its "what I said above".
When designing a Throttle, the designer can choose to behave (at least) two ways if the user changes direction with a non-zero speed selected.
option A which Lenz appear to use according to Bob's reports, as does the Uhlenbrock Fred which I own. Throttle will issue an "emergency stop" packet to the current decoder. Result, decoder stops instantly. All decoders will do this regardless of maker.
option B which appears to be used by Gaugemaster(MRC), NCE, certainly used by Digitrax (I own a couple of different ones) and Sprog. Throttle will issue a new "direction" packet to the decoder. Result, the decoder slows following CV4, changes direction, then accelerates following CV3 to the set speed.
Now, Ray has found some Lenz decoders which appear to be faulty if used on controllers which exhibit behaviour B. This is confirmed on some German forums. However, anyone owning a controller which have behaviour A will never see the problem because the controller issues the "emergency stop" packet if the user tries to carry out the reverse.
There seems to be a suggestion of a pattern in the choice of A or B above. European makers in A, and US makers in B. Its not total, but suggestion of a trend. Therefore if returning decoder to A&H, specify that they have to test on a system which is known to behave like B.
Edited by Nigelcliffe, 10 October 2011 - 22:46 .