How bad is this SB 10L bed?

[Richard King]

This is why i stopped posting about rebuilding subjects, and will take this as a hint to stop again, going forward.

The gages on the inverted Vee's were the point of illustration, to show where/how you keep track of deviation in an inverted way. It is applicable to the OP's inspection as contrast to leveling over the flats on top of inverted Vee's which are not a reference surface. They don't wear at the same rate or in the same way, as the rest of the inverted Vee, and provide no useful information.

The photo's i had available show the use with the lathe bed on the planer, and give some ideas to how the references can be used.

To answer your first Q, the planer tool is being adjusted, before using to plane the underside flat of the back way. But it does not matter what the purpose is. Concentrate on a method, any method, to keep track of deviation in the inverted Vee's, and then reference that to other cross sectional features in the bed.

smt

There are a lot of ways to skin a cat. After being on this forum for years I have learned how to turn the other cheek. I may have upset you a few years ago about headstock alignment and I apologize. I missed your information. I sometimes can be a bit of a pain I know, but I hate seeing info that as a professional rebuilder would not do. That's my opinion. Your way worked for you and you were trying to help others. We all need to learn from our mistakes and not get mad at each other, especially when all can share our lifetime of experiences. I would love to tour your shop and spend days learning about your wood working and machine shop. Stephen please continue to contribute in the Rebuilding forum. The forum has lost so many because people could not turn the other cheek. Charles has done an extremely fine job patrolling the forum and getting rid of the trolls.

Hang in here please as I have had to endure some rude people and still love to help pass on the lost art of my trade. I had to put a number of the trolls on ignore. It keeps my blood pressure down. Thanks again Smt. Richard King

 

[Gard]

Stephen, thanks for the post, your custom gauges for measuring the V ways looks to be an excellent solution. Just to be clear what I did was use V blocks that also reference to the sloping sides of the inverted V (not the tops). I then rested a heavy parallel on top of the V blocks to keep them parallel. Your idea is definitely better especially if the inverted V is not a perfect 90 degree.

Dalmatiangirl61, Several people agree the lathe bed even with 0.024" wear on front and 0.015" on the back is fine for making tractor parts. Most parts I have made in the past only need an accurate diameter over a fairly short length. So this seems like a reasonable option. The carriage is worn enough that it is now riding on the tailstock V and also pressing down on the lead screw so I think Turcite on it would be a good option.

Richard, thanks for the links, I have not yet done any work on the bottom of the carriage and I do not have a big heavy square so I figured the alignment of head stock to cross travel would be a quick check on if the cross slide was perpendicular. Looks like this can also be done with dowels on a surface plate. Clearly my setup will sag, hopefully about the same when I am measuring the front and back.

JST, interesting to think about the effect of the difference in height between the front and back inverted Vs. I may do some calculations to see what effect it has on diameter just for the heck of it, I have all the geometry.

Cyanidekid, and others, I can see your point, rebuilding this old lathe probably does not make economical sense for a business, perhaps I should not admit on this forum that this is a hobby with only occasional income for me.

Johansen, its interesting to think about how you attacked your similarly worn lathe. Not exactly by the book but I can see that I could probably make this bed better than it is using a good carpentry SE. Has it been working good for you after assembly?

John.k It's interesting to hear apparently someone is making money after grinding SB lathe beds. Anyone care to provide more insight on how long it takes and or what it costs to grind or plane a 4' bed?

Thanks for all the comments, interesting food for thought

 

[john.k]

A Hercus /SB 9 bed takes maybe one hour.....just so long as the bed looks nice,when its ground......Leo sets up the saddle by measurements taken before grinding.

 

[Gard]

Thats interesting, does the 1 hr include getting it set up in the planner? I guess like anything its faster after you have done a couple. Can you expand a little on what measurements Leo does before grinding and what he is setting up on the saddle?

 

[jim rozen]

"This is why i stopped posting about rebuilding subjects, and will take this as a hint to stop again, going forward."

Noooo! How else will us novices learn except by asking you these questions.... Promise I will never trundle any cast iron to your shop in emira!

 

[Richard King]

When I used to plane lathe beds I tried to remove the same amount off a the side by side V's and flats so it would come down evenly. Not change the geometry. If you take say .024 off the inside V and .010 off the outside V then the bottom of the saddle won't fit. It makes it easier to plane off the bed. I showed Keith Rucker how I was able to check the amounts removed from the V of his planner. Same principal we use on a planer or a grinder. Cash Masters PM Member did the work. He also first tested the bottom of the bed to be sure it would set on the mag chuck solidley. Also check minute 11 and 13 where he used an Idea I taught him. Drilled/reamed holes for a dowel pin to measure the amount removed.

 

[john.k]

the saddle is fitted with plastic pads in milled pockets ....as only SB9s are being repaired ,he has all the dimensions he works to..................far as I know,the smaller sized Chinese lathes use a similar system,where the saddle is not hand fitted ,but rides on splodges of plastic bearing material.

 

[john.k]

Im using a Landis 16x72 universal as a bed grinder.....the story goes that it was reconditioned by Landis in the US,and airfreighted both ways(its an ex air force machine).....anyhoo,its was a marvellously accurate machine......but ,how much cylindrical grinding work do you get ?......next to none......so in the big scrapping a couple of years ago,it was going to be scrapped........saving it cost me the then scrap price of $2500,then it cost $600 for a big crane truck ,and on top of that ,it spewed oily water all over the truck ,and that cost me $400 to get the truck cleaned......so the things cost me $3500 as it sits in my shed now,and ground 20 lathe beds .........for which Im supposed to be given a vintage motor bike ...which so far hasnt eventuated,because he wanted to give me a Jappa.....and I wont have Jappas ...I want a Matchless 500 that he has.

 

[johansen]

I just used my lathe today to turn a foot long 1" diameter 8620 bar down to a 13/16 12 tpi thread.
Turning it between centers with a wobbly eBay mt2 live center, I had to take a couple finishing passes to get the 10 inch long rod the same diameter at both ends and the middle to within .001". I was actually surprised it was that good. It's definitely more accurate now that I relieved the middle third of the front carriage turcite and I taped sandpaper to the bed and ran the carriage over it a few times to get better contact.

I doesn't matter if you don't have a perfectly straight straightedge to scrape a bed provided you keep the straight edge orientated the same way for all 4 V way surfaces. The v way cancels it's x direction variation and turns it all into vertical variation which doesn't matter near as much.

Anyhow I could definitely imagine lapping a bed back into spec if all you have is one flat straight edge and 4 V blocks and a test indicator.

The way you measure twist is set up an indicator under the straight edge in the middle of the lathe. Set the straight edge across the diagonal on the v blocks and compare the test indicator reading with both diagonals. The twist will be double the difference.

The matched 4 V blocks will be the most expensive part of this operation.

 

[john.k]

I set up using a precision level and a dial gauge ,and assume the extreme ends of the bed have no wear...if a bed has too much twist ,I put that one aside.There are plenty of spare beds in the collection.

 

[stephen thomas]

Just to be clear what I did was use V blocks that also reference to the sloping sides of the inverted V (not the tops). I then rested a heavy parallel on top of the V blocks to keep them parallel

Apologies, Gard; i could not quite see that detail.
I do think if there is much wear, a flat topped V-blockis unreliable, but then again, i have not tested that assumption.

Richard -Anyone should be able to question anyone else on procedures and methods, and comment when they think they see something going wrong. That's how we, and the trade advance. No problems with it.

I do not remember posting about aligning the headstock. I do remember that you objected to my response to a question about inspecting the cross-slide way for alignment (90 deg to faintly hollow). I showed photos over a pin in the dovetail, with an indicator swung in the spindle. I check before starting to scrape the non-gibbed cross slide way, then keep track as it progresses. (Swing the indicator and check the pin toward the operator, then slide the pin and check it toward the taper attach end) I don't remember if the indicator was a Blake or a simple DTI (have/use both), but apparently they read differently depending on position? I have tried to keep that detail in mind for other apps.

Jim - you are knowlegeable and intelligent. Across posts in various forums on multiple subjects you have often chosen to use those facilities to troll. Maybe it's your form of humor.

Be clear, objections and questions don't bother me.
They are an essential part of the process for all of us.
A few years ago as my wife and i decided we could "retire" , the process suggested an examination of how i wanted to spend time going into my 70's. Too much i have not accomplished, so little time in every day. So i now tend to skate around areas, people, etc that slow things down or detract from time spent actually doing things.

I can never afford to retire fully if i expect to continue many of my interests. Lately i have been doing a cabinet job that has periods of lost time for glue or finish to dry. My shop is too small to do other things. I have found myself perhaps binging too much on PM & I recognize & apologize that some of my proclivity to post, and to re-post old photos annoys some. Clearly i also miss details in others' posts, and will try to concentrate better before responding.

smt

 

[jim rozen]

Sub-fora here have strictly compartmentalized rules for behavior. Your comments smt are well-taken however.

 

[CBlair]

Time, boy whoever invented that one should have thought about it more. We all could use more of it, I have to admit, I have been looking around and realizing all the things I probably wont do. That is dangerous...

Charles

 

[Gard]

Thanks for the info on how to set up the planner, always interesting reading.

This post has been edited 7/20/22, I found an error in my calculations, I was previously calculating radius instead of diameter along the length of the bed. I corrected chart and equations, my apologies and hope that is the only one.

I decided to calculate the estimated error in part diameter caused by the measured bed flatness. Obviously there are a lot of additional things that will cause diameter to vary, just looking at height of ways and tilt between front and back ways for now. For the purpose of calculation I assumed a reference point of the rear inverted V at the right of the bed. The coordinates of the cutting point are calculated from lathe dimensions and cutting diameter. This provides an angle and distance to cutting point. At each point along the bed the new point coordinates are calculated from the change in height of the rear V and the change in angle caused by the front being different from the rear. I put the equations in a spreadsheet so I could easily change things like cutting diameter and lathe dimensions.

So I would be seeing the part diameter vary from 2 mils undersized to 16 mils oversized. That is a lot of material to file off. If I needed better accuracy for my antique tractor parts there are other ways to get it with the worn lathe but I still think its an interesting data point. This chart also provides some idea of what would happen with a shorter part, a part sticking out of the collet a 3 inches changes only about 0.3 mils across its length. A 15 inch long part mounted in the chuck would have a fairly constant taper of about 1.4 inches per ft so this could be fixed by adjusting the tailstock over if turning between centers. However for a 26 inch long part, its going to look ugly with no easy way to fix it without a traveling rest

I also measured the change in level of the carriage using a precision level and shims and also with an inclinometer, this data agreed quite well with each other and with the angle calculated from the bed flatness data measured at a point on the bed 4 inches from the left edge of the carriage. Its always satisfying to measure something in different ways and come up with similar results.
I will attach a screenshot of the spreadsheet showing the equations I used, I have no idea if I could attach the .xlsx sheet but I could send it to anyone if they want it.

 

[jim rozen]

"So I would be seeing the part diameter vary from 1 mil undersized to 8 mils oversized."

Yes, for a 35 inch long part.

You have discovered another point that old iron users understand: for parts of a reasonable length, the bed can be adjusted to keep the taper to within one thousanth inch, even for badly worn lathes. Your 0.006 taper can likely be reduced to much less. On a bench lathe this is accomplished with the bed adjusting screws, on a cast-iron base type with legs on the T/S side, by shimming the legs. Your calculations, accurate though they are, are a worst case scenario.

1) buy a planer, do it right.
2) pay somebody with a planer to do it right.
3) purchase a less-worn lathe.
4) work around the limitations inherent in a worn machine.

On edit - of course I should have realized the zero-error section was under the headstock - and the largest wear of course right hear the headstock - but the statement still stands that you could likely adust that 6 thou down to less than one.

 

[Gard]

Jim, you make some good points except the reference point, zero error section for the calculation is under the tailstock where the bed is not worn. Options 1 and 3 are out due to lack of funds, space and spousal understanding. After seeing the results from the calculation I am leaning more towards option 2 than 4.
I did the spreadsheet calculation along the entire length of the bed because I have that measured data on bed flatness. In reality the maximum length of part I could turn is probably closer to 24". It has a 48" bed.

The point about re-leveling is interesting. I would not of thought it would have that much effect but I have never played with it before. My other lathe (SB 14 1/2") requires messing with shims under the feet, not something easy to adjust. This lathe does have a leveling adjustment screw under the tailstock so if and when I get it back together that is something I can play with. I had used this lathe very little before starting to disassemble and fix things.

 

[john.k]

By simply having your cutting point above centre ,a worn bed can self compensate.....Ive worked with big (30ft bed ) lathes that had a 'story board' mounted behind that had lines of corrections to use to compensate for wear in the bed.....a powered topslide is another way to achieve good accuracy,no matter how worn the bed is.

 

[jim rozen]

You could have a brand new lathe like that (or, the same one you have now, completely re-worked "as-new") and unless you adjust those screws at the T/S end of the bed, it could cut worse than it does now. That adjustment, typically done via the two-collar method, is a requirement setting up any machine like that.

 

[dalmatiangirl61]

Im using a Landis 16x72 universal as a bed grinder.....the story goes that it was reconditioned by Landis in the US,and airfreighted both ways(its an ex air force machine).....anyhoo,its was a marvellously accurate machine......but ,how much cylindrical grinding work do you get ?......next to none......so in the big scrapping a couple of years ago,it was going to be scrapped........saving it cost me the then scrap price of $2500,then it cost $600 for a big crane truck ,and on top of that ,it spewed oily water all over the truck ,and that cost me $400 to get the truck cleaned......so the things cost me $3500 as it sits in my shed now,and ground 20 lathe beds .........for which Im supposed to be given a vintage motor bike ...which so far hasnt eventuated,because he wanted to give me a Jappa.....and I wont have Jappas ...I want a Matchless 500 that he has.

How are you using a cylindrical grinder to grind lathe beds?

 

[Gard]

I found an error in my calculation 3 posts up and have edited it. Depending on where I am working along the bed I should be able to get decent accuracy but not on a long part.

JST (and I think others) mentioned the the difference in height between the front and back was key to accuracy so I ran some specific cases in the spreadsheet. The first 3 lines have the front and back inverted Vs at the same height, if both are 0.025" low it only causes 1.2 mils error in diameter of a 1" part. This error is dependent on part diameter. When a slope is introduced between the front and back, the error gets much worse, increasing to over 15 mils. This error is nearly independent of part diameter.

Here is an idea, how about scrapeing the back V way down so it is at the same height as the front. Seems like this would be pretty simple to do and increase accuracy by over a factor of 10.
Below is another chart showing the same data as my first measure of bed flatness with averaged and interpolated data, I would need to lower the back inverted V by about 0.010" for a length of about 10", this means removing 0.007" from each sloping side of the inverted V.