Spindle Taper Maintenance

[Milling man]

They had a fixture with a quill set to the taper angle, and the quill moved in and out as the spindle was rotating at a slow rpm.

I assume most or all traveling spindle grinders do this?

You're right, spindle grinding on a milling machine; and at the factory - on a grinding machine, they are not very different - but only at first glance
To obtain a cone, it is necessary to combine two movements - rotation and linear movement. The accuracy of the cone will depend on how accurate these movements will be, as well as on several other features of the technology.
We'll hire from moving in a straight line. All of us, probably, have seen videos and photos of devices for grinding a cone on a machine many times on the Internet. Often this is some kind of linear motion unit mounted on an inclined or rotary table from a grinding machine - to rotate the linear path to the desired angle. As a rule, ball or roller guides are installed in such units. The length of the guides is quite short - hardly more than 40 inches. And now let's look at a high-precision grinding machine (I think that at a spindle manufacturing plant, final grinding is unlikely to be done on ordinary precision machines) - the movement of the part relative to the spindle is carried out on hydrostatic or aerostatic guides, which have an order of magnitude higher accuracy than ball or roller guides.
Now let's take a closer look at the rotation of the spindle. In a milling machine, it most likely rotates in ball bearings of ABEC7 accuracy class (I rarely met ABEC9 bearings in spindles, only in high-speed spindles), when on a grinding machine the rotation will be either in a spindle on hydrostatic bearings, or in a steady rest with sliding bearings, which at low rotation speed will provide higher rotation accuracy.
The grinding spindle of a high-precision grinding machine and the electro-spindle, usually used in portable grinding devices - I think it would be silly to even compare them
In addition to all of the above, do not forget that the grinding machine will be processed with coolant, which is much better than dry processing. Moreover, the coolant will not just be poured from a bottle, but will be filtered from metal particles and grinding material, and flow in a significant stream.
Therefore, we need to understand that grinding the spindle on the machine itself, where it is installed, is still not at all the same as processing on a grinding machine. But as LocherInc rightly pointed out above, there are new machine specifications and there is the real world And in this real world there are spindles that are not so lucky - and almost any grinding will make them better than they are.
In the lab where I work, there are two machines in which the students did not correctly fix the mandrels in the spindle, and they flew out during work. I would not hesitate to grind these spindles at least somehow and with something, rather than leave everything as it is.

 

[LocherInc]

You're right, spindle grinding on a milling machine; and at the factory - on a grinding machine, they are not very different - but only at first glance
To obtain a cone, it is necessary to combine two movements - rotation and linear movement. The accuracy of the cone will depend on how accurate these movements will be, as well as on several other features of the technology.
We'll hire from moving in a straight line. All of us, probably, have seen videos and photos of devices for grinding a cone on a machine many times on the Internet. Often this is some kind of linear motion unit mounted on an inclined or rotary table from a grinding machine - to rotate the linear path to the desired angle. As a rule, ball or roller guides are installed in such units. The length of the guides is quite short - hardly more than 40 inches. And now let's look at a high-precision grinding machine (I think that at a spindle manufacturing plant, final grinding is unlikely to be done on ordinary precision machines) - the movement of the part relative to the spindle is carried out on hydrostatic or aerostatic guides, which have an order of magnitude higher accuracy than ball or roller guides.
Now let's take a closer look at the rotation of the spindle. In a milling machine, it most likely rotates in ball bearings of ABEC7 accuracy class (I rarely met ABEC9 bearings in spindles, only in high-speed spindles), when on a grinding machine the rotation will be either in a spindle on hydrostatic bearings, or in a steady rest with sliding bearings, which at low rotation speed will provide higher rotation accuracy.
The grinding spindle of a high-precision grinding machine and the electro-spindle, usually used in portable grinding devices - I think it would be silly to even compare them
In addition to all of the above, do not forget that the grinding machine will be processed with coolant, which is much better than dry processing. Moreover, the coolant will not just be poured from a bottle, but will be filtered from metal particles and grinding material, and flow in a significant stream.
Therefore, we need to understand that grinding the spindle on the machine itself, where it is installed, is still not at all the same as processing on a grinding machine. But as LocherInc rightly pointed out above, there are new machine specifications and there is the real world And in this real world there are spindles that are not so lucky - and almost any grinding will make them better than they are.
In the lab where I work, there are two machines in which the students did not correctly fix the mandrels in the spindle, and they flew out during work. I would not hesitate to grind these spindles at least somehow and with something, rather than leave everything as it is.

I love the discourse. I of course wouldn't argue that a qualified OD/ID grinder would be more accurate at grinding a spindle shaft taper. We do that also here in the shop on a complete spindle rebuild. I also wouldn't argue that you could hold a more accurate cone. But "more" accurate doesn't necessarily mean the other is out of tolerance and just because a grinder is capable of holding such tolerances does not mean that the parts reflect that in the real world. It is easy to get caught up in what something "should" or "could" do vs the final product. Where the "most accurate" comment comes into play is really in radial runout. There is a reason factory tolerances for radial runout range anywhere from .0008"-.002" on the end of a qualified test bar for most machines. There is also a reason that when grinding a spindle on-site, you typically do not end up with more than .0000"-.0004" runout on the same test bar, varying with machine condition of course. Those spindle shafts are ground at the factory on extremely precise grinders as you outlined but are only held to the specified tolerance that is required. There is also a lot more to a spindle assembly than a singular shaft and taper. You're dealing with bore alignments in the housing, journal dimensions, ground clearances, shoulder runout, etc etc, all with their same range of specified tolerance. By the time you stack all these tolerances, it is completely normal to end up with a small but acceptable amount of runout. Now could they hold extreme tolerances at the factory and cut that down? Certainly, but there is the real world we talked about and that’s just not a plausible manufacturing scenario. When you grind a completed assembly, you are accounting for all of those small, stacked tolerances across multiple pieces at one location on the taper. The inherent runout of an ABEC 7 spindle bearing, your limiting factor when grinding in-place, is significantly more accurate than the stack of assembled tolerances.