What are the desired goals for fit in valve gear components? What is the desired new level of fit, and what is an acceptable level of fit for in service wear? How do we decide whether to work on the gear or accept it as is?

The valve gear on our tank engine feels pretty good to me, but there is a small amount of movement in pins and links. I am trying to decide whether to rebuild the parts/bushings/pins or whether to just spend effort on other tasks.

Machinery's Handbook has a section called "Allowances and Tolerances for Fits" which gives rules of thumb for press fits, shrink fits, and running & sliding fits for various diameters.

Diving deeper, the running fits are divided into nine classes from RC1 (close sliding fits intended to for the accurate location of parts which must assemble without perceptible play), to RC9 (Loose running fits intended to use where wide commercial tolerances may necessary).

For most valve gear joints of say 1.5" diameter between two parts that are both mounted on the frame of the locomotive and are unlikely to move out of alignment, RC6 would be suitable, resulting in a clearance tolerance of between .002 to .006 between the pin and bushing.

At the other end of the spectrum, the eccentrics used on Stephenson valve gear which are very large, and have a very high surface speed, and mounted on the driving axle which has lateral slop and can twist in relation to the frame, RC9 is appropriate, with a clearance range of .025 to .051 on a diameter of 15.75" to 19.69".

Use the RC fits guides to advise your decisions on how much clearance to give your various valve gear joints.

A primary requirement in valve gear is that it not be subject to binding up, as that can result in greater errors in timing than loose fits will. Also, as the old saying goes, it's better to hear 'um than smell 'um.

Greetings:
Kelly is quite correct. "The American Machinists' Handbook" by Colvin and Stanley is a wealth of knowledge, however you'll need an older edition as the later ones do not have the sections on "Railroad shop work", "Staybolt taps", "Rebuilding hub-liner of locomotive driving wheel by welding" and many other subjects applicable to our work.
My copy is the Seventh Edition, published in 1940. These can often be found on line at sites such as "Abe's Books".
J.David

In my extremely humble opinion, and without detracting from what either Kelly or J. David have said:

Valve-gear wear tolerance is somewhat less critical than 'power transmission' in main and side rods. The effects will really be most pronounced at the moment of valve reversal as far as cumulative lost motion is concerned, and this is much more a high-speed concern, unlikely to be severe at normal 'tourist' speeds.

The lost motion will throw out the steam-edge alignment, which in traditional practice might be timed to 1/64" or even 1/128" -- but again, this is probably a miniscule effect on a tourist engine.

What you're concerned with is longevity, and that means keeping up with appropriate tribology for the somewhat 'larger' fits, and recognizing when a loose fit starts to turn into metal-on-metal contact, galling, and the threat of seizing. I'd be tempted to have you hone a loose fit rather than properly bush it if you don't have the time, personnal, or money to 'do it correctly' right away...

A couple more..... working at Railfair back in.....1991(?) Uncle Pete brought in 844 and 3985 hauling long trains at track speed. No issues, no complaints. They sat on display and I was among those who brought supplies for servicing, helping arranging movement to get fuel and water, etc..... well, having the opportunity to get up close and personal with a lot of steam power at once I checked things out pretty thoroughly, and while wiping down the valve gear on the UP power I found I could shake it side to side about 4-5 inches at the bottom of the combination lever. 3985 had a blow in the front right corner of the inside firebox. All in all, it was a working locomotive, not a Swiss watch, and while there were a lot of things that could be tinkered with, they ran well and did heavy work and put on a fine show.

I was setting the valves on a Shay that had been in its share of OOPS moments, and found that of the 3 cylinders, one could be adjusted for the correct amount of lead on DC, one had twice the lead, and one had no lead (also no clearance at TDC, another story). It ran like a sewing machine. So, lesson learned, was not to sweat the small stuff.

At overhaul, decide what the sacrificial element in the gear will be: bushings or pins. I had a situation where hardened bushings and hardened pins were pressed into the rods, only to egg shape the holes in the rods when wearing since the rods were the softest things in the assembly. They are also the hardest to repair. When I rebuilt it, I used sintered bronze bushings and it makes taking up the lost motion an easier job with a small lathe and arbor press handy.

If it feels good and runs well, it's good. If it feels good and runs badly, time to set the valves as best you can until the next overhaul. Best wishes.

I certainly wouldn't claim to be as qualified as the previous four respondents to this thread, but here is what we hillbillies did at Texas State RR before the turn of the century: We had four medium size engines in rotation, and we handled rod and valve gear bushings the same way on all of them. We didn't fix anything unless it was broke !! Nothing sophisticated, we either took a bar of brass or bronze, or maybe recut a scrap bushing if we had something that would work OK. We would press the old bushing out and make the new one the same OD and length as the old one. Then we'd measure the pin and bore the new bushing a couple thousandths bigger. We'd put it in the freezer before pressing it into it's new home, then reassemble whatever we took apart.

Admittedly I haven't worked on as many engines as you fellas have, but every steam locomotive I've ever worked on had idiosyncracys unique to it. No matter how close you measured and cut that new bushing, it always ran hot and spit out gold colored grease until it decided it was happy. On average, by the time you dinkied the engine out to the first switch and back into the shop, that ID would be a good thirtysecond bigger than you wanted.

Some problems are a lot easier to live with, than they are to fix....

I appreciate all the good information in this thread but as long as we are learning and sharing, I would like to correct some terminology. The words "clearance" and "tolerance" are not interchangeable.

Clearance or allowance is the desired difference in size between two mating parts. It can be large or small.

Tolerance is the amount a component can vary from nominal and still be usable.

Parts machined to larger (loose) tolerances will assemble with greater variations in the nominal clearance. If you are aiming for small clearances, you had better hold tight tolerances.

Whenever I see something like this, it reminds me of Angus Sinclair talking about the steqam-flow arrows on patent diagrams. Mathematics and design, lapping and scraping, honing and tolerance are all meant to ALLOW the joint to work -- they do not determine how the joint itself CHOOSES to work... and the latter is the only 'real' thing that matters. If the thing runs hot -- as it so often does -- there is a reason you've missed. You can analytically try to find that reason and correct it, learning 'for next time'; you can wait 'til it beds itself in; you can remove it and check surface finish of bushing and pin and clear out to proper working clearance, but the joint is always right about whether it has been fit right.

We used 'allowance' as a term closer to the technical meaning of 'tolerance'. Tolerance is the amount of variation that can be tolerated in a fit to make it 'interchangeable' in assembly -- you normally see minus tolerance only in pin fit and plus tolerance only in bore associated with this. Allowance is a less scientific form of this -- you 'leave a little to be sure it will fit'.

Clearance, on the other hand, is precisely how much space is in the fitted joint. And as noted above, it may not be the mathematical subtraction of the diameter and tolerances...

What matters is not whether the part FITS.... it's whether the part WORKS. As a starting point for making parts that work, we try to get a good fit plus a couple thousandths "clearance". Then we go easy for awhile and cross our fingers. It ain't rocket science.