Electronic Boosters

[Kenton]

I've been thinking about a situation that seems to crop up quite often. It relates to DC and I know a similar problem may be handled differently is DCC but please let's keep the ideology of converting the world to DCC out of this - at least until the DC situation or solution has been fully explored.

 

The problem

We have probably all experienced it. A 2 axle loco crossing a point frog loses power and suddenly stops. The classic is on an insulfrog point but we have probably also seen it on electrofrog and even hand built though to a much lesser extent.

An un sprung loco becomes a 3 wheeler for just long enough to interupt power.

 

The mechanical solution

The first line of attack is obviously to ensure power collection on all 4 wheels and to ensure that the rails are level. This does go some way to ensuring the two wheels that are still on the point switch rails collect power but is no solution for the wobble induced on crossing the frog.

 

The electrical solution

Is there one?

 

My mind drifted to the idea that power is only required for a short duration. So would it be possible to design a (simple) circuit to sense the presence of power and to cut in with a backup, battery?, if that power is removed. Obviously, such a cut in would be very temporary, otherwise how would we stop?

 

Is such a thing simply ridiculous?

[noiseboy72]

Why not ?

A track independant supply (On board rechargeable battery) with a circuit to monitor the incoming track voltage and averages say, 1.5 seconds of values. Apply this figure to an on board speed controller. A technically complex system, but one that would work well.

 

Simple solution: A non polarised electrolytic capacitor across the motor terminals will supply a tiny power boost if the track voltage falters. As fitted to the old Hornby 3 rail kit.

[Oldddudders]

Isn't this what Relco purported to do? I think "ionising the air-gap" (whatever that means) was one of their straplines/claimed benefits. I used Relcos on my last DC layout, and I think they helped a bit - all points were insulfrog.

[Flying Pig]

It was discussed before on this thread. Unfortunately, it seems that it's emphatically not possible.

[Kris]

Use a capacitor.

[Guest baldrick25]

Use a capacitor.

 

Think you will find that the value needs to be so large, that the physical size would be prohibitive in any scale. You would need capacitors measured in tens of Farads , at a voltage greater than the expected line running voltage of at least 14 or 15Volts ( and thats cutting it very very fine).

The addition of a large capacitor in parallel with a DC motor can be also counter productive , in that it acts as a 'brake'. The motor , when its rotating and not receiving any track supply acts as a generator. anything connected across the motor terminals, be it lights, resistors or capacitors, will absorb the energy, adding braking to motor- now a generator.

OK the brave will say that capcitors don't dissapate power, ( not resistive ) but what happens is that it stores the energy produced , and its dissapated in the copper wire of the motor armature.

Want to demonstate the effect. Pick a small motor with a flywheel fitted , wire it to a power supply on the workbench and see how quickly it stops revolving when the supply is cut. Then repeat and connect a suitable electrolytic capacitor as its supply is cut , and watch the motor stop in less than a revolution.

To add an electronic flywheel to a small motor will end up with effectivly what is a UPS ( Un-interuptable power supply) and a lot of complex , circuitry.

Stick to a mechanical flywheel, on the fastest rotating components , usually the armature shaft.

[AMJ]

I have a Bachmann 0-4-0 Gas Mechanical that I have changed the DCC chip in to a DCC Concepts with stay alive. This have been programmed so that it will operate on DC as well. I think that it seems to run better with fewer interruptions as mentioned in the OP.

 

Might be worth considering fitting even if you are just DC. But please do not operate DCC fitted locos with Relco or feedback controllers.

 

I

[Jol Wilkinson]

Even if an electrical/electronic solution is possible, relying on poor mechanical systems such as Insulfrog points and rigid four wheel chassis in the first place will always be unsatisfactory.

 

Anyone considering using the Insulfrog, when there are alternatives readily available, should think again.

 

The question of a rigid chassis is a little more difficult if you use RTR products, but kit/scratch builders should always think about some form of springing/compensation for better pickup and perfromance.

[Kenton]

Thanks for all the thoughts so far.

 

Perhaps I need to reiterate that I was thinking along the lines of a switched circuit with polarity sensing and not along the lines of a simple capacitor. I accept that a capacitor in theory would be a simple solution but excludes itself on the basis of the size and possibly other factors.

 

On the issue of where we start from:

As for using insulfrog - I would agree the usual route is never to go down that route. However, we have to accept that Peco Settrack and all their insulfrog points do exist out there, are used by many and continue to be sold. One of the reasons for this musing is my present contemplation on using odd bits of this track in a project despite my better experience telling me NO. I am also certain that there are layouts out there with this already in place.

 

A flywheel on the motor is also not always a workable option and my experience with such additions has not always been much of an improvement - especially where a heavy loco is concerned.

 

I also have a rather indifferent opinion on springing. The old adage of a poorly sprung mechanism is far worse than a rigid mechanism - and most of my attempts at springing tend to be poorer than I would wish.

Yet both of these are mechanical rather than electronic.

 

I have a Bachmann 0-4-0 Gas Mechanical that I have changed the DCC chip in to a DCC Concepts with stay alive. This have been programmed so that it will operate on DC as well. I think that it seems to run better with fewer interruptions as mentioned in the OP.

 

Might be worth considering fitting even if you are just DC. But please do not operate DCC fitted locos with Relco or feedback controllers.

 

I

 

It is this possible solution that I see as going in the right direct and on the surface seems to offer a simple and practical solution.

I was not aware that a DCC chip could be operated under DC and was under the belief that the output would be poor and possibly damage the decoder.

[bertiedog]

All DCC chips should be 100% compatible with DC control, but will not improve the performance over a DC loco without a chip, they do not store power in any form.

 

It is not practical just to use large capacitors to store power, it is done with a PWM circuit by Lenz , which keeps the power going by efficiently pulsing the tiny amount of stored capacitor power for a few moments when contact with the track is lost. This is rather along the lines of a switched power supply, but increasing the frequency and adjusting the pulse duration to just keep the motor running for a few moments.

 

Large capacitors do exist and are used in constant lighting circuits, but work at lower voltages and small currents compared to motor demands.

 

If a large capacitor was used across a motor it would have to be in the range of 10 to 20 FARADS, and be able to stand a reverse voltage of at least 16 volts and a working forward voltage of the same. These power caps do exist, they are about the size of half a house brick!!

 

The tiny High Farad capacitors with very high values are extremely low voltage rated, usually 2 volts or less, if charged, they can discharge at a reasonable current, but the voltage would be far too low to work a motor.

 

What Lenz do is feed the low voltage to a switching circuit and voltage doubler, at a higher frequency, and then rectify it to supply power to the output, it works for just a few moments, dependant of the speed setting at the point of loss of power.

 

However there is another approach and that is to have a power supply onboard in the form of a storage rechargeable battery, and this could power the DCC or the DC output at point of loss.

 

You must remember however that at the point of loss of power, that you also loose the control of the train, and supplying power at that point causes a run away!!! Even suddenly turning off the power whilst running would trip the circuit and keep the train running in the worst situation.

 

Obviously the circuit could have a limit set on how long to maintain power, but working out the value would be plagued with problems!!! It would be fine as long as contact returns, but would be complex to implement in a satisfactory way.

 

hope this helps,

Stephen.

[Kenton]

What Lenz do is feed the low voltage to a switching circuit and voltage doubler, at a higher frequency, and then rectify it to supply power to the output, it works for just a few moments, dependant of the speed setting at the point of loss of power.

So does this mean the Lenz decoder would do the job for DC or not - some confusion here (Remember my OP was strictly discountin ANY conversion of a DC circuit to DCC. But what I am seeing hinted at is the addition of a (specifically this Lenz) decoder would operate on DC to produce the same over-runnning as on a DCC circuit?

 

However there is another approach and that is to have a power supply onboard in the form of a storage rechargeable battery, and this could power the DCC or the DC output at point of loss.

 

You must remember however that at the point of loss of power, that you also loose the control of the train, and supplying power at that point causes a run away!!! Even suddenly turning off the power whilst running would trip the circuit and keep the train running in the worst situation.

 

Obviously the circuit could have a limit set on how long to maintain power, but working out the value would be plagued with problems!!! It would be fine as long as contact returns, but would be complex to implement in a satisfactory way.

This is back along the lines of my original musing though I see the "rechargeable" battery as supply an unnecessary complication. I accept the over run timing issue but clearly do not see this as an issue. We would be talking in terms of a fraction of a second and would have no desire to stopo a loco on the "dead" section/frog - the whole point of the exercise being to avoid that situation in all circumstances. The event would only happen at slow speeds anyway and if the track power had been turned off the circuit would detect no power at the end of the "battery boost" period and shut down anyway. I could see that the detection of the power drop out being critical else a stutter would be observed. Also unless there were extra complexity in circuit detection the level of battery boost would need to be detected. But again I see this as not required as the "boost" period would be so small that the change in output would have little time to generate the de/acceleration required to be noticeable. A few pulses of say 3-6v probably for say 100ms would be enough to plough on through the frog? Control of the train is never really lost and there is never any chance of a run-away.

 

A circuit would have to be smart enough to detect a voltage existing and the lack of one, switching in the battery at the same polarity (obviously one polarity or the other) and of continuing to detect a voltage following the pulse period. Obviously if the pulse period passes and there remains non voltage no further pulses are made. To my simplistic mind this appears to be easy.

[Guest baldrick25]

A circuit would have to be smart enough to detect a voltage existing and the lack of one,

The voltage doesn't drop to zero,or even drop, it is the value of the back EMF of the motor, together with there being so much inductance around, an instantaneous drop to zero would produce an infinte inductive kick that would make a sensing circuit go potty. The voltage across the wheels, at the time the from supply interupts it , is very variable depending on the position of the commutator, and could be quite large. There is no circuit so no current will flow though. take the body off a loco and connect a scope to have some fun, preferably a storage scope.

Good luck with the experiments though, and do keep us updated how you do it.

[Kenton]

Good luck with the experiments though, and do keep us updated how you do it.

have nothing fancy like a scope - and probably wouldn't remember 45 years ago when I last used one, on how to use it.

 

This is after all me just pondering the possibilities a thought process around why no such thing exists rather than a concrete circuit proposal on how to practically do it.

 

I had no idea that the voltage would not drop to 0v across the wheels when effectively disconnected from the power.

[bertiedog]

The Lenz power maintaining chip requires full DCC control to function properly,,it will not work with DC control, DC merely bypasses the chip operation and passes a DC to DC control signal. As far as I know the chip will not keep a DC to DC signal, unless an expert on the exact circuitry knows differently

.

If the interruption in contact occurs, then detection is going to have to be as complex as the DCC power maiintaining chip, so why re-invent the wheel and just use a DCC controller and the power maintaining chip?.

To design a DC control circuit that detects a lack of current flow and then changes to a PWM type delivery to stretch the available power is a challenge, don't forget the power storage large value capacitors only work at about 2 volts max, so you are going to need to convert the voltage to higher frequency AC to be able to use voltage doubling or switched power supply circuitry to get the voltage to a level were it can run the motor, even for a couple of seconds.

 

Lenz do it by dividing the DCC signal into shorter pulses that stretch the signal from the DCC circuit to keep an already running,motor going. No motor is going to be maintained by under 2 volts at about 100ma, but can be maintained at 6-7volts at say 10ma for a few moments, as long as it is in pluses.to stretch out the meagre power available.

 

Stephen.

[bertiedog]

I think the issue of the speed of the loco may be the killer problem with a non DCC solution, the Lenz circuit knows the speed it is operating at by the nature of the DCC control signal at that point when power is lost suddenly.

 

Although an analogue voltage is an approximation of the control signal it is nowhere as accurate, and getting a circuit to work with a range of voltages and current drains would be a nightmare.

 

It could work with one hand picked loco, with a circuit tailored to the exact requirements of the motor, and over a pre determined range, but the thoughts of designing a circuit to respond to various locos and control voltage settings would be a nightmare.

 

The Lenz chip was a chance discovery, a by product of the design of the pulse train output of the chips, as they discovered that once running they could shorten the length of each pulse, but still keep the motor turning on amazingly low amperages.

 

Because the DCC control signal is already a string of pulses, it was fairly easy to find a way to keep the motor going with very short higher frequency pulses, based on detection of a sudden disconnection from the DCC signal, not voltage or amperage. It knows the signal has gone instantly, and then starts the process.

 

Hope this helps outline the difficulties, I really think an analogue version of the chip is going to be very difficult to build, after all there is no DCC signal to use as the tripping signal, it has to be based on the current flowing,, and this is going to vary with speed and load.on DC.

 

I do not want to damp the enthusiasm for developing an idea, but the problems are legion on this.

 

Stephen.

[Guest jim s-w]

Hi Kenton

 

Given a consistant set of track and wheel standards there is no wobble on a frog as all wheels are always supported by the rail. That is your solution for the mechanical option.

 

Hth

 

Jim

[Kenton]

The Lenz power maintaining chip requires full DCC control to function properly,,it will not work with DC control, DC merely bypasses the chip operation and passes a DC to DC control signal. As far as I know the chip will not keep a DC to DC signal, unless an expert on the exact circuitry knows differently

.

If the interruption in contact occurs, then detection is going to have to be as complex as the DCC power maiintaining chip, so why re-invent the wheel and just use a DCC controller and the power maintaining chip?.

Oh well that kills the Lenz chip option then

 

As said at the start "please let's keep the ideology of converting the world to DCC out of this" the main reason for this is that I firmly think DCC is not the answer to everything and is generally a waste of money for a small single loco layout. It is nice to know that a fulll DCC conversion/solution would work but it is not what was being proposed. There are thousands of model railway enthusiasts out there who don't care about DCC and frankly wish it would stop being offered as the solution to all the world's ills. I am not quite one of those. I do have 2 DCC controllers but their use is very limited as I find DC far more practical.

 

To design a DC control circuit that detects a lack of current flow and then changes to a PWM type delivery to stretch the available power is a challenge, don't forget the power storage large value capacitors only work at about 2 volts max, so you are going to need to convert the voltage to higher frequency AC to be able to use voltage doubling or switched power supply circuitry to get the voltage to a level were it can run the motor, even for a couple of seconds.

 

I come to this quite naive, it was really a layman's approach to why couldn't it be done.

 

As I see it the capacitor argument has also gone the way of the Lenz chip and don't see why you have revisited it here. Am I being just dumb here but a loco motor can be driven perfectly well off a small 6-9v battery - though not for any great length of time. I cannot see why this cannot be switched in and switched out for a short timed period - even with something as basic as a reed switch/relay. Equally detection of the lack of voltage (sorry, lack of back EMF) could be detected by some transistor magic.

[Kenton]

Given a consistant set of track and wheel standards there is no wobble on a frog as all wheels are always supported by the rail.

Even with peckups on all 4 wheels:

 

Consistent Peco track standards ?

Insulfrog point drop dead plastic frog?

Classic rigid chassis of most 0-4-0?

 

Most definitely a wobble.

 

Consistent OOSF/EM/P4 hand-built track,

(none of the plastic frogs)

Fully sprung (not 3 point compensated)

 

Most definitely no wobble.

 

I appreciate it is a good argument to go fine scale but what a heck of a load of fuss and bother when there are remains a significant majority of modellers who operate their "train sets" out of the box.

[Guest jim s-w]

Hi Kenton

 

It's nothing to do with finescale as the 00 standard works just fine, Peco is a system too. With an electrofrog point and the correct wheels there will be no wobble. The problem comes when a factor is changed such as a 00 kits built with wheels that are too fine.

 

Cheers

 

Jim

[bertiedog]

Forget the DCC (I was not promoting it!), forget the Capacitors, etc, all you left with is a Battery, and it will have to be at least 12 volts to allow for losses and the higher speed range, (assuming it should work at most settings of speed).It could be designed to work only at a crawl in some way.

 

Then the electronics have to detect a lack of current flow, not voltage, as there will be a voltage from the motor windings as the field collapses. The circuit has to work at a voltage of about 2/3 volts to about 10 volts, and at any of these voltages it will have to take over supply of power, and have a time-out before going back to the track, and if still no power on the track continue for a few moments longer etc.

 

It must therefore cut in at the voltage the failure occurs at, and continue or fade slightly until power returns.

 

As before it cannot continue ad infinitum as then a run away would occur.

 

The circuit must work in reverse as well as forward,and at various currents at each voltage dependant on load, Phew ........this is complex, very complex indeed..

 

Frankly I do not think an analogue circuit can do this easily, it is going to have to be a digital processor to do the calculation of the control voltage, and replacement of the missing current flow from the track. A small processor could be programmed to do this quite well, driving a D/A converter to a control circuit that regulates the power from the battery. The program logic could handle reversing etc., and the cut-off

 

But in the end this is all available of the shelf from DCC Lenz!! or use a bigger flywheel.............and live frogs with wheels wide enough for the track standards so that there is no drop or loss of contact on point crossways,......met by P4, 0r using 16.4 mm gauge etc etc.,

 

Stephen.

[Dagworth]

I don't see why a digital circuit should be needed. A fairly simple circuit with a 9volt battery feeding into a power transistor, the base of which is driven by a small capacitor charged up from the rails and which discharges fairly rapidly through said transistor. Charging of the cap can be via a bridge rectifier to protect it, with the output to the motor polarity switched by a latching relay dependant on the polarity of the track feeds.

The cap will only charge up to the voltage on the rails, and thus the power from the battery through the transistor could be matched to the loco's speed at power loss.

 

Andi

[Michael Delamar]

Kenton

 

A few days ago I would have agreed with you and be looking for alternatives.

 

I had a Hornby black 5 that was a bit noisy and growly at slow speed, would stall and wouldnt run smooth.

 

was suggested I try a dcc lenz standard chip, I already had a ncu powercab controller I bought a few years back but that has only been used on my O gauge diesel which still sits on a short stretch of track.

 

well I got the chip, under £20, popped it in, has a plug already there and it takes up very little space.

 

performance now, I can still operate and control the loco fully with the wheel like a normal controller.

 

but the performance is in a different league, itll run at a virtual crawl, it doesnt stall which was my biggest hate, the motor is now virtually silent and accelleration/deceleration is much smoother

 

Ive never wanted to go dcc, I think a lot of it was down to people banging on about it all the time, tends to put you off and put the blinkers on, I can find a better way etc.

and I dont really like the techy side of programming, its a bit nerdy which has put me off.

I dont even like talking about it much, I want to talk about the real railway and detailed models of that which run well first, controllers later.

not bothered if an exhibition layout has got a DCC sticker on the front of it, if its a big one its usally a bunch of blokes standing behind looking down punching in numbers in a controller everytime another train has to move.

people bore me with questions about it when operating at shows. "is this layout dcc?" "no"- "ok thanks" walks away. I dont get the attitude sometimes, however....

 

 

I dont want it to run loads of locos on one track with lights, it was the running quality I wanted, on a fast roundy roundy layout, I probably wouldnt bother but for slow speed shunting its brilliant.

 

I only have a small fleet of shunting locos' but they will all have this accessory in the future, and I think it should just be looked at as an accessory, same way if your adding extra detail bits,fancy gearboxes or better finer wheels. if your spending money to have all this, the accessory you plug onto the motor which will run it a lot better is worth having in my opinion.

 

Ive got the basic gist of the theory as to why it works, your trying to put 1-2 volts through the track, wheels, pickups to a motor, its got a chance of stalling.

 

with the extra voltage constantly running through them and the motor being controlled by the digital signal, its bound to run better.

 

as has been said,why try and reinvent the wheel.

 

Ive seen the light, people do bang on like its a cult that you should join and be converted to, but when you have a go and see for yourself, its worth doing it

 

Mike

[Kenton]

I still cannot see why the circuit would need to determine (and cater for) the range of speeds at drop out.

 

We are talking about a very short duration and a preset mid-range speed - 6v would be adequate. The actual visible change in speed of the loco would be almost imperceptible and even a minute lurch or slowing would be far preferable than the dead stop that is observed now.

[34theletterbetweenB&D]

...A flywheel on the motor is also not always a workable option and my experience with such additions has not always been much of an improvement - especially where a heavy loco is concerned....

For DC operation, a flywheel is easily the simplest solution to maintain movement. It 'knows' what speed and direction the motor is turning at all times, and can easily coast a model over a momentary power loss. But it has to be engineered, if it is to work satisfactorily. Especially at dead slow track speed where the problem of track power loss is most acute consideration needs to be given to flywheel mass as a proportion of the all up mass of the model, motor shaft speed, and thus the reduction gear ratio of the model, and the most efficient drive train possible so that kinetic energy in the flywheel is not wasted moving a very lossy drive train. So it may well be necessary to rebuild the model into a form to optimise the flywheel action if this is the path chosen...

[Kenton]

Assuming that there is enough space in there to fit a flywheel in the first place - typically the small 0-4-0 that might benefit often has the least space to fit one.