Stumbled on this video tonight. Thought he did a very methodical series of tests. Reminds me of some long posts we had in the past. Quite a few points & tests we touched on, but a few others that he showed we missed. But what has always nagged me on my lathe was the same seemingly random nature of results, just like him. Now that I have a granite plate I'm going to do better measurements. I still have a feeling my bar is out, but this might help share some deviation finger pointing between spindle (or in my case, MT5-MT3 socket adapter) issues
-
Scam Alert. Members are reminded to NOT send money to buy anything. Don't buy things remote and have it shipped - go get it yourself, pay in person, and take your equipment with you. Scammers have burned people on this forum. Urgency, secrecy, excuses, selling for friend, newish members, FUD, are RED FLAGS. A video conference call is not adequate assurance. Face to face interactions are required. Please report suspicions to the forum admins. Stay Safe - anyone can get scammed.
Test bar ...again
- Thread starter PeterT
- Start date
This was a good watch Peter. Maybe I'm more mellow this morning than usual, but I actually enjoyed the video.
Interestingly, his final conclusion was one of our initial suspicions. Basically, a premade test bar depends too much on two factors out of our control.
1. The perfection of the bar itself. Which seems to depend greatly on where it comes from. Some are clearly not great.
2. The perfection of the MT fit. Even a tiny misfit matters.
You are right, he did a few things we didn't. Maybe some of us thought about it, but if we shared it with each other, it didn't sink in.
I will probably check my spindle taper just for poops N giggles, but fundamentally, I'm not a fan of depending on the taper and never was. It's simply too hard to match both tapers perfectly, and too hard to mate them perfectly even when they are identical. Still, I can hope my earlier results were the result of a bar taper issue and that my spindle taper is fine. I can throw the bar away, but my lathe is not disposable!
In the end though, I am exactly where I was before. An MT test bar is a not a reliable way to test spindle alignment. The best tool is a two collar alignment bar mounted in a chuck and turned in place to form an extension of the spindle to validate spindle and bed alignment.
Again, thanks for sharing. The exception does sometimes prove the rule.
It's also good to watch a video that shows that it is indeed possible to produce a video that is worth watching.
Time to resurrect and finish my alignment bar project......????
While the video production imo was quite well done, my reaction to the mechanics is it's a study in how not do things. Some of the infractions didn't seem to matter (sort of like the broken clock that is right twice a day), but still you can't rely on the results if the process is flawed. What I saw/thought......
1. you do not rely on the edge of the surface plate (unless you've got some very fancy, special surface plate). They are not flat or square to anything.
2. He didn't (that I noticed) prove the V blocks). Perhaps negligible as may have done so in the past.
3. You don't drag an indicator along a shaft, no way of knowing its tracking the exact top the shaft. Instead, put the indicator on a surface gauge and sweep the work at different points
4. Same with test the bar in the lathe; dragging along can introduce error - i.e if test is checking Z/Y, and Z/X is out, Z/Y might be perfect but the indicator will still be out as it follows the curve of the work because of Z/X being out. More simply stated, the indicator will only show zero if both planes are dead on, and of course you can't know that.
5. Where is the precision level? Without using it first you have no idea if you sensing spindle misalignment or bed twist. (Perhaps stretching a bit give its such a small lathe, and sure, I get it, guys with a mini lathe aren't likely going to have a 199....but that is only way to know if there is twist and hence what the indicator readings are telling you)
6. Using the taper means you cannot differentiate between spindle taper error and headstock misalignment.
7. The better way.....so long as the ground bar is straight and the same diameter, hold it however you like...3 Jaw is fine. do a sweep at the headstock. Rotate the bar 180 degrees. sweep again. Move to the end of the bar and repeat. By interpreting the four results you will know exactly the state of the headstock alignment. Rotate 90 degrees and do the same in the other plane.
8. If what you really want is to check the spindle taper, you have to do #7 first to eliminate HS alignment as an error, then use the tapered test bar
9. If the concern is the spindle taper (vs HS alignment), forget all this crap, put a bit of steel in the three jaw and turn a 60 degree point that is then perfectly aligned....like machinists do everywhere.
Here's the little reminder I keep in front of me when doing #7 - perfect alignment is when a=b
(btw Peter, why did you think this off topic?)
1. you do not rely on the edge of the surface plate (unless you've got some very fancy, special surface plate). They are not flat or square to anything.
2. He didn't (that I noticed) prove the V blocks). Perhaps negligible as may have done so in the past.
3. You don't drag an indicator along a shaft, no way of knowing its tracking the exact top the shaft. Instead, put the indicator on a surface gauge and sweep the work at different points
4. Same with test the bar in the lathe; dragging along can introduce error - i.e if test is checking Z/Y, and Z/X is out, Z/Y might be perfect but the indicator will still be out as it follows the curve of the work because of Z/X being out. More simply stated, the indicator will only show zero if both planes are dead on, and of course you can't know that.
5. Where is the precision level? Without using it first you have no idea if you sensing spindle misalignment or bed twist. (Perhaps stretching a bit give its such a small lathe, and sure, I get it, guys with a mini lathe aren't likely going to have a 199....but that is only way to know if there is twist and hence what the indicator readings are telling you)
6. Using the taper means you cannot differentiate between spindle taper error and headstock misalignment.
7. The better way.....so long as the ground bar is straight and the same diameter, hold it however you like...3 Jaw is fine. do a sweep at the headstock. Rotate the bar 180 degrees. sweep again. Move to the end of the bar and repeat. By interpreting the four results you will know exactly the state of the headstock alignment. Rotate 90 degrees and do the same in the other plane.
8. If what you really want is to check the spindle taper, you have to do #7 first to eliminate HS alignment as an error, then use the tapered test bar
9. If the concern is the spindle taper (vs HS alignment), forget all this crap, put a bit of steel in the three jaw and turn a 60 degree point that is then perfectly aligned....like machinists do everywhere.
Here's the little reminder I keep in front of me when doing #7 - perfect alignment is when a=b
(btw Peter, why did you think this off topic?)
Last edited:
Omg! I have met mine enemy and stared him in the eye and he is me........
I noticed this too but concluded differently. It illustrates the danger of assumptions. My reaction was: "I didn't know those edges were square enough to do that!" Then I made a mental note to check mine some day cuz I doubt it's that good. But that would have been project 42y....that I will prolly never get to! LOL! So I'm glad you spoke up with the voice of experience.
My bad again. Whereas you challenge it (and rightly so), I was by this time being sucked in by the very warning I give to others "just because the video is well done doesn't mean its right". So I just assumed his V-blocks had been qualified.
Thank for restoring my previous lack of faith in YouTube videos!
You are right about the possible error. I have done what he did many times. But I always sweep the bar at the beginning and end of each longitudinal sweep to make sure I am sweeping it's top edge and I use a straight edge to guide my indicator. Again, I assumed he did that. But he certainly didn't mention the cosine error of indicating at different tangent slopes so it's quite likely he didn't do that.
Yes, and this time you have no straight edge to guide the travel so that potential cosign error is starting to growl.
Another bad assumption on my part. Sucked in by a well done video. Obviously someone that talented must have done that. I'm rather annoyed at myself right now.
Agreed.
Are you referring to that old somebody's dad method here? I kind of liked that method in principle, but I've seen so many crazy interpretations of it that I have lost faith in that approach. I never did sit down and figure out how to use that method properly cuz I'm quite happy with the two collar method. Regardless of whether or not you are endorsing daddy's method, do you have a written version of your process, or is it just intuitive to you?
Agreed
I thought he was evaluating the test bar, and it's ability to help evaluate Spindle Alignment, not evaluate spindle taper. The latter was an outcome of trying to do the former.
Nice!
I noticed this too but concluded differently. It illustrates the danger of assumptions. My reaction was: "I didn't know those edges were square enough to do that!" Then I made a mental note to check mine some day cuz I doubt it's that good. But that would have been project 42y....that I will prolly never get to! LOL! So I'm glad you spoke up with the voice of experience.
My bad again. Whereas you challenge it (and rightly so), I was by this time being sucked in by the very warning I give to others "just because the video is well done doesn't mean its right". So I just assumed his V-blocks had been qualified.
Thank for restoring my previous lack of faith in YouTube videos!
You are right about the possible error. I have done what he did many times. But I always sweep the bar at the beginning and end of each longitudinal sweep to make sure I am sweeping it's top edge and I use a straight edge to guide my indicator. Again, I assumed he did that. But he certainly didn't mention the cosine error of indicating at different tangent slopes so it's quite likely he didn't do that.
Yes, and this time you have no straight edge to guide the travel so that potential cosign error is starting to growl.
Another bad assumption on my part. Sucked in by a well done video. Obviously someone that talented must have done that. I'm rather annoyed at myself right now.
Agreed.
Are you referring to that old somebody's dad method here? I kind of liked that method in principle, but I've seen so many crazy interpretations of it that I have lost faith in that approach. I never did sit down and figure out how to use that method properly cuz I'm quite happy with the two collar method. Regardless of whether or not you are endorsing daddy's method, do you have a written version of your process, or is it just intuitive to you?
Agreed
I thought he was evaluating the test bar, and it's ability to help evaluate Spindle Alignment, not evaluate spindle taper. The latter was an outcome of trying to do the former.
Nice!
Pretty sure they are not, at least on any that isn't some extremely specialized thing where its highlighted as a feature. I have not seen or heard of the side surfaces calibrated. I doesn't make sense as flat and square, to surface limits is expensive to create, they are just going to do it for the heck of it.
Rolly Polly dads method? gawd no, lol.
Its not my invention, I learned it from the various people I learned to recondition machine tools from as the best method to check HS alignment. i.e. I don't think its a stretch to say its the method of choice by folks who most deeply understand checking (and correctly) machine tool geometry (those who recondition them). I've presented it here many times but either every knows this already, it was so dull eyes glazed over before the penny dropped (most likely!), no one read it or no one was paying attention
. imo it is intuitive that its the the best way to come at as you do not rely on the taper or alignment of the test bar.Written version, that is the written version!
so long as the ground bar is straight and the same diameter, hold it however you like...3 Jaw is fine. do a sweep at the headstock. Rotate the bar 180 degrees. sweep again. Move to the end of the bar and repeat. By interpreting the four results you will know exactly the state of the headstock alignment. Rotate 90 degrees and do the same in the other plane. (this assumes of course bed twist has been check and corrected, but any method would require that)
Last edited:
Danzo187
Member
interesting video, too many variables, editing blah blah.
all this pretty much goes away when you start making chips, now the other factors come in ,material,tooling,coolent speeds and feeds , rigidity of machine parts......on and on as long as the mic shows the magic numbers ....oh wait a minute ,what about pressure applied to mic thimble ,machined surface profile........oh crap I lost sight of my project of the day..making (6) flat washers +/_ .020/.030 is kloseanuf
That's more or less what I ASSUMED senior Rolly did. But like I said before I never closed that circle because too many other people had different explanations. I even recall seeing a spreadsheet to do the math. I no like it.
The one thing in the video that I did embrace was the movement of the bar due to the slop of the taper fit. That rang a bell for me and is the root of my reason for preferring the double collars. But as you say, chucking it and doing the math eliminates all that stuff. I wouldn't use the taper with math though.
Thanks for indulging me.
Dabbler
ersatz engineer
@Mcgyver great analysis of the possible sources of error
-- what I noticed is that his spindle has between 1-2 tenths runout at the MT, which amplified over the bar's length is another source of error....
The video was very good, but not comprehensive.
The youtuber is correct in saying this wrong angle is part of his problem... I have never assumed the headstock taper was used for accurate work - I *always* re-cut a 60 degree (soft) dead centre to turn between centres. I was taught it was wrong to do it any other way...
-- what I noticed is that his spindle has between 1-2 tenths runout at the MT, which amplified over the bar's length is another source of error....
The video was very good, but not comprehensive.
A quick scan of it suggests its not the same. Whats described above compares the position of the bar in the plane without rotating it, than takes the difference of that and compares it to the results 180 degrees apart. It absolutely works and is what reconditioners use. i've never take the Rollie thing too seriously as I see constant debates about it, scraps of paper involved and have not see any of the people I believe know what they are doing ever advocate it. I may be wrong because have not thought it through and am relying on hearsay, but that's my sense of it.
Still, the quickest of scans makes me doubt the Rollie method works .....given this claim (from here http://manuals.chudov.com/Rollies-Dads-Method-of-Lathe-Alignment.pdf )
The difference between the "near end average distance" and "far end average distance" is a measure of the misalignment of the spindle axis with the ways.
If you look at the image I posted, in the first sketch, you have a=b and perfect spindle alignment. afaik Rollie's result would say this was not aligned. I'd go to the mat on what I described as I know it works and understand it, think Rollie is quite different and I doubt it works but maybe I've missed something on it.
Last edited:
I hope you understand that I don't disagree with you. The bold ASSUMED is because it's a past tense reference. When I first saw it that's what I thought it was aiming at. I've since changed my mind. In fact I made a joke about the name because it is no longer a part of how I think on the subject. My humour isn't always obvious I guess.
All of my mentions (distant past threads and this thread too) of Sir Rolly say that I have seen too many different versions of it to be able to say that I agree with anything.
Your method makes total sense to me.
My two collars also makes sense to me mainly because I have no real faith in any bar I've yet seen. Maybe some day I'll run across a quality bar that is ground properly. I wouldn't care if it had a properly ground MT at one end or not cuz I wouldn't use it.
Are we really on different planes here? If so, help me see what I'm missing. But I really don't think we are......
If I had a nickel for every time I told someone that the taper should not be trusted, I'd be a rich man!
I've never in my whole life used the taper for any work I do. Other than that crappy bar I bought and a few adapters, I don't even own a tool that uses it. Prolly never will. That's not to say that others who do use it are wrong, just that I don't use it mostly for the very reason you explain.
Danzo187
Member
hmmmm say you have the perfect spindle ground to an absolute state of perfection ,but your headstock housing was bored to an alignment of +/_.0002 during production. just thinking out loud
Danzo187
Member
yeah way to many variables to consider ..what I look for in a lathe......chuck that faces the bed ,turning knobs so you can move those screw things ,a plug on the engine , the gripper thing that holds a drill chuck on other end of machine a shifter to change speeds , a VHS that shows how to run it.............maybe I spent to much time around the parts washer today
Danzo187
Member
I must have multiple personalities ,I find myself posting to my other self
The use I really like them for is with a collet adapter.....that is one nice thing to have. I had an SM with that set once upon a time, regret selling it. More recently I got a Habegger with a 5C spindle so have scratched the itch lol
I do think about that from time to time. But I have a Bison 5C D1-5 collet chuck that is awesome. It's my goto chuck for most of my work. I can't really see much value in adding a taper mounted collet holder.
Of course, no sooner spewed from my yap than I'll be hunting for something tomorrow...... Maybe to hold through stock bigger than 1-1/8! LOL!
What, you have never turned between centers?
Lots of different folks on here with different interests and different needs. Lots of room for those who just wanna make stuff, guys who just wanna fix stuff, guys who just wanna learn stuff, guys who wanna make their equipment better just cuz they can or cuz they need to, guys who just want one of everything, guys who love old iron, guys who chase perfection, guys who wouldn't use their machine cuz they would get it dirty, guys who can't find their machine but know it's in there someplace.
Believe it or not, there is even room here for guys like you with multiple personalities too! Most of us have that talent!
