Radio-Controlled Locos

[Scrat]

Concerning the article in the last TURNOUT.

A relay to cut off power in case of emergency (mosfet failure to short) is an option. But relays can also become troublesome....

I would like to suggest something similar to the well-known "crowbar-circuit".
The crowbar is normally used as overvoltage protection. If a specific voltage range is exceeded it shorts the supply, blowing the fuse.
This can be adapted to a loco.
If no power is commanded by the controller but (due to a failed MOSFET) power is present, the crowbar circuit will activate, blow the main fuse and bring everything to a standatill.


Failing MOSFET in general:
If using an H-bridge to power the motor, one failed mosfet is not enough.
Two must fail to short in order to get the motor to run.
Two MOSFETS failing at the same instance is a pretty rare event.
If one fails (assume forward, positive), forward negative still blocks and keeps the loco at standstill.
As soon as backwards movement is selected (switching on reverse positive and reverse negative MOSFET) the supply voltage will be shorted as forward positive mosfet is still activated. This will blow the fuse and get everything to safe state.

So if using an H-bridge (and a proper fuse) we already have a certain level of safety.


I would like to hear your opinion about this.

[Dave Noble]

Hi Holger,

Thanks, that sounds like a distinct possibility, but it's all new to me, I'm not into electronics at all. If it achieves the result we are looking for, it will do.

I'll do some investigating.

Dave.

[Scrat]

Hi Dave,

i hope to start a discussion here with members sharing their experience.
So we will not have to re-invent the wheel.

Holger

[Westernkid]

Over the years I have designed and built about half a dozen electric powered locos and I do not think a relay as a cut off switch is a very practical idea. One of my Locos Western Lady which was featured in Turnout some years ago as the loco with the split personality has six motors. The electrical load when pulling away can reach about 120 amps which would require an industrial sized relay. When I was first developing this loco twenty years ago I did try a heavy duty rated relay with DPDC as a reversing switch. It lasted about three minutes before the copper braids between the moving contacts and the terminals melted.
With regards to a Mosfet going short circuit and a runaway motor, I have never experienced this. As most speed controls have several Mosfets if one failed, probably the instantaneous short circuit load would totally destroy it leading to a Domino effect and blowing the others.

My suggestion for a reliable solution is for for a Kill Switch fitted in one of the battery supply cables to the Speed Controller.
Constructed very simply from a piece of 1/4" diameter brass rod to carry the electrical load and mounted between two Terry type clips to which the cables would be fastened.
It would be Bullet shaped, length as required but fairly short with a ring pull fitted at the blunt end, painted Red and installed such that the ring pull protruded from the rear of the locomotive.
It could simply be plugged in when the loco was powered up and a cord or chain fitted from the ring pull to the driving car.
In the event of a breakaway the rod would be pulled out of the clips giving an instant stop. Likewise in the event of a runaway motor the driver could simply yank it out.
If it became a standard then everyone would know how to stop a loco if the situation arose.

[Scrat]

I have never had trouble with relays as reversers.

My locos have wired controllers using sub-D plugs (with the screws removed) underneath the buffer beam. If the loco separates from the train, all I have to do is hold the handset. The plug will come off and the loco will stop.
On my class 216 I use the simple SPDT car relays. 5€ a piece
216 has two motors per bogie.
Controller limited to 120 Amps per bogie.
The relays are connected to form a kind of H-bridge.
If the loco breaks away, all relays de-energise, disconnecting the controller and shorting the motor.
This brings the loco to a stop quickly.
I never burnt one of those relays, even when using this emergency brake feature.

The bullet-shaped emergency stop works fine if the loco separates from the train but does not give any benefit when the loco is run via RC and a fault in the controller causes it to become uncontrollable.

I agree with you that Mosfets have a high MTBF and not all fail to short.
if one does, going in the other direction will reveal the fault quickly.
An H-bridge is (for me) sufficient as two Mosfets failing to short at the same instant is highly improbable.
Radio-controllet masterswitches are also not the big solution.
If the RC has a malfunction, one could use another channel to cut the power supply.
However if the receiver has a malfunction causing the controller to power the motors one cannot be sure that the cutout function still works as the receiver is already malfunctioning.

Let´s not make it too complicated.
One thing must be made sure while progeamming. The most failsafe RC is ised for RC planes.
There, stick down is motor off, failsafe position.
Left stick down for us would mean full speed in reverse.
So it must be assured to program the right way.

General issue with RC locos.
I have often seen this on open days here:
Long train, driver using RC and sitting on the last wagon.
How can he observe the position of points or watch out for hazards on the track?

[Westernkid]

I must admit that my thought were not with RC locos but with ordinary hand held controlled ones. I use a hand held control which has a kill switch wired in with the potentiometer and also plugs into a Din socket so either if the switch is thrown or the plug pulls out then the loco will stop.

It seems that the concerns are more with what happens if something out of the ordinary happens and I think a simple solution would solve the problem.

If the standard solution becomes a complicated technical issue just because of the RC requirements, then just as the Federations complicated Boiler testing through overkill, so they will make simple electric locos more complicated and another possible nail in the hobby.

[Scrat]

This is true.
We need to assess the rangers and their probability.
One can always construct a "what if" that leads to desaster.

I would like to get a group together and share experiences, perform a simple FTA, collect available data on mishaps and then try to come up with a sensible and safe procedure.

[Westernkid]

As I see it the biggest problem with RC is how in an emergency even with an H bridge do you insure that the power is totally cut off to the motors ? Even if there is a dedicated channel holding on a master control circuit I suppose you could argue that even that may go wrong with interference or such.
I have a nasty feeling that technology is going to lead us into problems.
The simple way is to bring in a rule that over a certain gauge, drivers must be in physical touching distance of their locos at all times after all,where is the point of having remote control at a distance. Yes it is necessary for Boats and Planes but for driving a loco which is the whole basic idea of the hobby, you may as well give up and go get a Computer game.

[Dave Noble]

Thanks for the feedback. I want to make it clear that we are not looking for problems here, we just want to assess whether there is a perceived problem, or not, and if so to put in place a simple solution to it. If we agree amongst ourselves that there is'nt a problem, then we don't need a solution.

However, we sometimes operate in public places, and some, particularly Shildon, are quite keen on doing things correctly, so we want to be able to keep them happy if they ask potentially embarrassing questions. It may be that we could argue that a system such as we are considering is not necessary, in which case we need a good argument to put forward.

On the other hand, it could be argued that it might just be a desirable thing to be able to stop a runaway loco, or to even stop it running away in the first place. I suggested a relay as that is far as my electrics go, but other solutions may well be better. I envisaged simply putting a normally open relay between the battery and motor, and keeping the relay energised (ie closed) with one channel of an R/C system, or by passing it's supply through the securely held hand controller. An R/C system conforming to the recommendations in Turnout (which is what most of us use anyway) should be sufficient to open the relay if there is loss of control or signal, or any interference. A brass pin as Clif suggests would also work.

There are certainly mixed feelings about relays, but I've just bought one rated at 100amps, about 2" cube and it only cost £8.50 delivered. Fitting it should be easy, though I'm not sure if I can test it as my little loco which will pull pretty big loads never takes more than 15 amps.

You make a valid point Holger about drivers on the tail of a train. I have only seen this once at our rallies, and I think someone must have had a word in the driver's ear because he doesn't do it now. Anyway, he was failing to conform to our recommendation that 'other safe operating methods eg driver awareness, must still be observed.

If we do insist on some method of stopping locos, it unlikely to be detailed, ie any system which is considered to be safe and reliable would be acceptable, so that it would seem that you have both solved the problem anyway, using relays or brass rods.

[Scrat]

Hi Dave,

I agree.

We do not need to be too precise as there are many ways of dealing with the problem.
If it works, it can be used.

for me the important issues are:

RC control of a NO relay:
How can be assured that if the RC goes haywire it is still reliable enough to shut off the relay? Not an easy point.
There are certain RC sets that need a kind of "lifesign" from the transmitter.
If this is not transmitted it shuts off. Highly implausible for the system to fail to full speed and still transmit valid lifesigns.
Will need investigation but sounds useable

The cutoff system should be located directly to the motors.
If the RC still works but the controller fails a system is needed that can disconnect the motors directly.
A good feature would be also to short the motors in order to reduce the stopping distance.
(like the one I use with my wired controllers)

I can post a block schematic if required.

Most RC systems have a configurable failsafe position. This is also a good tool to bring the loco to a stop if signal is lost.


PS:
we could run a few tests. I have 6 tesla coils of various sizes. They will jam any signal....

[Dave Noble]

You are quite right about the 'lifesign' signal, we have looked into this and it is part of the recommendations in Turnout, so it's covered.

And yes, the cut off must work directly on the motors, not just the signal to the speed controller. Shorting the motors might be unnecessary, a rapid rather than immediate stop is what is required.

A block diagram would be interesting.

[Scrat]

I am just putting together a "testing project".
The artillery car of the PZ201 armored train has been on my list for quite some time.
Now I have started.
I will show how I did the RC stuff.
As the RC system itself has a failsafe mode that switches off all controllers after not receiving any valid signal for some milliseconds I only have to take care about a possible failure of the motor driving circuit (4QD DNO)

[Scrat]

Going a bit further.
The 4QD controllers have a failsafe system.
If power is on but not commanded it will go to failsafe mode and switch off.
So the only critical path left would be the rc receiver or RC to 4QD adapter commanding full speed due to a hardware failure.
A simple stuck-at-high failure could hardly cause this as dynamic signals are used on the rc side.
So the output of the RC to 4QD adapter is the critical part.
I will not leave out a failure of the RC set that generates a seemingly valid input for the adapter or 4qd.
So a separate cutoff path is needed.
Either as previously mentioned with relays directly connected to the motor or at the ignition line of the 4QD.
The problem is how to avtivate this.
If the RC fails and generates a seemingly valid output how can be assured that it still works "good enough" to activate the cutoff?
A separate RC system is no option.
I had several emails with the technical folks at 4QD. They were very helpful so I would put the options of a controller failure to full power or "power applied undemanded" to almost negligible.

Does anyone have further knowledge about the receiver side? I would be interested to know if they also have and output readback functions to detect undemanded commands.

[Scrat]

This topic seems to have gone cold....

My armoured railcar (PZ201) is now ready to run under RC.
I used 4QD components.
Safety features as mentioned in my last post.
I addition the 4QD controller has an ignition function.
This can be coupled to a set of NO contacts on an additional relay and be used to cut motor power in case of emergency.
This should do the job.

[Admin]

Scrat, you say "the only critical path left would be the rc receiver or RC to 4QD adapter commanding full speed due to a hardware failure" and "I will not leave out a failure of the RC set that generates a seemingly valid input for the adapter or 4qd". BUT, you then correctly point out that "A simple stuck-at-high failure could hardly cause this as dynamic signals are used on the rc side". Doesn't this mean that the R/C input to the speed controller must be one of the following,

A legitimate signal from the transmitter, in which case the model does as intended.
A failsafe signal due to loss or corruption of the signal, in which case the model stops.
No signal, in which case the model also stops.

This surely leaves only the speed controller as the potential source of problems, which is where we started, though this isn't really an R/C problem at all.

If however, as you say, the chances of a 4QD controller's "failure to full power or "power applied undemanded" to almost negligible" there can't be a problem.

Lastly, couldn't any failure of the controller could be guaranteed safe by using 3 phase motors, where the speed controller is effectively the commutator.

[Scrat]

The failure of the speed controller is highly unlikely.

Most use H-bridge output so at two transistors have to fail to short circuit.
If one fails (let´s say one for forward) and you want to go backwards there will be a short circuit of the supply voltage, taking out the fuse.
So the first failure is noticed at once.
In addition, the 4QD have an "ignition" contact. This could be used as a backup safety function by adding a relay in series with the manual switch.
This relay can then remotely cut off the ignition and with it the power for the motor controller.
Problem solved.

I have been in contact with Steve at 4QD. He is very helpful when discussing technical details.