Measuring runout on a shaft relative to two datum surfaces 1

[PetkovStoyan]

Hi everyone,

I am struggling with one practical issue how to measure shaft runout, when two datums are specified simultaneously, having different diameter size. Can this be achieved only on a lathe with dial indicator? Is it possible to align both diameters this way, without using custom made jigs?

And in general, is it ok to use the lathe and dial indicator to align along a datum, even if it is single datum diameter? Is this practice legit and valid?

Drawing attached

https://files.engineering.com/getfi...-4246-9516-1899c1f8f3da&file=GD_T_Problem.jpg

 

[PetkovStoyan]

Thank you all for your replies, really helpful information.

 

[greenimi]

pmarc

If, by any chance, you are implying that CZ could be used to transform multiple tolerance zones into a single tolerance zone extending across multiple features of the same nominal size, then I would say UF (United Feature) instead of CZ should be used.

So, why would you use UF instead of CZ for the runout example?
Just trying to understand the differences and similarities between UF and CZ related with the runout geometric callout.
I cannot find good examples in ISO1101:2017 in the runout chapter (everything is pertaining to profile callout)

 

[pmarc]

So, why would you use UF instead of CZ for the runout example?

If, by any chance, you are implying that CZ could be used to transform multiple tolerance zones into a single tolerance zone extending across multiple features of the same nominal size, then I would say UF (United Feature) instead of CZ should be used.

 

[greenimi]

pmarc,

Yes, I understood the purpose, but I don't get the reason. What UF has and CZ does not?

 

[pmarc]

greenimi,

UF creates a single tolerance zone whereas CZ combines multiple tolerance zones together in orientation and location.

So if two (or more) nominally coaxial cylinders of the same nominal size are controlled for example with a total runout tolerance, the UF will create a single tolerance zone for the features, which, among other things that total runout regularly does, will add a control over the size variation between the features (not the absolute size control, but the size variation control only). If, however, CZ is used instead, there will be multiple total runout tolerance zones mutually coaxial with each other and with datum axis, but there will be no enforcement for the tolerance zones to shrink or grow together depending on the actual sizes of the cylinders.

 

[Burunduk]

Since CZ is discussed, can it be used to create a pattern from several different callouts, similarly to ASME's SIM REQT? Or does it apply only on a per-tolerance-indicator basis? If the former - is it under the condition that the different callouts are with reference to the same datum system, or without any datum reference?
ISO 1101 seems to only support per-tolerance-indicator cases with either multi-leader or the number of places 'nX' specifications, but it also directs for additional details to another standard that defines patterns, the access to which I don't have.

 

[pmarc]

CZ creates a pattern within a single callout only.

If one wants to create a pattern from multiple callouts, SIM needs to be added to the callouts (outside the tolerance frame).

 

[geesaman.d]

I design shafts like that (horizontal agitator shafts) and I use the same type of callouts.

Agitators typically have a working element / impeller that is overhung from bearings on the other end of the shaft. For practical reasons, we typically put a datum at each bearing location or connection point on the shaft, even if they are not the same diameter. Vendors who machine these can/should inspect by supporting the shaft at the datum locations (I imagine this is usually a combination of 3 jaw chuck and steady rest) and indicator at the toleranced locations. In our shop inspection area we have a fixture that rests the shaft on two roller bearings positioned at each datum location. We have spacers under the bearing support block to make up the difference in diameters, if any. Take care when building this kind of fixture that the roller bearing sets need to be exactly perp to the rolling axis or else the shaft will move axially as it rolls.

 

[Burunduk]

CZ creates a pattern within a single callout only.

If one wants to create a pattern from multiple callouts, SIM needs to be added to the callouts (outside the tolerance frame).

Clear now. Thank you for that!

 

[PetkovStoyan]

geesaman.d,

thank you for your reply. Do you experience any problems due to imbalance of the impeller? Since it is hanging on a relatively large distance in our case, some runout will always be present at the impeller, at least due to the hanging weight (shaft+impeller), even not taking into account the variable load at the end due to impeller imbalance or other factors. This produces vibrations, and eventually damages the bearing housing or mechanical sealing. Do you have similar problems and how do you deal with them?

 

[geesaman.d]

My method is to machine balance the impeller, specify a maximum runout of the impeller mounting surface of the shaft, and check the math to ensure the total assembled imbalance is less than the desired overall tolerance. I tend to balance the impeller to a very high degree, since it only costs slightly more when you've already invested in balancing on the machine. Then that leaves maximum tolerance for shaft runout. The imbalance of the shaft itself due to runout is generally insignificantly small.

Not performing the balancing operation on the (impeller + shaft) assembly allows replacement of the impeller later on as erosion and corrosion inevitably takes its toll. Some customers and agitator suppliers only balance as a final assembly - this is stupid, since it makes an impeller change very difficult.

I have seen big issues when aftermarket companies apply very thick and dense anti-erosion coatings to the impeller and add weight and imbalance without any controls. If coating is involved it needs to be considered in the overall final balance quality. Also, adding weight via coating increases shaft stress and reduces N/Ncr, both of which take chances on shaft failure, which is much worse than a worn impeller.

 

[PetkovStoyan]

geesaman.d,

what is the range of runout of the impeller mounting surface, and do you take sagging into account due to the weight of the shaft free end? Our tolerance is 0,05mm (as far as I remember, the drawing is not in front of me at the moment), but the sagging alone is more than that according to my calculations. So how can this be acheived? Unless it is supported on center at the free end I don't think this is feasible, but then this measurement will be irrelevant in my opinion
But on the other hand, sagging should always be downwards, and it shouldn't affect runout. Or should it? In general, how do you take into account sagging when checking runout in such cases?

 

[geesaman.d]

Sagging does not affect runout, if you measure with the shaft rotating.

Range of runout is as broad as possible. I wouldn't use 0.002" runout unless I was trying to meet some customer spec that is overly conservative at 200rpm.

David

 

[PetkovStoyan]

geesaman.d,

how do you calculate runout? Is there some standard methodology, or you have some internal procedures?

 

[geesaman.d]

I'm talking about the (weight of the impeller * runout) at the mounting surface, which creates a simple imbalance. I make sure this imbalance plus the impeller imbalance is acceptable for the rotating assembly.

 

[PetkovStoyan]

geesaman.d,

thank you for all the information. Really helpful

 

[PetkovStoyan]

Some follow-up, if anyone is still reading.
Using V-blocks as measuring setup creates one issue, since two of the controlled surfaces are conical. If you don't constrain axial movement during rotation in the v-blocks, you will get incorrect reading for runout on the conical surface. It seems pretty obvious, but I didn't consider that until I actually started the measurements. I think it's worth noting that.
If anyone can share any ideas on how to constrain axial movement on this setup, I will appreciate it. Just to remind - using two v-blocks, adjusting one of them (boosting its height) to compensate the smaller diameter, as 3DDave suggested.

By the way, since the total runout on the two datums relative to them (common datum A-B) was not specified, I looked for some info how to determine it. I found SKF's guidelines on this, general rule when you have a set of two bearings working together. You can check it here, click on table 1:

https://www.skf.com/id/products/rol...es/tolerances-for-bearing-seats-and-abutments

For "moderate speed and running accuracy" IT5/2 is specified for total runout. For most common applications, shaft seat diameter around 100mm or less, this would mean total runout of 0,01mm, even less. Isn't this a little bit excessive, too prohibitive for "moderate speed and accuracy"?

 

[geesaman.d]

My experiences with the SKF recommended tolerances:
1) They are designed for ground journals and bores. Many rolling bearing installations are manufactured this way, but certainly not all. I for one, have not dealt with many ground bearing bores/journals. I regularly stretch the tolerance zones a little bit.
2) They are often not provided with a scale for size. The tolerances on a 300mm dia bearing should not need to be the same as a 30mm dia bearing, but they often are.
3) They are specified for a bearing operating at maximum load, maximum speed, and/or maximum applied load. If you're well within the limits (my applications never exceed 1800rpm, and loads are such that L10 >> 50,000hr) there is conservatism in their tolerances.
4) They only apply to the portion of the part that actually touch their bearing. Holding those tolerances all over is likely overkill.

As for the taper features, this seems tricky. You could clamp the V-blocks at a fixed distance and let the part rest on the v-block edges, however this would only be suitable for light weight parts and few inspections. You could build a fixture with crowned cam roller bearings and adjustable angle for each pair of rollers so you can tilt them to make each bearing pair square to the taper surface. Then you might want to include stops on each end to constrain the piece from walking axially, which is bound to happen because the cam roller planes will never be dead square to the rotating axis.

Or get taper gauges and fit them to each end of the part for this check. If these tapers of are any particular precision, taper gauges are a necessity somewhere in the manufacturing process.

 

[PetkovStoyan]

geesaman.d,

Thank you for your reply.
Could you please clarify what exactly do you refer to as "ground journal"?

 

[3DDave]

ground - creating using a grinding process.
journal - the part of a shaft a bearing contacts.

geesaman.d - the datum features are cylinders. The controlled features are tapered.

Not sure how to restrain the part axially - I don't know what you have access to or what the entire part looks like.