Bearing Block and Quartering Discussion

[mmagliaro]

So as not to derail Spookshow's inquiry about the Key Imports 2-10-4, I am diverting the discussion about bearing blocks and quartering to here.

First, I will quote the posts on the topic so far, so we can all have a record of it here, and interested new parties can catch up when they find this topic:

Victor Miranda:

Hi spookshow,

the hitch is from imprecise quartering directly
and from loose bearing blocks indirectly.

if the bearings can move the axle can shift a little when
the quartering is off a wee little bit.  then it binds ... a little.

the 'a little' is best seen at slow speeds.  it also gives you an idea of how bad the
problem may be.
wedge a piece of paper into the gap of the frame and bearing block of the loco and see if you get an improvement.
often one side causes more bind and the other causes less bind.

if the problem is so big you can wedge both sides, the bearing block fit is most likely the only issue.



mmagliaro
Forgive the thread hijack, Mark, but this issue interests me because I am battling with a
Key/Yulim 4-8-4 at the moment (nowhere near the same quality drive as that 2-10-4 you have!)

These notoriously have (in my opinion) way too much fore-and-aft slop in the bearings, so as the
rods go around, the drivers wiggle forward and backward continuously, even though you may not have
a rod bind.

It seems to me that in the current one I'm wrestling with, I have the wheels quartered as best as my
"Eyeball Systems Quartering Device, Model 1000"  can do it.  Yet I can see that as a driver is pushed around by
a rod, sometimes the driver wiggles backward in its bearings rather than go around, and that can be bad news because
now the driver hasn't rotated that little bit, and the rod on the other side was counting on that wheel moving so it would be able to continue the rotation.   

I think that if there are little spots where the driver pushes backward or forward in its bearings instead of turning, because of this looseness in the bearings, it can cause binding problems even if the quartering is correct.

Maybe Victor has an opinion on this.   His suggestion about the little piece of paper intrigues me, becuase I was just about to try shimming the sides of the bearings to get rid of some of this back-and-forth play to see if that
helps.


Victor Miranda

Hi Max,
I am pretty sure you know the next three sentences, I want everyone on the same page.
the side rods and the axles must have the same center to center to run without binding.
the crank throws must  all match to run without binding.
the quartering must all match to run without binding.

of the problems, the quartering is the more common problem.
all problems are 'solved' by putting slop in the axles, usually by letting the bearing blocks roam.
So the brass loco makers get locos to 'go' by opening the bearing pockets.

I can't state the easiest fix.  Deciding what is the cause is the hardest part.
the bind at slow speed is the hardest because it means the 'problem' is very small.
so I try to find the error by removing the slop.... and do not trust the first attempt
take out the axle slop and check the siderods for proper fit.

I have found that some siderods are different center to center than others.
... or that the frame was cut by a distracted worker.

hohohoh! are you sure the driver is centered on the axle?

seriously, the issue here is that you have to both trust yourself and
go to the next item to be checked.
If the quartering seems good, check the axle/siderod center to center
if that seems ok, check for a skewed axle, then crank throws.... it goes on.

keep in mind the size of the error you seek.
If the loco does 5 smph and slows to 3 smph and then back to 5....
it is a darned small mistake.

victor

mmagliaro

I completely agree, Victor.  In my case, the speed can't really hold steady until the engine is up to about
12 mph.   Below that, it will run with a regular "lope" at one spot in the rotation, and it won't ALWAYS lope - sometimes it will go around.  That makes me even more suspect that it's the randomness of he wander bearings that is
at least contributing to this, if not causing all of it.

It seems to me that unless you get rid of the wandering bearings, it's very hard to even know if you have the
quartering correct.  When the bearings are firmly fixed, you can wait until the drivers rotate around to the "tight" spot where the motion binds.  Then you can wiggle the rods a bit with tweezers and see who is "stuck", which usually gives away which wheel isn't where it needs to be.

I'll try some shimming tonight and let you know what I find out.

[Lemosteam]

Max Victor's point about the drive being centered on the axle is a good one.  If you can find a way to locate off the bearing blocks and rotate th drivers within them, and zero out a dial indicator against the driver "tire", you can determin how much runout the driver surface has (dont touch the flange).  If the driver wheel is eccentric to the axle axis then the siderod pin will come along for the ride in the eccentricity.   Remember my failed experiment to directly press Kato Mike wheels onto an un-turned Minitrix axle?  That was an exaggeration of runout to an extreme.

My line of work is in the variational study of designs and how that variability affects quality.  No design can be manufactured perfectly each time and because of this, the design must take that manufacturing and sometimes assembly variability into account.  No manufacturing tolerance can be tightened beyond the capability of the manufacturer or beyond the worthwile costs of the part.

This is why parts have clearances- bearing blocks to frames, axles to bearing blocks, siderods to pins.  It is in the management of these varaitions that systems can even work properly.  This fact is exacerbated in mass-producton, but that is a different story as these models are made in very low volumes, so the chances of having worst-case "slop" or clearances are quite likely.

Anyway, these clearances can be shimmed and removed, but one must be careful not to over-restrict the intentional clearances or it may cause additional binding.

Victor's point about the siderod centers and sizes also brings into play the discussion of repeatabilty in the manufacturiing process. As an example, If the siderods are stamped then the variability will be higher than if they are etched, becuase the etching process has a higher level of repeatabilty because no physical tools are involved in the process (tools have clearances too!).

The most important thing to do when measuring is to have an unbiased feature to measure FROM, and that all of the contact surfaces are restrained whilst measuring.

And yes, particularly with siderod-driven models, the probability of all of the available clearances causing an issue is very high and quartering quality is tantamount, because angles are very difficult to control, particularly across and assembly process ( the making of the drvier axle assembly).

I know you are a tinkerer, but sometimes the measurements of what you are tinkering with can sometimes help understand the logical path to the problem.

Dissertaion complete.

[mmagliaro]

John,
Yes, it's all about tolerances and designing a mechanism to tolerate the slop that is in your expected manufacturing process.

The problem with these brass engines is that they are largely hand-assembled and I expect that there is a lot more room for error in casting of the parts than there is in a plastic production engine.

As for the drivers being centered... that could be a source of error, and the crank holes themselves may not be
at exactly the same distance from the center on opposite sides of the same axle.

===========================================
I shimmed the play out of the bearings on two of the drivers.  While this makes it look better running down the track, it does nothing (good or bad) for the binding issue.  I went through a whole second round of rods off, check running, put rods on front two drivers, check, adjust quarter, put back rods on, etc.   This engine has drivers and rods like this:

front   1   2   3   4

Rod between 1-2
Rod between 2-4.
Driver 3 is geared and is not pinned to the rod.
Driver 2 is geared and pinned so it drives wheels 1 and 4 around on the rods.


With the rods only on 1-2, it runs great on a roller stand.  As soon as I put on the 2-4 rods, the problem appears.
All the fine  adjusting in the world on driver #4's quartering can only make the problem "as best as it can", but it
never gets rid of the bind completely.

Therefore...
Most likely the driver position, crank holes, or rod holes are not quite measured out correctly.  That is, even if that driver axle
is quartered perfectly, it does not work, which means errors in the lengths, the driver being off-center, the crank holes
not being at the same distance from center, or a combination of these things.

I am not about to make a new driver.

Then... I made an interesting discovery.
Running the engine, all rods on, on THREE rollers on the roller stand (so that driver #4 just hangs in the air and goes around
on its rods), it runs perfectly   Indeed, it's as though that troublesome rod were not connected at all.

So... in the "good enough" department:
-File the bearing pocket on driver #4 a little deeper
-Add shim under the gear plate at driver #4

This holds that driver up "a little" so it puts marginal pressure on the rails.  In other words, leave the trouble-maker driver #4 as free to spin as possible, so that its errors don't bog down the rest of the mechanism.

Sure enough, while this doesn't make the engine run like an "A" (in Spookshow's parlance), it certainly runs
much much better, and can run an honest 10 mph, which it could not do before.   It also pulls much better.
Driver #3 has the traction tire (and I already shimmed it down to get a little more traction).   So getting driver #4
more "in the air" lets more weight bear on #3.   It pulled 25 cars up my test hill, whereas before, it could barely manage
14.

[victor miranda]

Max,
I find I am full of questions and puzzlements after I've read your post.

I am pretty sure your siderods from 2 to 4 are not on the same center
 as the axle center to center.

and it is not too far off...

Your solution works, so there is no need to flog the mechanism.
gotta ask this one though...
the rear driver isn't gyrating?

victor

[mmagliaro]

Max,
I find I am full of questions and puzzlements after I've read your post.

I am pretty sure your siderods from 2 to 4 are not on the same center
 as the axle center to center.

and it is not too far off...

Your solution works, so there is no need to flog the mechanism.
gotta ask this one though...
the rear driver isn't gyrating?

victor

I agree, Victor.  My first guess would be that the 2-4 rod distance, hole-to-hole, doesn't match the driver
axle length, so it hoses the motion.  I *could* precisely measure, then solder a brass filler into the rod holes, and
then drill new holes to correct it, but....   I am not about to invest that many more hours in this thing.
I am also repairing it for a customer who is on a budget, with further rules out lots of additional hours.

If I were to guess, just judging from how the #4 driver moves, I'd say the rod holes are spaced a little long.
I can see points in the motion where the rod that should be a-pulling hasn't quite come around enough to pull
in its crankpin yet.  And by the time it does, it's a little too late for the rod on the other side to keep from getting
stuck.

Driver #4 doesn't wobble or gyrate any better or worse than the others.  In other words, if it's doing anything
different than the others, I sure can't see it.

[sizemore]

Couldn't you just elongate the crank pin hole in the rod for the fourth driver?

The S.

[mmagliaro]

Couldn't you just elongate the crank pin hole in the rod for the fourth driver?

The S.

Nope.   When a driver is rod-driven, you can't just egg-shape the rod holes.  If you do that to alleviate
a bind at one point in the rotation, at some OTHER point in the rotation, that wider hole will not pull
or push when it's supposed to and the rod on the other side will end up binding.  No.  They need to be
round, and quite accurate.

The elongated hole thing works great if it is a gear-driver driver and the rod just goes around for show (like
a Trix K4).  But if the rod is really turning the driver, you can't get away with that.   If you file the the inside edge
of the hole to elongate it in that direction, you need to solder in a filler on the opposite side to "take up the slack"
as it were.  (Either that, or just make a whole new rod with the holes in the right place).

[victor miranda]

hi sizemore,
enlongating the hole creates another problem, as max says.

Given how small the hitch is, I would try it.
Please don't skip this next sentence.
Enlongating is part of the path to a creating proper siderod fit,
if you decide to re-position the hole.

The watch and clock repair people fix out of position holes by filing the out of position
pivot hole away from the  wear and then reaming the hole to fit a bushing.

That is what one does to fix these siderods.
it is a lot of work to make tiny bushings.

... and who pays?

it is usually better to  not start...

[mmagliaro]

hi sizemore,
enlongating the hole creates another problem, as max says.

Given how small the hitch is, I would try it.
Please don't skip this next sentence.
Enlongating is part of the path to a creating proper siderod fit,
if you decide to re-position the hole.

The watch and clock repair people fix out of position holes by filing the out of position
pivot hole away from the  wear and then reaming the hole to fit a bushing.

That is what one does to fix these siderods.
it is a lot of work to make tiny bushings.

... and who pays?

it is usually better to  not start...

I vote not start.

If I were to try it, it would likely create another problem that is worse than the problem I have.

If it were my engine, and I were going to spend an unlimited number of hours on it until it
ran perfectly, then I would fill in the hole with solder, drill a new one, test, narrow it down, file, etc.

Solder is too soft for a permanent fix, of course, but it's good enough to run for a few minutes to see if you've got
the hole distance right.

Once you get the position right, you fill and drill with some brass tubing or rod and drill the final hole.

Oh, I'll stop babbling. 

Victor, you and I (and others I'm sure) know how to narrow this down.  But these are time consuming techniques
that are prohibitive for this repair, especially given that it runs "pretty good" now, and
wouldn't run at all before (like most Yulim 4-8-4's... ha ha)