Problems with threading

[Pmedic828]

I am having problems threading as the existing threads that I have cut are being cut away on proceeding passes. I chucked a piece of HRS in the lathe chuck which extends about 1 1/2 inches from the face. I took a skim cut to make sure the stock was round. Set up the compound at 29 1/2 degrees, made sure that the cutter was square to the work and chuck. Marked the stock with an indicator mark to see if stock was slipping in in the chuck. Checked the machine for a 20 TPI thread. Zeroed the cross slide, advanced the compound about 3 thou and proceeded to cut a thread thread dial on #1 - everything looks good, thread matches up to thread gauge and rod mounted in lathe has a relief cut to prevent the cutter from crashing into the shoulder. After cutting the 1st thread, cross slide backed out and returned up to beginning of piece, cross slide returned to zero, compound advanced 2 thou, thread dial on #1, engaged half nut, cut same thread deeper - everything looks good, continued this another time with threads tracking -
now for the fun part, after re zeroing the cross slide, advancing the compound another 2 thou, thread dial on #1, engaged half nut and the threads that were already cut was eaten away due to cutter riding on crest of the already cut thread. Checked to see if rod slipped in or out or rotated in chuck - nope. This is the 2nd time I had this happen to me.
Mounted a piece of 1-1/4 already threaded rod left over from a bolt that I cut the hex head off of. Checked to see what pitch the thread was - it was 7 TPI, set threading knobs to 7 TPI, and left cross slide out a bit so cutter would track over the already cut threads. After 3 times, the cutter advanced over the crest of the threads which seems to indicate that the timing or advancement of the cutter is being influenced. it seems to be about 7 1/2 TPI as the cutter advances more than the existing thread - I have no 7 1/2 TPI setting on my lathe?
Does anyone have a comment - will make pics tomorrow if needed, but I don't think it is in the setup - the thread just seems to be increasing and decreasing in the linear distance for a specific thread setting. Comments please.... :thinking:

 

[Tony Wells]

Since this is all geared, about the only change could take place in the areas of keys an shear pins if you tool isn't moving in the post, or compound moving. Is the compound locked? Some cross slides have locking screws. If you never actually take a cut, after a number of "air passes" odds are that you have a gear slipping on a shaft or something.

At least that's my theory and where I would begin to look. There are probably other places to look, but that's where I'd start.

 

[John Hasler]

Is your threading dial slipping?

 

[Pmedic828]

I checked the threading dial, cleaned and remounted it - the compound is locked - the only thing that I noticed when taking the front plate off of the thread selection gears is that the tumbler (pin that locks in the selected hole, seems to be too far over against the next biggest gear - the corner of the next gear is rounded like the tumbler is riding up on this gear sometimes - moved tumbler gear over a slight bit so it tracks on the other gears in the middle - I can't see if this happens since after removing the front plate, i have no ability to keep the tumbler pin locked against the timing gears - I also noticed that I had some swarf under the half nut which I cleaned today - I will attempt to track the cutter while backed off from the cut from a known threaded stud mounted in the chuck = everyone I talk to, says that it is impossible to cut over the next forward crest of the thread but after watching it turning very slow, it seems that the cutter cuts slightly more than the 7 TPI, something like 7 1/2 since it takes about 1 1/2 inches to show the shearing of the thread.
I also engage the half nut on the same number as this is an odd thread number - it slips on 20 TPI also, weather I use the same # on the thread dial or any whole #.
Just a thought, if I never disengage the threading dial and run the carriage backward to start again, after multiple passes, it stays aligned but If I engage and disengage the half nut, it seems to increase it. Just some additional information.

 

[Tony Wells]

Have any end play in the lead screw?

 

[gheumann]

"Just a thought, if I never disengage the threading dial and run the carriage backward to start again, after multiple passes, it stays aligned"

That's a good clue. Something is wrong with your threading dial, OR you're not getting it precisely back in the same "slot". Typically the half nuts can be engaged at many more positions on the threading dial than where the numbers are. You have to anticipate a little and make sure you're re-engaging EXACTLY at the same spot on the dial.

As long as the tool is retracted when you run the lathe in reverse, AND you go far enough past the work to allow all backlash to be taken up before the tool engages the threads for the next pass, you CAN leave the half nuts engaged. And when you do this you're seeing no "creep" in the thread pitch which means the gearing is working properly. What you can't do is reverse the lathe with the tool still in the work thread. There WILL be backlash in the system and the tool won't be in the same place during the reverse cycle.

 

[chuckorlando]

Curious what your tool looks like? Maybe to wide and once deep it cuts over the thread?

 

[Pmedic828]

lead screw play is at minimum but seemed to maybe find the problem - the half nut lever seems not to travel the same distance every time when locking - while watching the half nut, it seems to bind around the lead screw but the lever travels only 2/3 distance most of the time - the only time it repeats is when the lead screw is not turning. Don't know how to adjust throw on the threading lever - placed call to tech support and awaiting reply - sometimes they answer, sometimes they don't - Will follow up and post the findings! Thanks to all who helped!

 

[gheumann]

ALL lead screws (and gear trains) have enough play that if you reverse the lathe without backing the tool out, the tool will NOTf be in the same place on the rearward path. It is the operator's responsibility to work around that by threading in only one direction. Lead screw play is not your issue. Now - if the threads of the lead screw or the half nut are munged up and you can't close the half nut all the way - that's different. But those re big, deep acme threads - I have a hard time believing it was sipping or you would have heard and felt it. But certainly if it isn't adjusted well it can make re-engagement at exactly the right spot difficult and that may well be your problem. Leaving it engaged the entire time and reversing the lathe with the tool backed out is a perfectly acceptable

 

[Tony Wells]

lead screw play is at minimum but seemed to maybe find the problem - the half nut lever seems not to travel the same distance every time when locking - while watching the half nut, it seems to bind around the lead screw but the lever travels only 2/3 distance most of the time - the only time it repeats is when the lead screw is not turning. Don't know how to adjust throw on the threading lever - placed call to tech support and awaiting reply - sometimes they answer, sometimes they don't - Will follow up and post the findings! Thanks to all who helped!

Bingo! That can cause the very problem you describe. The half nuts are grabbing enough of the lead screw to drive the carriage, even though they are not fully engaged. I would want the half nut lever to operate relatively freely. You are (or will) rely on a bit of "feel" so you know when they are properly engaged. If when you engage them, and it doesn't mesh to full depth, it is not going to follow the exact lead from previous cuts. When it feels that way, you should abort that pass. If you run this slow enough, and observe the thread dial, I am thinking that you will see that the index mark is not lined up properly, and not the same as when the half nuts are fully engaged.

Keep us posted!

Edit: I see your other thread where the problem is solved. Grats! I guess my solution is now only a suggestion to look for next time something like this comes up.