Using Capacitors to Simulate Momentum at Signal Stops

[fairflixt]

Hi Everyone,

I noodle around in my basement with different setups until I've annoyed the family up to the point of causing major strife and then pick it all up. As such there is nothing in my setup that's permanent and it also changes the cost/benefit calculus of spending big bucks on DCC equipment when I have perfectly good analog stuff to play with. I've been having fun with automating my setup and trying to replicate some of the automated layouts from the old World of LGB book like:
that alternates between the train on track 1 and 2, letting one run the circuit while the other waits in the station. I've also successfully combined it with an automated reversing loop at one point. It's been fun to figure out how to wire everything and plan a layout that incorporates the more interesting aspects of controlling trains automatically.

On the aesthetic side, however, it is pretty jarring when power is simply cut to one train as it enters the de-energized track in front of a signal. Same goes for start up on the other train that has now been energized. You get the wagons all crashing into each other or watch the loco starting up with the proverbial pedal to the metal. As such I was trying to think of ways to introduce more controlled slowdowns and startups.

My thought was to wire a capacitor up to the portion of the track that is de-energized by the signals. In my rudimentary understanding, the capacitor would then discharge power, slowly dropping power until the train stops (assuming the train won't blow through the signal) and when the track is re-energized, the capacitor will first draw down some of the power before and then slowly start passing power through too the energized track. (By the way, I'm saying power here because I don't know if it's voltage or current or both that are being affected.) Given my non-permanent setup, I'm looking for something flexible that can be applied in a number of designs where a train is stopped by the automatic signals. To my understanding, things like the sophisticated train controllers, Bridgewerks, Jumbo and others won't achieve this with multiple trains all on the same layout, but I'm happy to be proven incorrect.

An alternative I had considered is to use a bunch of 10153s with diodes to drop voltage over each segment prior to the dead section. This would at least make the slow down more controlled, but it wouldn't change the start up part.

Would appreciate any advice or thoughts you have. Also, is there any good way to calculate/guesstimate the size of capacitor to get? Currently was thinking 40000 uF, 25 VDC capacitors as per the automated block design from James Ingraham at automatedcontrols.org, but it would be good understand how far a train will go given a particular capacitor size.

Finally, is there anything that might damage equipment or be hazardous based on what I've described?

Thanks!

 

[Greg Elmassian]

Power is indeed that Volts times Amps, voltage times current.

Capacitors will be charged to a voltage, and will produce current (amps) into your load, the locos.

There are formulas to calculate this stuff, but each loco having a different "load" will make them work at different rates. I think you will find your sidings will be too short to make this easy to achieve.

You would probably be better off with the graduated application of diodes, which will give you a known voltage drop, and do this in several stages.

Again I think your sidings will be too short to do this smoothly.

By the way, the rule of thumb is 1.5 times the working voltage, if you will have 25 volts to charge the caps, you should use something like 36 volt capacitors. Capacitors get very unhappy when you exceed their rated voltage. You will need resistors to mitigate the inrush current or you will probably pop the breaker on your power supply

Greg

 

[fairflixt]

Power is indeed that Volts times Amps, voltage times current.

Capacitors will be charged to a voltage, and will produce current (amps) into your load, the locos.

There are formulas to calculate this stuff, but each loco having a different "load" will make them work at different rates. I think you will find your sidings will be too short to make this easy to achieve.

You would probably be better off with the graduated application of diodes, which will give you a known voltage drop, and do this in several stages.

Again I think your sidings will be too short to do this smoothly.

By the way, the rule of thumb is 1.5 times the working voltage, if you will have 25 volts to charge the caps, you should use something like 36 volt capacitors. Capacitors get very unhappy when you exceed their rated voltage. You will need resistors to mitigate the inrush current or you will probably pop the breaker on your power supply

Greg

Thanks Greg, that's very helpful.

I guess what you mean by graduated application of diodes would also work after the signal if I installed it in reverse order, but this would have to be a permanent install that couldn't be turned off if the train did not have to stop at the signal, right?

If the biggest issue for using a capacitor is different loco loads, then if I, say, keep the operating trains to single motor locos and maybe also ensure they are pulling similar lengths of rolling stock, I could conceivably avoid total catastrophe and try to use them??

My power supply is 24V, so I will look at higher voltage capacitors. Although, if the caps are incorporated into track power, they'll rarely see the max voltage given that I'm not generally cranking the dial to max. Would it be safe to say it's less of a concern then? Would it make sense to still build in a safety margin regardless?

Does a capacitor discharge at their rated voltage or does it change as it's drained? And if you're only supplying 24V can you charge the capacitor to higher voltage than that?

Appreciate the help, as always!

 

[PhilP]

Use 35V rated capacitors..
There are back emf, spikes, and other effects to consider.

Your capacitors will more than likely be polarised, so you will need to consider if you wish to have the ability to run in both directions.
Non-polarised capacitors are available, but higher values are a lot more costly.

PhilP

 

[GAP]

Thanks Greg, that's very helpful.

I guess what you mean by graduated application of diodes would also work after the signal if I installed it in reverse order, but this would have to be a permanent install that couldn't be turned off if the train did not have to stop at the signal, right?

If the biggest issue for using a capacitor is different loco loads, then if I, say, keep the operating trains to single motor locos and maybe also ensure they are pulling similar lengths of rolling stock, I could conceivably avoid total catastrophe and try to use them??

My power supply is 24V, so I will look at higher voltage capacitors. Although, if the caps are incorporated into track power, they'll rarely see the max voltage given that I'm not generally cranking the dial to max. Would it be safe to say it's less of a concern then? Would it make sense to still build in a safety margin regardless?

Does a capacitor discharge at their rated voltage or does it change as it's drained? And if you're only supplying 24V can you charge the capacitor to higher voltage than that?

Appreciate the help, as always!

Capacitor drain and charge time is termed a time constant.
This is a product of resistance multiplied by capacitance T=RxC
R will be the resistance through the motor.
This will give you some idea of what will be happening when the capacitor is discharged via a resistive load

https://www.allaboutcircuits.com/tools/resistor-capacitor-time-constant-calculator/

The capacitor will only charge to the applied voltage and no higher.
This gives a simple explanation

Charging and Discharging a Capacitor

Charging and discharging are the two main effects of capacitors. In this article, you will learn about charging and discharging a capacitor.

studyelectrical.com

The voltage rating of a capacitor is and indicator of the maximum DC voltage that can be safely applied to it. As Greg has said a rule of thumb is 1.5 times the working voltage, ie 24V is your working voltage (power supply max V, even though it will not generally be reached it just may at sometime) therefore safe voltage rating for your capacitors should be 36V min anything higher just adds to safety.

 

[Cobalt6700]

If you do use capacitors to achieve your goal, you would likely need a fairly large cap bank. For info, I use 6 4700uf caps in parallel (at 22v) as power storage in my locos, and that gets me about 2-3 seconds of 'run time' (depending on load) should track power be interrupted.

You would likely need to add a current limiting resistor to your cap bank - capacitors will draw as much current as they can when energised (they will also supply as much current as they can, this would make track shorts dangerous).

I can't remember the resistor value I use (I did work it out and write it down somewhere...). I also use a diode in parallel with the current limiting resistor. The combination limits the charge current, and other than a voltage drop across the diode, allows the cap bank to be discharged at the Locos current requirement.

I do agree with Greg - getting it to work on multiple different Locos on sh9rt sections will require different value cap banks to get the trains to stop at the right place. Diode droppers would likely be the way to go over caps, this going to require most of your track sections to be individually isolated and fed with power and a set amount of diodes.

If you are inside, I would suggest using something cheap and more flexible. An arduino based controller, controlling DC motor driver board/s and using IR sensors as inputs for train detection would allow you to set the acceleration and deceleration rate with a simple change of code - or even a potentiometer input.

This could easily be re-configuerd for different layouts.

I understand this requires you to be able to write some code, however controlling a PWM value based on IR sensor input is pretty simple in terms of coding, and something we could very well help you with.

 

[fairflixt]

If you do use capacitors to achieve your goal, you would likely need a fairly large cap bank. For info, I use 6 4700uf caps in parallel (at 22v) as power storage in my locos, and that gets me about 2-3 seconds of 'run time' (depending on load) should track power be interrupted.

You would likely need to add a current limiting resistor to your cap bank - capacitors will draw as much current as they can when energised (they will also supply as much current as they can, this would make track shorts dangerous).

I can't remember the resistor value I use (I did work it out and write it down somewhere...). I also use a diode in parallel with the current limiting resistor. The combination limits the charge current, and other than a voltage drop across the diode, allows the cap bank to be discharged at the Locos current requirement.

I do agree with Greg - getting it to work on multiple different Locos on sh9rt sections will require different value cap banks to get the trains to stop at the right place. Diode droppers would likely be the way to go over caps, this going to require most of your track sections to be individually isolated and fed with power and a set amount of diodes.

If you are inside, I would suggest using something cheap and more flexible. An arduino based controller, controlling DC motor driver board/s and using IR sensors as inputs for train detection would allow you to set the acceleration and deceleration rate with a simple change of code - or even a potentiometer input.

This could easily be re-configuerd for different layouts.

I understand this requires you to be able to write some code, however controlling a PWM value based on IR sensor input is pretty simple in terms of coding, and something we could very well help you with.

Hi All,

Thanks for the replies, it's giving me lots to think about and some good reading too!

It's very clear that a higher voltage capacitor won't hurt (if I go that route, which I'm starting to doubt) except for maybe the pocketbook.

R will be the resistance through the motor.

If I measure resistance across the contacts on the loco would that be an accurate measurement of this?

I also use a diode in parallel with the current limiting resistor.

It took me a bit to understand what you were doing here but that is a cool idea. This capacitor thing just keeps getting more complicated. How would I select a proper diode, lowest voltage appropriate for the max possible current?

If you are inside, I would suggest using something cheap and more flexible. An arduino based controller, controlling DC motor driver board/s and using IR sensors as inputs for train detection would allow you to set the acceleration and deceleration rate with a simple change of code - or even a potentiometer input.

I'm not opposed to some other solution and I like coding. Do I understand correctly, the Arduino would be installed on the locos, not to power the track section, correct?

 

[Greg Elmassian]

No, you cannot measure the resistance of a motor and get a real reading, since a motor is composed of inductors, not resistors.

The best way to figure the resistance is measure current and voltage at desired speed.

I gather you have never seen what a capacitor does when it encounters over voltage

https://www.nmra.org/sites/default/files/standards/sandrp/pdf/s-9.1_electrical_standards_for_digital_command_control_2021.pdf

Do not cut corners, get the 36 volt or higher capacity.

 

[Software Tools]

Quite simply, if just using a capacitor based solution would readily work on DC powered layouts, everyone would do it!

 

[dunnyrail]

Hi Everyone,

I noodle around in my basement with different setups until I've annoyed the family up to the point of causing major strife and then pick it all up. As such there is nothing in my setup that's permanent and it also changes the cost/benefit calculus of spending big bucks on DCC equipment when I have perfectly good analog stuff to play with. I've been having fun with automating my setup and trying to replicate some of the automated layouts from the old World of LGB book like: View attachment 306718
that alternates between the train on track 1 and 2, letting one run the circuit while the other waits in the station. I've also successfully combined it with an automated reversing loop at one point. It's been fun to figure out how to wire everything and plan a layout that incorporates the more interesting aspects of controlling trains automatically.

On the aesthetic side, however, it is pretty jarring when power is simply cut to one train as it enters the de-energized track in front of a signal. Same goes for start up on the other train that has now been energized. You get the wagons all crashing into each other or watch the loco starting up with the proverbial pedal to the metal. As such I was trying to think of ways to introduce more controlled slowdowns and startups.

My thought was to wire a capacitor up to the portion of the track that is de-energized by the signals. In my rudimentary understanding, the capacitor would then discharge power, slowly dropping power until the train stops (assuming the train won't blow through the signal) and when the track is re-energized, the capacitor will first draw down some of the power before and then slowly start passing power through too the energized track. (By the way, I'm saying power here because I don't know if it's voltage or current or both that are being affected.) Given my non-permanent setup, I'm looking for something flexible that can be applied in a number of designs where a train is stopped by the automatic signals. To my understanding, things like the sophisticated train controllers, Bridgewerks, Jumbo and others won't achieve this with multiple trains all on the same layout, but I'm happy to be proven incorrect.

An alternative I had considered is to use a bunch of 10153s with diodes to drop voltage over each segment prior to the dead section. This would at least make the slow down more controlled, but it wouldn't change the start up part.

Would appreciate any advice or thoughts you have. Also, is there any good way to calculate/guesstimate the size of capacitor to get? Currently was thinking 40000 uF, 25 VDC capacitors as per the automated block design from James Ingraham at automatedcontrols.org, but it would be good understand how far a train will go given a particular capacitor size.

Finally, is there anything that might damage equipment or be hazardous based on what I've described?

Thanks!

This is my take on what you have done, though no points involved at this time I have had the setup out twice. In both cases I find that a pretty slow speed gives a smoother transition to the proceedings with a glide rather than auto stop and Grand Prix start ups. As you can see having the Garden Vid second one start up before the first stops also helps the proceedings.

 

[Cobalt6700]

It took me a bit to understand what you were doing here but that is a cool idea. This capacitor thing just keeps getting more complicated. How would I select a proper diode, lowest voltage appropriate for the max possible current?


I'm not opposed to some other solution and I like coding. Do I understand correctly, the Arduino would be installed on the locos, not to power the track section, correct?

Apologies if I wasn't that clear about the cap bank protection, a picture would have helped. Diode wise I think I went for a 4A gen purpose.

As for the Arduino / other MCU solution - I was suggesting using the Arduino externally, leaving the Locos stock.

This would mean one MCU controlling the whole layout. As for power control, either separate motor boards, or use relays to select which track to power. I would suggest something like the Cytron MD13S for motor control - these can be had for around $10.

Dunnyrail's system looks cool - I think he is using magnets and reed switches for Loco detection? I can hear a relay going so guessing he is switching the power to that bit of track between diode dropper and off?

 

[dunnyrail]

Apologies if I wasn't that clear about the cap bank protection, a picture would have helped. Diode wise I think I went for a 4A gen purpose.

As for the Arduino / other MCU solution - I was suggesting using the Arduino externally, leaving the Locos stock.

This would mean one MCU controlling the whole layout. As for power control, either separate motor boards, or use relays to select which track to power. I would suggest something like the Cytron MD13S for motor control - these can be had for around $10.

Dunnyrail's system looks cool - I think he is using magnets and reed switches for Loco detection? I can hear a relay going so guessing he is switching the power to that bit of track between diode dropper and off?

Yes I am using the LGB EPL system with a magnet beneath the loco. Explained in here some time back by me and others.

 

[fairflixt]

Do not cut corners, get the 36 volt or higher capacity.

Understood, I won't.

For info, I use 6 4700uf caps in parallel (at 22v) as power storage in my locos, and that gets me about 2-3 seconds of 'run time' (depending on load) should track power be interrupted.

Why did you go for a bank of caps vs. a single. I assume cost or space, but wanted to make sure there wasn't some technical reason to do this.

Quite simply, if just using a capacitor based solution would readily work on DC powered layouts, everyone would do it!

Good point. DCC probably will be in my future, but for now, it's a distant future.

This is my take on what you have done, though no points involved at this time I have had the setup out twice. In both cases I find that a pretty slow speed gives a smoother transition to the proceedings with a glide rather than auto stop and Grand Prix start ups. As you can see having the Garden Vid second one start up before the first stops also helps the proceedings.

Hi Jon - these are helpful examples. I think I watched another of your garden EPL setups when originally getting started with the LGB EPL system that has driven me in this direction. Can I just ask some clarifying questions?

When you say glide here I assume you are stepping down voltage to the track section prior to the unpowered/stop track section with diodes? Are you stepping down over multiple stages or just the one? What are all the black square boxes after the reed switch?

Apologies if I wasn't that clear about the cap bank protection, a picture would have helped. Diode wise I think I went for a 4A gen purpose.

As for the Arduino / other MCU solution - I was suggesting using the Arduino externally, leaving the Locos stock.

This would mean one MCU controlling the whole layout. As for power control, either separate motor boards, or use relays to select which track to power. I would suggest something like the Cytron MD13S for motor control - these can be had for around $10.

Dunnyrail's system looks cool - I think he is using magnets and reed switches for Loco detection? I can hear a relay going so guessing he is switching the power to that bit of track between diode dropper and off?

I guess I should have clarified that, much like Jon (Dunnyrail), I am using reed switches and point motors with relays to do the signal switching and powering of tracks on and off automatically. Mostly it's LGB components.

So let's say I want to start messing around with your proposed MCU solution. Where should I go to start learning? I've been watching some Arduino tutorials and I feel relatively comfortable with the programming part, it would be more about putting together the electronics. Can you describe how the layout might look?

 

[dunnyrail]

Hi Jon - these are helpful examples. I think I watched another of your garden EPL setups when originally getting started with the LGB EPL system that has driven me in this direction. Can I just ask some clarifying questions?

When you say glide here I assume you are stepping down voltage to the track section prior to the unpowered/stop track section with diodes? Are you stepping down over multiple stages or just the one? What are all the black square boxes after the reed switch?

Those were my only 2 uses of EPL Automation. In fact the setup for both systems is the same 4 LGB point motors with added LGB switch units. I think there was another vid last Xmas of it running in the drive.

There is no stepping down of power, just the fact that for a short time both trains using the same amount of power.

The black squares that I think you refer to (left)on the vid indoors are LGB wire holders ref 10710, these come 5 in a pack. I do not know where I aquired them and always wondered what they were. I then kind of worked it out and used them on my last Xmas Garden setup for the locals.

 

[Cobalt6700]

Understood, I won't.


Why did you go for a bank of caps vs. a single. I assume cost or space, but wanted to make sure there wasn't some technical reason to do this.

Capacity / space. For example -a 28000uF cap is going to be about 50mm diameter by 120mm long - quite a beast. A quick look on mouser shows these at £40 each too - I think I bought 4700uF's for 70p each.

I guess I should have clarified that, much like Jon (Dunnyrail), I am using reed switches and point motors with relays to do the signal switching and powering of tracks on and off automatically. Mostly it's LGB components.

So let's say I want to start messing around with your proposed MCU solution. Where should I go to start learning? I've been watching some Arduino tutorials and I feel relatively comfortable with the programming part, it would be more about putting together the electronics. Can you describe how the layout might look?

Sounds like you are already half way there then!

Essentially, rather than using the reed switches wired direct to the relays, you wire it all to an MCU. I would suggest an Arduino MEGA - you can buy 'sensor shields' for these that allow 40+ inputs / outputs to be wired up using the standard servo plug 3 pin interface: Signal / +Ve / GND.

As for electrical connections - the reed switches can be directly connected to an MCU - my advise would be to feed the reed switches with +Ve and read the other side with a 'signal' line as a Digital Input - giving you a high state as active (reed switch closed) and low state as inactive (reed switch open).

Relays - depending on what you are using you have different options. I'm assuming that the relays you are using are 24v - at which point you will need to setup a low-side mosfet switch to allow an MCU to drive them.
My advise would be to pickup a/some relay boards designed for use with 5V logic - have a google for 'Arduino relay board'. These can be wired to Signal / +Ve / GND with signal being a Digital Output.

For motor control you will need 1 Digital Outputs for Direction, 1 Analog (PWM) Ouput for speed and a GND (for the MD13S I suggested). Whether you use 1, 2 or a number of motor boards is up to you - you may want to use one with layout power switched with relays or one for each track section - I'm not too sure what would be best. It is worth noting that you probably don't want to have 2 motor boards powering each other. Not something I have tried, but I could see that may not end well.

For adjustment of the speeds / stopping distances - something like 3 potentiometers. 2 which will allow you to vary the low speed / high speed setting, and another which will allow you to vary the accel / decel time. Again, if using a sensor shield, you set these up as voltage dividers and read the voltage on the signal pin (Analog Input), then 'map' this voltage against the variable you want to change (in Arduino code, this is pretty simple stuff).

If you are going down the Arduino route - I would suggest having a look at https://github.com/contrem/arduino-timer which is a great library for timing and sequencing events (I used this on my Arduino based point / signal control system) and GitHub - JChristensen/JC_Button: Arduino library to debounce button switches, detect presses, releases, and long presses which is useful library for doing things with many Digital Inputs - it has customisable debounce built in (which helps prevent false triggers - I used this with reed switches for a reversing loop) and allows you to preform actions based on event type - e.g. you may want to 'trigger' something on input change, or only do something whilst that input is in one state.

Hope that makes some sort of sense

 

[fairflixt]

Capacity / space. For example -a 28000uF cap is going to be about 50mm diameter by 120mm long - quite a beast. A quick look on mouser shows these at £40 each too - I think I bought 4700uF's for 70p each.



Sounds like you are already half way there then!

Essentially, rather than using the reed switches wired direct to the relays, you wire it all to an MCU. I would suggest an Arduino MEGA - you can buy 'sensor shields' for these that allow 40+ inputs / outputs to be wired up using the standard servo plug 3 pin interface: Signal / +Ve / GND.

As for electrical connections - the reed switches can be directly connected to an MCU - my advise would be to feed the reed switches with +Ve and read the other side with a 'signal' line as a Digital Input - giving you a high state as active (reed switch closed) and low state as inactive (reed switch open).

Relays - depending on what you are using you have different options. I'm assuming that the relays you are using are 24v - at which point you will need to setup a low-side mosfet switch to allow an MCU to drive them.
My advise would be to pickup a/some relay boards designed for use with 5V logic - have a google for 'Arduino relay board'. These can be wired to Signal / +Ve / GND with signal being a Digital Output.

For motor control you will need 1 Digital Outputs for Direction, 1 Analog (PWM) Ouput for speed and a GND (for the MD13S I suggested). Whether you use 1, 2 or a number of motor boards is up to you - you may want to use one with layout power switched with relays or one for each track section - I'm not too sure what would be best. It is worth noting that you probably don't want to have 2 motor boards powering each other. Not something I have tried, but I could see that may not end well.

For adjustment of the speeds / stopping distances - something like 3 potentiometers. 2 which will allow you to vary the low speed / high speed setting, and another which will allow you to vary the accel / decel time. Again, if using a sensor shield, you set these up as voltage dividers and read the voltage on the signal pin (Analog Input), then 'map' this voltage against the variable you want to change (in Arduino code, this is pretty simple stuff).

If you are going down the Arduino route - I would suggest having a look at https://github.com/contrem/arduino-timer which is a great library for timing and sequencing events (I used this on my Arduino based point / signal control system) and GitHub - JChristensen/JC_Button: Arduino library to debounce button switches, detect presses, releases, and long presses which is useful library for doing things with many Digital Inputs - it has customisable debounce built in (which helps prevent false triggers - I used this with reed switches for a reversing loop) and allows you to preform actions based on event type - e.g. you may want to 'trigger' something on input change, or only do something whilst that input is in one state.

Hope that makes some sort of sense

Thanks for sharing Cobalt! I'm sorry it's been a while for a response but with Thanksgiving and reading up on everything you're talking about, it took some time to come up with a plan. The Arduino solution seems like a fun project and a great set of items to throw on my xmas wish list. But that means it will be a while before I have any feedback to share.

In the interim, I bought some capacitors and wanted to test out the capacitor solution anyway. More as a project to mess around with given the likelihood it won't work well for a variety of locos and rollingstock. "Doing the math" on a current limiting resistor, I assume V=IR applies here and you select a max current you want to have for your DC voltage. I figured 1A would be plenty for the one and two motor LGB locos I'm using. So at 24V that would require a 24 Ohm resistor. Did I get that right?

I'll probably also make a few diode bridges to test a voltage step-down method which Jon (dunnyrail) successfully used to do a pretty decent looking slowdown.

So, thanks again to everyone for all the great ideas and I'll do my best to share anything I've done, if/when I get a chance to do it!

 

[Paul M]

It will be interesting to hear the results. Working things out mathematically never seems to convince me, especially as they seem to change the rules every now and then.

 

[dunnyrail]

I'll probably also make a few diode bridges to test a voltage step-down method which Jon (dunnyrail) successfully used to do a pretty decent looking slowdown.

Sorry F fairflixt I did not use Diodes to reduce power, just 2 locomotives running together for a short time slow them down as I referred to in post@14. To clarify that further the start Reed of the second train is before the stop isolation of the first one.

 

[fairflixt]

Sorry F fairflixt I did not use Diodes to reduce power, just 2 locomotives running together for a short time slow them down as I referred to in post@14. To clarify that further the start Reed of the second train is before the stop isolation of the first one.

You're right! I guess I was combining Greg's suggestion of a staged slowdown via diodes with your method of doing the transition at slow speeds to look more natural. I have a 5A power supply so running the trains simultaneously did not have an impact on the start/stop speed when two trains were running simultaneously vs. a single train.

 

[dunnyrail]

You're right! I guess I was combining Greg's suggestion of a staged slowdown via diodes with your method of doing the transition at slow speeds to look more natural. I have a 5A power supply so running the trains simultaneously did not have an impact on the start/stop speed when two trains were running simultaneously vs. a single train.

That could be why mine slow down as I am using a modest 1A LGB small controller.