Threading Problems on Colchester Student

[Stephen OsborneParticipant @stephenosborne30223]

I am having problems with metric threading on my Roundhead Student circa 1956 (a recent acquisition)

I realise that the half nut should be engaged permanently, but my results are appalling.

The first couple of passes seem to be OK but after that, the registering of the thread is way out, typically making cuts between the originals. After each cut, the saddle is returned by reversing the feed direction, leaving the half-nut engaged.

With the half-nut engaged, there is quite a bit of play/backlash on the saddle when attempting to turn the saddle handwheel, also there is a little end-float on the leadscrew during this operation.

Could this be the source of the problem?

I suspect a worn half-nut but have read that a replacement is not only outrageously expensive, but involves a major strip-down of lathe to get at the half-nut.

Does anyone have experience of this problem and it’s solution.

Many thanks in anticipation.

Steve

[Stephen OsborneParticipant @stephenosborne30223]

Erratic registering of thread

[colin hawesParticipant @colinhawes85982]

I don't know the exact design of this machine but the sort of things I would examine are : can the half nuts move? are there missing teeth in the gears? Has the shear pin partly failed? I would not expect end float on the leadscrew to be the cause of your problem. Colin

[HopperParticipant @hopper]

Halfnut slack or even leadscrew endfloat seems unlikely to cause that problem, as all the slack is taken up under cutting conditions.

Are you withdrawing the tool when traversing the carriage back to the start of the thread? Is tool firmly clamped in place and not moving?

If so, probably something amiss in the gear train, or maybe even the job is slipping in the chuck if jaws are worn etc.

[Dave WoottonParticipant @davewootton]

Hi Stephen

Had a similar problem some years ago trying to cut a metric thread on an imperial lathe, the lathe was in a maintenance shop, but somehow they forgot to maintain their own lathe!, from memory it was a very old Kerry.

Exactly the same thing occurred, I was reversing the machine at the end of each pass. when suddenly the thread flank became very thin scrapping the workpiece.

Turned out that reversing the drive to the saddle was causing the half nuts to open, but not enough to lose drive, this meant that the screwcutting drive was almost on the tips of the leadsrew and halfnut threads, because of the angle of the Acme thread leadscrew it was out of register enough to ruin the thread. A strip down and clean out of the apron cured it all, I think it was the first time it had ever seen an oil can. I think the general poor lubrication and the amount of swarf everywhere stiffened up the saddle movements possibly causing the problem.

Might be worth checking the endfloat on the saddle when the halfnuts are engaged before and after the saddle has been reversed a few times, just to compare. From memory it's not a horrendous job to remove a student saddle, you never know a drop of TLC might work wonders.

Hope this helps, good luck.

Dave

[old martParticipant @oldmart]

The Smart & Brown model A that I use has a new leadscrew nut and no endfloat in the leadscrew, but I wouldn't dream of leaving it in gear when reversing. Why? Because there is backlash between the spindle and the leadscrew through the gears.

Its easy to check, just set up for threading, turn the chuck by hand until the saddle moves and then turn the chuck the other way. See how far the chuck can be turned before the saddle starts to move. Using a dti will give the most accurate results. The finer the thread pitch set, the greater the backlash errors will be.

Edited By old mart on 17/06/2019 12:51:36

[Clive FosterParticipant @clivefoster55965]

+1 to what Dave Wootten said. Both his Kerry and your Colchester have single sided half nuts so its easier for the nut to be driven out of engagement from the leadscrew than more conventional double sided "clasp" designs.

Both breeds also have a known history of design revisions in this area suggesting that the system, as originally conceived, was less than ideal.

Half half-nut systems are more common than folk might think. Although they can work well and were attractive to machine builders due considerable cost savings in construction as the actuating and guide set-up can be much simpler the detail design aspects needed to get a system that works well for many years can be subtle.

I once knew a highly skilled machinist with a Kerry in respectable condition who considered verifying full half nut engagement before each threading pass part of the normal operating procedure.  As far as I know he never had an issue but he felt it essential to make sure there wasn't going to be one.  Not gonna argue with a man who considered cutting an acme screw and nut pair with about a tenth thou backlash routine.

Clive

Edited By Clive Foster on 17/06/2019 13:21:08

Edited By Clive Foster on 17/06/2019 13:21:43

[old martParticipant @oldmart]

The S & B also has a single element half nut with a bronze slipper on the other side of the leadscrew. I am always careful engaging the leadscrew nut fully as it can drive the saddle even if not aligned., This happens frequently when using the leadscrew for ordinary turning and doesn't matter in that application. If fully engaged, it has never slipped out of drive, even before making the new nut.

I disengage the leadscrew nut at the end of each pass, withdraw the threading tool and run the saddle back by hand, leaving the spindle running and not reversing. This is easier for me than reversing and withdrawing the tool simultaneously.

[SillyOldDufferModerator @sillyoldduffer]

Posted by old mart on 17/06/2019 13:53:45:

…

I disengage the leadscrew nut at the end of each pass, withdraw the threading tool and run the saddle back by hand, leaving the spindle running and not reversing. This is easier for me than reversing and withdrawing the tool simultaneously.

I take it the S&B only does imperial threads?

Cutting Metric it's often easier, or even essential, to leave the half-nuts engaged. And even on an Imperial lathe keeping the nuts closed avoids any possibility of misreading the thread dial.

Dave

[Nick Taylor 2Participant @nicktaylor2]

You say 'reversing the feed direction'.

Do you mean you are reversing the motor? Or do you mean you are using the stick on the gearbox to reverse the feed direction?

If you are touching the stick on the gearbox then this is your problem, you need to reverse the motor, if you bring any part of the gear train out of mesh then you risk loosing the sync between leadscrew and spindle.

Nick

[old martParticipant @oldmart]

The Smart & Brown model A that I use does imperial and metric threads, I have a 120/127 compound gear for it. I cannot understand why there is any difference between imperial and metric threading, and anyone who cannot read a thread dial shouldn't be using a lathe in the first place.

[SillyOldDufferModerator @sillyoldduffer]

Posted by old mart on 17/06/2019 15:13:26:

…

I cannot understand why there is any difference between imperial and metric threading, and anyone who cannot read a thread dial shouldn't be using a lathe in the first place.

When set-up to cut metric threads an Imperial lathe's spindle and lead-screw turn in the correct ratio because of the conversion gear so that's OK. Unfortunately the thread-chaser dial is driven by an imperial gear engaged with an imperial lead-screw. Using the dial successfully on a metric thread must have been pure luck I think.

Could be serious : depending on who is right, one of us won't ever be allowed to use a lathe again!

Dave

 

 

Edited By SillyOldDuffer on 17/06/2019 16:00:44

[old martParticipant @oldmart]

When cutting metric threads, I would always use the same number on the thread dial.

[Pete RimmerParticipant @peterimmer30576]

Posted by old mart on 17/06/2019 16:20:03:

When cutting metric threads, I would always use the same number on the thread dial.

Makes no difference if the machine is imperial leadscrew.

I think Nick has hit the nail on the head. I bet the OP is using the tumbler to reverse the leadscrew. You can't do that when cutting any thread let alone metric.

[Stephen OsborneParticipant @stephenosborne30223]

Gentlemen

Thank you for all your replies, most of which were very useful. Having come from the aviation industry I am guilty of ignoring the wise advice of my seniors "RTFM" – read the ******* manual.

Nick Taylor 2 wins the prize for pointing out that the only thing that should be reversed is the motor – as it says on page 9 of the manual. I have just cut a very nice 10 x 1.5 thread.

Many thanks to all for taking the trouble to reply

Steve

[BufferParticipant @buffer]

I found a new way to mess up work once. I was engaging the power feed and not the half nuts. I was so angry at not knowing why it wouldn't work I gave up trying to cut a thread in the lathe for about 5 years.

[Alan CharlestonParticipant @alancharleston78882]

Hi Stephen,

You don't need to keep the half nut engaged when cutting a metric thread on an imperial lathe if you have a thread dial fitted. The procedure to follow is:

After setting the tool to cut the correct depth, start the lathe in the forward direction and when the thread dial reads 1, engage the thread nut.

When the tool has advanced to the end of the thread, disengage the half nut and turn the motor off.

Withdraw the tool to clear the work.

Start the lathe in reverse and when the thread dial reads 1 engage the half nut.

Run the lathe in reverse until the tool clears the work.

Turn the lathe off but leave the half nut engaged.

Set the tool to the new depth and start the lathe in the forward direction.

Repeat till the thread is cut to size.

This method is useful because disengaging the half nut stops the tool a lot quicker than waiting for the lathe to run down giving a more accurate stopping point.

I don't know if it makes much difference, but the half nut on my lathe also has a bit of play in it and I always grasp the apron handle when starting a cut to apply a bit of force resisting the movement of the saddle and making sure any backlash is taken out of the half nut – don't get too enthusiastic though, a gentle pressure is what I aim for.

Regards,

Alan

[Clive FosterParticipant @clivefoster55965]

Alan

Interesting idea. Hafta say that I figured out something similar maybe 20 years ago and never convinced myself that it would actually work reliably. So I didn't try. Maybe I should have.

Way I see it its a question of controlling the errors.

Fairly obviously if you always engage the half nut so the drive always picks up at the same position on the bed with the chuck in the same rotational position it will always pick up the non-native thread correctly.

Potential errors are :-

1) Backlash in leadscrew. Running back to the same number on the dial each time locks the engagement point but theoretically actual engagement position in the leadscrew can be anywhere in the backlash region. Reversing back at least takes out all the drive slack. Something I didn't twig when I looked at the idea.

2) Spindle stopping position isn't controlled. Run down time will probably vary a bit as will the time and number of spindle turns between you dropping the half nut and hitting stop.

3) Number of turns you drive the spindle in reverse to pick up number 1 on the dial will probably vary.

Issue 1 probably isn't important in practice but I never saw a neat way to sort issues 2 and 3 reliably. Seems to me that changes in the number of spindle turns in run down and reverse back put an uncontrolled variation in relative rotational position of the spindle and screw for the half nut engagement position. I figured that errors adding up to more than half a leadscrew pitch shift would not be corrected if simply winding back to a bed stop before re-engaging the half nut. Effect would be similar to shifting a tooth in the drive train. Your run back under power to a dial number will clearly reduce errors.

I decided that the only reliable way would be to set a bed stop to define the saddle position when the half-nut was engaged and put a spacer between number 1 jaw on the chuck and the bed to lock the rotational position of the spindle on restart. This would seem to define everything but, obviously, with a random stop position there will be slight shifts of the engagement point along the bed. So long as the error is less than leadscrew pitch it would not matter. Looks like the foolproof technique would be to run lathe doing a dummy thread up to where the thread is to end then run back in reverse to the engagement point or thread dial number and reset the stop. All too much trouble.

Bottom line is of course whether or not such errors actually matter in practice. As it works for you it looks like a fairly major mess up is needed before it all goes wrong. Be interesting to know how much leeway there really is. I reckoned about half a turn of the spindle variation would kybosh things but that was my "wind back to stop" idea not your " finish up by reverse to dial number" technique which obviously takes a lot of potential error out.

Clive

[SillyOldDufferModerator @sillyoldduffer]

Posted by Clive Foster on 18/06/2019 09:26:16:

Alan

Interesting idea. Hafta say that I figured out something similar maybe 20 years ago and never convinced myself that it would actually work reliably. So I didn't try. Maybe I should have.

Way I see it its a question of controlling the errors.

Fairly obviously if you always engage the half nut so the drive always picks up at the same position on the bed with the chuck in the same rotational position it will always pick up the non-native thread correctly.

Potential errors are :-

…

Clive

Thanks to Alan I'm happier it's not just luck that puts the tool back into the thread with this approach.

However I'm glad Clive got in first because he's nailed my reservations about the method better than I could. Thinking about the effect of different metric pitch ratios on an imperial thread dial made my brain overheat! I'm still struggling.

Anyway, does the method work for all metric threads, diameter and pitch, or is it only effective for large diameter coarse threads where the errors don't signify much? (If the latter, it's still a good technique because lathes are rarely used to cut small diameter fine threads.)

Very educational this forum!

Dave

[Alan CharlestonParticipant @alancharleston78882]

Hi,

The idea isn't mine. I picked it up from this YouTube video.

He shows how to do it very clearly. I've used it on a bunch of different threads and it's never failed.

I think the reason it works is that by engaging the half nut on the same thread dial number when reversing, it engages on the screw at the same point as when it disengaged going forward i.e. it's the same as not disengaging the nut but stopping and reversing the lathe.

Regards,

Alan

[HopperParticipant @hopper]

I think with Allan's method outlined above, it is important that the lathe motor be switched off straight away so the thread chaser dial does not rotate by more than one revolution while half nuts are not engaged.

Otherwise, if re-engaging the halfnuts on the number 1, it could be one full revolution of the leadscrew out of synch, which with an imperial leadscrew cutting a metric thread will be out of line with the previously cut thread.

[Author]

Posts