Problems with our new GMD-1

[mark dance]

So i tried CV 54 and it did make them run better together but it also cut the top end speed to 33 smph. I always program my locos to go 30 smph at 50% and 60 smph at 100% so it looks like i still have some programming to do. anyone else with any ideas? i only need these locos to run together as they will only run with each other but the speed need to come up.

Jeff

I presume you tried beefing up CV 05 and CV 06?
Md

[Jeff AKA St0rm]

once i get them saved into JMRI that was going to be my next step.

Well i now have my GMD-1 running great. After you run the CV 54 auto set adjusting CV 2,5 and 6 brough the speeds right up to where i needed them.

[jdcolombo]

Mark and Jeff:

If you guys have a minute, can you read the values you have for CV's 52-56 and post them?  I'm curious as to how they differ from the defaults.

I should have told Mark at the beginning to try the auto-adjustment routine.  I completely forgot about it, because I like to play with CV's 

Thanks!

John C.

[Jeff AKA St0rm]

some differences between the 2 locos

Loco 1
CV 52 = 12 default = 32
CV 53 = 42 default = 140
CV 54 = 69 default = 48
CV 55 = 12 default = 32
CV 56 = 255 default = 255

Loco 2
CV 52 = 12
CV 53 = 39
CV 54 = 57
CV 55 = 12
CV 56 = 255

[peteski]

If these run in a consist, why are the values so wildly different?  What happens if you flip the locos around (make the other one the lead loco)?  Do they still behave themselves or do they start oscillating?

[jdcolombo]

If these run in a consist, why are the values so wildly different?  What happens if you flip the locos around (make the other one the lead loco)?  Do they still behave themselves or do they start oscillating?

Well, actually the values aren't wildly different.  They ARE a lot different from the defaults, but between the two, they aren't that different.  CV52, 55 and 56 is the same for both; CV53 differs by decimal 3, which is only about 1%; the biggest difference is in CV54, but I've found that that one takes pretty wide swings to make much of a difference in running characteristics; I'd bet that if Jeff set both engines' CV54 to something in the middle of the two values - say 62 - they would still be just fine.

It IS interesting how much these values differ from the defaults, particularly in CV's 52, 53 and 55.  These values are similar to what I used to get the best performance out of the FEF's motor using a LokPilot (non-sound decoder from ESU), though ultimately I switched to a Zimo.

I might have to do some experimenting with my own Loks . . .

John C.

[mark dance]

Mark and Jeff:

If you guys have a minute, can you read the values you have for CV's 52-56 and post them?  I'm curious as to how they differ from the defaults.

I should have told Mark at the beginning to try the auto-adjustment routine.  I completely forgot about it, because I like to play with CV's 

Thanks!

John C.

Loco 1: 52=12 53=77 54=63 55=12 56=128

Loco 2: 52=12 53=79 54=48 55=12 56=128

Defaults for both locos (I think...this is what I wrote down): 52=20 53=140 54=15 55=25 56=255

If these run in a consist, why are the values so wildly different?  What happens if you flip the locos around (make the other one the lead loco)?  Do they still behave themselves or do they start oscillating?

If I understand correctly these parameters are set automatically to the running characteristics of each motor/gear train.  This is what ESU's  procedure (I described above) does.  It would make sense from a control theory POV as it is trying to normalize/calibrate all the locos to a set reference.  I have run them in reverse with no effect. 

Note this morning when reading the variable I did notice an occasional slight oscillation but it dampened out quickly as opposed to going out of control. 

md

[mark dance]

Loco 1: 52=12 53=77 54=63 55=12 56=128

Loco 2: 52=12 53=79 54=48 55=12 56=128

Defaults for both locos (I think...this is what I wrote down): 52=20 53=140 54=15 55=25 56=255

Given the deltas, this is why I recommend Rapido auto-calibrate the locos automatically before leaving their shop. A 10 second procedure!

If I understand correctly these parameters are set automatically to the running characteristics of each motor/gear train.  This is what ESU's  procedure (I described above) does.  It would make sense from a control theory POV as it is trying to normalize/calibrate all the locos to a set reference.  I have run them in reverse with no effect. 

Note this morning when reading the variable I did notice an occasional slight oscillation but it dampened out quickly as opposed to going out of control. 

md

- edited position of quote tag, gfh

[CNscale]

I have three of these - two with sound, one without. All of them are a little bit tight in gauge, but not enough to cause any problems on my track, so I might just leave that alone. They run great, I've never had a locomotive that handles so well at slow speed. This is also the first time that I've had two identical mechanisms with different decoders, and I have to say that the motor control of the Loksound units is discernably better than that of the Digitrax, although the latter is still a lot better than my older Altas locomotives (with TCS decoders).

There is one problem though - one of the two sound units has a very noisy power train. The motor has a high-pitched whine even at the lowest speeds, and anything above half-throttle it squeaks and chirps like there's a chipmunk trapped inside. This is after running it for at least an hour in both directions to break it in. In fact I think the noise has gotten worse.

I've completely disassembled Atlas locomotives (to install decoders) without any problems, but after taking the shell off of the GMD-1 it is clearly a different beast. I'm hesitant to go any further and looking for advice - is this a problem that I'm likely to be able to resolve myself, or should I go back to Rapido on this? I'm reluctant to have to swallow the risk and cost of shipping (over and above the original cost + Rapido's last-minute 9% surcharge for Canadian customers) if there's an easy fix.

[Jeff AKA St0rm]

I have three of these - two with sound, one without. All of them are a little bit tight in gauge, but not enough to cause any problems on my track, so I might just leave that alone. They run great, I've never had a locomotive that handles so well at slow speed. This is also the first time that I've had two identical mechanisms with different decoders, and I have to say that the motor control of the Loksound units is discernably better than that of the Digitrax, although the latter is still a lot better than my older Altas locomotives (with TCS decoders).

There is one problem though - one of the two sound units has a very noisy power train. The motor has a high-pitched whine even at the lowest speeds, and anything above half-throttle it squeaks and chirps like there's a chipmunk trapped inside. This is after running it for at least an hour in both directions to break it in. In fact I think the noise has gotten worse.

I've completely disassembled Atlas locomotives (to install decoders) without any problems, but after taking the shell off of the GMD-1 it is clearly a different beast. I'm hesitant to go any further and looking for advice - is this a problem that I'm likely to be able to resolve myself, or should I go back to Rapido on this? I'm reluctant to have to swallow the risk and cost of shipping (over and above the original cost + Rapido's last-minute 9% surcharge for Canadian customers) if there's an easy fix.

Sounds like you have a dogbone come out of the motor. Mine was doing the same thing after i popped the truck out and tried to pop it back in. If you take the shell off and untape the decoder there are 4 screws 2 on each end to take off. If you take them out you can see the motor and the worm gear. Make sure everything is lined up. It is an easy fix that i did. Photo of the screws can be found here https://www.therailwire.net/forum/index.php?topic=35319.0

[CNscale]

Thanks. Removed the screws and opened everything up as you suggested, but everything was in order. At this point I think I've reached the limit of my own diagnostic and repair skills, so I'll have to talk to Rapido.

I do have to say that I always thought that my cantankerous Atlas engines were complaining about dirty track. Given how well these GMD-1s run on exactly the same track (in spite of this particular noise problem), and given their "hard-wired" decoder to truck connection, I'm inclined to think that the problem with my Atlas engines is poor truck-to-frame connectivity. Or maybe dirty wheels? Or maybe crappy decoders? Whatever it is, after playing with these GMD-1s, the other locomotives are just no fun anymore. With a GMD-1 I can couple onto a freight car without moving the car at all. I didn't think that was possible in N-scale.

[mark dance]

Hi all...an update.

Both my GMD-1s have experienced what appear to be decoder issues. Per the post above they were running beautifully once callibrated but all driving wheelsets were under guage. When the guage was corrected they began to hum and stutter again. They were both returned to Rapido.

Within 10 days they were back (thx Mike and Dan) however one displayed a slight stutter and auto callibration did not address it. I tried running both units' mechanisms in at full speed but both failed within about 5 minutes of continuous running and no longer run but lurch or don't move at all. They will be returned to Rapido Monday. I am looking forward to a diagnosis and path forward.

Might I ask how anyone else's ESU-equipped GMD-1s are performing?

Md

[Jeff AKA St0rm]

I have not been running mine since i got them speed matched as i do not run CN  . I will put them on today and run them for the day and see what comes of it.

[Puddington]

After due consideration I think Mark's locomotives are failing because he made fun of the Maple Leafs so......

Seriously; Mark's units are a mystery... Dan worked on them for hours and when they left he was satisfied that they ran perfectly... only to find they didn't and have "given up the ghost".... quite troubling.

We will get to the bottom of this. Next week when we get them back we will get ESU on the line and go through these units with a fine tooth comb. My concern is that while other (and we have over 2000 out there) haven't exhibited these issues, if Mark's did, then more could.... Mark's input has been invaluable and we will get to the bottom of this with ESU and, after CNY is over and everyone is back, the engineers at the factory.

N scale is different than HO (as we all know) The gauge issue we have seen is a direct result of the factory not understanding just how tight N scale locomotive tolerances must be - they now understand. Decoders, on the other hand, should be pretty much the same in HO and N scale. With ten years experience in DCC sound locomotives we are baffled at whet we have seen in Mark's units. I have complete confidence that ESU and our engineers will figure it out and I will report back here and on our website the results of what we find.

If others have issues, or, for that matter, no issues, please report that to us either by email or here. All input is helpful. This is a case where you can't have too much information.

We will make sure to take care of Mark (I see new units, specially decorated with "Go Leafs Go" on them) once we can assure ourselves that the quality of the units we ship him will equal that of the layout they will run on.

Thanks to everyone for their help and patience...

Mike

[mark dance]

After due consideration I think Mark's locomotives are failing because he made fun of the Maple Leafs so......

Mike

If not taking the Leafs seriously is a criteria for failure then units shipped anywhere near 90% of their own fanbase this year should be failing.

I am curious though Mike...you said Dan worked on the units for hours. What needed so much work? That sounds like serious tinkering which sounds like a "fiddly"  (rather than robust) design. In my experience, design margin (lattitude) is needed for production to run smoothly. If something fails QC you swap out a whole assembly *or ship the customer a new unit* rather than spending the time tweaking it. Tweaks have a bad habit of moving over time and unit to unit. And some times you just have to accept that a lemon will come off the line, especially if the design spec is really tight, and the lemons need to be scrapped as bad yield rather than trying to patch them.

Just curious...


Md