Keep-alive for DC?

[AndyID]

One of the nice features of DCC is it lets you add a "keep-alive" super-capacitor to you locos so that they keep rolling when power is temporarily interrupted. It's a bit like having a big flywheel on the motor.

 

So why can't we have super-capacitors capacitors on DC locos? You could strap a super-capacitor across the motor to keep the motor running for a reasonable amount of time but there are a couple of problems. S'caps typically have maximum voltage limits of around 3 volts and they cannot be reverse polarized. The positive terminal cannot be negative without damaging the capacitor.

 

However, in theory, you should be able to connect two s'caps back-to-back to make a non-polarized capacitor (protection diodes might also be required) and if you string several s'caps in series you can increase the maximum voltage across the combination (although there are some things to watch out for) . Unfortunately the series configuration reduces the total capacitance but s'caps can have a lot of capacitance, so that might not be a major issue.

 

Someone probably tried this before and found it does not work but I'm going to give it a shot. The bits should be here next week

 

 

 

Edited November 13, 2020 by AndyID
Typo, as usual

[Nigelcliffe]

Unless you add some control electronics, how are you going to achieve a precise start or stop of your loco ? 

 

With a "big capacitor" charged up, you turn down the DC controller to zero, and the capacitor discharges slowly.   Some time later the capacitor will get to zero, but during that discharge time it is powering the loco.   It will simulate a loco with lots of momentum and no brakes.

 

The big benefit of DCC Stay-alives is seen at low speeds.  The wheels don't stop when a microscopic bit of dirt interrupts the track pickup at low speed.  But not clear how that works with low DC voltages.  

 

Someone on RMWeb demonstrated a DC capacitor system a few years ago.  Lag in response to DC control changes was a limitation.  

 

I'm sure it can be done, but I think it needs a control system to place the energy store (capacitors) away from the speed control.  Even then, there is a problem to overcome;  does "zero DC track input" mean a pickup gap (so stay-alive needed) or does it mean "stop" ?   

 

 

- Nigel

[Hibelroad]

As above, the problem would be control, both when the loco is drawing its power from the track and when it’s using capacitor power only. With dcc the chip will take power from any source and then it adjusts the motor supply according to the controller setting. With DC control there could be some fancy electronics in the loco to do a similar job but this be expensive to develop and only end up as a downgraded version of dcc, so would not be worth doing. 

[AndyID]

10 hours ago, Hibelroad said:

As above, the problem would be control, both when the loco is drawing its power from the track and when it’s using capacitor power only. With dcc the chip will take power from any source and then it adjusts the motor supply according to the controller setting. With DC control there could be some fancy electronics in the loco to do a similar job but this be expensive to develop and only end up as a downgraded version of dcc, so would not be worth doing. 

 

It's really not so different from DCC or an actual flywheel. Unless it's transmitted by radio the DCC signal is lost when the loco loses contact with the rails. The DCC decoder maintains the last speed command for some time interval. The trick is to get the time constant right so as not to get too much or too little inertia (or momentum if you prefer).

[Crosland]

It's very different.

 

The energy stored in a capacitor is proportional to the square of the voltage.

 

With DCC the stay-alive is always charged to the full track voltage. On DC it can only be charged to the voltage on the rails.

[AndyID]

21 minutes ago, Crosland said:

It's very different.

 

The energy stored in a capacitor is proportional to the square of the voltage.

 

With DCC the stay-alive is always charged to the full track voltage. On DC it can only be charged to the voltage on the rails.

 

True, but the effect is still the same or at least similar. DCC maintains speed in the absence of control signal for an interval although it can keep the speed constant whereas the cap and the flywheel both slow as the energy is consumed. Despite the anticipated indignation I'm not criticizing DCC in any way. I'm just exploring the possibility of a more "brute force" method. Also note that it's posted in the non-DCC section

[DCB]

  The problem I foresee with super caps on DC is when something either shorts out or someone changed the direction switch on the controller before stopping the loco.  

[AdamsRadial]

There was a sort of keep-alive system called a free-motor back in the late sixties I read of (either RM or MRC, can't recall which). It was a small DC motor connected in parallel with the traction motor, and it had a flywheel on the output shaft. When you applied power to the loco the free motor started up first because it had less load on it. As it ran up to speed it drew less current and the traction motor then began to turn. With both motors running at whatever speed the controller determined the system was in a sort of equilibrium. When the loco passed over a dodgy track and lost connection with the controller supply, the back emf from the free motor's flywheel inertia was supposed to keep the traction motor turning enough to get it across the dead spot. I don't know how effective it might be, but it's one of these ideas I've often thought about giving a trial, especially now that motors are so small and comparatively cheap.

Edited November 13, 2020 by AdamsRadial

[AndyID]

5 minutes ago, AdamsRadial said:

There was a sort of keep-alive system called a free-motor back in the late sixties I read of (either RM or MRC, can't recall which). It was a small DC motor connected in parallel with the traction motor, and it had a flywheel on the output shaft. When you applied power to the loco the free motor started up first because it had less load on it. As it ran up to speed it drew less current and the traction motor then began to turn. With both motors running at whatever speed the controller determined the system was in a sort of equilibrium. When the loco passed over a dodgy track and lost connection with the controller supply, the back emf from the free motor's flywheel inertia was supposed to keep the traction motor turning enough to get it across the dead spot. I don't know how effective it might be, but it's one of these ideas I've often thought about giving a trial, especially now that motors are so small and comparatively cheap.

 

Thanks! Yes, I tried that once but it didn't work well at all. When running on "flywheel power" there was too much voltage drop and the traction motor immediately slowed down. It might work if the voltage applied to the flywheel motor from the track was stepped up somehow but that would require some electronics.

[pauliebanger]

Hi AndyID,

 

As discussed above, adding stay alive supercaps to DC models is frought with problems and cannot be achieved without some electronic control circuitry.

 

There is a relatively simple way to achieve supercap stay alive control on DC however by using a ZIMO DCC decoder in DC analogue mode. This definitely works with ZIMO decoders, see below, but as far as I'm aware, no other brand has this ability. Happy to be proved wrong if anyone knows differently.

 

Nigel highlighted the most significant problem on DC. How to differential between a lack of track power due to poor conditions (undesired) and the reduction in power due to reducing the 'speed' controller (desired). There would be little point in having a stay alive fitted which removes the operator's ability to bring the model to a halt where planned. ZIMO decoders are capable of this on DC analogue.

 

Additionally, some ZIMO decoders are equipped with all the required circuitry to connect directly to supercaps, thereby removing the need for additional components.

 

Have a look at this linked thread where this issue is discussed. The ZIMO decoder in the clip happens to be a sound decoder, but I believe the non-sound equivalents can also operate stay alive caps on DC analogue.

 

 

Best regards,

 

Paul

Edited November 13, 2020 by pauliebanger

[Dagworth]

Heehee, here we go again

 

We've done this subject a few times over the years, worth a re-read of a couple of old attempts.

This one was quite fun although it may not make complete sense as there was one very obnoxious chap who argued that black was white and then deleted most of his posts.

 

 

Around the same time this thread was also running 

 

 

Have a read and see where we got to in those threads, may save a lot of time and effort!.

 

Andi

Edited November 13, 2020 by Dagworth

[Pete the Elaner]

2 hours ago, Crosland said:

It's very different.

 

The energy stored in a capacitor is proportional to the square of the voltage.

 

With DCC the stay-alive is always charged to the full track voltage. On DC it can only be charged to the voltage on the rails.

 

Which when running at low speed (when you most need the stay-alive) is not a lot.

 

I like the suggestion made by Paul: use a Zimo with a stay-alive.

 

If you've heard that decoders don't perform well on DC then that is old news about old decoders & is similar to believing that cars can't do much more than 70mph because this was true of those built in the 1950s.

Modern decoders work well on DC & in many cases help to improve running, partly because they won't start until the track voltage is 6(ish) volts. This would also give the stay-alive something half-decent to work with.

[Ben Alder]

Aaaah, dear old Kenton, miss his somewhat  individualistic input to the forum......

Edited November 13, 2020 by Ben Alder

[AndyID]

3 hours ago, Crosland said:

With DCC the stay-alive is always charged to the full track voltage.

 

I doubt if that's the case. Single supercaps have a maximum voltage of around 3 volts. Presumably DCC decoders with keep-alive include a boost converter to allow the motor to be powered from the cap.

 

EDIT: I just looked at the caps decoders use and they are mostly electrolytic so no problem charging them to track voltage and no need for a boost converter. The snag is they don't store very much energy but it's probably enough in a lot of situations.

 

 

 

 

 

Edited November 14, 2020 by AndyID
More info

[AndyID]

4 hours ago, Pete the Elaner said:

If you've heard that decoders don't perform well on DC then that is old news about old decoders & is similar to believing that cars can't do much more than 70mph because this was true of those built in the 1950s.

Modern decoders work well on DC & in many cases help to improve running, partly because they won't start until the track voltage is 6(ish) volts. This would also give the stay-alive something half-decent to work with.

 

Thanks. I had not heard that although I do happen to have one and it is rubbish. I doubt very much that it supports any sort of short duration BeeGee's "stayin' alive" but as I said this isn't a criticism of DCC (although, somewhat predictably, many seem to be taking it that way). It's an experiment to see if I can take advantage of super-capacitors. It may well turn out to be a complete fiasco but at least I will have learned something. There are some aspects of it that really concern me but nobody has raised any of them thus far.

[Crosland]

10 hours ago, AndyID said:

 

I doubt if that's the case. Single supercaps have a maximum voltage of around 3 volts. Presumably DCC decoders with keep-alive include a boost converter to allow the motor to be powered from the cap.

 

OK, perhaps I should have said constant voltage. With supercaps on DCC it is invariably boosted back to the track voltage, or somewhere near so the effect is the same. SOmepeople add series supercaps to get the voltage tolerance but you then lose a lot on total capacitance.

 

[RAF96]

All the stay alive does in a DCC situation is keep the decoder running. The motor runs at the last good command speed and direction. If the track voltage is interrupted by a dirty spot or plastic frog the stay alive continues to power the decoder until the signal is restored. Without a stay alive the decoder dies on the dead spot then reboots when track signal is restored. If DC running is enabled  it can then shoot off at max chat until it gets a cohesive signal again.

 

Compare with DC where the track voltage is altering with the throttle setting and the stay alive merely acts like a reservoir of power emulating momentum.

[AndyID]

18 hours ago, Crosland said:

 

SOmepeople add series supercaps to get the voltage tolerance but you then lose a lot on total capacitance.

 

 

The experiment will use series caps to increase the voltage. The reduced capacitance is a benefit because the 1.0 F caps I'm using are probably a bit too much

 

I'm a bit concerned that I will not be able to make a non-polar supercap by connecting them back-to-back. It works with electrolytic caps but I can't find any references to doing it with supercaps. In theory it should work but it could well be a step too far.

 

The other concern is that the capacitor won't actually charge to the full track potential (the reverse biased cap should work like a sort of diode) which means that the motor voltage will immediately step down if track power is interrupted. There will be Schottky diodes across the reverse biased caps to minimize this but that might not be enough.

 

 

[AndyID]

18 hours ago, RAF96 said:

All the stay alive does in a DCC situation is keep the decoder running. The motor runs at the last good command speed and direction. If the track voltage is interrupted by a dirty spot or plastic frog the stay alive continues to power the decoder until the signal is restored. Without a stay alive the decoder dies on the dead spot then reboots when track signal is restored. If DC running is enabled  it can then shoot off at max chat until it gets a cohesive signal again.

 

Compare with DC where the track voltage is altering with the throttle setting and the stay alive merely acts like a reservoir of power emulating momentum.

 

The stay alive caps on the likes of Zinio seem to be much too small to keep the motor running for enough time to do much at all so I suspect you are right. I think there are some decoders that use supercaps and I would think they can continue supplying motor current for a reasonable amount of time. But as my knowledge of DCC is almost non-existent I better leave it at that.

[Nigelcliffe]

4 hours ago, AndyID said:

 

The stay alive caps on the likes of Zinio seem to be much too small to keep the motor running for enough time to do much at all so I suspect you are right. I think there are some decoders that use supercaps and I would think they can continue supplying motor current for a reasonable amount of time. But as my knowledge of DCC is almost non-existent I better leave it at that.

 

Paul's post on Zimo decoders, above, discusses supercaps, has a link to a video of a loco using Zimo decoder + capacitors, running on DC across a "party trick" bit of paper, whilst maintaining full control from DC.  All written up in a magazine last year.  Paul is very experienced in what works on Zimo devices.  
I've used 0.1F (supercaps in series) on Zimo decoders, which will run a loco for a yard or more (though any sensible user sets the CV setting to limit the no-DCC-signal run time).  It will also do a useful trick of nudging the loco along until it finds power if it were to stop on a dead spot as it goes to zero speed.     

Numerous other makers offer supercaps which can be added to decoders. 

For most purposes, 0.1F is over-kill, something a bit smaller will do the job.  

 

- Nigel

 

Edited November 15, 2020 by Nigelcliffe

[AndyID]

20 hours ago, Nigelcliffe said:

 

Paul's post on Zimo decoders, above, discusses supercaps, has a link to a video of a loco using Zimo decoder + capacitors, running on DC across a "party trick" bit of paper, whilst maintaining full control from DC.  All written up in a magazine last year.  Paul is very experienced in what works on Zimo devices.  
I've used 0.1F (supercaps in series) on Zimo decoders, which will run a loco for a yard or more (though any sensible user sets the CV setting to limit the no-DCC-signal run time).  It will also do a useful trick of nudging the loco along until it finds power if it were to stop on a dead spot as it goes to zero speed.     

Numerous other makers offer supercaps which can be added to decoders. 

For most purposes, 0.1F is over-kill, something a bit smaller will do the job.  

 

- Nigel

 

 

Thanks Nigel,

 

The video is very impressive!

 

My s'caps should arrive on Tuesday and it should not take too much time to determine whether or not I am up a gum tree   It has occurred to me that the overrun problem (train doing a simulation of the Largs event) might be minimized if the DC track controller not only giveth but also taketh away. That should not be too difficult to do although it might require a fair bit of current dumping. Hopefully wheels will not weld to the rails in the process.

 

Andy.

[AndyID]

So far so good.

 

I made a non-polarized 12 volt 0.25F cap from two banks of four 1F caps connected back-to-back. There are Schottky diodes across each bank to minimize reverse polarity. I strapped it across the controller output and tested a variety of locomotives and the results are pretty much as I expected. It's actually a lot easier to control the locomotives that I thought it might be. Slowing down and stopping at the correct spot doesn't seem to be too difficult.

 

It has not been tested extensively of course but so far the non-polarized capacitor technique seems to work as well with supercaps as it does with other electrolytic caps.

 

[Crosland]

<pedant mode> You should also have bleed resistors across each capacitor to equalize the charge (also help to discharge when the power is off) </pedant>

[Barclay]

In MRJ 109 was an article called 'The Electronic Flywheel' -  He worked in 7mm, used 2 x 470 microfarad capacitors, (36 volt, non-polarised) and with RG7 motors achieved a run-on of 1/3 of a loco length. This would be plenty to prevent stalling on a grain of dust etc. and not enough to require any kind of control circuitry to prevent running away. I assume the new capacitors are much less bulky. 

[AndyID]

4 hours ago, Crosland said:

<pedant mode> You should also have bleed resistors across each capacitor to equalize the charge (also help to discharge when the power is off) </pedant>

 

Good point. I should at least measure their voltages.