speed matching and others issues with digitrax decoder

There typically always there be differntial Forward/Reverse running characteristics. CV95 provides this however, you need to make sure that CV is actually in the decoder (like the ancient code in the DZ123Z0)

#2, ALWAYS do speed matching with ALL BEMF CV's OFF ! Otherwise, what you appear to see will likely not be the next time and certainly not in a Consist

BTW: I have tried to turn off all Digitrax CV's and can cleary show you on an oscilloscope they are still functioning. The CV's and data were also confirmed with a well known DCC Guru.

As Mark mentioned, we really haven't had much issues with speed differential, on a proper running loco (DC speed relatively linear and doesn't top out well below the nominal 'top speed'. Put the faster loco behind a slower loco in a Consist so the faster one is slowed by its load.

The BEMF (Back ElectroMotive Force), a.k.a. Load Compensation or 'help my poor, sluggish locomotive overcome its weakness'

This electronic sensing uses motor current changes for what would either be a same Speed Step or more typically for slowest Speed Steps to help start the loco in its typical worst case (overcoming inertia). Think of just starting when Speed Step 1 or 2 applied to the motor, to start it moving. Those very short pulses can't normally overcome the friction of the trucks and the static load of the rolling stock. So the motor current gets essentially is much higher. The BEMF calculations tell the motor output to increase the pulse On duration (like internally auto-increasing the Speed Step) until the motor finally gets moving and the motor's current then reduces and so does the BEMF compensation.

This concept was also design for a train transitioning from a flat to a grade, where the locomotive would require more energy. So BEMF function again auto-increases the 'internal' Speed Steps so the train appears to run the same speed up the grade and the opposite compensation when it goes down the hill until a level plane again.

Unfortunately, these algorithms were designed around HO and N scale locos, most with Permanent Magnet motors that draw much more current (especially the brush/commutator changes) compared to a Coreless motor or our tiny Z motors! The decoder algorithm can't ideally read our low current motors and therefore are not consistent in compensating.

As Mark has shown, most all motor perform with reasonably linearly. The slopes can be more variable but the CV5 (max) and CV6 (mid) can help the fastest to slow down to the lessor performer. And yes, the CV5 and CV6 numbers seem backword in labeling but someone had reason or a whoops for it.

Using a PWM [pulse width modulation] motor supply can greatly improve the low speed running, where most of the pure DC speeds very as motors just don't have enough magnetic field effect at those low voltages. With PWM, the motor gets the full voltage for say 1% 'On' (99% Off) but this typically happens at ~60 to 250 Hz in most circuits to ~10,000 to 20,000 with DCC decoders.

I have two ES44AC's that are now equipped with TCS AZL4 decoders, so this is a bit "off topic" but they ran at almost the same speed on the stock DC and then initially on DCC. But one of the loco's was not smooth, lights flickering and often stalling. I have spent hours now taking the poor performer apart and have finally got it running much more reliably, all of the things I tried could fill another thread! But now they are very different in speed with the same DCC settings! The poor performer is a lot slower and won't even move on throttle step 1, generally there are about 4 throttle steps between the two when at the same speed. This must be due to some drag in the mechanics. Anyway, it is time to completely tear down the slower problem loco and reassemble it, meaning separate the chassis halves. But obviously speed matching is a very tenuous thing with these locos.
As alway Jeff, your post are very helpful and I have tried to follow them to the letter. I may be creating my problems as I learn the art of working on these locos.