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

[peteski]

Tim, have you ever run into situation where the Kato 932090 wheelsets swapped into the Atlas/Intermountain HT-C trucks, caused the resulting truck to derail upon entering / curves? (12.75" radius)

I had exactly that problem with one of my Atlas/Intermountain HT-C trucks with Kato 932090 wheelsets (but no problems with the other one with Kato wheelsets).  I don't know what caused this truck to consistently derail (albeit, I didn't chase too much trying to debug it).  Because, just to test, I put the old IM wheelsets back in.... and the truck no longer derails.   Then found out that the noise problem was not in this particular set of trucks via this thread.

Those Kato wheels have much smaller flanges than other wheelsets. Is it possible that the derailing truck doesn't swivel freely in the chassis?  So when it enters the curve instead of swiveling it keeps going straight?  The small flanges aren't able to get it to steer it into the curve maybe and it jumps the track? Also did you check the gauge of both the wheelsets and the track?

[atsf_arizona]

Good morning, all,

I made a YouTube showing test method suggested by kiwi_bnsf (see reply #126) to determine if the mechanism itself is noisy or not.

I show testing these standard DC mechanisms :

* Intermountain SD45-2
* Atlas/Kato GP30  (mid-1990's era - still has the inner bearing block)
* LifeLike GP20 (with Ron Bearden inner bearing block removal technique)

The mechanism screws have been re-tightened to "standard level", after removing the trucks (so that the mechanism can sit directly on the test track).

As we can see, the IM SD45-2 and the LifeLike are basically dead silent at top speed 10-12 volts or so (the IM mechanism amazingly).

In contrast, we have identified that the 1994 Atlas/Kato GP30 is a growler. 

/>
---------

The point is, to be able to test the mechanism and drivetrain alignment for noise. 

If the mechanism itself is noisy / vibrates, solve that first;  only after that, should you move on to looking into the worm gear / gear tower interaction for noise. 

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

Safety-for-your-DCC-mechanism advice:

1)  Be aware that if you have a DCC decoder in the locomotive mechanism, the DCC decoder could be susceptible to getting fried in this schema if the vibration of this test causes the mechanism to slip and accidentally causes a short circuit.   IMHO if you have DCC decoder in the mechanism, it's probably best to use alligator clips or other safer method to apply power to the mechanism.  This test setup shown doesn't have a circuit breaker, so we *don't* want you to fry a decoder.


Some advice on mid-1990s Atlas/Kato drives:

2)  I also would like to point out this thread seems to confirm that mid-1990s Atlas/Kato drives seem to be prone to 'noise-introduction' if dis-assembled/re-assembled.  If you have one of those mechanisms and it's acceptably quiet as it is, I suggest leave it well enough alone.   If you do decide to (or need to) dis-assemble one of these 4-axle locos, then you may well be signing up for noise debugging upon re-assembly, and likely to need to do the Ron Bearden bearing block removal ( http://u18b.com/wp-content/uploads/2013/03/Bearingblock.pdf ) to try to return it to acceptable quietness.  Note:  As this thread has shown, we've found that we're not been able to guarantee, even with the bearing block removal, to always be able to return a Atlas/Kato drive to quietness after re-assembly.  My 2 cents. 


Variations on a theme:

3) If you want to check a mechanism without removing the trucks (to avoid a partial 'dis-assemble/re-assemble), you do have the option of testing the locomotive on the track, inverted, as shown here:

/>
The above is a standard DC mechanism. 

Note: *don't* press down.... it is possible to short out the  underside solder points of the  printed circuit board on the frame  (don't ask me how I know this..... I learned the hard way and burned out a LED).

If you have a DCC decoder in your mechanism, I'd definitely use alligator clips instead, because inadvertent vibration-caused 'slip' by this setup resulting in accidental short circuit, could fry a expensive DCC decoder board. We don't want that.

I myself do (*carefully*) use this inverted method testing with the trucks on, because I can then immediately pickup the mechanism, apply alligator clips, and test the vertical up/down fit of the trucks (testing for binding or noise caused by vertical engagement of the gear tower and the worm).
If you do put the loco inverted on the trackon this way, be careful.   


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

[atsf_arizona]

Those Kato wheels have much smaller flanges than other wheelsets. Is it possible that the derailing truck doesn't swivel freely in the chassis?  So when it enters the curve instead of swiveling it keeps going straight?  The small flanges aren't able to get it to steer it into the curve maybe and it jumps the track? Also did you check the gauge of both the wheelsets and the track?

Unfortunately from a debugging standpoint, as stated in reply #131:    this formerly troubled truck, I went back and replaced in the original IM wheelsets to see if the problem went way (it did), and then have not returned to this particular truck with Kato wheelsets.    Your suggestions and that of Tim Benson much appreciated, someday I may need that info to debug a future troubled truck.

[BCR751]

Yep.  1.04 Mbit/sec = approx 104KBytes/sec.  For a 166 MB file, that equates to 27.7 minutes transfer time.  Keep us posted if you make progress on that with your service provider.

My phone call to our ISP produced some interesting results.  Seems they unilaterally dropped our upload speed from 5Mbps to 1Mbps citing changes in their payment plans which, according to them, they can do without notice to their customers.  So, we were paying for 5Mbps and now, for the same money, we're paying for 1Mbps.  There is a "silver lining", of sorts, here because another provider just completed an install of brand new fiber optic cable right past our house.  Guess what we're going to do

Anyway, back to the topic.  Yesterday, I vowed to get to the bottom of this C-630 motor/noise problem so I began using that tried-and-true scientific troubleshooting method called "Process Of Elimination".  I started with the loco fully assembled and removed pieces one at a time,  I got right down to the bare frame with just the motor in it and still had the problem with the motor vibrating and losing speed at low RPM.  I took the motor out of the frame, removed the plastic cradle and applied voltage.  Without any containment from the frame, the motor started literally jumping around as voltage was applied (I wish I could post that damn video).  So, obviously, I have a bad motor.  Question now is what to do about it.  I've heard it's not really worth trying to repair it since they cost so little.  Should I try and get into it to see if it can be fixed or simply buy a replacement? 

Doug

[atsf_arizona]

So, obviously, I have a bad motor.  Question now is what to do about it.  I've heard it's not really worth trying to repair it since they cost so little.  Should I try and get into it to see if it can be fixed or simply buy a replacement? 

Doug

My knee-jerk reaction = buy a replacement motor.   I just checked the Atlas parts site, it's part number 540100:    http://www.atlasrr.com/pdf/PartsPDFs/NParts/NC630Loco.pdf

   http://shop.atlasrr.com/search.aspx?SearchTerm=540100

http://shop.atlasrr.com/p-2157-n-motor-flywheel-assembly-scale-speed.aspx    cost is $27.60 + shipping.

There's another guy on that auction site that also has the proper motor: 

http://www.ebay.com/itm/SD7-SD9-SD24-SD26-C628-C630-SLOW-SPEED-MOTOR-ATLAS-N-Scale-/371827692645?hash=item5692a85465

So we're talking about the time value of your debugging, vs. installing a $27 + shipping replacement motor. 

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

BUT BEFORE you do that..... this could lead us into another (somewhat niche?) area of useful questions for our inquiring minds.....

Anyone have tips on how to do straightforward basic model RR locomotive electric motor debugging?

• How hard/easy/worthwhile is it to debug a bad model railroad locomotive electric motor?   
• What are the basic things to check?  How?  What tools / instruments?
• What's the parameters of a "good" vs. a "bad" model RR loco electric motor?
• What are the symptoms that indicate it's fixable, or should just replace it?

This might well be applicable to debugging the "growler" Atlas/Kato drives (the middle of the 3 example mechanisms shown in my first YouTube video in reply #136 above).

Others here clearly have torn apart motors and seem successful.  Would be interesting to see how the slot car electric motor blueprinting info in replies #106 and #112:

http://slotblog.net/topic/51489-blueprinting-slot-car-motors/

Could be "tuned down" to our model railroad needs. 

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

P.S.  I do wonder what folks like Victor Miranda might say about this electric motor topic, he of "let's see how long it takes to burn out a model RR electric motor" fame....

" a motor discussion." : http://forum.atlasrr.com/forum/topic.asp?ARCHIVE=true&TOPIC_ID=55311   eight pages of wisdom

Also of interest: 

"Fixing electric N scale motors?"  :  http://forum.atlasrr.com/forum/topic.asp?TOPIC_ID=65763 
"3-pole vs. 5-pole":                       http://forum.atlasrr.com/forum/topic.asp?ARCHIVE=true&TOPIC_ID=23681
"  another 3 vs 5 pole test (hint: I am not pleased)":       http://forum.atlasrr.com/forum/topic.asp?ARCHIVE=true&TOPIC_ID=43689

There's other threads as well, all found by Googling:  " how long does a model railroad electric motor last victor miranda "

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

 

[u18b]

The confirmation of a bad motor is to

1. Replace the suspect motor with a new one, or one from a quiet loco.
Is the suspect locomotive now quiet?

2. As an extra check, install the suspect motor in a known quiet loco.
Is this loco now noisy.

3. And as one last check, (overkill) install the suspect motor back into the original chassis.
Is it back to noisy again?

Bob Horne and I did this to one of his pesky problematic atlas b40-8.  All these steps confirmed it was bad motor.
Bent shaft, unbalanced flywheel?  Who knows. But we isolated the problem to the motor

[narrowminded]

The most likely source of a motor only problem is the flywheels and a rig that holds the motor stationary so you can rotate the flywheel and check it with a dial indicator would be the way to inspect for that.  Absent a dial indicator a cruder rig might work that just held a stationary arm up next to the flywheel and rotate it checking for clearance changes as you rotate it.  A good time to don the optivisor, too.  A piece of standard paper is about .003".  Does it slip between at the loose point?  Two thicknesses?  Could you see the change from a galloping horse? ... in the dark?    If there is runout try to see the motor shaft and see if the run out is the same on the shaft.  And bad enough, a Sharpie held in place would mark the high spot as you spin it.

With all of that said, if runout is found, I'd inspect for a bent motor shaft from pressing the flywheel on in a poor or improperly used rig,  OR for one that was loosely fitting and got loctited on off center, OR if it's the hex pocket and out of balance, that the pocket got pushed in off center (maybe the reason for the bearing removal trick).  There are more possibilities and you should never underestimate the creativeness of an underpayed, undertrained, employee with the wrong tools and big hammer.   But to me those are the most likely sources of trouble in an assembly like this.

Now, for a balancing idea. While I've never done this I could envision a rig that would allow for a trial and error method of dynamically balancing it.  Two pieces of music wire or such cut to equal length, shorter than the circumference of the flywheel by maybe an eighth or a quarter diameter so there is a void.  Wrap them around a mandrel at a diameter smaller than the flywheel so that they can be slipped over under spring tension with the void being the weight differential.  Slipped in place with the gaps opposite should be balanced and the gaps together on one side would be the max weight that you could apply.  And as I type this, I think the same might be accomplished with a spring that fits snugly to the flywheel diameter, placed on a mandrel, and then split out two matched sections with your Dremel and a cut-off wheel.  Or with your lathe you could...  Anyway, the point remains and should be clear.

The way it would be used is to put a reference mark on the flywheel (Sharpie), slip the two rings on with the gaps displaced about a quarter of the way around and start it up.  Any better? Worse?  Then maybe take the pair as they are and rotate them together into a different quarter.  Any better?  Repeat this until you find the best quadrant, maybe add a new reference mark, and then start  moving one or the other and watch for a change.  If it got better, move a little more and try again.  If it had gotten better but then got worse, move it back and try the other ring.  Moving the gaps closer together adds weight to the opposite side and splaying them apart reduces that weight.  A few fiddles and you should find the sweet spot.  The heavy spot is centered between the two gaps.  Add a mark and that would be your drill point. 

By the displacement and knowing the diameter and gap, calculating the weight of each and then their displacement, then the weight of the flywheel material and the volume you would be reducing with your chosen drill, you could pretty well calculate what depth to drill at what diameter.  You could also go hit and miss eyeballing the wire gap and guessing your drill equivalent and then just drill a little, test, reset the wires which should logically follow the resulting change from drilling, drill a little more and test, and at some point the wire gaps will be opposite each other and the drilled point will be at 90 degrees from the opposing gaps with the motor running smoothly.  That's the balance point with the wire gaps cancelling each other out on a balanced rig.  Removing the rings should leave you with a balanced shaft.  If you overshoot with your drill this should all show up on the opposite side and the same method could be used there to chase the re balance as needed.  If it wandered a little from your point you could move the drill point a little to correct.  I think it would work and think it would not be as clumsy as it sounds describing it. 

And one more thought, if you want to try a rig for fun, smash various small sizes and lengths of solder and slip them under an O ring that fits the wheel and will hold them in place.  Reference mark, move, and see if you can find it that way.  I think the split rings make a more usable rig but for a way off, once in a while rig, this might get it done, too.

Hope that helps or at least afforded a good giggle.

[Point353]

So, obviously, I have a bad motor.  Question now is what to do about it.  I've heard it's not really worth trying to repair it since they cost so little.  Should I try and get into it to see if it can be fixed or simply buy a replacement?

It might be worth contacting Atlas to see if they would replace the bad motor (at no charge) if you returned it to them.
The postage for a USPS priority mail small flat-rate box is about $7, so, even if you had to pay for shipping in both directions, it would still be less than half the cost of buying a new motor. 

[muktown128]

You can also ship it USPS first class in a padded envelope for maybe half that cost.
I shipped a Kato Mikado replacement wheel set to someone that way recently for under $3.
There is a weight difference, but I would explore that as a lower cost option.

Scott

[rrjim1]

   Anyway, back to the topic.  Yesterday, I vowed to get to the bottom of this C-630 motor/noise problem so I began using that tried-and-true scientific troubleshooting method called "Process Of Elimination".  I started with the loco fully assembled and removed pieces one at a time,  I got right down to the bare frame with just the motor in it and still had the problem with the motor vibrating and losing speed at low RPM.  I took the motor out of the frame, removed the plastic cradle and applied voltage.  Without any containment from the frame, the motor started literally jumping around as voltage was applied (I wish I could post that damn video).  So, obviously, I have a bad motor.  Question now is what to do about it.  I've heard it's not really worth trying to repair it since they cost so little.  Should I try and get into it to see if it can be fixed or simply buy a replacement? 

Doug

The problem I have ran into is who ever manufactures the motors for Atlas uses some kind of epoxy to balance them. I have received motors from Atlas that this material had came loose and was still in the plastic bag. If this happens you have a choice, send it back and hope to receive a good motor or balance it yourself.

I just gave a Kato motor to a guy on another forum, the shipping was $2.65 in a padded envelope.   

[BCR751]

It might be worth contacting Atlas to see if they would replace the bad motor (at no charge) if you returned it to them.
The postage for a USPS priority mail small flat-rate box is about $7, so, even if you had to pay for shipping in both directions, it would still be less than half the cost of buying a new motor.

Too late.  I've already ordered a replacement motor from Atlas    Thanks for the post anyway.

And, for those with a reasonably stocked junk box, I thought I'd post a photo of the loco test setup I'm using.  A couple of old analog meters, a re-purposed project box, some chassis-mount connectors and Bob's your uncle.  The connections on the left end of the box are from the power source and the ones on the right go to the test track.  With this setup, both current draw and "throttle" voltage can be monitored simultaneously.  Real easy weekend project



Doug

[atsf_arizona]

Too late.  I've already ordered a replacement motor from Atlas    Thanks for the post anyway.

And, for those with a reasonably stocked junk box, I thought I'd post a photo of the loco test setup I'm using.  A couple of old analog meters, a re-purposed project box, some chassis-mount connectors and Bob's your uncle.  The connections on the left end of the box are from the power source and the ones on the right go to the test track.  With this setup, both current draw and "throttle" voltage can be monitored simultaneously.  Real easy weekend project



Doug

Serious "science" at work.  What a hobby. I love it. 

[peteski]

Those are some vintage multimeters Doug!  Bakelite cases an all . . .  they probably have jeweled movements too! I love it. Those are even older than the old Simpson multimeter I used in my first job as an electronic technician.

[jagged ben]

Fwiw...

I usually do my frame-on-track hookups like this:

[ Guests cannot view attachments ]

Its considerably less risky for light boards and decoders.  For easy plug and play decoders (i.e Kato hood units) I swap out the decoder for an old Kato light board whose LEDs have been cannabalized. 

Note I've removed the Unijoiner so its not in the way.  Obviously one could make a somewhat more ideal permanent setup with other track.

[atsf_arizona]

Jagged Ben,

Thumbs up. 

Agreed, much safer for the loco, light board, decoder, etc. 

(Don't ask me how I know that...... I found out the hard way when placing my
loco inverted on the track..... pressed a bit too hard on the loco...... puff of smoke as the underside of the LED light board touched the frame
and burned out the LED....)