Hi all,
Since I'm now in the testing phase of the mainline of my layout build, I thought it'd be a great time to get out some of my motive power, clean the wheels, and run the cobwebs out at the same time. Now I'm in the process of speed matching a few units I intend to MU, more so I can recall how consisting and Ops programming is done than anything else, to be fair ... it's been a couple of years since my layout room had a functional layout.
Anyhow, I have a couple of Master line series DCC Atlas GP9s from the same production run, no sound. Box suggests they have an NCE decoder, but I don't rightly know. They've always been decent runners in the past, and after cleaning the wheels they still are. Not fantastic; you have to really crank up CV2 to get them going at speed step 1, but they're steady and reliable runners once they were broken in. They're great on the small trainshow layout I built to display my wife's trees for sale, where they'll happily circle for hours and hours.
After setting momentum to zero and matching CV2, CV5 and CV6 for each in preparation to speed match in Ops mode, I sent them running for 20 minutes or so to warm them up. When I started to speed match, I immediately ran into an issue. Namely, for each, running long hood forward was significantly faster at speed step 1 than short hood forward. The discrepancy was reversable, too. Reverse each engine, and the long-hood one would catch and overtake the short-hood one in about the same amount of time as when they were running the other direction.
Now I know a solution would be to just run them consisted nose to tail. I'd prefer to run them short-hood to short-hood, which was proto on my modeled line, but needs must and all that. Both their long-hood forward and short-hood forward speeds are comparable between them, and I'm sure I can get them close enough matched to run successfully as a consist if I run them facing the same way.
But as the long-hood forward was faster for each, such that speed step 1 long-hood forward was somewhere around 30-40% faster than short-hood and that was consistent for each locomotive (ball-park, I didn't clock it to see exactly what the difference was), I wondered if there was perhaps something buried deeper in the CVs that I don't know about. Like maybe a CV setting some percentage multiplier depending on direction? Or is there perhaps such a thing as a separate speed curve for reverse and forward? Just curious if there's some tidbit of knowledge I've missed or forgotten about.
Any N guys with experience with the Atlas Master line, I'd love to get your opinions, especially if you've encountered this before.
For those who might be tempted to suggest upgrading to new decoders ... I hear you loud and clear, and that is on the long-term goal list. But I ain't there yet. I'm sure it's a lot easier than I imagine it to be for the drop-in decoders, but to be honest, I'm a big chicken when it comes to popping off the shells on these tiny N scale machines. No particularly good reason for that apprehension; I'll happily rip into a computer's hardware to replace cards and boards when they tank. I think it's just the small size of the components in N that spook me. I look at my engines, and I look at my fingers, and I just know that somehow they won't play nice together. I'll get over this nervousness eventually, but for now, the stock decoders are going to have to stay. That's a problem for after the major build phases of the layout are completed, at any rate.
Thanks,
-Jason
jkovacs5I wondered if there was perhaps something buried deeper in the CVs that I don't know about. Like maybe a CV setting some percentage multiplier depending on direction?
On NCE decoders CV95 is reverse trim. First, I would check and make sure it is 0 to start with (it should be) to make sure it didn't somehow get a value that is causing the issue, then you can adjust it to try to compensate for the difference. Values 1-127 add speed in the reverse direction and values 129-255 add speed in the forward direction.
Thanks Robert. I did some digging after posting last night, and by process of elimination determined the Atlas engines have the AN12A0 decoder. Found the manual with the full CVs available, including torque kick and depth which sounds like it would help with my very high CV2 setting just to get the power rolling. I saw the reverse trim CV95, but didn't know how it functioned.
With now three tiers of adjustments, what is the best practice? Set CV2 and CV95 to 0, then try to get smooth speedstep 1 running with torque kick and depth, then CV95 to balance out the forward/reverse speed?
jkovacs5With now three tiers of adjustments, what is the best practice? Set CV2 and CV95 to 0, then try to get smooth speedstep 1 running with torque kick and depth, then CV95 to balance out the forward/reverse speed?
I can't really help there. I've only got one loco with an NCE decoder and I've never bothered with the CV116 and 117, I just bump up CV2 and was happy with it.
I am not where I can look at my system, but I know that with my Kato, I had to work with the motor trim CVs (can't remember numbers).
I was surprised that the same model locomotive, each with the same decoder, could run different speeds. It took a while of adjusting the motor trim.
The CVs 2, 5, & 6 did not adjust to speed match for me.
I believe motor trim CVs were 66 for forward and 95 for reverse. I believe that the setting 128 was normal, with 129 up to 255 is faster and 127 to 1 is slower.
Of course, at my age, everything I've told you might be wrong.
York1 John
Interestingly, the decoders on these Atlas GP9s units have CV95 for reverse trim, but no forward trim. Or rather, since values of 1-127 make the engine run faster in reverse and 129-255 make it run faster forward, they consider it both reverse trim and forward trim all in one.
And it looks like a pair of Atlas C420s I have (in my modeled railroad's actual proto paint, no less) have the older Lenz decoder, which has acceleration/braking trim that the manual says affects the CV3 and 4 momentum settings but not the steady power settting. CV5 and 6 aren't used for top and mid speeds, and I'll have to use a speed curve for speed matching. Think I'll have to pick up a PR4 and JMRI before I tackle that set, though.
CSX Robert I can't really help there. I've only got one loco with an NCE decoder and I've never bothered with the CV116 and 117, I just bump up CV2 and was happy with it.