Model Railway Electronic Inertia

[dkvfx]

Hi

I posed a question on the NewRailwayModellers forum as follows:

New Railway Modellers Forums - Model Railway Discussion Forum For Beginners • View topic - Electronic inertia

"
As you know there are often moments when your loco loses power momentarily and results in less than smooth running.

Some locos have a flywheel fitted to help with this, but these are, as far as I can tell, practically impossible to retro-fit to locos.

So I've been thinking for a while. Surely there is an electronic solution to this with the fitting of capacitor circuits within the loco. As the loco travels, the capacitor charges. If, for example, the loco finds a momentary deadspot on some points, the capacitor discharges through the motor to continue the travel.

I would imagine this would require electrolytic capacitors which are polarity sensitive, so if you need both forwards and backwards smoothness, then some proper circuitry would be involved.

I also think that because I cant find any mention of this idea online, then there are inherent problems, or I'm not using the right language in my searches.

Any thoughts?
"

I've had a less than promising result, but I havent been convinced it cant be done.

The problem of polarity switching could surely be resolved using diodes, even if you use 2 capacitors for the different directions of travel.

Now I'm not particularly electronic savvy, although I have in the past created some sophisticated circuits, my basic knowledge is all gone. For me, electronics is not like riding a bicycle.

Any thoughts anybody?

 

[blueroomelectronics]

Don't most DCC controllers support inertia?

 

[dkvfx]

This is for good old fashioned Analogue.... Sorry should have mentioned that as it was pointed out that DCC can do this

 

[blueroomelectronics]

I had an old set of books from the 1970s called Electronic circuits for model railways. They included several throttles with inertia, I no longer have them but your local library may still have a copy. They used a capacitor to set the decay time. The simplest used an SCR and was a neat design (had a switch for speed nota pot)

 

[dkvfx]

I think Ive got that book somewhere. That circuit was built into the controller. Im planning to fit the circuit into the loco itself.

 

[crutschow]

I think Ive got that book somewhere. That circuit was built into the controller. Im planning to fit the circuit into the loco itself.

The problem likely is the size of the needed capacitor.

How much voltage and current does the loco take and how long would the capacitor have to supply power?

 

[dkvfx]

A quick look says that the typical draw is 100 to 200mA. The voltage is anything between -12v and +12v depending on the controller setting.

Locos with tenders either have the motor drive in the engine or the tender leaving space in the other section for some smoothing electronics connected by a small cable.

The amount of time the capacitor has to supply depends on the speed the loco is running, but whatever it is, it doesnt need to be long... just moments.

the nice side effect is, when you shut off power, the loco will slow to a stop rather than stopping dead

 

[crutschow]

The formula is C=(I*t)/V where C is the capacitance, I is the motor current, t is the time, and V is the voltage drop during that time.

So for example for 200mA and a 5V capacitor voltage drop over 0.1s, the required capacitance is 4,000µF.

Edit: Since the voltage is plus and minus and large capacitors are polarized, you need to connect two capacitors back-to-back with a diode across each cap to prevent reverse voltage.

 

[ljcox]

The problem likely is the size of the needed capacitor.

How much voltage and current does the loco take and how long would the capacitor have to supply power?

If you do the maths, as I did many years ago, a capacitor charge/discharge does not represent true inertia.

I wrote down the equations of motion of a train and designed an analogue computer to simulate them.

The basic circuit & maths are attached.

I have more info available if anyone is interested.

Note that x is the distance travelled by the train, x' is its speed & x'' the acceleration. Fm is the motor force & Fb the brake force.

a & b are friction coefficients.

 

[crutschow]

If you do the maths, as I did many years ago, a capacitor charge/discharge does not represent true inertia.

That may be true, but he's not primarily interested in true inertia, just something to keep the train moving smoothly when the voltage is momentarily interrupted by poor track contact. A capacitor in the loco should work fine for that, if there's room for one large enough.

 

[ljcox]

That may be true, but he's not primarily interested in true inertia, just something to keep the train moving smoothly when the voltage is momentarily interrupted by poor track contact. A capacitor in the loco should work fine for that, if there's room for one large enough.

My thought on that issue was to apply a high, current limited voltage across the tracks so any brief breaks in continuity are broken down by it. The 12 Volt would be fed via a diode in order to prevent the HV being applied to the controller.

 

[crutschow]

My thought on that issue was to apply a high, current limited voltage across the tracks so any brief breaks in continuity are broken down by it. The 12 Volt would be fed via a diode in order to prevent the HV being applied to the controller.

How would a diode prevent the high voltage from being applied? A diode stops reverse voltage, not forward voltage.

 

[dkvfx]

will have to look at that when my brain is less hurtable. But like the guy says. True inertia would be great, but not the main problem.

So, its possible?

 

[MikeMl]

You would have to store energy in a supercapacitor, and then use a PIC-controlled SEPIC converter to keep the motor voltage at what ever level existed just before the dead spot.

 

[dkvfx]

sounds interesting, but for a luddite like me, i'll have to look at the three things I know nothing about!!!! It looks pretty complicated as the chips wont be constantly supplied with power. But thats just because of my lack of knowledge and my brief look on Google for "PIC-controlled SEPIC".

but you have hit on the problem of outputting the steady voltage.

 

[MikeMl]

I was once hired to design such a system. A SEPIC converter is a special type of Switch-Mode Power Supply (SMPS) which can make a regulated output voltage which either is higher or lower than the input voltage. It can also hold the output voltage constant even though the SuperCapacitor is discharging. The SEPIC can also step up the rail voltage to a higher voltage so that the SuperCapacitor is fully charged even though the train was running very slow before it hit the dead spot.

You need the microcontroller to provide the sampling of the rail-to-rail voltage prior to hitting the dead spot, and to manage charging the SuperCapacitor while track power is available. When the voltage disappears, the uC sets the output voltage of the SEPIC to match what existed just prior to the interruption. All it can do at that point is hope that the rail voltage is reestablished before the SuperCapacitor runs down. The motor gets the rail voltage if it is present, and gets the SEPIC output if it is not.

 

[ronv]

10,000 ufd 16 volt caps are not that large & I bet it would give you some benifit.

 

[ljcox]

How would a diode prevent the high voltage from being applied? A diode stops reverse voltage, not forward voltage.

That is quite right. The diode blocks the HV from the power supply but not from the tracks.

See drawing.

I have not tested this idea. I don't have a Model Train.

But logic tells me that a HV must breakdown any oxide film.

The only question is how high?

I would try 50 Volt initially.

 

[crutschow]

That is quite right. The diode blocks the HV from the power supply but not from the tracks.

See drawing.

I have not tested this idea. I don't have a Model Train.

But logic tells me that a HV must breakdown any oxide film.

The only question is how high?

I would try 50 Volt initially.

HV will break down an oxide film but you scheme does not apply high voltage to the film which occurs between the track and the train. The voltage on the track stays essentially constant whether the oxide film opens that connection or not.

 

[peufeu]

Higher voltages would create more sparks, which would deteriorate the rails and wheels quicker.

Online Sketch: Untitled Sketch by Anonymous Author

First, rectification, then an inductor to prevent huge sparks when the wheel reconnects to the track after breaking, then big capacitor.

Then you need to control the motor's direction, this can be done with an integrated 4-mosfet H-bridge. Not represented is a sensing circuit which uses the last seen voltage on the rails to choose the motor direction.