> I'll look more carefully, but I think one difference is that the
starting voltage is different between the two models. I would guess
this relates to some difference in the lighting circuits. The effect
is that the actual voltage applied to the motor is offset between the two
models. Also, the E-8's have different offset or starting voltage running
forward compared to running backward due to the constant lighting circuit.
At least this is my impression. Maybe someone can check this out more
carefully and post results?
Since I have dealt with this problem, I'll respond with my experiences.
I bought my first E-8 (UP #928) about 3 years ago when they first came
out. (super nice too!!!). Loved the mars light, etc.
This winter, I while calling all over the country to get interiors for my
Rivarossi UP cars, I found a hobby shop that had a UP #949, so I bought it
(since my era is 1985-1995). I then traded #928 for a #951 that Bill
Gripp had. When I got them both I experienced the same problem in that
the loco runs faster in reverse than forward (for the same voltage
setting), causing the "rear" unit to push the forward unit, and with the
swivel coupler mount, this causes some derailments on crossovers.
I opened up the units, removed the circuit board and figured out how the
lighting works. Here is an ASCII schematic (does not apply to Mars light
models, but it would be similar.
T | | | |
R | +---bulb--+ |
A | | Motor
C +---------|<|----------------+ |
The extra diodes may be carried over from the Mars light circuit, but what
this does is provide about a 2v drop in the forward direction and a 0.7
volt drop in reverse (resulting in higher reverse speeds).
What I did to make it work...
I moved the wire to the bulb from one point on the circuit board to a
point where the bulb runs in reverse (basically flip the input/output of
the diode circuit). I don't specfically remember which hole I moved it
to. (I'm at work and the E units are at home...)
Now my units run back to back with no problem, lights on in one direction,
off in the other. Converting these to DCC would eliminate this problem as
If you would like more information please e-mail me direct at