TimberSurf’s Model Railway
Modelling tips & Model Railway Controls Guide
The Lumsdonia Railway
I would have once agreed with you, up to the advent of DCC. All need for a flywheel eliminated, as electronically synthesised inertia does the job more effectively, especially in the dead slow scenario where there is very little inertia from a mechanical flywheel, and the track supply remains permanently at full power so stalling is eliminated (and for locos with poor pick up there's capacitor charge storage available too). There are further advantages in the way of reduced motor bearing wear, and the potential for greater loco weight: instead of a brass flywheel and airspace about it for clearance, the void that the flywheel would have taken up can contain a greater volume of yet more dense ballast. I have removed flywheels and the redundant motor shaft from several of my smaller models to increase the loco's all up weight, to very good effect.
I remember some of the motors of the era were still three pole, then they developed five pole, then skewed armatures. All of this was to reduce the "cogging" effect as the armature rotated between the poles.
In the DC days the flywheel helped overcome some current pick-up irregularities but with DCC a break in the current causes more headaches, especially with sound, since everything has to come back to life after a power interruption. Here is where keep alive capacitors earn their keep. With the flywheels at least everything keeps moving and you minimize the time that the power is actually interrupted.
However, if there are no interruptions in rail power to the loco pickups, neither flywheels nor capacitors are necessary. With my electrically bulletproof trackwork, I get along without having either.At least if you intend to use BEMF. A flywheel will always work against it, making the running characteristics erratic and unpredictable. Check with the decoder manufacturers, they will most likely tell you the same.
In short, no programmable acceleration/deceleration curves do not render flywheels irrelevent because they serve different purposes.
Flywheels is a secondary solution to get over problems you're better addressing directly. A free running chassis with good pickups on clean track doesn't need one.
If pickup is lost loco will continue to run following previously set instructions*** until either capacitor is discharged (fractions of a second to 30 seconds, depending on capacitor size and capacitor technology used) or pickup is restored. As the capacitor voltage drops away the loco will slow a bit, then eventually stop.
*** I said "instructions". This means that if you've told the loco to come to a stop, and it is following internal pre-set momentum, the loco should continue to follow that momentum and will stop.
With one maker of DCC chips (Zimo), if the loco comes to rest without pickup, then the loco will attempt to move along very slowly until it finds current again, and then stop. This is to try to prevent a loco stopping on a dead bit of track. ( I have a test loco where I can demonstrate this at will, it does work ).
With one maker of Chip and stay alive components (Lenz, with the Gold chip and Power-1 module), the loco can sense DCC signals from either one rail, or even through paper from both rails. It does this through a capacitance trick. Consequently, a Lenz chip with the Power-1 module can tell if it is on the rails or not, and can receive changes to instructions if on the rails but without full pickup. If off the rails, the decoder stops the loco even though there is power in the capacitor module. With all other systems, the loco will continue to run until the capacitor discharges (This can be a problem - think what damage you can do to a loco in 30 seconds of running without any control over it ! - that might be a reason to not fit a capacitor system which can run for 30 seconds.).
Capacitor technology, size and storage time. There are broadly four types of capacitor (and circuit) available for stay alive. Each has different size and storage times.
The simplest, cheapest and least time per unit volume comes from Electrolytic capacitors. You might see half a second with a huge bank, but really its less than 1/10th of a second. Makes a big difference to microscopic "stuttering" and consequently running quality, but can't do "tricks". 16v rated capacitors are smaller per unit of energy stored than 25v. There are regular arguments about whether you should use 16v or 25v rated components; fundamentally this comes down to the track voltage from your DCC system.
Next in storage time per unit volume come tantalum chip capacitors. These are often rated below 3v, so you need a bank of them in series to get up to DCC voltages (at least 16v, some would argue for 25v). They are slightly more compact than electrolytics.
Next comes "gold caps" or "super caps". Again usually need a bank in series to get the required voltage. The storage can easily be seconds to half a minute or more. The capacitors inside the TCS KA1 and KA2 are half a dozen super-caps in series plus a few other parts for controlling in-rush current, a KA2 can run my test loco for half a minute.
Finally, there are gold caps with step-up voltage circuits. This is what Lenz have in their Power-1 module. They use a low voltage capacitor and electronics to push the voltage up to that required by the chip. Makes the circuitry more complex, and in practise most devices like this are specific to certain decoders.
Of course we are limited in the diameter of flywheel we can install, especially in smaller and therefore, lighter locos. In addition, the requirement for a flywheel is usually to overcome slow speed "stalling" or jerkiness, when a flywheel will be rotating slowly and therefore it's effect will be far less than at higher motor speeds.