Twisted lathe bed - Portass Model S

Hi Steve I wouldn't try to bend the castings as you are likely to cause more problems. If you Can put some pictures on to this site to give us a view on what needs doing.

David

Depends how much the twist is. If it is measured in thousandths, you should be able to bolt it to a strong base and pull it down flat. The welding sounds rather ominous though...

maybe the heat from welding has caused the twist?

can you mill, or gently grind off the steel that has been welded to casting? (odd thing to do to weld together two quite different ferrous materials. (Never had any success when in my youth i tried )

If it is not too bad, it may be possible to cure it when the lathe is bolted down. Lots of us believe in having a bed without twist, to avoid turning tapers inadvertently.

Howard

Steve as David says photos will be very helpful to give an opinion on h ow to proceed also your location is useful there may be a member close by willling to give hands on advice.

As a very long ago Portass S owner I'd say it was a lost cause.

Profile of the bed will make it hard to un twist it correctly, the gap is too weak and it will wander off in all directions rather than simply untwisting the working part if you try to operate via the feet. The headstock having been welded down is probably evidence that someone has tried to adjust things before. I have helped straighten out a lathe bed twisted across the gap. Much harder than expected.

Bed on mine wasn't wonderfully true, albeit nowhere near as bad as yours. Never checked it properly but things were far enough out for the longitudinal feed nut to have worn significantly oval. I always, jokingly, said that the feed screw guided the cross-slide rather than the bed shear on that side. Which wasn't a complete exaggeration.

Probably have to re-machine the bed to sort things any sense. Even if its sorted its still an old cheap, low end lathe not capable of useful work by modern standards. Realistically all you'd gain is the satisfaction from fixing it.

Clive.

True that. On the other hand, the previous owner presumeably was using the machine for something so it might be worth putting it back together as is and clamping it down on a piece of six-inch steel channel etc and see what happens. It might be useable enough for rough work, polishing or whatever. (Or to move along to the next owner!)

As suggested, I would bolt it down, preferably to a large section box section. It will pull down over that length.

I'd stick it in a big bonfire to relieve the stresses (and you of the colour scheme, mercifully), then machine it flat again. You'd have to compensate for the resulting slight vertical offset of the tailstock relative to the headstock though, possibly by shimming under the tailstock.

Not convinced about trying to bolt it down to a steel lump - you'd end up with a strange, highly stressed assembly. Whether or not it reversed the current distortion would remain to be seen.

I have an identical Portass S, even down to the self-inflicted paint scheme. I always took care to clamp it firmly at the main foot and leave the tailstock foot bolts loose - precisely to avoid distortion.

Murray

The main issue with simply bolting to s strong flat base to pull things straight again is that the amount of rectification twist applied along the bed will be inversely proportional to its stiffness. The basic structure is an I beam of varying section so the untwist will run along the neutral axis. Need to consider the leverage effects of applying force via the feet too. The weakest point is in the gap where the thinner centre section of the beam is relatively deep. The bed proper is of varying depth and of asymmetric section due to the wider shears so the neutral axis will be at an angle. The various asymmetries mean that the bed will want to shuffle sideways and upwards a bit too. Not much but significant in machine tool terms.

If the original cause was simply due to bolting down to a very stiff, relative to the lathe bed, flat bench with a twist in it then simply pulling things down to a strong flat plate should, eventually, reverse the effect. During the process the bed will of course be unnaturally stressed but thats hardly likely to affect its function on such a small machine. I think it unlikely that the original bench was simply twisted. Probably some flatness and tilt errors too so its anyones guess whether you can simply reverse the effect.

If its age distortion of an un-stabilised casting then reverse twisting probably won't correct the effect as the mechanism is different and where the error lies has more to do with mass distribution rather than stiffness. Probably need to do as Murray suggests and use a bonfire to remove most of the residual stresses before re-machining.

Proper procedure is to fit a long straight bar in place of the spindle, clamp the the headstock down securely to your surface plate and map out the bed bend both vertically and horizontally relative to the spindle centre-line. Preferably use a thick wall tube rather than a solid bar to minimise droop. Once you have a map you can decide what to do. If it is ageing twist you may be able to correct it on the main bed by clamping two heavy section bars to the ways at headstock and tailstock end and twisting.

Were I to attempt the job I'd simply replace the main bed with a heavy, bevelled edge, plate Hardinge style. Starting by machine the tailstock feet flat and fitting appropriate spacers before doing the main milling job. I'd probably not bother trying to de-stress or straighten first, figuring that its stable how it is, but would check for truth after milling and make appropriate adjustments to the tailstock end feed before a final skim.

More work than simply re-twisting but guaranteed to work and far less futzing about if a simple re-twist doesn't go right first time.

Clive.

Edited By Clive Foster on 15/11/2017 12:19:38

Edited By Clive Foster on 15/11/2017 12:21:27

Edited By Clive Foster on 15/11/2017 13:12:30

Call me unsympathetic or lacking vision, but I would take it down the scrappie and then find another lathe that didn't have a twisted bed!

I have a query.

Have you checked the twist on the ways as well as the twist under the feet? (Your initial statement could be read as either way) If the ways are twisted, you need to care. If it rocks on its feet with no twist to the ways, who cares? You can use a wedge to stop the rocking.

Given the age of the lathe, if it's the ways that have the twist, I'd recommend that it's recycled.

Regards

Richard.

Edited By richardandtracy on 15/11/2017 12:47:17

Steve

You have to map relative to the lathe spindle centre line. Fundamentally all the matters is that the bed is parallel to the spindle in all planes and that the tailstock taper is concentric with the spindle centre at all positions along the bed. Within appropriate tolerances and limits of course as perfection is impossible (unless you are dead lucky).

Everything else is pretty much in the wind and can be shimmed. packed or machined as desired so the beast sits properly on the bench.

Clive.

Steve

Don't know of any references but, bearing in mind I've never done the job myself, a couple or three methods spring to mind.

You do need a accurate bar long enough to reach the end of the bed mounted up in the headstock bearings.

1) Probably the officially correct way is to clamp the headstock end to a surface plate with the tailstock feet (just) clear of the table. Fit a guide to the surface plate so you can slide the dial indicator mount along parallel to the test bar. Find a block close to the centre height of the lathe plus half the bar diameter to sit the indicator base on when measuring off the bar. Record the height of the test bar at suitable intervals whilst sliding the indicator mount along the guide. Repeat on both sides of the bed after removing the block. Lot easier to make the measurements on the bar if you have an elephants foot on the indicator, a disk maybe 1/2" to 3/4" diameter, instead of the usual ball tip. With a ball tip you need to go straight down the middle of the bar so the guide has to be just right. Elephants foot sits on the centre even if you are bit out. Difference in readings for both sides will give you the twist.

2) Clamp the lathe down on a bench by the headstock end so the tailstock feet are clear of the surface. Lightly pack the tailstock feet for stability. Put the indicator in an L shape carrier so that its tip is above the bar centre line when the upright is against the bar. Arrange suitable base so it can be sat on the cross slide and held down with the upright touching the side of the bar. Slide along and record measurements at suitable intervals. Repeat for other side. This way will also give you a good indication of any sideways bend in the bed.

3) Probably the way I'd do it. Arrange the bar so it can be slid though the headstock bearings. May need a bush for the front one if its smaller than the back. Fit the tailstock and measure the relative offsets of the tailstock quill and the test bar are suitable intervals along the bed. Obviously the test bar is slid though the headstock to meet the tailstock. Lots of ways of measuring the offset. Quick'n dirty is to butt the bar up against the end of the tailstock quill and hold a vernier or digital calliper at a shallow angle so it spans the gap. Better is to sort out some sort of overlapping tongue or a pair of pins in suitable clamps.

Method 3 can be lightly modified to give an excellent test for straightness by making a sleeve with different size bores each end one side a sliding fit on the test bar, the other a sliding fit on the tailstock quill. If you can engage the sleeve on both and slide it back and forth things are probably close to right. This is an exceedingly exacting test. Sliding with thou oversize bores at bot ends would probably be better than new.

Clive.

Hi Steve

Just the removing the spindle will do. Essential as the test bar has to slide through the bearings as you move down the bed. Important that there is negligible play or shake.

The nice thing about this particular method is that it gives you the error between spindle axis and tailstock axis directly. Obviously if the tailstock axis and spindle axis are in line then the bed is true, assuming that the tailstock belongs to that particular lathe.

There are numerous ways of measuring the relative offset. If you have a decently sharp tailstock centre you could do a quick and dirty first run through using that and an a finely graduated rule to measure how far out things are. If your eyes are good or you have a magnifier you can probably estimate to better than a 1/2 mm (about 20 thou), maybe even 1/4 mm.

Clive