wouter,
No, I haven't checked the JMRI speed table for this decoder recently. With EasyDCC I do normally have a start voltage set, and mid and full speed, often evenly divided with full speed at 100%. On a few of my faster locos, full speed is reduced. Momentum is set for 4 or 5 accelerating and braking. It seems the JMRI roster info needs to be checked for that.
In 28 speed step mode, wouldn't the JMRI be sending "speed step nn", and the decoder using its value for that speed step to control the motor? I don't think there is a way JMRI could tell these decoders the actual duty cycle the decoder should be setting.
If the number of speed steps is mismatched, or the JMRI speeds are 0 for start, mid, and full, wouldn't F0 still control the lights? I can't even do that on LE103s with JMRI, but can with EasyDCC plug in throttles. The speed table values seem more likely to be the source of trouble with the Zimo decoder, where I could control F1, but not train speed.
The recent discussions of what may be wrong are interesting and educational. Thanks for helping me, guys! I have always set these decoders up in EasyDCC as Class 2, which supports 28 speed steps. Even in the late 1990s, the decoders I used supported that. There was no reason to use the coarser 14 speed steps on any of my locos. Notice 1477 read back showed it is configured for 28/128 speed steps. I believe all my roster entries are set for 28 speed steps, so that should match, but I will confirm that.
Don Weigt
Connecticut
