What's new
What's new

Holbrook T13 - Lathe Spindle Plane Bearings

Lurk

Cast Iron
Joined
Feb 10, 2014
Location
Peak District UK
Time for another head scratchier

I am setting up a plan for bedding the lathe headstock to the newly scraped lathe bed. In order to ensure the head stock is in-line with the bed ways for spindle alignment I have to be able to take some meaningful measures against a test bar in the spindle (well I think thats the plan). However, the headstock bearings in which the spindle runs are white metal and initial assessment showed 0.035" of lift at the chuck when the bearings halves were clamped down (obviously excessive wear and no shims between the halves). I am trying to avoid re-pouring babit or machining new bearings.

My query is as follows -

From what I have read to date, the vertical lift at the chuck should show something of the order of 0.002" for a 3" dia' bearing journal - by the same token, other specifications suggest a tolerance of between 0.001" & 0.002"/inch of diameter - i.e. between 0.003 and ).005" between journal and bearing ID - which is going to give far more lift surely ?

Then I get to the dimensions given by Dr Schlesinger for tool room lathes - up to 0.0004"/12" - on a test bar. I do not understand how that can be achieved with a bearing described above - stationary you would get one result - once running the shaft would align within the bearing oil and give you another figure surely.

Any thoughts or suggestions much appreciated ?

Thanks in advance
Mat
 
At .035 wear, your running the risk of gears meshing funny, labyrinth seals rubbing etc. I dont see how re-babbitting is avoidable. On most plain bearing machinery, at rest, the spindle lays dead parallel to its dynamic condition, unless it is belt driven. However during a rebuild you most likely would not have belts on. I have only worked on a few plain bearing lathes, but many grinders, I've never run across an issue where the the dynamic alignment was much different than the static, just a little bit higher as the shaft lifts.
 
I would say it depends on the top speed of the lathe a lot. The lower the speed the tighter the clearance you can run. I would think .001" would be my preference. Actually a retrofit to roller bearings would be something I would look into if the idea is to use this lathe more than restore it.
 
I would say it depends on the top speed of the lathe a lot. The lower the speed the tighter the clearance you can run. I would think .001" would be my preference. Actually a retrofit to roller bearings would be something I would look into if the idea is to use this lathe more than restore it.

Hi Brian, lathe top speed on the gears is 600 rpm - while slow is adequate for my turning requirements - I plan to run the lathe via a VFD so top speed is unlikely to get past 850 rpm. Roller bearing retro fit would be a big mod' maybe later down the line as my engineering skills develop. The bearings are lubricated via wicks at the lowest point of each lower half of the bearing. No seals and no oil bath swashing about near to the bearings.

Im not far off being able to seat the head casting back on the bed - at which point I will re-test the bearings and check what the clearance is around the spindle, lift etc etc. Then I guess its time to make a decision and move forward.
 
I subscribe to no clearance except that needed for an actually round journal to freely rotate in an actually round plain bearing. It is after all a machine tool , not a wheel barrow

I gleaned such an outlook from what highly regarded makers as Pratt & Whitney provided on their plain bearing Model B lathes - external adjustment screws - both spindle bearings - that let you squeeze them down to practically nothing for the fussy jobs and open them up for the rough and tumble work

Screws (with screw driver slots) plainly visible in Thumbnail
 

Attachments

  • PW B Adjustment screws.jpg
    PW B Adjustment screws.jpg
    92.4 KB · Views: 255
I subscribe to no clearance except that needed for an actually round journal to freely rotate in an actually round plain bearing. It is after all a machine tool , not a wheel barrow

I gleaned such an outlook from what highly regarded makers as Pratt & Whitney provided on their plain bearing Model B lathes - external adjustment screws - both spindle bearings - that let you squeeze them down to practically nothing for the fussy jobs and open them up for the rough and tumble work

Screws (with screw driver slots) plainly visible in Thumbnail

It does seam strange to me to have a clearance of the size which would result in the spindle effectively being able to 'move' - So I do tend to agree with the less is best approach - with the consideration that it has to be able to rotate at the desired speed, with the lubrication used and under the loads required without running 'too hot' - which makes for quite a subjective tolerance :-)

Thanks for your observations John.
Mat
 
Hydrostatic bearing clearance is a balancing act, heavily dependant on the oil viscosity and top rpm. On my Boley 4LV, a really thin oil is specified, barely thicker than diesel. The plain bearing clearance is adjustable, and should be set for 0.02mm, which is only 0.0008 in inches.
Here is my write up from scraping that in...
Boley 4L

What oil is specified for your spindle?
Mark
 
I wonder if bronze bearings would be a possibility? Depending how complicated the babbit ones are, you might be able to make drop in replacements.

I have an old handbook that shows 1900 planer set-ups with one showing a lathe head-stock and tail-stock having their ways plained in tandum. The fixturing held the castings from the spindle and quill bores so the features would be co-planer. I wonder if, until you could address the bearing issue, you could align the head-stock using the bare bearing seats, and then once the new precision bearings are in place, make any fine adjustments for parallelism to them and not the ways?

IMO unless you have a good spindle in your head-stock, I don't see any other way to align it to the ways, other than how it was done from the factory.
 
There is very little in the way of information available for the lathe - it dates from the 1950's and mine seams to be one of the last made of this model. The owners manual is a hand written affair dealing with screw cutting gearing and relieving attachments. No mention of lubrication anywhere.
So far I have not found any tags on the machine which might provide further clue.

The headstock has a small sump which splash lubricates the gears, it is filled via a msall port at the front and holds maybe 2 pints at best. The plain bearings are fed via oil wicks on the lower portion of the bearings - in turn fed by gravity from ports filled by the operators (I assume each day / shift).

max speed is 600rpm. I am looking to improve on that if possible.

I still have some work to complete before I can sit the headstock back down on the bed and then test fit the spindle within teh worn bearings and get a better measurement of contact and wear. Round they are not, maybe once, no longer.
 
I have seen a grinder run so far out of maintenance that the labyrinth seals friction welded to each other, .200 deviation from center on the pulley side, and as much as I dreaded having to rebuild it, it was not that bad, there are many things to reference true center off of. Does the machine use inserts or is the babbitt poured in place? If they are poured in place the headstock must come off and after pouring, peening and cleaning it would be line bored. I dont know Holbrook's too well but on american machines the head sits on the tail stock ways so you can set it down on a boring mill with the v way on a ground rod that is already dialed in. The object is to bore it as close as you can with intentions to scrape in the final alignment. However, if you have the ability to test fit the spindle while it's still set up in the boring mill you can check that all of the gears etc. Are lining up. If they are go ahead and bore out to the final dimension, ie. shaft + clearance. However if you cannot test fit, it is best to leave it size on size to the spindle diameter as it makes the scraping go much faster, ideally you would leave the centerline a few thousandths high, so you can scrape it into alignment.
Bronze spindles are best used when you have clean environments and are going to run low speeds with high loads as it will sustain boundary lubrication (metal to metal contact) much better than babbitt. Babbitt has high embedability so any dirt that should enter the system gets pushed into the softer babbit causing no harm, but in the event of metal to metal contact, can wear more as it is so soft. Also a babbitt spindle tends to gall and stop the spindle when things go horribly wrong, which is a good thing, bronze liners dont give any symptoms and every one I've ever rebuilt the spindle shaft had considerable scoring. The right way to go about a machine like a lathe is to align the spindle to the gears and other respective parts in the headstock and then scrape the headstock base to bring it into alignment with the bed. And on a quality machine like a holbrook if you do the spindle right it should be pretty darn close to lined up with the bed as it was most likely right once. As far as clearance, it is directly tied to viscosity and speed. You cannot reduce clearance without running thinner oil or cutting back speed. On most plain bearing spindles oil enters in only one place and it must make a full revolution, which means the clearance must be such that the film thickness can support a full revolution for the given speed and viscosity. If you want to choke up clearance to reduce spindle deflection you need to look into a multi point oil entry system like Landis did with Microsphere bearings. In actual operation spindle deflection means very little because roughing cuts deflect rolling elements too, you always take a finish pass for accuracy. Hydrodynamic bearings, when properly set up, hold true center with great accuracy, and it is why they are still used on large grinders and high accuracy cnc grinders. A hydrodynamic spindle has no moving parts besides the spindle and it offers tremendous load carrying capabilities, superb finish, unmatched damping and long life provided you give it oil. It is a misnomer to think a rolling element is all around better, it definitely has a few advantages like the RPM range and ease of installation, but it also must be much larger than a plain bearing of the same load capacity.
A few years ago we were rebuilding a 1942, 14x72 cylindrical grinder which had a babbitt spindle and the lube system failed causing it to seize, one of our maintenance guys called Landis, (now Fives Landis) and asked them if they could provide any insight on converting it to tapered roller bearings "like a modern machine" making an assumption, the engineer at Landis laughed and told him that a bearing big enough and accurate enough to replace that babbitt spindle would cost more than the machine, and went on to warn him that without the hydraulic cushion of the hydrodynamic bearing the 30" wheel could explode if it started to chatter. They are still using hydrodynamic and hydrostatic spindles on their finest cnc grinders today, many of which they build in conjunction with Cranfield Precision.
All that being said, converting that lathe would be cost prohibitive, you would be time and money ahead buying a different machine. I would rebuild it just the way the factory made it.
 
600rpm max is fine for larger parts but it seems rather low for smaller work. Do you have a cross section drawing of the spindle? Maybe converting it to angular contact bearings wouldn't be that hard? When my workshop in France is up and running again this summer I could probably help out with that project. Modern bearings are really very good so normal grade high quality bearings (SKF etc) would be fine for this application and not cost much at all.

Luke
 
I was interested to find, tearing into an automobile turbocharger, that it was using plain bearing and expecting to see up to 20,000 rpm. In its defense it was a much smaller bearing, but the interesting cheat was that it used an intermediate piece. That is, between the turbine shaft and the housing was a separate part that looked somewhat like a ball bearing cage. Perhaps 1/4" wall thickness and drilled with radial holes. The theory apparently was that it divided the SFM between shaft and housing with the intermediate piece free to rotate at some intermediate speed. All fed with pressurized oil, of course.

No help in this instance but a bit of trivia on more modern use of plain bearings.
 
I have seen a grinder run so far out of maintenance that the labyrinth seals friction welded to each other, .200 deviation from center on the pulley side, and as much as I dreaded having to rebuild it, it was not that bad, there are many things to reference true center off of. Does the machine use inserts or is the babbitt poured in place? If they are poured in place the headstock must come off and after pouring, peening and cleaning it would be line bored. I dont know Holbrook's too well but on american machines the head sits on the tail stock ways so you can set it down on a boring mill with the v way on a ground rod that is already dialed in. The object is to bore it as close as you can with intentions to scrape in the final alignment. However, if you have the ability to test fit the spindle while it's still set up in the boring mill you can check that all of the gears etc. Are lining up. If they are go ahead and bore out to the final dimension, ie. shaft + clearance. However if you cannot test fit, it is best to leave it size on size to the spindle diameter as it makes the scraping go much faster, ideally you would leave the centerline a few thousandths high, so you can scrape it into alignment.
Bronze spindles are best used when you have clean environments and are going to run low speeds with high loads as it will sustain boundary lubrication (metal to metal contact) much better than babbitt. Babbitt has high embedability so any dirt that should enter the system gets pushed into the softer babbit causing no harm, but in the event of metal to metal contact, can wear more as it is so soft. Also a babbitt spindle tends to gall and stop the spindle when things go horribly wrong, which is a good thing, bronze liners dont give any symptoms and every one I've ever rebuilt the spindle shaft had considerable scoring. The right way to go about a machine like a lathe is to align the spindle to the gears and other respective parts in the headstock and then scrape the headstock base to bring it into alignment with the bed. And on a quality machine like a holbrook if you do the spindle right it should be pretty darn close to lined up with the bed as it was most likely right once. As far as clearance, it is directly tied to viscosity and speed. You cannot reduce clearance without running thinner oil or cutting back speed. On most plain bearing spindles oil enters in only one place and it must make a full revolution, which means the clearance must be such that the film thickness can support a full revolution for the given speed and viscosity. If you want to choke up clearance to reduce spindle deflection you need to look into a multi point oil entry system like Landis did with Microsphere bearings. In actual operation spindle deflection means very little because roughing cuts deflect rolling elements too, you always take a finish pass for accuracy. Hydrodynamic bearings, when properly set up, hold true center with great accuracy, and it is why they are still used on large grinders and high accuracy cnc grinders. A hydrodynamic spindle has no moving parts besides the spindle and it offers tremendous load carrying capabilities, superb finish, unmatched damping and long life provided you give it oil. It is a misnomer to think a rolling element is all around better, it definitely has a few advantages like the RPM range and ease of installation, but it also must be much larger than a plain bearing of the same load capacity.
A few years ago we were rebuilding a 1942, 14x72 cylindrical grinder which had a babbitt spindle and the lube system failed causing it to seize, one of our maintenance guys called Landis, (now Fives Landis) and asked them if they could provide any insight on converting it to tapered roller bearings "like a modern machine" making an assumption, the engineer at Landis laughed and told him that a bearing big enough and accurate enough to replace that babbitt spindle would cost more than the machine, and went on to warn him that without the hydraulic cushion of the hydrodynamic bearing the 30" wheel could explode if it started to chatter. They are still using hydrodynamic and hydrostatic spindles on their finest cnc grinders today, many of which they build in conjunction with Cranfield Precision.
All that being said, converting that lathe would be cost prohibitive, you would be time and money ahead buying a different machine. I would rebuild it just the way the factory made it.

That is very helpful insight, thanks for taking the time to type it. Over the next couple of weeks I will get the headstock sat on its original scraped pads and on the lathe bed (newly scraped) - I have to make up a test bar to see where abouts the spindle is now and having cleaned up the spindle bearings (dried oil) I will get a better idea of the wear within - after which I can move forward. It is my understanding that they are poured in place - I will atach a photo to this thread.
Thanks Mat
 
I was interested to find, tearing into an automobile turbocharger, that it was using plain bearing and expecting to see up to 20,000 rpm. In its defense it was a much smaller bearing, but the interesting cheat was that it used an intermediate piece. That is, between the turbine shaft and the housing was a separate part that looked somewhat like a ball bearing cage. Perhaps 1/4" wall thickness and drilled with radial holes. The theory apparently was that it divided the SFM between shaft and housing with the intermediate piece free to rotate at some intermediate speed. All fed with pressurized oil, of course.

No help in this instance but a bit of trivia on more modern use of plain bearings.

I have been reading forum chat on scraping large generator & power plant plain bearings running with huge loads and at several thousand rpm - I then came across a reference to the use of such bearings by RR in jet engines in the 20,000rpm high load environment and after years of service showing next to no signs of use. As you observe, they crop up in all manner of situations and each is tuned accordingly with lubrication and clearance to suit the conditions. The higher end stuff as I understand it runs in an oil bath with pressure feed into the bearing.
Mat
 
Images of teh headstock & bearing

IMG_6402.jpg
IMG_6404.jpg

First image is the smaller bearing end - second is the bearing nearest the chuck.
At the minute I have not image of the assembly with spindle in place showing gear mesh etc. As it stands the castings is now cleaned and painted - awaiting bedding on to the bed - which requires a test bar - the spindle nose is (i think MT 4.5) and I have to make a bar to suit - once I have that I can see where everything lines up and assess if the existing bearing can be adapted by dropping the spindle a little further without it adversely affecting gear meshing.
 

Attachments

  • IMG_6354.jpg
    IMG_6354.jpg
    95.6 KB · Views: 200
Last edited:
As a Non Expert,I would suggest-are you certain the front bearing has worn down? I would have thought it would wear upwards due to tool load,in which case you could simply let the cap down. If the front has come down,couldn't you let the back down to level the spindle! I've had correspondence with Richard Anderson,and he told me Holbrooks had fitters that went out to deal with problems like yours. I imagine these are the kind of things they did,using dummy spindles.
 
There is nothing plain about hydrodynamic bearings.

However, those bearings did nothing to create that gap on their own. What is the condition of the shaft? Crooked? Cracked?
 
There is nothing plain about hydrodynamic bearings.

However, those bearings did nothing to create that gap on their own. What is the condition of the shaft? Crooked? Cracked?

Initial visual inspection shows the spindle journals to look fine - not scored or pitted or galled - same goes for the bearing faces. From what I have read about other Holbrook lathes - the bearing halves were shimmed when new - mine had none fitted. The measurements taken when I first looked the lathe over - were quick and 'rough' and showed 0.035" of lift at the chuck. Later, having removed the headstock cover and found a good quantity of brass fine swarf in the oil sump (splash lube of the gears) - I believe his to be from the gear change shifters.

As stated above, it may well be that the major wear is in the tops and they can simply be dropped down. I should know more later this week when I get to shift the headstock into approximate position and drop the spindle back into the bearings. I will try and get some better dimensions and a view of the contact via some micrometer blue on the journals.

Mat
 








 
Back
Top