Lathe Bed Restoration

[4GSR]

it has to be like a Rockford planer each pass is the width of the cutting tool

Not necessarily.
That's ideal if you have a real planer to mount the bed down to and plane. Small cuts about .025 to .030" per pass is sufficient. Depth of cut just enough to get 100% cleanup. This may require making several passes to do. The problem doing this method is, you have to come back and scrape out most of the tool marks and establish your bearing points per square inch. This is where a scraped cast iron straight edge comes in handy. The straight edge does on have to be the length of the bed for this method. Just long enough to make good impressions and pay attention to the markings so you don't loose coplanarity of the surfaces involved.

 

[tq60]

That Atlas is not worth "real money" or time required to do what has been discussed but the top can look ugly and can be cleaned up relatively cheaply by stepping outside the usual places.

The lathe is about the same size as an auto engine head.

So check arround for engine machine shops as there are still some arround that have surface grinders for truing heads.

After locating some take the bed only to them and discuss your needs so they are clear to only remove enough material to have a consistent surface and also that the surface is parallel to the bottom end to end.

Have you checked this yourself to insure the bed is not bent?

Check thickness of bed and it should match your wear indication but if same thickness end to end the bad could be warped.


Sent from my SAMSUNG-SGH-I337Z using Tapatalk

 

[kdtop]

Have you considered planing the bed? I am not familiar with the Atlas bed but possibly you could use the unworn inside vertical sections of the bed as a reference and build a sled to slide on these surfaces.

I recently planed my south bend v-way bed using a sled made with the tailstock base, sliding on the relatively unworn tailstock v-way with excellent results. I only did the front v-way since that was where most of the wear was.

The whole operation took maybe 20 mins, after the sled was fabricated. With the flat ways you could get by without the compound on the sled. Stone or lightly file after the planing to debur and scrape in some 1/2 moons to complete.

Here's a video of the process (not mine):

I have been thinking more about your sled option. I can't see for sure, but is your sled riding on a dovetail below the plane of that top surface? In my situation (see crude diagram posted before), I have inner tailstock ways, and outer carriage ways. So I can't plane down the entire surface flat because part of that surface would be supporting the sled.

But perhaps I could leave the inner ways at original height (preserving the support for the sled), and just use a planing action like you show to drop the outer ways down to a consistent depth -- say 0.010". This would accentuate the step that is already there, but should be fine.

KT

 

[martik777]

Sorry I missed your diagram and post #10 from: https://www.hobby-machinist.com/threads/repair-worn-1930s-craftsman-atlas-ways.91748/

My sled is riding on the a flat portion that is slightly lower than the tops of the 2 outer carriage V-ways and on an inner v-way that is the same height. I think you only need to plane the top ways "C" as the other parts, B and A will have very minimal wear. You can check this with a straight edge or with a dial indicator attached to your sled.
I used a brazed carbide toolbit but HSS should work too. The whole process took less than an hour including a little hand scrapping to clean the ways up. I didn't bother building up the saddle since I did not have any engagement issues.

 

[kdtop]

Quick update of progress so far.

I decided to try to grind the ways myself. I have been taking a machining class at the local community college, and they have a surface grinder. The bed travel was less than the length of the ways, but we thought we could get full reach with 2 set ups, doing about half each time. And I planned to scrape in the final product. So if there was a slight mismatch between the two runs, I could correct with scraping. They have a large surface plate that would hold the ~40" length of the ways.

Of course it didn't go as planned...

The overall ways have a bottom contact surface on each end, each about 10 inches long, where the feet are normally bolted on. These were ground flat first, with the top working surfaces being placed downward, in good contact with the magnetic chuck. After that, the unit was flipped to start work on the top ways themselves. But approximately the middle third of the bottom surface didn't have any contact the magnetic chuck. Imagine a bridge with a support on each side and an arch over the river. The height of that arch was only about 1/8" so it didn't really trigger me to realize the problem initially. But it was enough of a gap to prevent good fixture and work holding. The grinder was a large DoAll brand, and it has the feature of coolant. But the class apparently thinks high-flow coolant splashing is too messy and doesn't have it enabled or working. And instead I was instructed to spray water from hand-held bottle. The grinding took 2 days, and it was only on the 2nd day that I was told about the spray bottle. As you might imagine, this ended up not being a good solution. With the heat of the grinding, the middle, gapped section ended up raising up through heat-warping, and was excessively ground. If I had a table stroke length equal to the full length of the ways, I would have noticed it grinding more in the middle. But because I was doing it half and half, with the ends of each half being approximately in the middle, it was impossible to detect this initially. End result is that with later inspection on a surface table, the center was about 15 thousandth's low!

On the advice of the class teacher, my next plan is to take the ways to a local machine shop which has a Blanchard horizontal rotary surface grinder. He thinks that the cost won't be excessive. We'll see. If it is, perhaps this will be the impetus to buy that better lathe!

In retrospect, I am sure experienced machinists here could have told me that this wasn't going to work. I am counting this to part of my education. Let's just hope my lesson isn't too expensive.

I'll post progress as I go along.

KT

 

[Weldingrod1]

Blanchard will be pretty rough relative to normal surface grinding.

Sent from my SM-G892A using Tapatalk