Lathe Threading - Spindle & Axis Sync (Resolved)

[hebs]

Good day Centroid community!

Is anyone able to give me a steer please?
I am not getting good repeatable results when threading. I am trying to run M25 x 1.5 external threads and have had approximate (clean-ish) success once out of about 20 attempts (so luck I guess). Each thread has been run between 8-12 passes and seems to 'unsync' on about the third pass and destroy the thread.

I have tried the following with no noticeable effect:
-reduced RPM,
-adjusted parameter 240 to between 0.5 & 4 times the thread pitch,
-checked backlash on the Z,
-checked the encoder mechanics (1:1 on the spindle and no slop),
-tried different g-code G32 vs G76 vs G92,

I have a converted Sieg SX4 (Grizzly 0720R) using ClearPath servos and a Linde optical encoder.

I have attached my report and some G-Code.

Here are links to:

a video showing the cycle fail, https://www.dropbox.com/s/w93iq2l75p4fe ... d.MOV?dl=0

a video of my encoder setup, https://www.dropbox.com/s/z4t5l2xxh0s2y ... s.MOV?dl=0

a couple of photos of results, https://www.dropbox.com/s/2nvgwqbzy99r1 ... HEIC?dl=0 https://www.dropbox.com/s/kkf00uydsfei0 ... .HEIC?dl=0



Can anyone see any issues that I am not?

Many thanks,

Keith Hebs

[Dave_C]

Keith,

I had the same issue! I tried everything in the book but as it turned out, I had noise causing the Index pulse to false register and so the thread would off track ever so often.

I replaced my encoder with a Centroid encoder with a proper cable and threads have been perfect every time from that day forward.

Dave C.

[martyscncgarage]

Do tell us the make and model of your encoder, how you terminated the cable, and show us the wiring in your cabinet.
Encoder cable drain wire should be carried all the way through and soldered to the DB9 connector shell and secured to the mating DB9 on Acorn.

Marty

[martyscncgarage]

From Dana Centroid's R&D and Manufacturing Engineer who has a good grasp of lathe's:
"I have to agree with Dave about noise. It is losing position. I just looked again and see you responded too. You and Dave are on the right track. Looking at the g-code, both are running the G96 -CSS. This may also be the problem also. Have him do G97 - RPM to see if that might be the problem. With the G96 - CSS when pulling away from the part the spindle speed has to slow down and speed back up when coming back down to the starting X. This may be causing it to lose position in X and Z because of the short distance in X. The X and Z can not keep up with the changing spindle speed."

As a test, remove CSS.
Marty

[hebs]

Thanks gents, I will check the wiring and report back. I will also try again without CSS enabled.

[hebs]

Disabling CSS had no effect. I will change the DB9 connecter and solder it properly, It is currently a screw terminal version.

[hebs]

No luck

- I shortened the cable
- I shielded it with a foil tape/wrap
- I double shielded it with a copper braid
- I used metal DB9 connectors - connected to the shield
- The shielding was connected only on the acorn end to the drain (not the encoder end - to avoid any loops),
- I insulated over the top of the shielding with electrical tape
- I protected the whole thing in a nylon braid
- DB9 soldered, each connection insulated with heat shrink
- I checked for shorts
- I even threw on 2 ferrite cores

I still get the same poor results.

The encoder is a Linde RSI 503 - photo attached.

Photo of electrical cabinet is attached - a little messier than normal


I am stumped!


Where can I purchase a centroid encoder and cable? I can't find them on the website.

[tblough]

What is the foil wrap and copper braid connected to? It should be grounded on at least one end.

[cncsnw]

What kind of spindle drive are you using?

Set up a threading cycle to dry-run, or run at a larger diameter than the stock in the spindle. When the program gets to the threading cycle, stop the spindle drive, and start turning the spindle by hand. Does every pass correctly start at the same place?

You will want to program a fairly steep thread lead for this, so it doesn't take forever to hand-turn it through 8-12 passes.

Next, belt a stationary exercise bike to your spindle, so you can test it at higher RPM and with actual cuts

[hebs]

Tom - as previous post: the shielding is connected to the drain of the DB9 connector at the acorn end only. Connecting it at the encoder end as well would create a ground loop.

Marc - The motor is a 1.5kw DC brushless motor (https://www.arceurotrade.co.uk/Catalogu ... less-Motor)

I will have a play at manual speeds - I might even be able to get it as low at 50 rpm with the spindle after that
(the cross trainer will be a bit tricky to belt up to it)

PS - thank you for the material on your website, it has been very helpful.