Lead screw or thread dial not right?

Dabbler said:
I know you guys are talking backlash, but the machine is brand new, with replacement parts. so no wear on the half-nut. the gears should all be meshing correctly, so backlash from traditional sources should be minimal. When I do the forward/reverse test on my 41 year old 12"Taiwanese lathe or my 40 year old 15" LeBlond lathe, you cannot tell the difference in the forward or reverse scratch impressions.

The forward/reverse test is only to test the half-nut engagement, nothing else.
Click to expand...
Totally agree with you on wear as the machine is new. Not completely sure about the rebuild, things could have gone awry there, due to an error or some alignment of all the gear shafts, or there's still some hidden damage present. It's still detective work at this point... Best we can do is suggest some diagnostic tests that might give experienced people like you a clue.

Pardon my ignorant questions, but if the lead screw pin was broken, or the lead screw can move along its axis, couldn't that give different cuts in the work piece, or yield inconsistencies? I mean, something's wrong, so we can't assume everything is working perfectly.
 
Threading forward and then back is not a valid test of anything. There IS backlash in a brand new lathe. Gears in particular NEED backlash, there is probably some backlash just about everywhere, and there are a lot of steps along the way from beginning to end. That will give the results here, by design. It's an invalid technique for threading because the position of the is NOT repeatable when switching directions.

Backlash is NOT intermittent, and is not related to the dial position. When the carriage is moving forward, it's all taken up. When the cariage is backed up (properly) and run again, all of the backlash is again taken up. Then the position of the tool is repeatable. The results of this stopping and backing up without withdrawing test demonstrates a properly working lathe.

WobblyHand is quite correct, "something" is obviously not right, and you can't "assume" anything. What needs to be done though is to apply a logical and systematic thought process to the whole situation.

The lathe WILL cut perfectly good threads, and repeatably, provided that the dial is ALWAYS engaged on the number 1. That right there tells you that every gear in the gear train is proven for proper and continuous engagement, and the lead screw are PROVEN to be repeatable and consistant, although it's thread pitch and pitch accuracy are still up in the air. The half nuts are repeating their position and and engaging correctly every time, so long as the dial is on the same number... There is ONE thread on the leadscrew. They can't engage wrong based on the dial position. What we know right now is that something goes wrong ONLY and ALWAYS when starting on a DIFFERENT number that SHOULD be valid. The whole power flow through the machine has been proven to make presentable threads. The cross slide, compound, and carriage ways and gibs have been proven to be quite adequate to cut presentable threads. Adding drag to the carriage lock has already been proven not to be necessary to cut presentable threads. These are not places to chase this problem.

What has been proved this far? Something IS wrong, that's proven. But an awful lot of the lathe it's self have already proven to be NOT wrong. The gearing and lead screw and the half nuts have been proven to repeat every time, IF the same number is chosen every time. The actual -measured- thread pitch of the thread being cut, has that been 100 percent proven yet? The condition and tooth count of the threading dial gear, has that been absolutely proven and thoroughly inspected yet? Nothing is amiss with the lathe's ability to cut threads, something is amiss with the timing of the starting point of the cut. There are limited items that could affect ONLY that aspect of threading.
 
@Jake M is saying what I've been saying. I just cut a thread on my PM-1340GT (no wear) which has an ELS and this is a 10 TPI thread @ 100 RPM and I ran it in 1.9" and just reversed the lathe as it has a VFD, the difference in the lines is the backlash. You fellows who don't have any backlash sure can save some time (if you don't have a threaded spindle) as you can thread in and out and do it in half the time.

IMG_3759.jpg

Turned off previous passes, adjusted Gibs on half-nut (really too tight) and made another pass.


IMG_3760.jpg


@jareese When you installed your thread dial, did you have half-nut engaged, then sync the dial on a whole number?
 
Last edited:
@xr650rRider No one is saying that one can thread forwards and backwards... All systems have some backlash and some worse than others. I have an ELS on my lathe, might be interesting to try this. I suspect it would look a bit like the picture in #93. In my case, I'd guess it was primarily due to the lead screw and carriage interface. The test was more to bound the problem, not a suggestion of how to run a lathe.

As @Jake M stated, something's not right, but a lot of things are most likely correct. It's a fault isolation problem, where we are trying to make tests that will isolate to a section, and eventually to a faulty assembly or component. There's a couple of ways to do this, we could assume that everything is perfect and find where it deviates, or we could say its not working and work backwards from that point, imagining if the components were good, what would we see. If the behavior is unexpected at that point, it may be the fault. Either way is valid, although one approach may bear fruit earlier. Beats me which approach it is at this point.

If we can think of some more tests to try, then maybe we can solve this puzzle.
 
That comment about threading both directions was sarcasm. But if you don't see any backlash then it's possible. An ELS can't eliminate slop in a mechanical system. The encoder just doesn't move until the backlash is taken up, then the servo turns the gearbox but again it has to take up the slack before any motion is seen.

@jareese probably feels like he's on a snipe hunt when he does a test and told that's not correct.

My last question from above.
@jareese When you installed your thread dial, did you have half-nut engaged, then sync the dial on a whole number?
 
xr650rRider said:
That comment about threading both directions was sarcasm. But if you don't see any backlash then it's possible. An ELS can't eliminate slop in a mechanical system. The encoder just doesn't move until the backlash is taken up, then the servo turns the gearbox but again it has to take up the slack before any motion is seen.

@jareese probably feels like he's on a snipe hunt when he does a test and told that's not correct.

My last question from above.
@jareese When you installed your thread dial, did you have half-nut engaged, then sync the dial on a whole number?
Click to expand...
Our friend @jareese has been having one heck of a difficult time getting this poor lathe back up and running. It must be terribly frustrating. I definitely get the reference to snipe hunting, my friends tried to catch me on that one, long, long ago.

Have to say, that's a good question about the dial.
 
Check the tooth count on the threading dial. Someone here will know what it should be, just not Me.
 
He indicated that he did not replace the threading dial, the standard gear for imperial threading on my lathe with a 4TPI leadscrew is a 16T. I would remove the threading indicator, check the gear for damage and also put a dial gauge on the gear hub/shaft and check the TIR. It is very possible that something could have been damaged given that he needed to replace the leadscrew. If the thread dial checks out OK then the next step would most likely be to check damage to the half nuts. On higher TPI leadscrews I have seen issues with half nut engagement just before or after a line causing a similar issue. The issue of threading misalignment only occurs when he does not restaart the threading on the same number/mark, it would imply that this is not a movement of the leadscrew if the misalignment does not occur with the same number.

I have fitted a proximity stop system to the 1440EV lathes, in that case you no longer use the threading dial, the half-nut stays engaged until the thread is completed.
 
Sorry if I have missed this but are the teeth on the threading dial gear an even number?
 
WobblyHand said:
Our friend @jareese has been having one heck of a difficult time getting this poor lathe back up and running. It must be terribly frustrating. I definitely get the reference to snipe hunting, my friends tried to catch me on that one, long, long ago.

Have to say, that's a good question about the dial.
Click to expand...

At this point I’d be tempted to flop it on its back and see if that helps.
 
You must log in or register to reply here.
Back
Top