Electronics - Feedback Controllers
Some manufacturers do not recommend using feedback controllers with N gauge models. I have seen such advice from Dapol and Gaugemaster. This is a very sweeping statement, and surely depends how much feedback gain the designer builds in to the controller circuit. Using very high gain ensures very good speed regulation, but tends to be unstable. If you set the speed knob to slow and the loco stalls, the controller could ramp up to a full 12V or more. This would not be a good thing, but I am not aware of any actual controller that does this. Most practical designs deliberately limit how much feedback they apply, so they only partially compensate for speed or load changes, but avoid destructive extremes. You can think of this as having the user set the speed knob to a certain setting, then allowing the controller to move it a little up or down from there. You will need to use your own judgement on this issue - as far as I know, no manufacturer publishes any specific details or test results.
I am using a feedback controller design for my new layout (with limited gain), and during testing I wired one up with a toggle switch to turn the feedback circuit on and off. I then tried out a variety of locomotives and operations both with and without feedback to compare the effect. The results revealed a few benefits that I had neither anticipated nor read about.
Before trying feedback for myself, I had anticipated improved speed regulation on gradients and curves, and better results from an inertia option (there is not a lot of point in having inertia with a conventional controller, as normal load variation tends to overwhelm the inertia effect). These benefits are real and useful, but some others proved to be at least as valuable:
I am using a feedback controller design for my new layout (with limited gain), and during testing I wired one up with a toggle switch to turn the feedback circuit on and off. I then tried out a variety of locomotives and operations both with and without feedback to compare the effect. The results revealed a few benefits that I had neither anticipated nor read about.
Before trying feedback for myself, I had anticipated improved speed regulation on gradients and curves, and better results from an inertia option (there is not a lot of point in having inertia with a conventional controller, as normal load variation tends to overwhelm the inertia effect). These benefits are real and useful, but some others proved to be at least as valuable:
- sticky valve gear on steam locos. Many steam models have mechanisms that stick a little at certain points in the rotation of the wheels and valve gear. You can see this when driving the loco slowly. The feedback effect reduced this and gave much smoother low speed operation. It was especially visible on an oscilloscope - you can see the power increasing and decreasing during each revolution of the wheels.
- slow running through pointwork, especially with short wheelbase locos. Most small 6-wheel steam locos tend to hesitate a bit when running through pointwork as slight imperfections or limited power pickups miss a few power pulses from the controller. The feedback circuit compensates for this, so my little Dapol 14xx 0-4-2 now rolls across double slips and crossings as if it really did weigh 40 tons instead of 30 grams.
- voltage drop. The feedback circuit detects the speed loss due to voltage drop on a larger layout and partially compensates for it.
- reduces differences between locomotive types. A feedback controller makes widely different types of locomotives and motors perform more consistently. The controller designer implicitly or explicitly assumes a certain set of ideal loco characteristics. For example, if running at 9V then it should produce 4V of back EMF (about typical for a modern N gauge loco). If the loco is an older, power-hungry design, it will run quite slowly at 9V and might only produce 2V of back EMF, so the feedback circuit gives it an extra volt or two. A overly sensitive loco such as a Dapol 14xx will be running at a scale 100mph on 9V and return about 6V of back EMF, so the feedback circuit throttles it down. This is a critical feature for me, since I need to leave my main line controllers on a fixed speed setting that has to work for all my locos. Note that the speeds will still not be identical or even ideal, but close enough to be usable.