Jump to content
 

Keep-alive for DC?


AndyID
 Share

Recommended Posts

I ran some tests with five of the AVX 1F caps in "normal" series (pos to neg to pos etc.) feeding them from a 12 volt supply. As expected there is some variation in the voltages on the individual caps but they are all well below the rated 3 volts. The highest was less than 2.5 volts.

 

Then I discharged the series combination through a resistor and reversed the connection to the 12 volt supply. Apart from the fact that the voltages on the caps were now negative there was very little difference at all. The discharge rate was the same.

 

These are double-layer supercaps which are nothing like conventional electrolytic capacitors. As far as I can determine the electrodes are symmetrical and the designated polarity is arbitrary. I assume the manufacturer assigns a polarity to differentiate these devices from traditional non-polarized capacitors. I don't think you would want to use them to pass a relatively high frequency current but in our situation we are only reversing the polarity after a considerable time.

 

More testing required in case there is a big "gotcha" hiding in the weeds but so far so good.

  • Like 5
Link to post
Share on other sites

On 21/02/2021 at 21:43, AndyID said:

I'm thinking five caps might be better as it will provide a bit more voltage margin at twelve volts.

I've just been having a play with Smokey Joe and was wondering the opposite...

 

It runs so much better with the caps attached; temporarily unplugging them returns it to the jerky heap of junk it used to be, and reminds me why we never used to run it all those years ago. So I'd like to make it permanent.

 

Instead of an attached wagon, three caps would easily fit in the loco. Four might, but with more of a faff. The thing with it is that it's so fast the voltage for the practical top speed is only about 4 to 5 volts; any higher and it's rocket propelled and in great danger of flying off at the bends. So under normal circumstances three caps would do fine, as you would never want to apply 12v anyway. Hell, two would be fine.

 

Except, of course, if 12v was accidentally applied, then there would be a strong possibility of the magic smoke being released.

 

I'm not enough of an electronics person to know if there's any way to limit the maximum voltage inside the loco to prevent this. A voltage divider with low value resistors across the motor and capacitors, perhaps? Two opposing zener diodes in series?

 

I'm going to try some more in an old diesel next. Plenty of room in there for all the capacitors in the world; turn off the power and it'll keep going for hours...

 

 

  • Like 2
Link to post
Share on other sites

Hi Phil,

 

I was thinking along the same lines for a L&Y Pug. The idea was to have an area that could only be supplied by a controller that was limited to six volts or even three volts which is not hard to do. I even bought some 3 and 6 volt motors but I have not got around to fitting them in the Pug. At 3 volts we could get away with a single cap and it would actually reduce the "overrun" effect because of the larger current drain.

 

We might end up with something that's just about as good as radio control for a lot less expense and complication (actually that's what got me thinking about this in the first place). That these caps turned out to work so well for this was a pretty big fluke :)

 

It's definitely not a good idea to over-stress the caps beyond their rated max. voltage, and it's a good idea to stay a bit below that to increase their life. As you say it should be possible to prevent over-voltage on the caps with Zener diodes and some sort of fuse or non-linear resistor (a filament light bulb?) but that might take up a fair bit of space too.

 

I'll see what I can come up with but someone might have a better idea in the meantime.

 

Anyway, I hope you are having fun. Of course it's not quite the same as landing a vehicle on Mars but it is exploration nonetheless :D

 

Cheers from snowy Idaho!

 

Andy

  • Like 2
Link to post
Share on other sites

On 24/02/2021 at 02:59, AndyID said:

Of course it's not quite the same as landing a vehicle on Mars but it is exploration nonetheless :D

 

The descent of Curiosity was known as "eight minutes of terror" because it took eight minutes for the signals to get from Mars to Earth to indicate whether it had worked or not; if I give Smokey Joe a blast of full power it feels like that's how long it takes for the little blighter to stop without crashing into anything...

 

I'm going to make a new set of caps for Smokey Joe that fits in the loco so I moved the caps into an old Hornby diesel instead. This raises a new problem; running is super smooth when it's going, much smoother than without, but starting is not so good.

 

I think it's because it's just a primitive three pole ringfield; it stops slowly enough that I think it nearly

always stops with the motor stalled, so it won't start again either without a shove or, occasionally, a quick blip of high power, but that's no good for small movements.

 

This motor was pretty ropey anyway, it is prone to stalling or just buzzing away to itself at low power even after new bushes and a clean, so I'm tempted to try one of these cd motor conversions to see if that improves matters.

 

I love watching it going though. Seeing an 80s model with inadequate pickups crawl smoothly over complicated pointwork as if it wasn't there is amazing.

 

On 24/02/2021 at 02:59, AndyID said:

or non-linear resistor (a filament light bulb?) but that might take up a fair bit of space too.

 

Flickering fireboxes for DC as well as DCC? :-)

 

Edited by Phil Himsworth
  • Like 1
Link to post
Share on other sites

  • 1 month later...
On 01/01/2021 at 05:21, AndyID said:

This is where I ended up.

StayAliveCap.jpg.57e4ec8771452748d7eec004449b3c69.jpg

 

At first I thought I would need eight 3 volt caps for 12 volts, but it only takes four. The combination of four 1F caps in series produces an effective capacitance of 250,000 micro-farads which will allow any locomotive to cruise through a dead frog without any hesitation. Using pure DC, crawl speed over a long distance is pretty impressive too.

 

These caps are 8mm in diameter and 12 mm long.

 

This method of connecting electrolytic capacitors back-to-back to create a non-polarized capacitor has been around for a very long time but it's usually used with AC where some amount of charge oscillates between the caps. This is a bit different and there could be some degradation of the caps over time. I have not noticed any but I have not conducted a lot of tests.

 

Oh, and Happy New Year!

 

 

 

 

 

 

Hi Andy I hope this discussion is not closed, as I am clearly coming late to this topic. The performance you achieve is impressive congratulations. 

May I ask about the 1 ohm resistor in series with the capacitor & motor.  If a motor is drawing say 0.25 A at 12 V this 3W so it is seems to me quite a beefy resistor which could get hot. Has this been a problem? I presume is to determine the time constant of the keep alive action. Can the resistor be moved to just be in series with the capacitors and not the motor?

Cheers Tom

Link to post
Share on other sites

1 hour ago, OldTom said:

 

Hi Andy I hope this discussion is not closed, as I am clearly coming late to this topic. The performance you achieve is impressive congratulations. 

May I ask about the 1 ohm resistor in series with the capacitor & motor.  If a motor is drawing say 0.25 A at 12 V this 3W so it is seems to me quite a beefy resistor which could get hot. Has this been a problem? I presume is to determine the time constant of the keep alive action. Can the resistor be moved to just be in series with the capacitors and not the motor?

Cheers Tom

 

Hi Tom,

 

Still active as far as I know :D

 

No problem with resistor heating. The dissipation is I squared R which is around 65 milliwatts.

 

Cheers,

Andy

  • Like 1
  • Informative/Useful 1
Link to post
Share on other sites

Of course! Silly me! Thanks for the quick reply.

 

Now I really look forward to hearing of any of your updates on this scheme, and trying this out myself (once I have finished wiring my new layout that is!).

 

Cheers,

Tom

 

 

 

Link to post
Share on other sites

On 18/04/2021 at 12:39, OldTom said:

Of course! Silly me! Thanks for the quick reply.

 

Now I really look forward to hearing of any of your updates on this scheme, and trying this out myself (once I have finished wiring my new layout that is!).

 

Cheers,

Tom

 

Hi Tom,

 

I have tried it on several more locos. No apparent problems so far.

 

I should really come-up with a controller that giveth and also taketh away the stored charge. That would prevent embarrassing "overruns". It's possible to do the same thing with an emergency brake that just shunts a lowish value resistor across the controller output but I think most people would prefer a single speed control knob.

 

There's also the possibility of a low voltage version that would only require one or two s-caps. That might be useful for small locos that don't have a lot of space for the caps. The motor would have to be low voltage too of course. 3V might work and that would only need a single cap. Of course the track voltage would have to be limited. I have a half-baked idea about an unstoppable Airfix Pug :)

 

Cheers,

Andy

  • Like 1
Link to post
Share on other sites

On 23/04/2021 at 02:03, AndyID said:

 

Hi Tom,

 

I have tried it on several more locos. No apparent problems so far.

 

I should really come-up with a controller that giveth and also taketh away the stored charge. That would prevent embarrassing "overruns". It's possible to do the same thing with an emergency brake that just shunts a lowish value resistor across the controller output but I think most people would prefer a single speed control knob.

 

There's also the possibility of a low voltage version that would only require one or two s-caps. That might be useful for small locos that don't have a lot of space for the caps. The motor would have to be low voltage too of course. 3V might work and that would only need a single cap. Of course the track voltage would have to be limited. I have a half-baked idea about an unstoppable Airfix Pug :)

 

Cheers,

Andy

 

Thanks for the update Andy. A single cap with a low voltage motor has a lot of volume efficiency merits. One would need to prevent accidental supply of an over voltage if also using other locos though.  Perhaps this protection could be done with a Zener diode circuit on the loco?

 

May I ask if you have thought how the system would respond to a feedback controller, since it would seem this might not work because it never gets to read the back emf. Of course it would likely not be necessary to use a feedback controller but I would be worried that there may be undesirable effects in the controller itself if used, in which case you would have to be careful to use a simple dc controller with cap equipped locos.

Cheers Tom

  • Interesting/Thought-provoking 1
Link to post
Share on other sites

5 hours ago, OldTom said:

 

Thanks for the update Andy. A single cap with a low voltage motor has a lot of volume efficiency merits. One would need to prevent accidental supply of an over voltage if also using other locos though.  Perhaps this protection could be done with a Zener diode circuit on the loco?

 

May I ask if you have thought how the system would respond to a feedback controller, since it would seem this might not work because it never gets to read the back emf. Of course it would likely not be necessary to use a feedback controller but I would be worried that there may be undesirable effects in the controller itself if used, in which case you would have to be careful to use a simple dc controller with cap equipped locos.

Cheers Tom

 

Hi Tom,

 

Yes some sort of protection would be a good idea. Zeners or maybe just a current limiter at the controller.

 

Feedback in the usual form would not work but I don't think there would be any adverse effects if the controller did have a feedback feature. There is another way to implement feedback by sensing the capacitor's discharge rate when the current from the controller is interrupted but that would require a special controller.

 

Cheers,

Andy

  • Like 1
Link to post
Share on other sites

1 hour ago, AndyID said:

 

Hi Tom,

 

Yes some sort of protection would be a good idea. Zeners or maybe just a current limiter at the controller.

 

Feedback in the usual form would not work but I don't think there would be any adverse effects if the controller did have a feedback feature. There is another way to implement feedback by sensing the capacitor's discharge rate when the current from the controller is interrupted but that would require a special controller.

 

Cheers,

Andy

Hi Andy,

Thanks for your response.  I can see that the a special purpose feedback controller could be in principle be implemented but I think it takes the complexity to orders of magnitude more if using a micro controller. My guess is that a feedback controller wouldn’t really give you anything above S-cap keep-alive for slow speed running anyhow. Do you agree? My worry was whether the keep-alive would actually damage a feedback-controller, but thinking about it, as you say it is unlikely to damage anything.

Cheers Tom

Link to post
Share on other sites

1 hour ago, OldTom said:

Hi Andy,

Thanks for your response.  I can see that the a special purpose feedback controller could be in principle be implemented but I think it takes the complexity to orders of magnitude more if using a micro controller. My guess is that a feedback controller wouldn’t really give you anything above S-cap keep-alive for slow speed running anyhow. Do you agree? My worry was whether the keep-alive would actually damage a feedback-controller, but thinking about it, as you say it is unlikely to damage anything.

Cheers Tom

 

Hi Tom,

 

Whether it would be worth the extra complexity probably depends on the type of layout. Feedback helps to maintain speed when the load changes due to curves and gradients. It's probably of little benefit on a level, linear sort of layout with decent power distribution.

 

It could be done with a micro controller but it would also be possible to do it with a purely analog circuit although that would still be more complicated than a basic controller.

 

It's also possible to do it with an analog circuit that continuously senses the current and modifies the output voltage to effectively cancel out the voltage drop across the motor's internal resistance. The output is smooth DC. That works without caps and it should work with caps too.

 

Cheers,

Andy

 

 

  • Like 1
Link to post
Share on other sites

12 hours ago, AndyID said:

 

Hi Tom,

 

Whether it would be worth the extra complexity probably depends on the type of layout. Feedback helps to maintain speed when the load changes due to curves and gradients. It's probably of little benefit on a level, linear sort of layout with decent power distribution.

 

It could be done with a micro controller but it would also be possible to do it with a purely analog circuit although that would still be more complicated than a basic controller.

 

It's also possible to do it with an analog circuit that continuously senses the current and modifies the output voltage to effectively cancel out the voltage drop across the motor's internal resistance. The output is smooth DC. That works without caps and it should work with caps too.

 

Cheers,

Andy

 

 

Andy,

There are some interesting ideas there! However, I would have thought the most pressing issue that arises with a keep-alive is when the locomotive stops. If it stops in a state where the track-wheel-pickup interface is dirty and non-conducting then ‘the big-friendly-giant’s hand’ is required to push start the loco. Many many years ago I spoke to the late great Manchester modeller Syd Stubbs, who told me all his DC equipment including his self wound motors was at much higher voltage (I think possibly 24v?) because it had less problems with electrical pickup than the then standard 12v. My thought is that if the loco was stopped at a non conducting state (non conducting at 12v or even the suggested single super-cap 3v supply) then a circuit in the loco might nevertheless be able to detect say a 24v or even 36v pulse without damage, and use this to initiate some remnant capacitor discharge to get the loco to move.

 

I don’t understand what the damage limitation constraints are on super capacitor supply voltages so wouldn’t know how to implement that though!

Cheers Tom

Link to post
Share on other sites

On 28/04/2021 at 03:26, OldTom said:

Andy,

There are some interesting ideas there! However, I would have thought the most pressing issue that arises with a keep-alive is when the locomotive stops. If it stops in a state where the track-wheel-pickup interface is dirty and non-conducting then ‘the big-friendly-giant’s hand’ is required to push start the loco. Many many years ago I spoke to the late great Manchester modeller Syd Stubbs, who told me all his DC equipment including his self wound motors was at much higher voltage (I think possibly 24v?) because it had less problems with electrical pickup than the then standard 12v. My thought is that if the loco was stopped at a non conducting state (non conducting at 12v or even the suggested single super-cap 3v supply) then a circuit in the loco might nevertheless be able to detect say a 24v or even 36v pulse without damage, and use this to initiate some remnant capacitor discharge to get the loco to move.

 

I don’t understand what the damage limitation constraints are on super capacitor supply voltages so wouldn’t know how to implement that though!

Cheers Tom

 

Hi Tom,

 

I've never actually tried it but it should be possible to apply a high voltage pulse to the track to start conduction. It would not do the caps any harm as long as the pulse was of short duration and fed from a high impedance source (through a resistor).

 

Of course the best solution is to stick batteries in everything and either use radio control or send a control signal through the track. The supercap flywheel idea is a bit of a cheap and cheerful compromise :)

 

Cheers,

Andy

Link to post
Share on other sites

11 hours ago, AndyID said:

 

Hi Tom,

 

I've never actually tried it but it should be possible to apply a high voltage pulse to the track to start conduction. It would not do the caps any harm as long as the pulse was of short duration and fed from a high impedance source (through a resistor).

 

Of course the best solution is to stick batteries in everything and either use radio control or send a control signal through the track. The supercap flywheel idea is a bit of a cheap and cheerful compromise :)

 

Cheers,

Andy

Hi Andy,

Thanks for your opinion on the possible use of high voltage spikes. That means more fun experiments are possible!

 

Yes I completely agree that radio control with batteries is ultimately the way to go, not least because you are then in the blissful position of not having to do any track wiring at all! There is at least one good thread on rmweb about this. I myself would eventually (once I have perfected the 48 hour day!!) like to do something with this and Arduino. But as you so rightly say and have yourself very nicely demonstrated, supercaps are a quick compromise which will work more or less straight away with existing legacy models. Plus one of the main things for me is I like the idea of playing with all the different technologies just for the fun of it, and in the end that’s what hobbies are for! 

 

On that theme, I have just been looking at how the old pre-WW2 Trix and Marklin AC motor systems worked. Because they had an AC motor using field coils without permanent magnets, they had to have a solenoid system to reverse the polarity of the field coil. This solenoid was activated by high voltage spike in the Marklin system (and a drop to zero voltage in the Trix system). So higher voltage spikes were also used then.

Regards Tom

Link to post
Share on other sites

  • 1 year later...
On 22/02/2021 at 00:49, AndyID said:

I ran some tests with five of the AVX 1F caps in "normal" series (pos to neg to pos etc.) feeding them from a 12 volt supply. As expected there is some variation in the voltages on the individual caps but they are all well below the rated 3 volts. The highest was less than 2.5 volts.

 

Then I discharged the series combination through a resistor and reversed the connection to the 12 volt supply. Apart from the fact that the voltages on the caps were now negative there was very little difference at all. The discharge rate was the same.

 

These are double-layer supercaps which are nothing like conventional electrolytic capacitors. As far as I can determine the electrodes are symmetrical and the designated polarity is arbitrary. I assume the manufacturer assigns a polarity to differentiate these devices from traditional non-polarized capacitors. I don't think you would want to use them to pass a relatively high frequency current but in our situation we are only reversing the polarity after a considerable time.

 

More testing required in case there is a big "gotcha" hiding in the weeds but so far so good.

Polarity of Supercapacitors 

 

Hi Andy, Apologies for the year long delay (wiring my new layout). I have eventually got round to obtaining some supercapacitors and intend to experiment with your approach. However your circuit drawing  is currently not showing on the website presumably because of the losses on the rmweb. So because of fear exploding polarised caps I wanted to double check with you your conclusion from your experimentation, which I believe was that the supercapacitors could effectively be treated as non-polarised. Could you confirm this?

 

On this issue you might be interested in the reference to this supercap polarity available in the link:

 

https://www.eaton.com/content/dam/eaton/products/electronic-components/resources/technical/eaton-supercapacitor-application-guidelines.pdf

 

Which in its ‘Polarity’ section says:

QUOTE ‘Eaton supercapacitors are designed with symmetrical electrodes, meaning they are similar in composition. When a supercapacitor is first assembled, either electrode can be designated positive or negative. Once the supercapacitor is charged for the first time during the 100% QA testing operation, the electrodes become polarized. Every supercapacitor either has a negative stripe or sign denoting polarity. Although they can be shorted to zero volts, the electrodes maintain a very small amount of charge. Reverse polarity is not recommended, however previously charged supercapacitors have been discharged to -2.5 V with no measurable difference in capacitance or ESR. Note: the longer they are held charged in one direction, the more polarized they become. If reversed charged after prolonged charging in one direction, the life of the supercapacitor may be shortened.’ END OF QUOTE

 

This seems to imply that issues with their supercaps only really occur with negative charging when the supercapacitors are left in a prolonged negative charged state, in which case the consequence seems only to be reducing the component life (it’s not clear whether the performance is also affected by this). Anyhow there are no BANGS!! I am assuming that all supercaps are similar.

 

Cheers Tom

  • Thanks 1
Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
 Share

×
×
  • Create New...