QSI’s explanation
about Back EMF in their decoders.
There
has been a lot of discussion about back EMF in decoders since a model train
manufacturer in the U.S. Mike’s Train House, MTH, threatened they would take
action against any decoder manufacturers that were using, what was reported at
the time, “Back EMF”. This led some decoder manufacturers to take some action
for fear of a large fine.
QSI
decided to disable the back EMF that was released in locomotives from Broadway
Limited after their first loco. Many modellers bough BLI locos, most knowing
that the Back EMF “was not there” or disabled.
QSI
has announced an upgrade chip to the locomotives and this has generated a fair
debate on why, what etc can be done. Some modellers
were so disappointed with BLI and QSI that they refused to purchase any more
models.
What
this upgrade chip does has not been explained yet, and at US$35, some are
peeved at this considering they have bought a model with what some have called,
something missing. Whatever each modeller’s
understanding on the back EMF issue is, will depend what they do etc.
Gerry
Pruss from QSI wrote this on the Yahoo QSI group on
28 Feb 06.
There
seems to be some confusion in regards to the use of BEMF in QSI equipped
locomotives. Hopefully I can clarify the matter in this note.
All
QSI equipped HO locomotives use BEMF voltage to measure the speed of the
locomotive. One or more pre-production models of each of these locomotives are
calibrated at QSI to map the BEMF voltage to scale MPH.
The
BEMF to SMPH mapping is used in a variety of ways in all locomotives,
including:
(1) To determine the amount of Doppler shift.
(2) To determine whether the locomotive is
quickly decelerating so
that the brake squeal
should be played.
(3) To determine that the locomotive is
moving or stopped.
BEMF
is also used to determine the chuff rate in steam locomotives.
All
the above uses of BEMF are present in all QSI equipped HO locomotives and have
never been disabled.
What
has been "disabled" is the use of BEMF to try to maintain a constant
locomotive speed. QSI equipped models which support Calibrated Speed Control
try to maintain the BEMF which maps to the SMPH corresponding to the DCC
throttle speed step. For example, if the throttle setting were at 10 the
locomotive would attempt to maintain a BEMF which maps to 10 SMPH.
It
is this Calibrated Speed Control feature which has been "disabled".
The legal
issue does not really have anything to do with BEMF. It is the Calibrated Speed
Control. It does not matter if BEMF or a optical or
mechanical cam is used to measure the speed. The legal issue is maintaining a
constant speed based on a throttle position communicated digitally to the
locomotive.
As
far as you model owners are concerned, the Calibrated Speed Control feature has
been REMOVED from the firmware. QSI deliberated whether or not to keep this
feature in the firmware in a disabled state to be enabled later by some
mechanism, but our lawyers advised against this. If one of you clever guys
actually discovered the enabling mechanism and posted the mechanism on the
internet, then it would be as if we had not disabled the feature in the first
place.
From
my point of view as a software developer, the Calibrated Speed Control feature
is DISABLED. The feature exists in the source code for the firmware and I can
enable/disable it at will for each firmware build by defining/not defining a
compiler flag. When the compiler flag is not defined, the source code for the
feature is skipped by the compiler and the executable code for the feature does
not exist in the resulting firmware module.
It
is often hard to communicate subtle distinctions such as the difference between
"disabled at compile time" and "removed from the firmware",
but hopefully you can see the difference.
Will
future QSI equipped locomotives include Calibrated Speed Control?
I am
told not until the legal issues are settled.
However,
the issue of Calibrated Speed Control is becoming less and less important to
QSI. We believe we have a much more prototypical control mechanism in Regulated
Throttle Control (RTC). Calibrated Speed Control is roughly equivalent to
automobile Cruise Control, where the automobile attempts to maintain a constant
speed up hill and down hill, regardless of load.
Do any modern Diesel or Electric prototypes have Cruise
Control?
RTC
operating under Analog DC has been available in all version 6 firmware modules.
With
the Quantum Upgrade Program, RTC operating under DCC will also be available.
Gerry
Pruss
QSI
--- In
From what
you have said, it sounds like the "upgrade" chips will not close the
loop on speed control via BEMF. Is that
true?
Although I
like the idea of the RTC feature, I fail to see how it will provide accurate
prototypical control without closing the loop on speed control. Simulated inertia isn't going to help the
guys trying to run accurately at low speeds when the motor control is running
open loop without BEMF.
Gerry Pruss
replied
In
both RTC and Speed Control, BEMF is used in a PID (Proportion, Integral, Differential) motor control loop.
In
Speed Control, a fixed target BEMF is specified by the throttle setting (DCC
speed step). The PID motor control loop continually adjusts the power to the
motor (PWM) based on the last BEMF measurement. If BEMF(measured)
< BEMF(target) then the PID loop increases the PWM. If BEMF(measured)
> BEMF(target) then the PWM is decreased.
In
RTC, the PID control loop is the same. However, the target BEMF is not fixed.
There is a second parallel control loop that adjusts the target BEMF according
to the load the locomotive is experiencing. The overall result gives
prototypical behavior such as the locomotive slowing down when going up a hill
or speeding up when going down a hill.
The
overall result is also very apparent when running locomotives in a consist. Without RTC, you must carefully tweak the speed
tables and V Start and V High of each engine so that they don't fight each
other while coupled together.
Otherwise
you see behavior like one locomotive doing all the work and the other being
dragged along. With RTC, two or more locomotives with quite different motor
profiles cooperate with each other.
You
have to look carefully at the couplers between a pair of locomotives to see the
slightest tendency of one locomotive to pull or push the other.
This
probably raises more questions in your mind that it answers. Feel free to ask
more questions regarding RTC.
I
will answer as I find time.
Gerry
Pruss
QSI
Further
notes from Gerry Pruss
Don,
In
listing the advantages of RTC, I forgot to mention the "steadily creep
along over minor obstacles like turnouts" issue and the "jackrabbit
start" issue.
Under
both RTC and Speed Control, a locomotive encountering a minor obstacle will
attempt to push past the obstacle. Under Speed Control the locomotive will try
to supply more and more power up to maximum or until it starts moving again.
Under
RTC the locomotive will supply more power up to a point at which the 2nd
control loop tells it to lower it's speed.
The
result is that for minor obstacles RTC makes the locomotive overcome them
smoothly. For major obstacles, like placing your hand in front of the
locomotive, RTC makes the locomotive initially strain against the obstacle, but
gradually give up.
Under
Speed Control the locomotive will continue to strain against the major obstacle
until its motor burns out.
RTC
also minimizes jerkiness in engines with binding problems under slow speeds.
The binding is like a minor obstacle to overcome.
As
for jackrabbit starts, operating under RTC eliminates the slot car behavior
that you sometimes see under STC and Speed Control when you increase the
throttle rapidly from 0 (even with
CV3
set to 0).
Gerry
Pruss
QSI