Best Of *Science* of solving N scale diesel mechanism drivetrain vibration noise?

[atsf_arizona]

Hi, all,

FYI, I updated my posts on page 11 with a updated Google Drive URL link to the Atlas / Atlas-Kato / Intermountain loco parts cross-index spreadsheet described on page 11 of this thread.  The spreadsheet resulted from this long discussion below which started in Dec 2016.  Here's the spreadsheet link which you can view/download from Google Drive:

https://drive.google.com/file/d/1A6VJR3uFGOLfr50-GreWia_oe4UCCxaL/view?usp=sharing

The purpose of the spreadsheet is to answer one of our the major discussions that arose in the thread:

• What are the correct replacement part numbers to use in these N scale Atlas / Atlas-Kato / Intermountain locomotives?
• What parts are swap-able between locomotives?
  (focusing on motors, motor saddles, worm/shaft assemblies, bearing blocks, trucks)
  

For whatever reason, the original 2017 January link on page 11 stopped working, but the above updated link should work for you / all of us now .
(8:40PM EST 12/22/19 - uploaded new version that fixed the URLs to point to Atlas' new location of loco Parts diagrams)

Enjoy. Happy Holidays 2019 and Happy New Year 2020!

=======================

Original thread I started way back then:

=======================

Let's have discussion in this thread, about the *Science* of solving drivetrain vibration-caused noise in a N scale Atlas, Kato, and Intermountain diesel locomotive split frame  mechanisms.

Asking our experts here for advice, especially those who have installed many aftermarket plug-n-play DCC decoders in your locos and have performed this dis-assembly and re-assembly many times.

————————

1) Why is it that when I dis-assemble  / re-assemble N scale diesel split frame mechanisms from Atlas, Kato, and Intermountain, no matter how carefully I do it……..  sometimes, the previously silent mechanism now has some small “coffee-grinder” sound introduced into it.    In a few locos, a large amount of “coffee-grinder” noise?   For these, I can’t seem to find and resolve the cause, no matter what I do.

I’m well aware of long-standing fixes and internet forum treatises on the “Bearden bearing block fix”, the “SD60 truck fix using Kato wheel sets ( http://forum.atlasrr.com/forum/topic.asp?ARCHIVE=true&TOPIC_ID=42265  )”, the need for proper lubrication of the motor and worm gear bearings, proper orientation of the worm gear bearings, the brass fingers that hold the worm bearing blocks in place, etc.……  (and really thank you all for that info).

What did I get out of alignment upon re-assembly, that is causing that sound?  Thankfully, it doesn’t happen every time, sometimes I get a great result.   

Here’s a mild and a major examples of what I’m talking about, I dis-assembled / re-assembled all four of these locos using the same technique and care:

-------------------

/>
1st (good) and 2nd (noisy) are Intermountain SD40-T2 and IM SD45-2 mechanisms, respectively.
The 2nd IM SD45-2 noisy mechanism, I've also replace the IM wheelsets in the trucks with Kato wheelsets, to try to eliminate that as a source of noise.
I dis-assembled / re-assembled both mechanisms, using exactly the same techniques and care for each.

-------------------

/>
1st (good) and 2nd (noisy) are Atlas B40-8 mech’s with Atlas C628 trucks for my home-built RSD-15s.
I dis-assembled / re-assembled both mechanisms, using exactly the same techniques and care for each.
This noisy 2nd mechanism is especially disappointing, because I have to custom-build and grind clearances to fit the trucks and Model Power RSD-15 shells, I can't just go out and buy a quiet-running replacement.
The noisy 2nd Atlas B40-8 mechanism was perfectly quiet before I modified it.  The build of the 1st mechanism went perfectly, runs quietly.


I’ve been doing dis-assembly / re-assembly of N scale split frame diesel mechanisms for 10+ years now….. so I’m not a neophyte at it, I believe I take excellent care in trying to maintain the proper alignment upon re-assembly, including the worm bearing block orientation. 

The worst offenders for me are original Atlas/Kato drives from the mid-1990’s.  Seems like when I disassemble a formerly silent running Atlas/Kato mechanisms from that era, upon re-assembly, I never seem to be able to get the silent running characteristics back.     

There must be something more happening here.   I carefully test each step of the re-assembly, listening for the sound of the whine.   

What I am missing?  What is the foundational knowledge of what the root cause(s) are, and how to engineer around them.   Assume the bearings/worms are all lubricated properly.

What causes that whine?  It is drivetrain vibration caused by mis-alignments somewhere, leading to the metal mechanism acting like a “sounding board” that is causing the “coffee-grinder”?

For reference, here is disassembled shot of the noisy IM SD45-2 mechanism:

   

http://www.pbase.com/atsf_arizona/image/164747597

I feel like I’m missing some fundamental knowledge of what and why.   Right now I (and I suspect most of us) do “trial and error”….. but I don’t know how to scientifically (and with what tools?)  determine exactly what the specific reason for the vibration-caused "coffee-grinder" sound in a specific noisy N scale diesel split frame mechanism. 

After years of “trial and error”, despite finding and applying the before-mentioned Bearden fixes for 4-axle diesels, worm bearing alignment, worm bearing fingers, etc..... and searching for this knowledge, I’m still  not finding these specific answers.    Because of this re-assembly problem, I am hesitant to ever dis-assemble any of my locos' mechanisms - I never know whether it's going to be quiet or not upon re-assembly.   

Those of you who have large fleets of plug-n-play DCC-decoder equipped locos have found the answers to the above? 
 

————————


2)  What *scientific diagnostic tools* are available to us, to isolate and identify the cause of noisy “meat-grinder” N scale diesel split frame mechanisms?

Are there any specialized audio tools, using stethoscope, laser alignment tool, miniature rotating mechanism balancing techniques, or other tools to isolate and identify the cause of noisy “meat-grinder” N scale diesel split frame mechanisms?

For me, trial and error is taking too long and too much effort.  My basic problem is, I can’t determine the root cause of the “meat-grinder” sound.

——————————————————


3)    Finally, what is the *science*, the engineering skill set or discipline, behind *designing* and *mass-producing* really silent running split frame Atlas and Kato N scale  split frame miniature diesel locomotive mechanisms?   

For Kato in Japan or the Atlas Chinese factory, what do they do to deliver in mass-production, such silent running N scale diesels mechanisms?

Do they have some “factory drivetrain balancing” assembly techniques, diagnostic tools, or other “secret sauce”?

What is the name of the engineering design science behind engineering and manufacturing silent smooth running miniature electric motor drivetrain mechanisms.  Where are documented, the fundamental principles and expertise in this branch of engineering? 

What are the Google search keywords to find information on the engineering discipline and expertise that does this design work?

After years of searching for this knowledge, I’ve still not found these specific answers either.   

——————————


As you can tell, with some time off over the 2016 Holidays, I finally took the time to document and ask these long-standing questions of mine.

Let’s have a good discussion on this topic here, which will greatly benefit all of us.    We will store the answers for posterity.

I know we have here some stellar excellent N scale locomotive tuners over the years ( you know who you are     )  who I hope will be willing to shed some light on this issue.

Happy Holidays, all.

[randgust]

John, you've done most of the basics.   I particularly noted that you've got indicator marks all over the place for motor and part orientation (what way the motor was, which truck was which, which way the truck was pointed, on and on.    I've become absolutely paranoid on that stuff when I disassemble, numbering bearing blocks and putting arrows on them, identifying drive shaft and direction, everything.  Looks like you have too.

And you're not nuts.  Bluntly I haven't seen much of it on six-axles, but I'm still plagued with it on older Atlas Classic GP's, even with the bearing block fix.  I have noted that the new Atlas universals on current production GP's are a big improvement, so much so I finally bit the bullet and replaced two mechanisms I've never been able to quiet down with the newer ones.  I'd concluded that those old hex-nut universals were the worst part, and I've also concluded Atlas must have agreed with me.

You're getting some real racket particularly without adding shells.   Mine usually don't go nuts until after I add shells and you can see or jostle anything.

I'm still of the school that fingers and ears are pretty good tools but I use a long pin or needle to prod inside a mechanism to feel for vibration and also to move/brace/shim things to see what difference it makes.

You know we've both grown up as IT guys, and the rule on problem solving is always 'reduce the variables'.    About the only thing I don't hear you discussing is leaving out a worm during reassembly to try to determine which end/truck is producing the racket.

The other technique I've done is to leave the trucks off and clip to the frame halves for power.  You can reach in the bottom of the frame with a piece of stripwood and introduce independent load/drag to the worm gears (as well as push and prod the worm bearings around) while it is running to see if changing anything in terms of harmonic vibration.   Object there is to see if universals, flywheels, worm bearings, anything in the top drive chain is doing it, and you can do that with the frame fully assembled and running (and even in the shell).  Dragging the stripwood on the worm itself with the truck off, while changing direction, etc., lets you independently test each end of the drivetrain under independent load.

Oddly enough, I've discovered that those 'fingers' on the worm bearings sometimes have to be looser than you'd think and that relationship isn't as 'true' as I used to imagine. 

Back in the day of Rapido GP's and cup gears I was messing with shimming with paper strips and stripwood to try to get things aligned quieter and this was the only way those darn things would run - inside the shell.

We're assuming the trucks themselves aren't doing it, but most times it won't make the racket unless you're introducing some load on the drive train itself.

If that still doesn't do it, pull the driveshafts out and just reassemble with the motor in the frame halves and try it.   One thing that does happen, whether you like it or not, is that nylon/plastic motor carrier gets worn/distorted as you remove it in and out of the frame halves repeatedly.   If you get down that far and the vibration is still there, that's about all that is left to blame if the motor was originally quiet when you started and you haven't dropped it on a concrete floor to whack the flywheels.

[atsf_arizona]

Great ideas, Randy, and worthy of further experimenting, which I will do over time, and report back here.

Further comments welcome, especially on:

• "Here's where there's documented experience / expertise on designing / debugging miniature rotating electric motor mechanisms for silent operation"
• "Here's my experience after putting in upteen numbers of plug-n-play DCC decoders", or
• "Here's what tools are avail to scientifically determine where the problem is"?
• "Here's how Kato or Atlas do the design engineering / manufacturing to produce whisper-quiet mechanisms in mass production"

Thx.

[daniel_leavitt2000]

I hate, HATE the Atlas HTC trucks in general. I think It was a sloppy, poor design that really hurt the SD series of engines. A few of the major gripes:
They are ugly and do not look like the prototype.
Too many small gears add friction where one larger gear would have worked better like the Atlas/Kato Flexcoils (and later the GE trucks as well).
The gears do not mesh well.
The on center worm and pivot leads to the trucks jumping more than the offset Kato Flexcoil.
Fit and finish are pretty bad.


One thing I noticed from your video is that the buzzing correlates to the motor speed and not the wheels. So I would start there.
1. With the trucks off does it still buzz when powered? If so, you are looking at binding at the motor, cradle, universal or worm.
2. If it is quiet with no trucks, does it make noise with one truck? What about the other?

If I had to guess, I would bet the motor cradle is just a bit unseated. Not enough to be noticeable, but enough to throw the universals off. My second suspicious would be that the bronze bearing springs are pushing the worm and universal off center just a tad and that would cause binding against the other frame half. This would also explain noisy Kato and Atlas Flexcoil equipped engines.

[atsf_arizona]

Daniel, nice to hear from you, and thx, more good ideas for me to try.  I will report back here after doing so. 

All, in case you haven't seen the Ron Bearden multi-page thread on fixing the Atlas HTC trucks, here is the link, direct to the page where the wheelsets are discussed: 
http://forum.atlasrr.com/forum/topic.asp?ARCHIVE=true&TOPIC_ID=42265&whichpage=14
and another here on TheRailWire:  https://www.therailwire.net/forum/index.php?topic=35167.0  )

Tons of good info.  On my noisy IM SD45-2 above, per this Ron Bearden Atlas Forum link.....  I did replace the stock IM wheelsets on that loco with the Kato 932090 wheelsets. 
(Hint:  buy them from MB Klein here:  http://www.modeltrainstuff.com/Kato-N-Low-Profile-Geared-Wheel-Sets-p/kat-932090.htm  )

Excellent article on how to replace those wheelsets in the IM / Atlas HTC / SD60 truck is here:  http://mysite.mweb.co.za/residents/grela/atlaskatotruck.html   by Andre Kritzinger of South Africa)....

Unfortunately, replacing the wheelsets, didn't solve the buzzing problem on that IM SD45-2 mechanism. 

So yes, it "appears" the problem is alignment up in the drivetrain.  I will consider how to methodically apply the tests already given to me in this thread, to these noisy mechanisms.

            (JS spoiler alert:   Go forward to page 6, reply #76 to see what we finally found the problem to be)

-------------

Relevant side topic: 

Surely the engineers that designed Kato, Atlas mechanisms, or similar mechanisms in other fields/products, must have some deep expertise / knowledge / tools they used, documented somewhere.   

In fields such as modern medical small implant pumps, or small watch making, or small mechanical devices.... somewhere there must be education or engineering expertise that can be read on the internet, that applies to our model RR mechanisms?

• Does anyone have any experience or tools that can be used to identify where a mis-alignment might be, or the sound originating from, in a small rotating mechanism?  I can't believe that "trial and error and see what works" is the only way?   
• What sort of engineering skills and engineering disciplines do Kato (Japan) or Kader (China  https://en.wikipedia.org/wiki/Kader ) design engineers come from?

[u18b]

John,

Having been a performance nut for almost 30 yrs..... here are a few observations.

As far as the science goes, quietness (assuming a clean loco with no trash) comes from (not all of these apply to disassembly/reassembly issues):

1.  perfectly matched bearings (in some bearings, the hole is a few microns too big- allowing for wobble)
2.  perfectly straight shafts.  We can never assume.  Bob Horne and I spent a lot of time with one of his B40-8s.  Final conclusion was it had a slightly bent motor shaft.  There was no other possible explanation.
3.  Motor shaft and worm shaft in perfect alignment (especially in shorter locos).  This was the primary issue I discovered in disassembly/reassembly issues in four axle locos--- thus leading to my inner bearing block removal article.  I ALWAYS remove the inner bearing block in my locos.   Not all believe in this (as has been discussed in various threads), but my own results have been overwhelmingly and spectacularly positive.
4.  perfectly ROUND/balanced parts.  To my it is odd when a 6 axle loco with a u-joint is noisy.  One possible explanation is the u-joint piece itself is not perfectly round and balanced.  When there is a problem in either, the spinning action will create wobble.  Anybody with Samhongsa brass locos notice the sprue marks/bumps on the drive shaft.  Those drove me crazy.  I cut them all off making sure the shaft parts were as smooth as possible.  Maybe it didn't matter, but I did it anyway.
5.  bearings that are perfectly 90 degrees to the shaft.   When manipulating things, and a bearing is twisted even slightly, friction and noise go up.
6.  Bearings that are perfectly secure.  When they are loose, they can rattle.  Horrible examples include the old Kato GP50  (got quiet as soon as I locked the bearings down), some Kumata brass locos (even Kumata would place expoxy to lock those worm bearings down in longer locs like E units), and yes, also some Atlas locos with plastic worm bearings where the bearing was too loose.
7.  gears that float too much.   The greater the precision of gear mounting, design and plastic quality, the better.   The HTC truck by Atlas has been mentioned a couple of times in this thread.  Yes, I traced a huge noise/performance issue in Atlas SD50s and 60s to a gearcase that allowed that top gear to float too much.   I could be wrong, but almost all of the complaining came from the first year or so of those locos.  I have not heard the same complaining in years.   I think those first run locos had a poorer quality plastic that changed over the years.   
8.  bearings that are symmetrical.   I never did any measurements with a calliper, but there are times when working with some Atlas and Kato plastic worm bearings that I could SWEAR that the bearing outside was not perfectly square.  How else (besides damage) can you account for a bearing that is loose in one direction, but rotate it 90 degrees and now it fits more snug in its holder?
9.   parts that are undamaged. Obviously cracked gears are a  nightmare.  Kumata locos are also prone to having their u-joint cups crack over the years.   But aside from the obvious, I think there can possibly be a subtle issue.    I'm convinced that part of our problem with the move in the last 20 years to plastic worm bearings is that the process of removing the parts and re-installing them can place pressure on the bearings.  If the inner race of the bearing is damaged in ANY way in this manipulation, then the loco is not going to run as well.  It is NEVER a good thing to force something since this kind of damage can result.  By the way, another example of parts that get damaged over time is the motor mount itself.  That plastic motor saddle that Atlas and others has used is perfect when new.  But on your FIRST disassembly, you damage to tips of the four prongs that hold it to the frame.  I doubt that this may cause much noise.... but changes in those prongs can potentially effect the perfect alignment and orientation of the motor.

To be honest, I think one of the very best bearings ever made for our N scale locos was the combo bearing Kato put in the locos made for Atlas in the early days.  Y'all remember that bearing that was a phosphor bronze/brass ball (or doughnut) mounted in a plastice housing?  The housing allowed the bearing to self-align.  I miss that bearing.

10.  Motor properly locked down.  Ajin has done a TERRIBLE job in this regard with some of their later locos in N.  Check out my discovery of how bad the brass little Joe is.   I am shocked that loco got out the door.    Most little Joe's run terrible.  By the time I get through with them, they run like Kato locos.

11.  LATERAL slop in worm shafts and u-joints.  I have found that noise can come from worm parts moving toward and away from the motor.  Sadly, I think this is a problem with some long 6 axle locos like: SD50, SD60, SD40T-2, SD45-2.  Normally we think of noise coming when a shaft or part moved up and down in reference to the motor shaft.  But if a shaft can move IN and OUT of a flywheel (for example), then noise can happen there too.  Over the last almost 20 years, there have been people who just cannot believe that my Inner Bearing Block removal works.  They try it with poor results.  On many occasions, I have had people sent me their loco so I can look at it.  99% of the time, the person has not paid enough attention to the placement of the hex nut that goes into the flywheel.  It must be placed perfectly so that it ALMOST bottoms out inside the inner wall of the flywheel.  Hex nut too far out on the worm shaft- a bind is produced.  Too far in, and too much lateral slop  is allowed.

3, 5, 6, 8, and 9 above may all apply to disassembly/reassembly issues.

This is a lot.  I'll probably think of more later.

[tehachapifan]

One thing I've noticed is that bearing blocks are not perfectly square in the inner "tread" area (if you think of it as a square, pulley-type wheel). And/or, the smooth vs. louvered-looking surface of that tread area can affect how things align depending on how they're orientated in the bearing block holders in the frame.....or if they are not all orientated the same way as each other.

[atsf_arizona]

Ron Bearden (U18B), more great tips (your 30 years of experience much appreciated).  Thx for chiming in.

One thing so far.... seems all of us still have to do "trial and error" to locate the problem. 

Is there no way to instrument these small mechanisms?  How do the Japanese and Chinese factories do quality checks?

[u18b]

Ron Bearden (U18B), more great tips (your 30 years of experience much appreciated).  Thx for chiming in.

One thing so far.... seems all of us still have to do "trial and error" to locate the problem. 

Is there no way to instrument these small mechanisms?  How do the Japanese and Chinese factories do quality checks?

A calliper would help in a couple of this items I mentioned.

Max is a HUGE believer in a volt/ammeter to indirectly tell if something is wrong.  I have often thought I should follow Max in noting the current draw before and after I work on a loco.

But yes, you are correct.  Trial and error (and LOTS of time) have been my method over the years.

[tehachapifan]

Ron, you posted your initial reply just before mine and noticed you did touch on the bearing symmetry too. I think this might be a culprit. I notice that bearing blocks may appear perfectly square as you look at them from the sides with the holes but that this in no way indicates how symmetrical the inner surfaces of the bearings are. I recently had a problem after reassembling a loco, where it was binding and where a truck actually kept falling out, until I flipped all the bearings 90 degrees. I think it's vital to note this orientation of these blocks during disassembly (as in, is the smooth or louvered-looking side facing out when you take apart the frame). I also notice that these bearings will sometimes rotate themselves if you lift a worm a little as you are manipulating everything into place.

[atsf_arizona]

It occurs to me that I should take my perfectly running mechanism (still assembled of course), and check it side by side, the play / tightness / alignment with my noisy mechanism.  See if I can find a difference.

Randgust's ideas of using a use a long pin or needle to prod inside a mechanism to feel for vibration is a good idea I've not been told of before.   As well as try gently prying / shifting the worm or the bearing blocks while the mechanism is running, seeing if a difference occurs. 

Randgust's and Ron Bearden's input that:   " Trial and error (and LOTS of time) have been my methods over the years"  puts my mind to rest, at least I know that others have been down this same time-consuming and frustrating road before me. 

I still have to believe that  design engineers of miniature mechanisms must have alignment and diagnostic tools of some sort.   (Laser?)
How else could they do design work and manufacturing quality checks? 
What would I Google to try to find internet info on that topic / engineering discipline?   (I know, I am harping on that point). 

[jagged ben]

Do the factories assemble everything on a jig that keeps things aligned?  When I reassemble by hand, I have a tendency to splay the bottoms of the frames apart, with everything assembled inside, in order to slip the trucks in.  That can't be good for alignment of the motor with everything else.  I'd imagine the factories have a better method of holding the trucks in place during assembly.

[jagged ben]

I'll also say this:  it's hit and miss with factory assembly sometimes, especially Atlas China.   And I've had plenty of locos go back together just fine, especially Katos.  I'm not an obsessive about loco noise though,  as long as it's not noticeably affecting performance.

[atsf_arizona]

Jagged Ben,

I wondered that too, do the factories use jigs? 

I've noted that compared to mid-1990's Atlas diesels, today's Atlas N scale diesels are almost always whisper quiet right out of the box.   What does the factory do different now, that they didn't do back then?

How does Atlas manage to get it's 4-axle diesels to be whisper quiet from the factory, but not require the "Bearden fix"?   Is it because their assembly components are brand new?  Or is it also because they have some sort of alignment quality control procedures. jigs, or ?

And if they do get a bad batch of locos that don't run quietly (and it's not due to lack of lubrication), how do the manufacturing engineers identify the problem and then systemically fix it?

Also,  today's Kato SD70Ace's not only run whisper quiet, but they also will COAST better than most of the other non-Kato diesels we've seen.  And we all remember that original Kato U30Cs and SD7/9s from 1992 or so (25 years ago!) would also COAST.   Why?   What does / did Kato do in assembly, that makes them so much different?

In the answers to these questions, I'm hoping we all will find the answers we're looking for in this thread. 

Comments welcome.

[jagged ben]

Jagged Ben,

...

Also,  today's Kato SD70Ace's, not only run whisper quiet, but they also will COAST better than any Atlas or Intermountain diesel I've ever seen.  And we all remember that original Kato U30Cs and SD7/9s from 1992 or so (25 years ago!) would also do this.    Why?   What does Kato do in assembly that makes them so much different than Atlas or Intermountain?

That is not assembly, that is truck gearing design.  Take any dirt-free truck out of a Kato tooled since at least (and including) the first Dash-9s and try rolling it by itself.  It will roll nearly as free as an ungeared freight truck.  You can't say the same for any other manufacturer.  Most of them will not roll freely at all, while some will roll a couple inches if you apply then perfect amount of force.  My conjecture is that Kato just has built much more play between the gear teeth.