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

[ncbqguy]

John Sing wrote:

"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".   

John brings up an interesting topic.
In my 25 years in the Model Railroad Industry I have worked with a number of factories that produce our trains.   I have a great deal of respect for their knowledge, expertise and professionalism in the creation, research, design, engineering, manufacturing, production and assembly and maintaining the business involved in bringing models to us. 

However, all of them exhibit serious weaknesses that all of the importers struggle with to some degree or another.....

Almost none of the people involved in the factories are Model Railroaders...and even those that are do not fully appreciate how we use their products.   When I was with Kato the Kato USA staff went out of the way to bring visitors from Japan to Operating Sessions at local Model Railroads to attempt to educate them that the cars and locomotives are not just run on a loop of track with all one brand of track, equipment, etc..   Model Railroading in Japan and Europe is radically different than in North America. First of all, most North American Model Railroaders are adults over the age of 35 whereas in Japan the main market (although diminishing rapidly because of electronics) is a teenage kid.  The demographics in Europe are closer to ours.  Second, we accumulate our trains and keep them as opposed to Japan where the kids outgrow them.   Space restricts permanent layouts and large collections in Japan and that is somewhat true in Europe as well.   I have some locomotives that still run that I purchased in the late 1960s.   Our use patterns are different than Japanese and Europeans as well....we like to run long, heavy trains and expect locomotives to run well with other brands and pull their share of the load. Operating sessions can run from a couple of hours to constant running all weekend at a show.  Having layouts where rails have to be replaced due to wear and locomotives need replacement wheels, pickup wipers, motors, etc. due to worn out parts is not unusual.  The concept of producing parts for repair purposes on a long term basis is not well understood by the factories and this makes it more difficult for the importers to stock repair parts for models not in current production.

All the importers get excellent customer feedback at shows, telephone, email and snail mail contact, and by monitoring forums.  The staff at the importers understand and appreciate the input, but incorporating customer experience into product is an uphill battle.  Engineers like to reinvent the wheel with each project.   Even within a company a proven design will be ignored and something different tooled, even against the request of the importer client.   The often cited Japanese concept of Kaizen....continuous quality improvement....seems to only apply to things that were working perfectly well and deftly ignores well documented problems.   

Some solid design concepts were proven many years ago but are ignored even by the company that originally came up with them. Micro-Ace has built models of just about every Japanese prototype steam locomotive using the same engineering pioneered in the revised Con-Cor J3a Hudson of the early 1970s....a simple, bullet-proof design that pulled well and has yet to be equaled after decades of more sophisticated and complicated US steam engine models brought in by every importer.   Even the superior split frame diesel mechanism and low friction electrical pickup concepts are sadly not universally used.  Not to say there are no new ideas that may be better...but so far these successful concepts have not been improved upon.

Model Railroad products require a degree of precision in design and manufacture equaled only by the Medical and Aerospace industries.  Add to that the overall degree of sophistication of the Model Railroad consumer (rivet counters) makes it a difficult business to be in.  So next time you are at a show talking to the representatives of the importers, realize that they, like most of you in your jobs, don't have total control over the process and try their best to bring in products that satisfy the consumer and do want to hear your constructive comments.   

Charlie Vlk
Railroad Model Resources
     

 

[NorsemanJack]

I was merely suggesting that maybe "something moved"- hence a slight alignment issue.

I still don't think that lube present or not present on new trucks would do what you've described.... but of course I might be wrong.

Looking forward to your report when you get some grease to add to the trucks.

Reporting back.....

I obtained fresh bottles of both LaBelle 102 (medium/gear lube) and LaBelle 108 (light).  I completely disassembled both 25 year old E unit mechanisms and applied lubricant as follows.  Keep in mind, those LaBelle lube dispensers have the real nice long metal "needle point" applicator, that makes is possible to apply the smallest of drops direct from the bottle.

a) one small drop of LaBelle 108 on each motor output shaft between motor and flywheel
b) wiped each worm shaft with a microbrush moistened with 108 in the areas where they interface with the block bearings
c) using a moistened microbrush, wiped the side of each block bearing with LaBelle 102
d) wiped inner bore of flywheels with 102
e) applied a miniscule drop of 102 on each trucks gear train and ran them back and forth on a test track to spread to other gears
f) after partial reassembly, applied a drop of 102 to each worm gear

Both units are now whisper quiet, the rival of any brand new Kato in my fleet.

This is just a theory to consider:  I am beginning to suspect that the reason some units run quiet before disassembly, but noisy afterwards, is that "dry" plastic components will not slip or float as well as they may want to in order to "find home."  Hopefully that explanation is clear.  The plastic worm bearing blocks and the drive shaft "nut" that engages the flywheels are designed to "float" a bit within the mechanism.  It may be that fresh units, from the factory, readily enable this and once they find home through a small amount of running, will stay quiet over time.  Take them apart, and you restart the process.  To Ron's earlier point, although I did not disassemble the one mechanism, I did remove the worm covers and install new trucks.  This would almost certainly result in some movement in the worm bearing blocks (unless I had the worm and mating truck gear teeth perfectly aligned).  Just a thought.....

[mmagliaro]

The day Kato (and others) stopped using the plastic bearing blocks that had the bronze spherical bearings inside them was a big step backwards.  (Now they are just simple plastic blocks with a hole in the middle).  The spherical bearings could rotate inside the block, and would guarantee that the shaft would "find its home" as you put it, and then run with no wobble.  Too bad they don't use those anymore!

[peteski]

John Sing wrote:

"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".   

John brings up an interesting topic.
In my 25 years in the Model Railroad Industry I have worked with a number of factories that produce our trains.   I have a great deal of respect for their knowledge, expertise and professionalism in the creation, research, design, engineering, manufacturing, production and assembly and maintaining the business involved in bringing models to us. 

However, all of them exhibit serious weaknesses that all of the importers struggle with to some degree or another.....

Charlie, thank you very much for this insight into what goes on behind the scenes at the model railroad manufacturers.  You confirmed my speculations of why even new models are not using proven mechanism designs, and why Kato for example tries to "outdo" themselves with "better and better" more advanced mechanisms which are mechanically superb, but actually too delicate and failure-prone in actual use.  It is too bad that more model RR design engineers do not follow the most logical path of using the best, but not over-complicated designs.


Norseman Jack: You seem to only use oils when you are servicing your models. Not that there is anything inherently wrong with that, but IMO while the light oil will work well inside shaft bearings, it will not provide adequate lubricant film on open surfaces which require lubrication. When I tune up my models I use both, oils and greases. To me oils are best suited for rotating shafts, while grease type of lubricant is best suited for lubricating worm and gear teeth and other slow-moving parts.

I use the Woodland Scenics lubricant set.  Light oil on motor and worm shafts and on any other gear shafts.  Then I use the Teflon Grease on the truck pivots and on warm and gear teeth.

Of course in our tiny models (where forces applied to the mechanisms are minuscule) using proper viscosity lubricants is not an absolute requirement (especially for the slippery plastic gears), to me using the proper lubricant for a specific application makes the  most sense.

I would also like to nominate this thread for the "Best Of" status.  It is long overdue.

[NorsemanJack]

Noseman Jack: You seem to only use oils when you are servicing your models. Not that there is anything inherently wrong with that, but IMO while the light oil will work well inside shaft bearings, it will not provide adequate lubricant film on open surfaces which require lubrication. When I tune up my models I use both, oils and greases. To me oils are best suited for rotating shafts, while grease type of lubricant is best suited for lubricating worm and gear teeth and other slow-moving parts.

I use the Woodland Scenics lubricant set.  Light oil on motor and worm shafts and on any other gear shafts.  Then I use the Teflon Grease on the truck pivots and on warm and gear teeth.

Just to clarify.  I used Labelle 102 on the gears, which is marketed as "gear lubricant" (medium viscosity) and also contains Teflon (i.e. PTFE).  I'm very conservative with lubricants, and concerned that a cure of "too much," or "too thick" is worse than the disease.  To be honest, these are the first Kato locomitives that I've ever applied even a drop of lubricant to; and I have many dating back over twenty years.  As I suggested earlier, I believe that the main reason those retrofitted mechanisms were making a "slight" coffee grinder noise was that I used factory replacement trucks that likely had zero lubrication of any kind.  I guess the proof's in the pudding!

https://labelle-lubricants.com/shop/labelle-102-is-an-extremely-versatite-gear-lubricant/

[peteski]

Thanks NorsemanJack.  I would still use Labelle 106 grease for things like truck pivot points.  I guess it is just a preference.

[NorsemanJack]

Even the superior split frame diesel mechanism and low friction electrical pickup concepts are sadly not universally used.  Not to say there are no new ideas that may be better...but so far these successful concepts have not been improved upon.

Model Railroad products require a degree of precision in design and manufacture equaled only by the Medical and Aerospace industries. 

Charlie, thank you very much for this insight into what goes on behind the scenes at the model railroad manufacturers.  You confirmed my speculations of why even new models are not using proven mechanism designs, and why Kato for example tries to "outdo" themselves with "better and better" more advanced mechanisms which are mechanically superb, but actually too delicate and failure-prone in actual use.  It is too bad that more model RR design engineers do not follow the most logical path of using the best, but not over-complicated designs.

Yes, thanks Charlie for weighing in.  You likely understand more about this business than almost everybody on the forum (combined).

That said, I'm guessing that your comments regarding the "superior split frame design," are relative to some of the manufacturers who still use more archaic solid frames (e.g. the recent BLI F units, which are the topic of lively treads on several forums).  I would contrast this with Kato's non-split frames in the E units, PA units and latest F units (among others).  Compared to the split frames, I consider these superior and a great "next step" in design evolution.

I'm guessing the comparisons with Aerospace and Medical might have been a bit of hyperbole.  I spent decades in those realms, and the day that our trains require those levels of precision and quality control will be the day that the hobby disappears because nobody will be able to afford anything.

[peteski]

Yes, thanks Charlie for weighing in.  You likely understand more about this business than almost everybody on the forum (combined).

That said, I'm guessing that your comments regarding the "superior split frame design," are relative to some of the manufacturers who still use more archaic solid frames (e.g. the recent BLI F units, which are the topic of lively treads on several forums).  I would contrast this with Kato's non-split frames in the E units, PA units and latest F units (among others).  Compared to the split frames, I consider these superior and a great "next step" in design evolution.

Well, the "non-split-frame" design is really going back to the early days of N scale. Early locos used that design.

Kato's version of that design is vastly superior over the early designs.  They use many advanced features (some from their split-frame design) like wire-free electric pickup using the springy electric pickups for trucks, a "power module" cradle housing the motor, light board and electric pickup strips and a screw-free design.  BLI and other manufacturer's versions of the non-split frame is usually a rats-nest of wires, and uses screws to hold items together.

[NorsemanJack]

Kato's version of that design is vastly superior over the early designs.

Yep, exactly my point.  Those Kato E/PA/F one piece mechanisms are a marvel of modern engineering.  Simple/precise/reliable = great value.

[atsf_arizona]

Hi, all,

I updated my posts above with a new Google Drive URL link to the Atlas / Atlas-Kato / Intermountain loco parts cross-index spreadsheet on page 11 of this thread:

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

For whatever reason, the original 2017 January link stopped working, but the above 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)

Please let me know that the link works for all of you, if there's a problem I'll keep trying to fix it.

Below is excerpt of what that spreadsheet is about.   

Happy Holidays 2019 and Happy New Year 2020!

==========

One of our the major discussions has been:

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

To answer that question, I collected the part number information from the Atlas and Intermountain parts websites, and created an downloadable Excel spreadsheet which has drop-down Sort menus allow you to find and correlate part numbers across almost all the N scale Atlas/Kato, Atlas, and Intermountain SD45-T2 / SD40-T2 locomotive types. 

This allows you to answer for yourself :

*By locomotive type,  what part number for  motor, motor saddle, trucks, etc?

What became apparent when building this spreadsheet, is there are some notable "exceptions" to what you would expect to be a compatible part. 

• Example:  according to the instruction sheets, the DCC-ready Atlas GP30/35, GP7 and GP9 use a different motor saddle than the commonly used p/n 480001.  I verified and found this to be true in a Atlas 48082 Santa Fe GP7 #2692 some time ago.  I swapped in a older fast-speed motor into that GP7 DCC-ready mech, to speed match the loco with my older fleet - and found that the standard motor saddle p/n 480001 plastic fingers are spaced *differently* than the GP7 DCC-ready mechanism motor saddle holes.  I had to "force fit bend the plastic fingers" to match the GP7 DCC-ready mechanism (at the time, I had no idea these particular newer DCC-ready mech's used a different motor saddle).

These exceptions are especially worth knowing to save time and effort, not to mention possibly being cause of a hard-to-shoot bug in a noisy locomotive mechanism.

Here's screen shots of the spreadsheet, sorted by locomotive.  (Downloadable link is at bottom of this post). 

The bolding and colored cells are some of these 'unexpected exceptions' to the part number you would expect to be compatible.

(P.S.  Updated the Google Drive spreadsheet copy on 12/22/19 to reflect new URL prefix that Atlas started using for the PDF parts diagrams, compared to the screenshot)



Here's the link, clicking it should take you to Google Drive screen where you can view / download the spreadsheet:

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

(8:40PM EST 12/22/19 - uploaded new version that fixed the URLs to point to Atlas' new location of loco Parts diagrams)

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

I trust the above is helpful to us all.  I haven't attempted to update the spreadsheet since Jan 2017, but I trust there's still value in it.

Happy Holidays 2019 and Happy New Year 2020!

[u18b]

I wrote and submitted an article over 5 years ago to N-Scale Magazine on how to diagnose loco noise.

Maybe they will publish it one day.

Thanks John for your great work on this.

[atsf_arizona]

I wrote and submitted an article over 5 years ago to N-Scale Magazine on how to diagnose loco noise.

Maybe they will publish it one day.

Thanks John for your great work on this.

Ron, good to hear from you.  Yes, I really hope Pam and N Scale Magazine would do that .   Thx for all you do for our N scale knowledge base.
Happy Holidays.

[MK]

Happy Holidays John and good to see you here.  That spreadsheet is very helpful and will come in handy for any future work or research.  Thank you!

[randgust]

John- do you have any similar referencing of the change in the GP-series (any B-B Atlas) of the 'current issue' universals for part number and usability?    I haven't been able to tell which are which other than I know the new ones (which are more like a Kato design than the old hex-nut) are WAY better and quieter.  Current GP7/9 Master universal part number is shown as 9480003 but no image.    If that were the newer universal, that's probably the single biggest improvement to noise prevention in those yet - I ended up retiring most of my Classics just to get the new chassis with that slow-speed motor and that universal, and they are really quiet.

[atsf_arizona]

John- do you have any similar referencing of the change in the GP-series (any B-B Atlas) of the 'current issue' universals for part number and usability?    I haven't been able to tell which are which other than I know the new ones (which are more like a Kato design than the old hex-nut) are WAY better and quieter.  Current GP7/9 Master universal part number is shown as 9480003 but no image.    If that were the newer universal, that's probably the single biggest improvement to noise prevention in those yet - I ended up retiring most of my Classics just to get the new chassis with that slow-speed motor and that universal, and they are really quiet.

Hi, Randy,

1) No, I've no knowledge of a newer universal, until you pointed it out. 

I see that Spookshow's N Scale Locomotive Encyclopedia has picture of this new universal which he says was started to be implemented circa 2016   http://www.spookshow.net/loco/atlasgp79.html

At the URLs  below, the hex nuts on the flywheel end of the worm shafts were removed and replaced instead with U-joint connectors:

http://www.spookshow.net/loco/files/atlasgeepmod4.jpg
http://www.spookshow.net/loco/files/atlasgeepmod3.jpg


2) I just now checked the most recent Atlas GP9 that I have, road number #702 GP9 Santa Fe Zebra Stripe, which were announced in mid-2014 and released in 2015 July.  Thus it probably has the older universal.  For what it's worth,  I opened the box and looked at the loco parts diagram in it. That parts diagram still says part number 480003 as the universal.

3)  I'd be interested to know if someone who has a Atlas GP7/9 run released after 2016 (or any Atlas loco released after 2016), has the above newer universal... and if so, if their jewel case has a newer parts diagram with a newer part number for this universal.

See this URL for when various Atlas N Scale loco roadnames/locos were released:  http://archive.atlasrr.com/n-mcloco-archive.html
The N scale Atlas GP7/9's are listed here:  http://archive.atlasrr.com/NLoco/arc-ngp7.htm )

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

4) It wouldn't be the first time that Atlas or any other company has a "slightly newer, slightly different part" assigned to the same part number.   
I've found this to be the case with motor saddles, part number 480001.  See below.   

I learned that when swapping fast speed motor (came in 1995 era mechanism) and scale speed motor (came in DCC ready 2014 mechanisms) between Atlas GP7/9 frames, I found:

-  Keep the same motor cradle (Atlas PN 480001) with the same *frame/mechanism*! 

   - i.e. when swapping motors between these two generations of GP7 frame/mechanism....
   - ==>Keep the motor cradle with the frame/mechanism it came with.  <==
   - just swap the motor / flywheel assembly itself -  i.e. snap the motors out of the cradles, swap just the motor / flywheel assembly

-  Even though old / new parts sheets specify the motor cradle as the *same* PN 480001  ( I checked )

-  I found my old 1995 era version of motor cradle p/n 480001 tabs, are spaced *slightly wider* than 2014 DCC ready mech's motor cradle

-  I.e : the  older 1995 cradle's  "tabs" are slightly wider apart than the newer 2014 DCC ready mechanism holes. (these are the tabs that go
   through the mechanism holes, and hold the motor in place).

-  The older 1995 mechanism's slots are wide enough that it doesn't care, it accepts the newer 480001 cradle's *narrower* tab spacing w/ no problem

-  BUT the *older motor* cradle p/n 480001 from Fast Speed motor 1995 era Atlas mechanism.... DOESN'T EXACTLY FIT into newer 2014 DCC ready mech.

-  I was able to "cram" this older fast motor 'wider spaced' cradle into the newer 2014 mechanism, because the plastic tabs would bend inward far enough without breaking (the motor cradle plastic is flexible enough)
-  But it's far from ideal fit, the older fast speed motor cradle fit is hard to bend into place, takes lots of 'finger grease' to force it into place
   on the newer 2014 mech...

-  Dis-assembling this loco mechanism with the ill-fitting motor cradle later on is to be avoided if possible - while it could be done, it'd  be
   problematical and the 'finger grease' bending the tabs would have to be redone.   (This mech is in my Atlas ATSF GP7 #2692 )

-  SO =  keep your Atlas GP7/9 motor cradle with the generation of mechanism that it came with !

Not surprising this could happen given 20 years between the manufacture of these two parts.

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

5) I also found that the frames of the non-DCC Atlas GP7/9 is slightly thinner than the DCC ready GP7/9 mech
(I was putting an Atlas GP7/9 mechanism under a LifeLike GP18 (original run) shell).

   I found the old 1995 era GP7 mech is slightly thinner than DCC ready GP7/9 mech

   - This is one of original 1995 GP7 mechanisms without the 'jog / notch' at the top
   
         (see: http://www.spookshow.net/loco/atlasgp79.html )

   - Thus, use an old 1995 era GP7/9 mech to fit under LL GP18 shell with only very minor grinding.  The newer mech would've been noticeably more work to get it to fit.
 
         Thx to a Gene Maddox post on old Atlas forum that put me onto this knowledge

- The newer DCC ready mechanism, is just a bit *too wide* for the LifeLike GP18 shell. 
   
   - So don't bother with newer GP7/9 Atlas DCC mechanism if you're trying to do this swap
        - get an older GP7/9 non-DCC ready Atlas mechanism for your LL GP18 shell to sit on


There's other things I learned in doing this swap.  Ping me if interested in knowing what those were. 

There turned out to be a lot more little time consuming things than "just slide a scale speed GP7/9 Atlas DCC-ready mechanism under the old LifeLike GP18 shell and you're done."  Nothing hard, but it took some time to:   

a) re-orient the scale speed motor for proper polarity direction of travel, in the non-DCC-ready older GP7 mech   
b) find + file down many little spots to get LifeLike GP18 shell to fit on the older Atlas GP7 mechanism
c) find and fix the inevitable "buzzing" that often happens when dis-assemble/re-assemble a N scale split frame locomotive.... i.e. what this whole thread is about.  I needed to "Beardenize" the GP7 mechanism, for example.

==========

I hope these tidbits help.