good now. from wath i can see the problem was a timing problem. i dont know how mach3 manage the threading sequence but from wath i can see, if the spindle slows down a little bit while threading because of a change in load, probably mach3 expected an index pulse that is comming a little too late, it then assume that it missed it and realign the z speed accordingly. so i end up with a z axis that is one turn of the spindle too late and eat the thread on the shaft.

i had to cut this thread the same day i upgraded from a previous version of mach3 and assume the upgrade was the problem. so if cutting acme thread use the low gear on the lathe and light cuts because it takes more power then the v thread of the same pitch.

thank you for your help

and by the way, where did you get the info about that g32 thing? is there a secret mach3 bible!
Title: Re: thread problem
Post by: Hood on January 17, 2009, 04:48:06 AM
Not sure how Mach through the parallel port reacts to a slowing spindle, I was under the impression that it would slow the axis to compensate but as I use the SmoothStepper I cant test. Actually I couldnt really test anyway as the spindle on my lathe would t' get affected by even taking a full DOC in a threading op.
  I can confirm however that the SS does react to spindle speed, this was proved to me when a key inside the headstock had sheared and I didnt know, I started a thread and heard the spindle slowing and was about to hit the e-stop but noticed that I heard a change in axis motor speed so let it continue. It did all 10 passes and I did a feedhold and inspected the thread and it was perfect. During the threading the spindle was at 1000rpm at the start of each pass but by the end of each pass it was down to 200 or so rpm and to say the least I was very impressed with the way the SS handled things.
Hood
Title: Re: thread problem
Post by: WoodyCam on February 07, 2009, 03:37:09 AM
Hi there,

I've just had exactly this problem cutting a M14 x 2 thread. As soon as the cut reach sufficient force to slow the spindle by more than 20rpm (from 400) then the next cut would advance the X axis into the work giving a taper to the thread, and the next cut would do the same if I left it. I'd then stop it, rewind the code and run again. It would then do loads more cuts until again the depth was sufficient to slow by more than 20rpm and then the X axis would start advancing in again, making the spindle slow more...

My lathe is small so I ended up with 0.04mm deep cuts and about 40 passes! But got there in the end. However, it shouldn't really do this so I wonder where am I going wrong?

Parallel port operation
No missed steps that I've ever noticed
Mach3 R3.042.020
Single index pulse with very clean signal - never seen any glitches

No problems at all air cutting 40 passes, the macro seems fine
No problem with very very light cuts
Just X axis moving (in the DRO and in practice) in G76 threading when the cut slows the spindle more than 20rpm

So, why would the X axis move if the spindle slows >5%? Its doesn't make sense?

Also, the pitch is perfect. The Z axis copes fine with the change in speed as Hood mentioned. Just that X axis move - I don't get it.

Thanks if you can point me in the right direction, otherwise, threads take soooo many passes!

I'll try the "turn diag" and version ..43.0 when I get the chance.

I also found feed-hold during threading a little hit and miss. Is there guidance when you can and can't use this - obviously not during the cut!

Thanks for your help!

Woody.
Title: Re: thread problem
Post by: RICH on February 07, 2009, 10:34:54 AM
Woody,
I'll fool with threading hopefully today and use V3.042.20. I want to test out the "PUNNY" lathe here anyway.
Can use the PP port or the SS, slow the spindle down, and with new steppers installed and backlash down to
.001 on the Z can now try threading at a variety of spindle speeds probably out to 1300 rpm.

I don't have an answer for you, but, just quickly dry running a few times, motor running and steppers disengaged i was able to see the DRO's change at rpm decreases of less than 5% and really go bonkers if you varied the
rpm at certain times of the thread cycle. I wouldn't put much credibility to the above as it wasn't done in a disciplined way. One thing I did notice though, was, that during the "threading" part of the cycle, if the spindle slowed down,...... the x axis didn't change / stayed at it's position.

Later,
RICH
 
Title: Re: thread problem
Post by: Hood on February 07, 2009, 10:39:26 AM
Are you sure you dont have a taper set for the thread, thats really the only way the X should move while in the threading pass. Can you attach your xml and I will see if I can find anything amiss therel
Hood
Title: Re: thread problem
Post by: WoodyCam on February 09, 2009, 05:00:51 PM
Hi Hood,

Thanks for looking - well, I checked carefully, and taper was set to 0, and its cuts perfectly on very light cuts but anything that slows the spindle more than 5% on a cut then on the next cut, the X-axis moves in slowly over the cut - its totally repeatable for me. I think this is coming from mach3 because when it does this strange behaviour with the X axis, it does not miss any steps at all. It comes right back on the next cut to the right place, then repeats exactly the same behaviour. I'll see if I can make a video of it.

I wonder if its linked to the index pulse timing signal, but the rpm info is perfectly smooth as far as I can see.

I'll dig out my xml and Gcode when it stops raining and I can reach the workshop!

Thanks Rich -grateful for your experiment (how did you do that!??). Yes, it only seems to do it with real cutting on the lathe under load. Its OK air-cutting too.

tubular2000 - what version did you use that worked OK for you in the end and do you think it was just the light cuts that fixed it?

Thanks, really appreciate it.

Woody.
Title: Re: thread problem
Post by: Hood on February 09, 2009, 05:37:31 PM
Do you have the "Use Spindle Feedback in Sync mode" checked on Ports and Pins, Spindle Setup page?
Hood
Title: Re: thread problem
Post by: RICH on February 09, 2009, 07:46:22 PM
WOODY,
I just have a Sherline lathe that is CNC, so with finger pressure or usng the manual spindle rpm adjustment
you can play with slowing the spindle down and see what happens. Don't even think about something like that
with some other lathe. Even a punny lathe can hurt you!

The X axis shouldn't move at all when once the threading starts. I think i am correct in that all variations
in cutting, when spindle slows down, are taken care of in the following Z axis feeds.  It will be a few days before i can experiment in an intelligent way.

RICH



 
Title: Re: thread problem
Post by: keithmech on February 09, 2009, 09:28:25 PM
I have the same problem.My workaround is to add a dwell of a second at the start of each cut.
Kind of like waiting to engage the half nuts.Seems to work ok.I bought another ss for
the lathe but have not had time to install it.
Title: Re: thread problem
Post by: WoodyCam on February 10, 2009, 10:39:41 AM
Hi Hood, yep, that's checked and working - doesn't work without that - its all over the place without that checked. Z axis is near perfect no matter what happens with the spindle speed when checked, so I know that is working well.

Keith, I can't add that delay specifically as I'm using G76 (unless I change the macro??), however, this already has about 1 or 2 secs delay before each cut. Also, I feel the problem starts at the point the spindle slows, not before the start of the next thread cut. Its really counter-intuitive!

I'm starting to think it may be a bug in this version, either of the macro or Mach3...

As I said, everything is fine until the spindle slows more than about 5%. 3% or 4%, no issues but when the cut is sufficiently deep to slow it 5% then the very next cut exhibits this change in X axis, totally repeatably, in the next cut, and the next and the next but with no loss in actual position and no loss in steps. So it must be being commanded under Mach3 control otherwise the X DRO would be wrong, but it is absolutely accurate.

This is with and ORAC lathe and a 1/3hp motor (2000rpm) with VFD down to about 400rpm, so there is not much torque and no closed loop control of spindle speed - that is a future project...! With that in place I think this problem would not show up.  But, it should not be moving the X axis during a cut.

Thanks for your input!

Woody.
Title: Re: thread problem
Post by: Hood on February 10, 2009, 10:57:32 AM
As a test it might be worth getting the macro to output G32 and see if it happens still.
Hood
Title: Re: thread problem
Post by: keithmech on February 12, 2009, 07:13:49 PM
did this get sorted out?
Title: Re: thread problem
Post by: WoodyCam on February 13, 2009, 05:52:18 AM
Not yet Keith, I have some homework to do to give Hood a little more info on my setup and a few experiements. Itching to get into the workshop... Day job just always seems to get in the way!

Woody.
Title: Re: thread problem
Post by: RICH on February 14, 2009, 10:34:16 PM
HI All,
Did some thread cutting on my Punny  ( Sherline ) lathe today.
All the threads were done using the SmoothStepper, 20 TPI  @ 29 deg, varying the depth of cut,
and one case spindle speed also, using the wizard / M76. A single slot disc is used. You don't see a change in the rpm reading
when using the SS but you do see a change in the feed rate. The feedrate varies according to the spindle rpm.
Note the .009" passes made at 150 rpm and a change in feed rate of approx 18%. You could hear the spindle
slowdown and watch the Z change. The pitch dosn't vary as shown in the pictures and frankly not much
difference between the threads. During the 60 pass cuts i varied the spindle rpm more than 10 times  ( i can'y say
 how much because i prefer to watch where my fingers are ) and if you look at the cut / blued trace lines you can't even tell.
The threading works great.

When I  get a chance i will try the equivilent using G32. ( but if i remember it behaves the same way ). Never once did the x axis change movement.

RICH
Title: Re: thread problem
Post by: WoodyCam on February 16, 2009, 04:46:00 AM
Hi Rich, yeah, they look great - thanks for the info. I think your SmoothStepper is doing very nicely. What version of Mach3?

This is very cheeky of me but is there any chance you could try a metric 2mm pitch thread if you get a moment? Could it be a metric thing for me?

I've cut a few 1mm pitch threads, gently, with no issues over the weekend. It only seems to be with the deep, over 1mm pitch threads that problem shows up for me. When I get the time, I'll do more experimentation.

I'm stuck with my trusty parallel port...

I did look at the M76 macro code and that can't vary the X axis during a thread. It also outputs good G32 code when run in diagnostic mode, and the taper "bit" is set to zero. So, I'm still hunting for the reason!

Cheers,

Woody.
Title: Re: thread problem
Post by: RICH on February 16, 2009, 07:33:17 AM
Woody,
A M6X1 is close to a 1/4-20 in terms of total thread depth cut
   M14X2      "          9/16-12             "                                 

I can try the m14x2 but think it's going to be a stretch for the punny lathe ( don't think the spindle motor
has the power to do it.  Maybe I'll try at a small pass like  0.002".

The m6x1 should be no problem though and just to check if it's a metric thing ( but frankly Hood would have
picked on somethnig like that a long time ago ) as i haven't had to do any metrics.

The PP works just as well as the SS. The reason for me using the SS is because my one computer can't be used
with the PP ( the PP pulse signal sucks / and the steppers just won't run correctly).

It's not cheeky to ask at all as i am sure others would gladly do the same.

Back some 7 months ago i spent a lot of time trying to cut thread's on the lathe. Threading was actually broke in MACH. Art spent a lot of time and in particular implemented in M76 and G32 the ability to adjust for spindle
speed problems. In the future ( maybe /whenever that is ) you will be able to use the SS via an encoder with steppers thus the Z would be slaved to the spindle down to a deap stop. Software can only do so much and
is just a comprimise for mechanical ills within reason.

I believe that there will always be threading problems for users. NOT BECAUSE OF MACH, but due to the users
mechanical / machine side of threading. Threading is one time which really shows just how good your overall machine is.Accuracy is important as a thread is sort like a go  no-go gage. Over all it will show up on either side of the spectrum
namely small / fine pitches  and the other end deep. If you can't do in between, you will never do the extremes. That's my nickle on threading.

MODIFIED: RUNNING MACH3 LOCKDOWN VERSION 3.042.20

RICH 
Title: Re: thread problem
Post by: Hood on February 16, 2009, 03:29:49 PM
Its not a Metric problem, I can confirm that as I have only ever cut one non metric thread and my threads come out fine and I have done quite a lot :) Nowadays its all with the SmoothStepper but a lot were done with the parallel port.
 Below is a thread I did a few days ago, its the first one I have done since the servo went on the spindle, its working well :)
They are 18 X 1.5 and 16 X 2 threads, the close up is of the M16
Hood
Title: Re: thread problem
Post by: keithmech on February 16, 2009, 09:40:46 PM
I wonder if the problem is in release 42 20? I struggled today installed the smooth stepper
had a very frustrating time with every thing I touched.Oh well .Then I tried the newest release 43.000
and it seems to work fine.I also found that the c11 board from cnc4pc may have an issue.It would cause the computer
to freeze if I had it connected.I ended up connecting my pulse signal directly to pin 13 on the
ss port 2.As a side note a good idea to check the debounce value for encoders/index on the ss config.
Title: Re: thread problem
Post by: RICH on February 16, 2009, 10:16:05 PM
I don't think 3.042.20 is the problem. In fact I just cleaned off everything and it was a fresh install along with
the SS latest plug in. Could be wrong, but, if there is some poop i usualy have the knack to step in it!
RICH
Title: Re: thread problem
Post by: WoodyCam on February 17, 2009, 06:06:22 AM
Thanks Hood, they look rather nice! I can achieve that but with VERY gentle cuts - 0.03mm diametrical depth per cut at about 400rpm. What depth and speed did you use?

Do you think running SmoothStepper makes any difference to this potential issue?

Thanks Rich, I'm tending to agree at the moment!

Woody.
Title: Re: thread problem
Post by: Hood on February 17, 2009, 06:37:46 AM
I did these threads in 5 or 6 passes but then again my lathe is a bit bigger than yours and Rich's, I could do a thread in one pass (if the tip would stand it ) and the spindle wouldnt slow,  the spindle servo is 12.5Kw.
  I did threads before with the parallel port, heres a  pic of two with the pp and 2 with the SS when I was testing the Beta plugin for threading with the SS. (http://www.machsupport.com/forum/index.php?action=dlattach;topic=8040.0;attach=6447;image)
 The SmoothStepper is definitely more responsive than the PP for threading but the PP  can do very nice threads.

Hood
Title: Re: thread problem
Post by: TrevorH on February 25, 2009, 11:59:58 AM
Has a solution to this problem ever been found?  I am using .020, and it is giving me very similar problems.  If during a cut the spindle slows down very much at all the z speed compensates just fine, but the next pass is cut very unusually.  The next pass will be cut at what appears to be a constant taper with the z axis very slowly moving from a depth that is outside of the thread to the final depth that the pass should have been cut at.  This movement can be seen both physically, and on the DRO.  I have also had a problem where the next pass after a slowed down pass will be completely outside of the thread in the air.  I am not losing steps, and the pitch is always spot on, but the depth is very messed up.  I have run a test piece of code posted elsewhere in this forum that is a g-32 code that has a very shallow thread cut in 80 some passes.  I have run it three times, and the results are perfect, and repeatable. 

I am very confused as to what is going on here.  I don't have a very good understanding of what the g76 is commanding while it cuts at all the diffirent depths to achieve a thread, but something seems to be amiss.  I am pleased to hear that you can change the output from the threading wizzard to a g32, or what I would understand to be a longhand version of the code, I will try that as soon as possible to see what happens.  Thanks
Title: Re: thread problem
Post by: WoodyCam on February 27, 2009, 08:05:41 AM
Hi TrevorH,

I still have the same problem, but like you no missed steps and no errors coming up anywhere in Mach3. But, when the spindle slows more than about 5%, the "X" axis is then messed up in the next cut. Z axis is fine for me. Its 100% repeatable which makes me think it is in the coding somewhere. The solution for me is lighter cuts, with which I can then make perfect threads, even with "normal" spindle speed change.

So, I believe there are now four people posting with a similar issue with this Mach3Turn version. I still have more experiments to do when I get time and I will post them back here.

Looking through the G76 macro, I can't see how that is a problem, all it does is repeatedly call G32 and work out the correct parameters for each call. It can't make any direct X axis movement requests. So, I think there is a problem with the G32 macro with some setups.

However, others do not see this problem, esp with smoothstepper. So it could be a combination of settings or setup which then allow this error to occur.

Hmmm... Any ideas out there? No matter what the setup, the X axis should not move with a non tapering thread!

Cheers,

Woody.
Title: Re: thread problem
Post by: TrevorH on February 27, 2009, 10:01:44 AM
Woody, and others if you haven't allready check out another thread titled"Problems threading on the lathe"  For some reason it seems to have alot more interest, and some great discussion took place yesterday.  Last night I posted the reaults of some test cuts that I did, and I am awaiting feedback and suggestions.  Thanks