Hex Turning - Runout Question

[Darren Collins]

Over the past few years, our team has moved to using many of the hex bored gears, sprockets and wheels that seem to be ubiquitous within the FIRST community. There is one challenge that we have encountered and I’d appreciate some advice. When turning down a hex shaft to round the end for a bearing, we’re having difficulties ensuring that the round end is co-axial with the hexagon. We have tried both a 3 jaw chuck and a hex collet in various lathes, but our tolerances are still limited to the runout in these holders. For many applications in our drives, this small runout is a non-issue. However, we are finding some slight binding in some of our gearboxes due to this runout. I’m curious if other teams have encountered this issue and how they have addressed it. I’m also wondering if some teams have chosen to just avoid using hex stock for some of their precision gearboxes. I would like to add that I am not a machinist and that I may just be missing something fundamental.

Thanks in advance for any input.

[Aren Siekmeier]

Originally Posted by Darren Collins

Over the past few years, our team has moved to using many of the hex bored gears, sprockets and wheels that seem to be ubiquitous within the FIRST community. There is one challenge that we have encountered and I’d appreciate some advice. When turning down a hex shaft to round the end for a bearing, we’re having difficulties ensuring that the round end is co-axial with the hexagon. We have tried both a 3 jaw chuck and a hex collet in various lathes, but our tolerances are still limited to the runout in these holders. For many applications in our drives, this small runout is a non-issue. However, we are finding some slight binding in some of our gearboxes due to this runout. I’m curious if other teams have encountered this issue and how they have addressed it. I’m also wondering if some teams have chosen to just avoid using hex stock for some of their precision gearboxes. I would like to add that I am not a machinist and that I may just be missing something fundamental.

Thanks in advance for any input.

Shouldn't this be just as good as the runout from turning anything else while held in a 3 jaw chuck? If the chuck is not holding the part concentric, the turned diameter won't be concentric, regardless of the shape being held in the chuck (round or hex).

2175 has also turned down the ends of Vex hex shaft for gearboxes, and doesn't usually have issues. However, I wouldn't be surprised if the few issues had something to do with shaft concentricity. There hasn't been any effort to measure these sorts of things until very recently.

Curious to hear what some experienced machinists have to say about this.

[weberr]

You can get runnout in a lathe. First, most brand new made in china lathe chucks have a TIR (Total Indicator Readout) of .005", or just over the thickness of a sheet of paper. Compounded with long shafts, chucks that have been abused, chucks that have had a wreck with, bent jaws, etc. that TIR can grow to over .010"

You can easily indicate your shaft and adjust it to run true; You can tap the shaft on the side that is high, or you can use think material such as aluminum foil, paper, plastic to put under the jaw near the high side of the shaft (indicator), etc. Personally, I would check the chuck to make sure it is running true first, as it may be dirt, a burr, etc. on the chuck mounting surface that making the chuck run out.

Personally, I do like to "grind in" the bearing journal on the hex shafts using a whirly jig with a C5 Collet on a surface grinder. But then again, I am the son of a gage maker.

You might be having the gears tooooo close together. I have found that the HEX broached holes and the shafts do have some slop; infact we have tried many different suppliers of the hex stock until we found one that was tight, but not tooo tight.

Good Luck!

[346CADmen]

Have you checked you chuck? 3 jaw chucks, like any product, can be made in a variety of ways (how they work) and quality. Put a nice new dowel pin in your chuck and barely clamp down on it. Place an indicator on the tool holder and check the run out of the dowel, tighten the jaws and check again. The same? Acceptable? Also run down the pin axially to verify alignment there.
Collets can have similar issues. Especially when they are sprung or being used on over or undersized stock.
I apologize if this is below your level of experience, but perhaps others will find it useful.

[Andy Brockway]

The problem may be with the lathe or chuck. I check the run out of every piece I turn prior to cutting the first chip. I am fortunate to have a set-true tight chuck so I can correct the run out prior to starting my machining.

In the three jaw chuck you can sometimes adjust the piece more towards center by tapping with a soft (plastic) hammer while checking run out. Best is .001 TIR but I have also made parts at .002 TIR.

It is interesting that the hex collet is also causing trouble. Check the mating surfaces between the lathe, collet closer and collet for grime/debris. Check the run out of your piece after chucking. Possibly you can tap to adjust to center.

[asid61]

There are independent-jaw 3-jaw chucks available in places. Personally, I'm a fan of just using a 5C hex collet, but when I need to use a 3-jaw and have runout issues I just regrind the chuck jaws a bit.

[Chris is me]

I'd say the problem is probably in your tooling as well. Do the same problems exist if, for example, you use the three jaw chuck on a round shaft and turn down the end of it to a smaller diameter? Is the total runout the same if you use a chuck versus your collet? Is your tailstock (if being used) aligned properly?

Also, just to be sure, where do you source your hex stock? Some suppliers might be better than others here. We've always had good success with the Vex stuff, but do keep in mind the Vex stuff is undersized slightly, so maybe that will affect how the collet grips on it if the collet is otherwise finicky? I'm not an expert on hex collets though.

I would avoid an independent 3 jaw chuck if you can, just because they are such a pain to get lined up and easy for students to misuse. I would make sure your 3-jaw is in good working order first, and if necessary make some adjustments as asid alluded to.

[RoboChair]

Originally Posted by Darren Collins

Over the past few years, our team has moved to using many of the hex bored gears, sprockets and wheels that seem to be ubiquitous within the FIRST community. There is one challenge that we have encountered and I’d appreciate some advice. When turning down a hex shaft to round the end for a bearing, we’re having difficulties ensuring that the round end is co-axial with the hexagon. We have tried both a 3 jaw chuck and a hex collet in various lathes, but our tolerances are still limited to the runout in these holders. For many applications in our drives, this small runout is a non-issue. However, we are finding some slight binding in some of our gearboxes due to this runout. I’m curious if other teams have encountered this issue and how they have addressed it. I’m also wondering if some teams have chosen to just avoid using hex stock for some of their precision gearboxes. I would like to add that I am not a machinist and that I may just be missing something fundamental.

Thanks in advance for any input.

You could in theory use a 4 jaw independent chuck if you want it dead accurate. But your problems are likely from other causes either in set up or asking for too tight of tolerances that are not realistically attainable, we're making FRC robots after all, not aerospace components.

Originally Posted by Aren Siekmeier

Shouldn't this be just as good as the runout from turning anything else while held in a 3 jaw chuck? If the chuck is not holding the part concentric, the turned diameter won't be concentric, regardless of the shape being held in the chuck (round or hex).

You are correct, if using a 3 jaw chuck you should never really expect better than 0.003" TIR.

Originally Posted by weberr

You can easily indicate your shaft and adjust it to run true; You can tap the shaft on the side that is high, or you can use thin material such as aluminum foil, paper, plastic to put under the jaw near the high side of the shaft (indicator), etc. Personally, I would check the chuck to make sure it is running true first, as it may be dirt, a burr, etc. on the chuck mounting surface that making the chuck run out.

You might be having the gears too close together. I have found that the HEX broached holes and the shafts do have some slop; in fact we have tried many different suppliers of the hex stock until we found one that was tight, but not too tight.

You can try shimming it if you really want it perfect as is suggested in bold.
I also agree with the closeness of your gearing, how many thousandths are you adding between your centers? FRC convention is to add 0.003" between centers.

Originally Posted by Andy Brockway

It is interesting that the hex collet is also causing trouble. Check the mating surfaces between the lathe, collet closer and collet for grime/debris. Check the run out of your piece after chucking. Possibly you can tap to adjust to center.

A lot of hex collets aren't made to great tolerance on the TIR, ones that are can be very costly and won't do you any good if your spindle is not in great condition.

[Cory]

If you're using a quality collet and you're having trouble there's only 2 possible sources of error

1) your spindle nose/collet closer taper is jacked up
2) your hex stock isn't straight. This is probably the culprit, especially if you don't have this issue when turning round shaft. We've had some pieces from VEXPro that have had this issue, but have not seen this problem in awhile.

If you're not using a quality collet, get one from Hardinge or Lyndex-Nikken. McMaster's 3223A24 and 3223A25 are Lyndex-Nikken and have no appreciable runout in our experience. Well worth $35 ea.

[346CADmen]

Another thought here, how long is the work piece?
Centrifugal force can be an enemy when turning a shaft/extrusion.
Be sure the non-work end is properly supported or the off-balance force may magnify problem post machining.

[skypatrol]

I don't know what you're using the shaft for, but you might be able to just swap the round bearings for hex bearings.

[AustinSchuh]

When designing gearboxes, we add 2-3 thou to the c-c for each reduction. This loosens up the tolerances on the rest of the components. What c-c are you using for the reduction that you see binding?

[Al Skierkiewicz]

We came across some hex shaft that was not concentric in the past. If you took a close look at the flats you would see that they were not all the same size. We first noticed an issue when we sent the shaft to a sponsor for turning and it came back offset. I found that knocking off the corners of the hex shaft until the turned part just meets a single flat showed that not all sides were the same. As I remember, aluminum hex was really bad for this problem. You get what you pay for.

[Darren Collins]

Thank you for all of the insightful responses. I will take a closer look at the hex stock we are using and check for uniformity and straightness.

I looked at the designs on our gearboxes and found that the center to center spacing only had an additional +0.001". We will begin using +0.003".

I'll also purchase a higher quality hex collet and ensure that our workpiece length is minimized.

[InFlight]

Darren,
We buy all our Hex stock from the Seattle Online Metals Warehouse. We've had no issues with their 7075 Hex stock. We usually pick up 6 foot lengths of it. We could do a will call pickup for you any time, if you would like to have a shorter drive to our shop.