Atlas GP38 with factory dcc - sluggish performance

[Dwight in Toronto]

Just purchased a new Atlas N scale GP38 dcc loco (Ontario Northland livery; I believe it has a TCS non-sound decoder factory installed).

Did a cursory set-up on the programming track with my NCE PowerCab …  just enough to get it running/breaking in on my test loop (ie - programmed loco #, ensured momentum is off for now, established 28 speed steps).

Front & rear lamps worked fine in fwd & rev, respectively.  But the loco didn’t even begin to budge until speed step 16.  Went back to the programming track, entered the ‘motor control’ segment, saw that cv2 was ‘0’, so entered ‘1’ as a starting value.  Set cv5 to 254, and cv6 at 125, again as starting values.  Nothing changed - still dead as a doornail ‘til speed step 16.  Decided to let it toddle around for 10 minutes or so, and found that I could slow it down to a slow crawl at SS 10.  Any lower than that, and it would stop dead, and would require minimum of SS 14 to get moving again.

So, I progressively increased the values for cv2 in increments of 10.  Long story short, at cv2 = 50, I could get the loco to just barely start crawling at SS 6.  After running for a bit, I can slow it down to a crawl at SS 1, so, it’s kinda working, but using cv2 = 50 is throwing me for a loop.  I’m used to ESU sound decoders in the rest of my fleet where SS 1 effortlessly sets every loco on a smooth slow crawl, with cv2 values of 1, 2, or 3 max!  I shudder to think what I might have to set cv2 at in order to make this loco move at SS 1!

This is the first Atlas loco I’ve ever owned, so I’m looking for commentary from others who are more familiar with this brand.  Is cv 2 = 50 a typical value for getting an Atlas GP38 to start moving?  Or might something be amiss (heavy-handed grease job?  Old, dry lube? Just needs to be broken in?). 

Thanks in advance for any feedback.

[samusi01]

If it’s actually new in box, it may be several years old (Atlas lists it as a 2015 run). I would recommend a disassembly and thorough cleaning, followed by appropriate levels of lubrication.

As for the decoder, if it is a TCS decoder, I’ve had mixed luck. Most are good enough but I have had a few that don’t respond correctly to the full speed table. Not as bad as Digitrax (or the Lenz decoders atlas was using: those get replaced immediately) but certainly not at the level of ESU.

[mu26aeh]

If it was factory equipped, probably a NCE decoder.  I don't think Atlas ever used TCS. Lenz/NCE/ESU

[signalmaintainer]

If it was factory equipped, probably a NCE decoder.  I don't think Atlas ever used TCS. Lenz/NCE/ESU

This ^^^^^^. It's more than likely a Lenz decoder. And it probably needs a good cleaning out of the factory gunk in the mechanism.

Then replace the decoder with a Digitrax DN163A0 or NCE equivalent.

[Dwight in Toronto]

Thanks guys.  The decoder “manufacturer i.d. code” is 127, which is “Atlas Manufacturing” so no insight there as to whose decoder is being used (which I find kinda unexpected and unusual - thought that the decoder manufacturer would have their specific i.d. ‘burned’ into the device).  Fwiw, “decoder version” is 036. 

The loco is indeed new-in-box.  The label says “Item # 40 002 297, N GP38-2 Locomotive, Ontario Northland, Road # 1800”.   Ah ha … just noticed on the side of the box … “NCE The Power of DCC”, so that’s no longer a mystery. 

Anyway, my plan all along was to put an ESU decoder and speaker in this thing, but didn’t think that it would become a rush job.  In the meantime, I’ll pull the shell and see what’s in there.  Think I’ll be sticking to Kato, but it was the Ontario Northland livery that forced this Atlas purchase. 

[Dwight in Toronto]

I just had a look at the decoder installs section of TRW, and there’s a good tutorial by Rick Brodzinsky, putting an ESU 73100 into an Atlas GP38-2.   

Check out Rick’s opening commentary …
“The NCE board was not functioning well (needed >30% throttle to move, then minimal speed control, and no response from decoder even on programming track)”.  Seems like there’s a veritable history of shoddy NCE decoder performance.  How companies get away with this crap is so annoying. 

[John]

I've had other decoders do that .. but as previously stated, take it apart and clean out all the gunk grease the factory puts in there - a light dab of oil and then reassemble .. also reset the decoder to make sure it's brain is working right ..

[signalmaintainer]

I just had a look at the decoder installs section of TRW, and there’s a good tutorial by Rick Brodzinsky, putting an ESU 73100 into an Atlas GP38-2.   

Check out Rick’s opening commentary …
“The NCE board was not functioning well (needed >30% throttle to move, then minimal speed control, and no response from decoder even on programming track)”.  Seems like there’s a veritable history of shoddy NCE decoder performance.  How companies get away with this crap is so annoying.

The most recent Atlas locomotive I purchased (2019) came with a factory equipped NCE decoder. It persistently dumped CV settings. After a few conversations with NCE, I decided I was engaged in a fool's errand and replaced the decoder with a Digitrax DN163A0. Works like a champ.

[rrjim1]

Kato locos have a faster speed motor than a Atlas. With the same brand decode my Atlas locos need 10-15 speed steps more to run the same scale speed as a Kato. I really like the Atlas scale speed motor and most of my Kato locos now have a Atlas scale speed motor.  My Atlas locos do move at speed step 1 but really slow.

[peteski]

Atlas used decoders made by Lenz and NCE.  I believe the early ones were Lenz, then later they started to use NCE. The early Lenz decoder was very basic bare-bones decoder. But reading the manufacturer's ID CV they will show Atlas because those decoders are specifically made for Atlas.  Of course Atlas also used sound decoders made by ESU, but none were ever used in the GP38s.

If the loco is new old stock, then the grease in the mechanism is probably hardened. Running the model at speed for 10-15 minutes would have likely loosened the grease somewhat, but a thorough disassembly and cleaning (and re-lubrication) is a better option.  Also when putting it back to together, there is no need to lubricate the plastic gears in the trucks.  They are made from self-lubricating slippery plastic.  But I do drop small amount of oil in all the gear shaft bearings. You do need to lubricate (using light oil) the motor bearings, and warm bearings.  I also use Teflon grease on the worm.

[Dwight in Toronto]

Just curious - having to use unusually excessive values for cv2 might be unconventional, but I suspect that there is no potential harm being done, right?   I mean, the corresponding voltage values are readily within the intended operating parameters with respect to both the loco motor and the decoder, correct? 

I guess that the main downside is that, with, say, a value like cv2 = 100, there is then a much narrower “band” left over which cv’s 5 & 6 can “do their thing”.  I dunno.   Why is it so much to expect things to work out-of-the-box?

[John]

If it's a mechanical issue (friction) then the load on the motor will be higher --- and more power from the decoder .. I would do the teardown and cleaning -- why take the chance

[peteski]

Just curious - having to use unusually excessive values for cv2 might be unconventional, but I suspect that there is no potential harm being done, right?   I mean, the corresponding voltage values are readily within the intended operating parameters with respect to both the loco motor and the decoder, correct? 

I guess that the main downside is that, with, say, a value like cv2 = 100, there is then a much narrower “band” left over which cv’s 5 & 6 can “do their thing”.  I dunno.   Why is it so much to expect things to work out-of-the-box?

Well, if the loco sat unused for years, and it is overlubricated (which is routinely done at the factory) with grease that hardens, then it will have difficulty running.  Unfortunately this is typical with some old-new-stock older models.  If they only adhered to the loco lubrication recommendation that less is more. Lubrication instructions included with models often state just that, but that advice is somehow ignored at the factory. They slather the model with oil and grease, which in storage often even oozes out onto the outside of the shell and into the jewel box.  It's crazy!

None of the values in the speed table CVs (3 or 28 point) by itself will not cause any damage to the motor or other parts of the loco.  What you're doing is equivalent to a DC loco with the same problem having to be ran at at least half throttle setting. But since the motor has to overcome the extra friction, it can get very hot, and possibly overheat after extended run.  However that heat could soften the grease, making it run better (but when the grease cools down, it might harden again).

[Dwight in Toronto]

Completely disassembled the loco - there was just a negligible amount of factory grease in the vicinity of the worm gears (and it was by no means old/sticky/hardened in any way).  Put a tiny amount of LaBelle oil in the usual places, a careful smear of fresh grease on the worms, and buttoned it all back up.

Unfortunately, nothing changed … pathetic start-up performance remains. 
Just ordered an ESU 58751 Micro DCC Direct Atlas Legacy sound decoder; here’s hoping that will set things right. 

[peteski]

Completely disassembled the loco - there was just a negligible amount of factory grease in the vicinity of the worm gears (and it was by no means old/sticky/hardened in any way).  Put a tiny amount of LaBelle oil in the usual places, a careful smear of fresh grease on the worms, and buttoned it all back up.

Unfortunately, nothing changed … pathetic start-up performance remains. 
Just ordered an ESU 58751 Micro DCC Direct Atlas Legacy sound decoder; here’s hoping that will set things right.

Well, that eliminates problem with old grease. Did you check of the trucks roll freely?  Did you verify that the motor armature is also spinning as freely as other motors you have?  Would it be possible for you to remove the decoder and run the loco on DC?  Even a fresh 9V battery would be a good DC power source, but you would have to get creative about how to connect the motor leads to the frame halves.

One thing I did not see mentioned: When you were testing it with the decoder, was the top speed what you would expect it to be? Well, it will be slower than similar Kato locos due to the fact that Atlas uses "slow speed motors".