Opinions on throttle direction changes while moving?

[mmagliaro]

I have been refining and testing my wireless-enabled DC throttle a lot over the last few months. 
https://www.therailwire.net/forum/index.php?topic=55335

Yes, the project has continued,
and I'm very close to showing more details and talking about prices in the event that anybody wants one.

This is an operational behavior question:

If you are running a train and hit the direction control, the simplest thing that can happen is it just does what you say,
the train jerks into reverse, probably derails, and this isn't very kind to the motor or gears.   I have seen throttles that try to be
"nice" and simply refuse to do a direction change unless the train is stopped.  I have implemented mine that way (but could easily
remove this "protection")

What I do is the following:
If you flip the direction switch while moving above a certain speed (very very slow), the train does not reverse.  As soon as you slow
down below the safety threshold, THEN it does the reverse.

Exercising it myself, it *seems* to be intuitive.  When you are slowing down, you might jump the gun and hit the direction switch,
and this prevents that from causing a disaster.  When you stop and start up again, your train goes the other way, and since the direction toggle is pointing that way, it goes the way you set it.

But would people find this "weird"?   If you are rolling along pretty fast and hit the direction, it ignores you.  When you stop,
maybe you don't pay attention to your direction switch, or you forgot that you changed directions, so you might be unhappily
surprised that you start back up going the other way.

Safety?  No safety?   Would people prefer that it just does exactly what you say, and if you are careless and hit the direction
while going 50 mph, the wreck is on you?  Or would you rather it ignored the direction change and did it when you stopped?

[nickelplate759]

Max,

My first thought is that if the users throws the reverse switch at an unsafe speed the train should slow down, come to a stop, then reverse.
My second thought is that might be convenient, but might also be wrong.

Thinking about it like a car:

1. on my electric car, if I try and throw it into reverse while moving, it will yell at me (audible alarm plus error message on screen) and ignore my request.  The forward/reverse switch (stalk on the steering columnt) isn't sticky, so it just returns to it's normal neutral setting.

2. on my last manual-transmission car, the gears would grind and it wouldn't go into reverse, but it sounded awful and probably was bad for the transmission.

3. on an automatic transmission, I think there's a lock-out (on newer cars) so that you can't throw it into reverse while moving forward, but it's been so long since I had one that I don't really know.   This seems to be closest to the behavior you describe, but not exactly the same, since your reversing switch would at least temporarily not match the direction of travel.

I like #1 above best for a model train throttle.  You'd need a sprung center-neutral reversing switch.

[Bill H]

Max:
I have seen a number of operators who swear when switching a busy yard that may take an hour or two of work, that it works best to keep the throttle at very low speed and just use the direction switch to make their moves. Of course, their switch engines all have a it of momentum built in, which might not be the case with your throttle or? Just a thought…

Kind regards,
Bill

[Lemosteam]

Is there a way in your electronics to have the direction switch act as a throttle-down and then allow the reverse polarity overt the span of a few seconds?

[C855B]

My first thought is that if the users throws the reverse switch at an unsafe speed the train should slow down, come to a stop, then reverse.
...

That's the way DCC works, even with momentum set to minimum. I like it that way. What I don't like is the idea of switching direction with the throttle as an "execute later" operation subject to manually throttling down to zero. That's a big fat opening for confusion.

With ceramic magnets in DC motors, it is alleged - possibly accurately - that repeated sudden polarity reversals under load weaken the magnet. Flywheels in this case are a load, too. IOW, "Don't do it!".

[jagged ben]

Max,

My first thought is that if the users throws the reverse switch at an unsafe speed the train should slow down, come to a stop, then reverse.
My second thought is that might be convenient, but might also be wrong.

Occasionally I use an old Aristocraft Train Engineer radio throttle and this is how they've implemented it.  It is intuitive and sometimes nice when switching.  This behavior gets my vote.

I believe the slow down and speed up on the Aristocraft is a constant volts per second, i.e. the faster you're going the longer it takes to slow down.  Which makes sense and is how I would do it.

Of course options would be cool, but I know it's more work and not necessary.   But namely:
- Option to turn the feature off or choose another behavior.
- Option to change the speed of slow down and speed up

[mmagliaro]

Hey now... slow down automatically, then reverse and ramp back up to the current set speed...
I had not even considered that.  I could implement this.

But this opens the door to more issues.  It's things like this that make me still wonder if it should do ANY of this,
and if it wouldn't be better to make the operator responsible for not wrecking their own train.   There will always be somebody who thinks it should work some other way. 

These are how I think it should behave:

1.  What if it starts slowing down to do the reverse, and you change the speed UPWARD while it is doing that? 
I would say it just ignores you, and after it gets to a stop and changes the direction, it ramps back to whatever speed you currently have set.

2.  If you move the speed down lower than where the automatic deceleration is,
 it jumps to your new lower speed and then keeps automatically decelerating.  And of course, if you keep turning
the knob down, it's just keeps going down as fast as you turn.

3. What if it starts slowing down and you hit the direction back to where you were?  i.e. you hit reverse, then you hit forward?
It just ramps back up to the current throttle setting.

Complications:
4. When it hits zero, it does the direction change and starts ramping up to wherever the throttle setting is.
But what if you change the speed in the meantime?   I mean, it's straightforward enough to say that moment to moment
it looks at your throttle setting and just keeps ramping in that direction until it gets there. 
But if you don't even have a momentum behavior on, and you want instant response, how does it know to "quit" and just go straight to your throttle setting like it normally would?

And then, as one post mentioned, there is the whole question of "how fast" to ramp down before doing the direction change.
I don't want the special cases and adjustments to get out of hand.   So I will just choose an arbitrary fixed deceleration rate for this.  If you are going faster, it will take longer to stop.  (i.e. the constant "volts per second" that nickelplate759 suggested)  If you really need to change direction NOW, you should be stopping the train yourself.

[Tom L]

Max:
I have seen a number of operators who swear when switching a busy yard that may take an hour or two of work, that it works best to keep the throttle at very low speed and just use the direction switch to make their moves. Of course, their switch engines all have a it of momentum built in, which might not be the case with your throttle or? Just a thought…

Kind regards,
Bill

I do a lot for of switching at sloooow speed and I frequently do this.  My throttles have center off toggles for direction, so I set a slow speed and when changing directions I pause in the center off position and then move it to where I want it.  I sometimes (try) to emulate things like stopping prior to coupling or stop to pickup/dropoff a crew member to throw a switch, so I often move the toggle to center off to stop and then back to proceed in the same direction.

Tom L
Wellington CO

[mmagliaro]

Other quote from Bill:
"Max:
I have seen a number of operators who swear when switching a busy yard that may take an hour or two of work, that it works best to keep the throttle at very low speed and just use the direction switch to make their moves. Of course, their switch engines all have a it of momentum built in, which might not be the case with your throttle or? Just a thought…

Kind regards,
Bill
"

I do a lot for of switching at sloooow speed and I frequently do this.  My throttles have center off toggles for direction, so I set a slow speed and when changing directions I pause in the center off position and then move it to where I want it.  I sometimes (try) to emulate things like stopping prior to coupling or stop to pickup/dropoff a crew member to throw a switch, so I often move the toggle to center off to stop and then back to proceed in the same direction.

Tom L
Wellington CO

I am not using a center-off toggle for direction.  Just a toggle.

But I could allow the operator to just keep the train moving very slowly and kicking the direction switch.  They would have to turn off the speed knob to actually stop, however.     I would just have to relax my threshold for where I will allow a direction change.  Right now, I arbitrarily said, "If it's > 10% of throttle, don't allow it".  But at 10%, most DC engines don't move.   And if I made it more like 30%, then you could get your train moving and "gently" jerk it back and forth by hitting the direction.

What happens if you are barreling down the main, and you put that direction switch in the middle?  Does the train jerk to a halt?
See, this is what I mean when I say, the more I try to implement some sort of automatic behavior, the more there could be
people who would rather it work some other way.

I could put a center-off toggle in my hand-held, and have it set the speed to zero in the center position.  Now, you could use
it just as you describe.  And it would be up to YOU, the operator, to decide what a safe crawling speed is.

[NtheBasement]

Personally I prefer that the throttle does what I say instead of questioning whether or not what I say is correct.

On my Power Cab if I set a momentum value it decels to a stop then accels in the other direction, which makes sense.  That's the throttle though; can't say how any particular decoder handles it if you set CVs 3 and 4.

[Rasputen]

When dealing with magnetic uncoupling ramps that are quite a distance away, I have found that keeping the throttle at a low speed and cycling the direction switch makes it much easier to get the couplers into the delayed position.  On some areas of my friend's layout, you cannot reach the cars in order to uncouple them by hand.

[mmagliaro]

Personally I prefer that the throttle does what I say instead of questioning whether or not what I say is correct.

On my Power Cab if I set a momentum value it decels to a stop then accels in the other direction, which makes sense.  That's the throttle though; can't say how any particular decoder handles it if you set CVs 3 and 4.

What does it do if you do NOT have any momentum set?  Does it jerk to a halt if you change direction while it's moving?
(I do hear you on having it just do what you say, even if what you say is going to wreck the train.)

[Ed Kapuscinski]

When I have a device I want it to behave predictably and consistently. I think that's one of those things from the Universal Principles of Design book I read all those ages ago.

How does that apply here?

Well... do you have momentum built into throttle changes? Like, if I rip the throttle all the way to 100%, is there momentum built in, or does it just max the voltage?
The reverse button should do the same thing.

If you're building a direct control device, make it a direct control device.

[peteski]

On my Power Cab if I set a momentum value it decels to a stop then accels in the other direction, which makes sense.  That's the throttle though; can't say how any particular decoder handles it if you set CVs 3 and 4.

It is not the throttle - when you select momentum values, the Power Cab actually writes new values into the decoder of the addressed loco's CV3 and 4 using programming on main mode.

Max, since your throttle is microprocessor controlled couldn't you just make several options of the behavior available as a user setting?

To me most sense would make to have gradual slowdown when switching directions until the model ls stopped, then gradually speed up to whatever the speed control knob is set to.  If the direction is set back to the original direction before model comes to full stop then just ramp the speed back up to the current throttle speed setting.  The ramp period should be handled like momentum (acceleration/breaking) in DCC decoders.  Basically it is a linear ramp.  So the ramping time will appear very short when the loco is traveling sloe, and much longer to get from zero to full 100% speed.

That way the slow switching will nto be much affected by the delay, but when the loco was traveling faster, the gradual ramp down then up will be noticeable.

[mmagliaro]

First, for Ed's question:
The controller has a switch.  You can run it with or without momentum.  Without momentum, yes, if you wing the dial all the way around to full, your train will go from 0 to full speed immediately.  With momentum on, it will take about 30 seconds to go from 0 to full.
But I don't see what this has to do with errantly (or even deliberately) changing directions while the train is moving.
Are you saying that if momentum is off and the user does this, it should just instantly jerk and change direction?
And if momentum is on, then I could use this scheme of slowing down to a stop, switching direction and ramping back up?

Peteski:
----------
Everything you describe is exactly how I implemented it (today), right down to the linear ramp.
So if you are running at 7 volts, and you hit the switch, it takes about 5 seconds to slow down to 0, then 5 more to speed back up.  If you are at 12 volts and do this, it takes about 8 seconds down and 8 seconds up.  If you are really crawling at, like
3 volts, it only takes about 2 seconds down, 2 seconds up.

"Making user options available" now means you want an interface directly to the computer, like a DCC decoder's CV settings.
I really was not planning on doing that.   I want to keep this thing very simple, with as few controls and variables as possible.

Now... of course, since it is an Arduino that can be plugged into a PC with a USB cable, I *could* eventually write a "configuration" app to run on the PC that would let you tweak a variety of settings and then upload them to the device.  That might be a nifty cool thing to add.... LATER.