Threading issues

[mcbr549]

I cannot cut a decent thread. This has cropped up in the past 6 weeks or so. I have a Grizzly G9972 with an Acorn setup. I have a Calt GHS 38 rotary encoder with a flex coupling mount. I have a GS20X VFD driving a 2hp DC motor. I just did the spindlebenchtest again to make sure it was getting the correct inout for RPMs

I've checked the physical things, no slack in Z axis couplings tight and so on.

I tried it with a program generated in Fusion 360 and one in Intercon, same result

I'm thinking it's in the encoder setup but not sure where to look

Report attached.
Any input will be welcomed.

Mike Canada

[tblough]

For now, stick with Intercon until you get the threading straightened out. One less variable to factor in.

Before we can help debug this issue, it would help to have some photos, or even better a video, of the problem. Trying to guess what is not a decent thread is tough. What makes you think the encoder is the problem? After threading, is Z in the same position as before the threading operation? Is your encoder connected to the spindle 1:1?

[mcbr549]

I didn't video the process but here are a few pictures. It cuts the proper lead in the initial pass, it may take a few passes correct then it misses the lead. The amount it misses seems to be random.
Yes, the encoder is mounted on a one to one ratio. However, the encoder went out ( quit sending info to computer) so I replaced it with a new one, a CALT GHS38 2500p/r. I wired it into the DB9 connector. It seems to be working correctly.
I reran the spindle benchtest program and it accepted the input.

I tried all the options on threading in Fusion 360 like angle infeed, straight infeed, etc. then I programmed it using Intercon....same result.

[tblough]

There are two things that will cause this. Either the encoder index pulse is not repeating at the same point on each revolution of the spindle, or your tool position is changing.

Since you've replaced the encoder and the problem is still happening, then the probable cause is the tool position changing. Once you rule out physical movement, like the toolpost twisting, then you are down to loosing steps on the Z axis.

This is normally caused during the rapid moves after the end of the thread when the cutter repositions to start another pass. Use your feedrate override so slow down the rapid moves and see if the problem goes away. If so, slowing your acceleration down in Setup for the Z axis should fix the problem.

[Dave_C]

There is a possible third reason and this was the one that caused my threading to look and act just like your threads in your photos. Same missed lead-in!

Electrical noise causing false index pulses!

I could not solve the problem until I bought a better encoder and cable from Centroid. Threads have been perfect ever since!

Just saying this might be the cause,

Dave C.

[cnckeith]

and always used closed loop axis motors on a lathe for best results.
they can accelerate harder/faster to keep up with the threading.

[mcbr549]

Thanks to everyone. I have a few things to look into. The original encoder and cable came from Centroid. It quit working a few days ago and I replaced it with the CALT. Maybe I should go back to it.

I'm using a closed loop stepper on the Z axis.

I will recheck the physical side of everything. I have doubts about anything major here as it completely misses the lead, sometimes splits it.

I'm going to go to a timing pulley setup with a serrated belt just to make sure.

I will try slowing down the feed rates as well.

Thanks,

[Dave_C]

I'm going to go to a timing pulley setup with a serrated belt just to make sure.

Hum, did not know that you did not already have a timing belt type setup. If the belt you have slips at all then the lead is going to change for sure.

Dave C

[martyscncgarage]

If your new encoder is 2500ppr, then you should have 10000 set in your spindle encoder count field. For some reason, based on your report it appears to have 3600???

You MUST belt the encoder to the spindle at 1:1 with a timing belt.
The index pulse MUST be precisely predictable.

Many encoders are not designed to take a radial load. They will be damaged. I have a hunch this is what did the Centroid encoder in.
Would like to see a picture of your spindle encoder setup.

Marty