DGE-Railroad
The Orchard Line

Reporting back with some positive news!
So I pulled the decoder, hooked it up to the ESU programmer and delved into the motor control section.
I'd originally intended setting CV54 to enable auto-tune, popping the loco on the rolling road and hitting F1.
I found this page though, on the very informatinve Streamlined Backshop site that reports occasions when the auto-tune doesn't help.
There's a deal of helpful information around the various CVs and the way they're exposed in the programmer, so I decided to compare the authors recommendations with the decoder settings
CV49 - DC Motor PWM Frequency: RV (Recommended Value) = 40Khz, (Decoder Value) DV = 40Khz
CV10 - Back EMF Sampling: RV = 8, DV = 8
CV56 - Regulation Influence: RV = 255, DV = 255
CV51 - Slow I value: RV = 0, DV = 0
CV52 - Slow K value : RV = trial and error, DV = 15
CV124? - Adaptive Regulation Frequency: RV = OFF, DV = ticked
CV54 - K value : RV = probably less than half of what is set, DV = 50
CV55 - I value: RV = <10, DV = 100
I reasoned I could try bringing the relevant CVs (124, 52, 54, 55) into line with the recommendations and if things didn't improve, trust to an auto-tune.
I had also thought about switching off the BEMF completely as Alan suggested, as a test. I decided to keep that in my back pocket as a last resort
Well it has indeed made an improvement - nice and smooth from a standstill to maximum speed and back again. There is still a slight lag of the front truck when starting from a standstill but I'm putting that down to the extra drag put on it, or a difference in the motors. I may look to test the drag theory at some point by removing the Shay drivetrain and checking again, but if it's behaving okay on the track...meh.
Thanks for all of the fantastically useful and knowledgable input - it was indispensable!
Darren
So I pulled the decoder, hooked it up to the ESU programmer and delved into the motor control section.
I'd originally intended setting CV54 to enable auto-tune, popping the loco on the rolling road and hitting F1.
I found this page though, on the very informatinve Streamlined Backshop site that reports occasions when the auto-tune doesn't help.
There's a deal of helpful information around the various CVs and the way they're exposed in the programmer, so I decided to compare the authors recommendations with the decoder settings
CV49 - DC Motor PWM Frequency: RV (Recommended Value) = 40Khz, (Decoder Value) DV = 40Khz
CV10 - Back EMF Sampling: RV = 8, DV = 8
CV56 - Regulation Influence: RV = 255, DV = 255
CV51 - Slow I value: RV = 0, DV = 0
CV52 - Slow K value : RV = trial and error, DV = 15
CV124? - Adaptive Regulation Frequency: RV = OFF, DV = ticked
CV54 - K value : RV = probably less than half of what is set, DV = 50
CV55 - I value: RV = <10, DV = 100
I reasoned I could try bringing the relevant CVs (124, 52, 54, 55) into line with the recommendations and if things didn't improve, trust to an auto-tune.
I had also thought about switching off the BEMF completely as Alan suggested, as a test. I decided to keep that in my back pocket as a last resort
Well it has indeed made an improvement - nice and smooth from a standstill to maximum speed and back again. There is still a slight lag of the front truck when starting from a standstill but I'm putting that down to the extra drag put on it, or a difference in the motors. I may look to test the drag theory at some point by removing the Shay drivetrain and checking again, but if it's behaving okay on the track...meh.

Thanks for all of the fantastically useful and knowledgable input - it was indispensable!
Darren
Last edited: