Hello everyone, I’m in the process of re assembly of my milling machine after some pretty in depth repairs. I’m currently set up to preload my spindle bearings and would like to know how everyone else that has been here has gone about measuring the preload. I have done plenty of research however there seems to be a multitude of ways to do this. I have all the tools and instruments at my disposal and am familiar with bearing preload on rear differential pinion gears but not very familiar with this setup. Just curious how you guys would approach the task at hand. Thanks in advance
-
Welcome back Guest! Did you know you can mentor other members here at H-M? If not, please check out our Relaunch of Hobby Machinist Mentoring Program!
Milling spindle preload
- Thread starter Pevehouse
- Start date
The first questions are what mill is it, and do they have any direction as to what the target preload might be? The spindle design it's self, bearing type, and expected loading of said bearings will all be factors.
You're right, there are ALL KINDS of instructions all over the internet, "most" of which are probably workable, but for anything beyond a very light preload, they're taken out of context, and get a wrong result, they're first hand tested by somebody that beat their stuff together with a claw hammer and a bent screwdriver, turned it on once, and it didn't blow up, or it's good advice that's "misquoted" by someone with good intentions but remembered the list wrong, and in the end has never touched a bearing in their life...
If you're gonna hit a bullseye, step one is to have a target...
You're right, there are ALL KINDS of instructions all over the internet, "most" of which are probably workable, but for anything beyond a very light preload, they're taken out of context, and get a wrong result, they're first hand tested by somebody that beat their stuff together with a claw hammer and a bent screwdriver, turned it on once, and it didn't blow up, or it's good advice that's "misquoted" by someone with good intentions but remembered the list wrong, and in the end has never touched a bearing in their life...
If you're gonna hit a bullseye, step one is to have a target...
Ha, love your take on that. It’s a Chinese mill, a total piece of **** but it’s what I have right now so I’m trying to make the best of it. It’s a typical quill with a through spindle which is based on kind of an R8 Bridgeport design. Using Timken taper bearings packed with Kluber isoflex NBU15. Here a few pics of where I’m at but to answer your question there’s no manual and no specification.
Attachments
I’ve been tinkering with it. I tightened the nut down to seat the bearings and races as I would normally do with this set up. Then backed off the nut. That’s where I’m at. Ready to start my adjustment. I do have a mitutoyo tenths indicator if that would work better. Just kind of through this setup together
Also, I have a very precise inch pound torque wrench if we are looking at trying to hit a rotational torque specification like zero inch pounds rotating. But I assume we are gonna go a lot more in detail than that
Both my lathe and mill have a fine thread to adjust the preload. I adjust the nut (split type) just by hand really tight, back off the nut, tighten the nut just finger tight and tighten the 2 lock bolts. After that, I measure no play and turning and milling is accurate. I check how the spindle rotates before and after tightening. In general, there is no real difference.
Just finger tighten a nut that has a fine thread puts a lot of preload on the bearings.
Just finger tighten a nut that has a fine thread puts a lot of preload on the bearings.
These won't want to go by torque the way a pinion set does. They've got a "crush spacer", that after it's "crunched enough", kinda acts like a spring, so you essentially set a "maximum" preload and the spacer maintains it. These on the other hand, have no allowance to allow for bearing heating (which happens first) and spindle heating (which follows the bearings but only in due time).
Being tapered rollers, and assuming that they are "decent" or better bearings (not "junk", but most certainly not high end) that means that they probably don't want a lot of preload. They do want to be "pushed in" with excessive preload, just like tapered wheel bearings on a trailer, or on an older rear wheel drive car. Any time one of the pressed races is disturbed, you want to run them down until they barely turn by hand, then turn them, then release the tension. (If you have not moved any press fits, it's not necessary, but it sounds like you're on that already).
I'm not gonna pretend to be the engineer who built this thing.... But I'm going to say that I suspect (don't know, but suspect) that these do not want that much preload. Probably a little more than a "light" preload, but not much. And they're greased, not oiled... I'm gonna speculate it's probably in the "one thousandth or below" range that's ideal. But I'm speculating based on photos, I have NO IDEA what materials or specific bearings are in there to calculate pounds of preload, so in the end you will HAVE TO TEST, but this will get you close- And use judgement- If this feels tighter than it used to when you hand spin the quill, back it up some regardless of what I've said... it MAY be that this spindle is more rigid than I give it credit for, or I've mis-estimated the angle of the rollers by more than an acceptable margin...
You can't really measure preload in a home shop environment, so the way to "infer" it is this. Start at the loose position, where you can measure a few thousandths of end play. (Axially, in line with the spindle.) Then, incrementally, work down until you can "hit" zero end play exactly by turning the nut in carefully controlled amounts. It'll be small amounts.
Next question- What's the thread pitch of the spindle where that spanner nut tensioner ring screws on? (I don't care about diameter, just the thread pitch). The next step is going to be to divide that nut into useful increments.
Huub Buis got in there in front of me, and that advise is not wrong... But if you're used to automotive rear ends and wheel bearings, those are kinda a different animal, they don't necessarily feel the way you'd expect these to. (Or maybe you've buggered with enough different ones to have a feel. I dunno...). The way I'm doing it (measuring as close as possible and calculating as little as possible, but still as much as necessary) is a good way to quantify and develop some "feel" for what's going on.
Either way, in the end, short of a proper engineered solution, you MUST watch the bearings for heat or chatter, (although "slightly warm" after significant operation is a good thing) and be prepared to adjust accordingly. Working out a known (even if approximate) amount of "nut turn" to "preload change" is a valuable tool to react to any abnormalities, should they occur.
Being tapered rollers, and assuming that they are "decent" or better bearings (not "junk", but most certainly not high end) that means that they probably don't want a lot of preload. They do want to be "pushed in" with excessive preload, just like tapered wheel bearings on a trailer, or on an older rear wheel drive car. Any time one of the pressed races is disturbed, you want to run them down until they barely turn by hand, then turn them, then release the tension. (If you have not moved any press fits, it's not necessary, but it sounds like you're on that already).
I'm not gonna pretend to be the engineer who built this thing.... But I'm going to say that I suspect (don't know, but suspect) that these do not want that much preload. Probably a little more than a "light" preload, but not much. And they're greased, not oiled... I'm gonna speculate it's probably in the "one thousandth or below" range that's ideal. But I'm speculating based on photos, I have NO IDEA what materials or specific bearings are in there to calculate pounds of preload, so in the end you will HAVE TO TEST, but this will get you close- And use judgement- If this feels tighter than it used to when you hand spin the quill, back it up some regardless of what I've said... it MAY be that this spindle is more rigid than I give it credit for, or I've mis-estimated the angle of the rollers by more than an acceptable margin...
You can't really measure preload in a home shop environment, so the way to "infer" it is this. Start at the loose position, where you can measure a few thousandths of end play. (Axially, in line with the spindle.) Then, incrementally, work down until you can "hit" zero end play exactly by turning the nut in carefully controlled amounts. It'll be small amounts.
Next question- What's the thread pitch of the spindle where that spanner nut tensioner ring screws on? (I don't care about diameter, just the thread pitch). The next step is going to be to divide that nut into useful increments.
Huub Buis got in there in front of me, and that advise is not wrong... But if you're used to automotive rear ends and wheel bearings, those are kinda a different animal, they don't necessarily feel the way you'd expect these to. (Or maybe you've buggered with enough different ones to have a feel. I dunno...). The way I'm doing it (measuring as close as possible and calculating as little as possible, but still as much as necessary) is a good way to quantify and develop some "feel" for what's going on.
Either way, in the end, short of a proper engineered solution, you MUST watch the bearings for heat or chatter, (although "slightly warm" after significant operation is a good thing) and be prepared to adjust accordingly. Working out a known (even if approximate) amount of "nut turn" to "preload change" is a valuable tool to react to any abnormalities, should they occur.
Yes I understand exactly where your going and I do have a temperature gun on hand to keep and eye on spindle temperatures. I do have a pretty good feel for these things and my plan was to shoot for zero play. It makes me feel better however knowing you’re here and willing to help where you can, if indeed I run into anything. Thank you. The spindle thread appears to be about 18 threads per inch if that helps
If it's a Chinese mill, it's probably metric. If you're gonna bank on this, you probably want to match it up to a thread gauge. A bolt or a screw with a known thread pitch would suffice in this situation, albeit a little awkward. If the nut is deep enough (has enough threads) for a positive ID, that'd be valid too. And on a
First, (going with 18tpi) turn threads per inch into inches per thread, which is 1/18, or 0.05588. Make that into thousandths, it's 55.555 thousandths of an inch per thread.
Use the spanner notches in that nut to divide 4 sections. (I think there's four, right?...) Then each "quarter" would be 13.75 thousandths of an inch advance per quarter turn. That's a bad number, so I'll insert a fudge factor here where it'll divide out to "who cares", and call it 0.056 thousanths per turn (half a thou liberty there), and it's 14 thousandths per quarter. When you're in the range where you're "close", mark a spot (sharpy is fine, no scribes needed), use a ruler or estimation to mark out those 14 divisions on one of the spanner slots. If they "look" equal- They're equal. Each of your marked divisions will represent 0.001 inches (one thousandth of an inch) of travel. That is the resolution you want to make the smallest meaningful adjustment, without going to the point of getting no returns. Two marks is too big to be considered one "step" of adjustment, and half a mark is pushing the resolution of your adjustments finer than you're going to see any results from.
If it turns out to be a metric 1.5mm thread (very close to 18tpi), the math is exactly the same except it's already in "distance per thread", and it works out (again with slight rounding, done early so it divides away nicely), that's 0.060 per turn, or 15 divisions on each quarter of the nut. to get the 0.001 target resolution. Or any other number you "science" if it turns out to not be one of those two.
If you've got it together and it's good- I'm not suggesting that it can't be. But now or in the future, that's a good way to calibrate your "click elbow" for doing them freehand.