Mystery bind in wheel/rod motion

[Steamin]

Richard,
Thanks for the analysis! I was trying to figure out how to interpret those graphs. My gut had pegged axle 1, engineer's side, as the trouble spot, I'll measure the crank pin radius' to confirm.

When I did the measurements, all the axle boxes were at the bottom of the frame resting on the pedestal binders. No suspension components have been reinstalled. So the axle centerlines should have all been at the same level. But next time I do this I will block the boxes to keep them from moving.

Steamin

[RET]

Hi Steamin,

An easy way to check the crank pin radius would be to use a depth gauge to measure from the wheel tread to the crank pin on both sides. You know (or can measure) what the wheel diameter is, so calculating the crank pin radius should be pretty easy. The important thing is to see what the difference is between the two sides. If necessary, a part drawing in a CAD program should help because of the accuracy of CAD.

By the way, blocking the axleboxes so they are at the middle of their travel is better because the valve and connecting rod linkages will be close to their running position.

Keep letting us know how you make out.

Richard Trounce.

[Builder01]

A coned wheel tread might make using a depth mic or gauge a little challenging.

[Steamin]

Here's what I did to measure the pin centerlines; bolted the wheel down to the table, using a center in a collet centered the axle under the spindle and zero'd the DRO. Then move over to the pin and align to center; then move back to axle center and re-center again. Repeat several times until the pin looks in line and the axle center is unmoved. Because I was not able to revolve the wheel around the axle center line, the pin centering process moved the whole wheel and axle centerline each time.

indicating the pins from axle center

Removed the center, put the edgefinder in the collet, touched off the left edge recording the DRO number, moved over and touch off the right edge recording the DRO number.
Calculating the pin center from left (DRO reading + .100 edgefinder + half pin diameter), and from right (DRO reading - .100 edgefinder - half pin diameter) gives these numbers:

Firemen's side: 1.154 and 1.153
Engineer's side: 1.162 and 1.160

Both firemen's numbers should be the same, same also the engineer's numbers, so you can see +-0.002 seems to be my accuracy limit with this method.

"Aha!" I thought, it looks like there is 0.010 difference in the pin centers. "Finally a measurable error for this issue" I thought to myself.

According to the print, the pins are to be 1- 5/32" from wheel center (1.1562)

[Steamin]

Then I read Richard's suggestion to measure the pin from edge of the wheel. That sounds like a simpler way to measure, and would double-check my numbers. Measuring from the tapered wheel tread sounded tricky, so I bolted a parallel to the wheel to measure from the wheel edge:

Measuring pin from wheel edge

Yielding the following measurements:
Fireman's: 2.008 and 2.011 the second time
Engineer's: 2.009 and 2.010 the second time

Both wheels measure 6.935 and 6.934 diameter.

Huh. These numbers indicate both pins are at the same distance and do not confirm the previous measurements of 10 thou error.

[RET]

Hi Steamin,

I see there are center drilled holes in the ends of axle #1. What measurement would you get if you use number drills to see which one just fits in the bottom of the center drill hole in the axle (using the drill shank) and then use a dial caliper to measure over the outside of the crank pin and the number drill? From your numbers, it would seem that the engineer's side crank pin radius is somewhere between .008 and .010 too large, but the fireman's side is pretty close to what it should be. .008 to .010 is a little less than what you see on the line graph, but not by a lot.

The taper on the tread would introduce an error when using the depth gauge, but it shouldn't be a lot if the edge of the depth gauge is roughly in the same place on both sides. Typical wheel tapers are 3 degrees or a little less and using the one in 60 rule (1" rise in 60" run is 1 degree of angle), if the tread is .240" wide, 3 degrees gives .012" difference in height over the full .240" distance. You should be able to position the depth gauge a lot more closely than that.

The limitation seems to be more on the ability to measure accurately than anything else, which for relatively small errors is not surprising.

Don't give up.

Richard Trounce.

[Greg_Lewis]

I'd take Ret's suggestion a step farther and machine up a center that just fits into the center drilled hole in the axle, but such that the taper doesn't extend much out of the hole. Then machine the rest of the center pin round, not tapered. This way you can set up the whole assembly between centers like you did the axle in post #21 above. (And if the drivers are too large to clear the table, do it in the lathe.) Thus you have a good hold on everything and a setup that will not change from one measurement to another. You can measure as Ret suggests, from the far edge of the center pin to the far edge of the crank pin. It won't matter what the measurement is as long as all the drivers are the same.

[Harold_V]

Hi Steamin,

I see there are center drilled holes in the ends of axle #1. What measurement would you get if you use number drills to see which one just fits in the bottom of the center drill hole in the axle (using the drill shank)

In a pinch that might work, but drill shanks have two issues that complicate things. First, they are not the same size as the drill itself. The shank diameter is slightly reduced due to the slight taper in the body diameter of the drill. They do that so contact of the margins of the drill is minimized (a method of reducing friction).

The second problem is that drill shanks, being soft, tend to not be round. The are usually deformed from having been used.

While they are not often found in the home shop, drill blanks are highly recommended by this old guy. I have at least two complete sets of numbers, letters and fractions (1/64-½"), plus a large number of spares. Drill (or reamer) blanks are precision ground, and hardened just as if they were intended to be used as a cutting tool (62 Rc or harder). They retain their roundness and straightness, and are held to exceedingly tight tolerance. Unlike drill rod, they are straight and round.

Spare drill (or reamer) blanks are quite useful for making precision pins.

Drill blanks are generally held a couple tenths under nominal, while reamer blanks tend to be held a couple tenths over nominal. If one must limit their acquisition, drill blanks are likely a little more useful.

H

[Steamin]

I wasn't satisfied with my depth mic setup, I don't think the way I clamped the parallel to the wheel gave a reliable and repeatable method.

This morning a better setup came to mind, using the surface plate and height gauge. The clamped parallel is only to keep the wheelset from rolling, and the one pin resting on the parallel stack.

Alternative method to measure crankpin radius

The height of each wheel was measured and found to be L1:6.934 and R1:6.935. Assuming (yeah, I know) the centerline of the axle is in the center of the wheel, that gives me axle centerline height of 3.467. Taking the axle centerline minus 1/2 the pin diameter, I set up the stack of parallels to that height, making the crankpin centerline the same as the axle centerline and then read the height of the opposite pin. Turn the wheelset around, adjust the parallel stack for the new pin diameter and read the height of the opposite pin.

Results:

• L1:3.467 - 0.304 = 3.126 R1 measures 4.924
  R1:3.467 - 0.301 = 3.165 L1 measures 4.925

Only 0.001 difference, that's pretty close. Certainly not the 0.015 error the rod and pin have.
=============================================================================================
Reading the board, I combine Richard and Greg suggestions for another measuring method, taking Harold's advice about drill shanks, and give this a try.

Since I still had the setup on the mill table I with a center in the chuck of the rotary head, I used it to measure over the center and the pin.

A better setup to measure crankpin height

The ground center measures 0.499 diameter and remember each pin is a different size. Subtracting half the center diameter and half the pin diameter gives the Pin centerline from axle center.
Results:

• L1:1.709 - 0.2495 - 0.304 == 1.1551
  R1:1.708 - 0.2495 - 0.301 == 1.1566

The pin centerline from the drawing calls for 1-5/32 (1.1562). Again, the numbers seem pretty close to print.
===============================================================================================
I end up with two repeatable methods to measure pin centerlines; and both agree with each other (L1 is 0.001 closer than R1).

I think the method Richard and Greg suggest is slightly better--measuring over pins; my method of measuring on the surface plate relies on closely machined wheel diameters and assumed identical axle centerlines.

At this point I am ready to conclude the crankpin radius' on driver #1 are on print and not a major factor to the bind.
Might be time to press the wheels off again, and crankpins, to check quartering.

[Bill Shields]

check the quartering with everything in place...do not take things apart.

there is a good possibility that one of the wheelsets is correct.....

[Steamin]

My previous attempts to measure the quartering were not effective, or reliable. My mentor Bill, a retired master machinist, was here last night and we went through my setup again re-measuring and checking my previous work.

Since I had convinced myself wheelset #1 had a quartering issue, we proceeded with pressing the wheels off and the pins out using the hydraulic press.
A piece of shafting was made with a light press fit and the wheels pushed on by hand, with a bit of effort, using light oil lube. The wheels were kept in their same back-to-back orientation. The pin holes were lined up and a dowel pin pushed in. No additional effort was required to align both pin holes to allow the dowel to connect both holes at the same time. (This confirms previous readings that the pin radius for both are identical).

The only dowel I had which was correctly sized was a piece of a broken one, so don't be alarmed by the roughness of its appearance. The body and other end are still smooth.

Hub/Pin/Keyway alignment check

Both wheels have the keyway directly opposite the pin holes, the quartering offset is done by the axle. We hold the stack up to the light and look through the keyways cut in the wheel....they are perfectly aligned.... We could see no offset or shifted difference with the square holes.

It seems clear any quartering issue is not with the machining of the wheels.

I turn my attention to the axle.
A better method to determine the centerline of the center was used. Instead of using a scriber point to pick up the tip of the center and then measuring the height, the height gauge was used to touch the top of the center then calculate the center point.

Measuring the keyways for quartering

We do the calculations and set the adjustable parallel up to put the center of the key at the same height of the center point. The centerline of the centers are still in line and the table/DRO reads zero. We move over to the other side and measure the front and back edges of the key. If there is any angular misalignment, these readings would be different.

The front reads: 0.196 and the back 0.186. "Ah ha!" I exclaim, "Finally the error is found!"

Bill says "Let's re-check some things" and finds the adjustable parallel height is wrong. Maybe it shifted when I tightened the screws.

We measure again and... front: 0.192 back 0.192. No quartering issues here.

So far I cannot find fault with the original machinist's work.

Time to re-check the newest machine work -- mine. I'll have Bill help me re-measure the rods and new Bushings.

[RET]

Hi Steamin,

It looks as though you and Bill have pretty conclusively eliminated axle #1 as the problem. Perhaps you can check axle #2 the same way and although #2 and #3 seem to be alright, it would be interesting to do #3 as well.

The binding has to come from somewhere and its just a question of elimination to track down the problem. Two heads are often better than one, so keep on working at it.

And here I thought (from the graph) that I could tell where the problem was! Oh, well, c'est la vie. Sorry about that.

Richard Trounce.