MT3 Collet Chuck In lathe - runout

[pontiac428]

To add to what @pontiac428 said, put something six inches or so long and reliably round in a collet, and see if the runout increases as you move away from the spindle. The taper inside the spindle could be perfectly round at every station, but still point whatever is in it off to the side.

The reason I wouldn't do this is when you do, you immediately have to confront the reality that ER collets, especially with plain nuts instead of bearing nuts, are not as reliable as we think they are. Prove the taper concentric first, then shatter your dreams by measuring 6" of stick out from a trued collet holder- you will get different results with every loosen and re-tighten of the nut. Bearing nuts take the spring-twist out of the equation and work much better at clamping with consistent concentricity, but it's a four-piece system- work, collet, holder, and nut. Start with the holder. It should be true. Then look at the combination of parts.

 

[Huub Buis]

Add some pental black inkt on the chuck taper. Push the taper in the chuck and rotate it a few degrees back and fort. Than remove the taper and check the scratch marks on the chuck taper. This will give you an idea of how good the taper fits.

Second, place the taper chuck and rotate it a few times completely. Check the scratches for a deep scratch that goes around. That could indicate a burr or high spot. I you find a deep scratch that doesn't go all around, the taper of the chuck or lathe is crooked.

The runout of both my Chinese lathes and mill is less than 0.01 mm.
My MT3 ER32 collet chuck (bought in DL, Paulimot) has a run out of 0.01 mm when placed in the lathe.
My MT2 J7316A collet chuck (bought in NL, HBM) has a run out of 0.01 mm when placed in the mill.
To me, this run out is to expected when using hobby tools.
My self made vice model rotary table using self made CA bearings and ER32 spindle, had (after assembly) a run out of 0.015 mm on the ER32 collet holder. I will test it again in a few months but i am quit happy with 0.015 mm.

 

[Ben17484]

Thanks for your input everyone. Whilst testing out the suggestions you’ve all made here, I think made a discovery that is a lot worse. I’ve no idea what I was doing differently before, but now when measuring the inside of the spindle, I can see the same runout there:

I’m not sure how I managed to see a zero readout runout before, but this now checks out.

That’s a whole lot of runout, right? Like a really bad amount? The runout doesn’t jump and it’s consistent when rotating. I believe this indicates a bent spindle rather than burrs etc?

What direct effect will this have on making parts on this lathe?


Sent from my iPhone using Tapatalk

 

[Manual Mac]

You inspired me to check runout on my SB9.
your Smart & Brown SAB is basically the same lathe?
mine is ancient (1938) and has been beat pretty bad by a high school for many, many
years before I got it, but MT3 spindle runout is .00025
i use a ER32 collet chuck & the runout is the same.
I’ve been tempted to try & smooth the MT3 taper carefully with a MT3 reamer, as I can feel some slight rough spots, but this is close enough for me & I don’t want to possibly make it worse?
The collet chuck is a fairly cheap import, but I use a 3/8-16 drawbar to tighten it.
to be fair, I haven’t checked it 6” out yet.
I agree, .08mm equals about .003in which seems like too much.
I suspect I haven’t helped you, but my figures are just for referance.
hope you sort it out.

EDIT: you posted while I was writing.
maybe a nick or something deep in the taper? could a MT3 reamer (used Very Judiciously) be tried?

 

[Ben17484]

I’m going round and round on this. One thing suggested was to use a gauge pin to measure. I don’t have any of those, but I’m fairly sure I’ve seen someone do runout checks using a end mill. So I’ve tried that. I put it in the original collet holder for the lathe (3C??) and get 0.005mm runout on the end mill shank.

This then made me wonder if the cheap dial test indicator I was using first off was throwing things off, so tried using a dial indicator directly in the spindle:

Is this method of measuring ok? It reads about 0.005mm.

Is my method of measuring just no good? Is my DTI just junk? I tried this latest set of tests with two different Dial Indicators, and they more or less read the same.


Sent from my iPhone using Tapatalk

 

[pontiac428]

I got good results reaming the taper on my old lathe. It was a job worth the price of the reamers, and the results were good- no more slipping and concentricity restored. Did the tailstock too. There were threads about doing the job recently where others shared their good results.

There's runout, which is concentricity, and there's deflection, which is support. If you are seeing runout now where you didn't before on the free spindle, check it again with deflection load by pushing the spindle with your thumb in all directions with the dial indicator on. Deflection can be fixed with bearing torque if all else is well. But it can screw with your measurement, so test for it.

 

[mksj]

A runout of 0.005mm is very good, well within the specifications when you add up all the interfaces and variations. I would still recommend seeing if there are any high spots or burrs in the MT3 taper pf the spindle, but that TIR is well within an acceptable range. The other thing I like to check is the variation in TIR as you move away from collet (skew), I would use some 1/2 or 3/4" stock.

 

[Ben17484]

A runout of 0.005mm is very good, well within the specifications when you add up all the interfaces and variations. I would still recommend seeing if there are any high spots or burrs in the MT3 taper pf the spindle, but that TIR is well within an acceptable range. The other thing I like to check is the variation in TIR as you move away from collet (skew), I would use some 1/2 or 3/4" stock.

Thanks mkjs.

If 0.005mm is correct, I’m very happy with it. My concern at the moment is that my readings change across indicators and methodology’s. I’m still learning all of this, so want to be sure that what I’m seeing is correct and not a side of effect of how I’m actually measuring.

I’ve ordered an 8mm gauge pin which will arrive tomorrow. If I mount that in one of the lathes original collets (again, I think these are 3C, but may be proprietary I’m not sure) and measure using a dial indicator as I did in my last post, can I assume that this measurement represents how good the internal spindle runout is (Unless the collet closer and internal thread have both worn in exactly the same way which seems extremely unlikely)? Is there anyway reading with a dial indicator as pictured above will give incorrect readings (excluding co-sin errors)?


Sent from my iPhone using Tapatalk

 

[Jake M]

This collet chuck does not appear to register on what I think is the primary register surface of the lathe. Nor does it appear to register on the spindle taper. Any runout from the lathe spindle, or the interface between the lathe spindle and the collet chuck would have to come from the spindle threads themselves, or from the flat face at the right end of the spindle threads, as you stand in the operator's position. I suspect it registers there anyhow. I'd start with that flat face, as I'm not 100 percent sure that this was intended by the lathe builder as a registration surface. I'm also not 100 percent sure that it wasn't a registration surface, but I'd start by verifying that face.

If indeed the collet chuck does register off of the threads, and the right hand most face of the spindle... I would expect that right there would be some concentricity issues which may well not be repeatable from one installation to the next, as both of those items will make a great reference in the axial direction, but not so perfect in the radial direction. Not "horrible", but not really that good either.

 

[WobblyHand]

Thanks mkjs.

If 0.005mm is correct, I’m very happy with it. My concern at the moment is that my readings change across indicators and methodology’s. I’m still learning all of this, so want to be sure that what I’m seeing is correct and not a side of effect of how I’m actually measuring.

I’ve ordered an 8mm gauge pin which will arrive tomorrow. If I mount that in one of the lathes original collets (again, I think these are 3C, but may be proprietary I’m not sure) and measure using a dial indicator as I did in my last post, can I assume that this measurement represents how good the internal spindle runout is (Unless the collet closer and internal thread have both worn in exactly the same way which seems extremely unlikely)? Is there anyway reading with a dial indicator as pictured above will give incorrect readings (excluding co-sin errors)?


Sent from my iPhone using Tapatalk

Using a dial indicator that way may not be reliable. The vector decomposes into a vector in the correct dimension along the dial indicator axis and one perpendicular to it. The perpendicular force is trying to bind up the mechanism and may greatly increase the force needed to measure properly. This can give false or misleading readings. A dial indicator is made for measuring along it's plunger axis with little or no side force.

A DTI is more appropriate for spindle runout measurements. I would recommend that the Noga indicator holder not be extended so much in the horizontal plane, it can be deflecting from it's own weight. Make the Noga folded up and more vertical, it will resist deflection a little bit better. Also the probe tip should nearly be parallel with the taper to minimize cosine errors.