keep alive has no effect

[chaz]

I have just installed a keep alive circuit in an O gauge 0-4-0ST.

 

This has a V3.5 ESU Loksound decoder (the 1A one) driving a Mashima motor on an ABC gearbox.  The loco worked well before I made the change although it was a bit more prone to stall, being an 0-4-0. The added circuit uses a 25V 3300uf capacitor (the biggest value that would fit into the space available) with the usual 100R resistor to limit the charge current and a 1N4001 diode to bypass this for the discharge path. The added circuit was wired carefully to the DC side of the two rectifier diodes, #1 & #3 counting from the nearest edge of the PCB. CV29 is programmed to turn off the DC capability.

 

The circuit has no effect at all. Both the motor and the sound stop as the loco is lifted from the track. Putting a meter across the capacitor shows that it is reaching full charge. Leaving the meter connected after the loco is lifted from the track shows a very slow discharge.

 

What silly mistake have I made?

 

Chaz

[tender]

Either way if you expect the motor to noticeably run after lifting from the track you're going to need a much bigger capacitor.

[chaz]

Is the diode in the right way? Sure you have the correct pads for solder on?

 

Yes, I think so. You know what it's like when you check again and again. But yes to both of those.

[chaz]

Either way if you expect the motor to noticeably run after lifting from the track you're going to need a much bigger capacitor.

 

How much bigger? According to stuff I have read 4700uf should make a model pretty well stall proof, so I would have thought 3300uf would be acceptable. Could you tell me please what size you think I should use?

The fact that no charge is being drawn from the capacitor suggests that either the diode is in the wrong way round or the joints are not conducting but I have checked and neither of these appears to be the case....

[Nigelcliffe]

3000uF would be a few tenths of a second at best. Whether you can see it on lifting a loco is uncertain, but it should improve running. If you expect a second or more of running, you need well over 20,000uF. I have 5000uF in a 7mm 0-4-0, and that is fine. ( space isn't an issue if willing to use expensive capacitors. Cheap electrolytics are not very space efficient. ).

 

Diode should be arranged so that current flows from capacitor positive, via diode, to decoder when there is no power from track.

 

But if your meter is showing the voltage holding up on the capacitor for more than a second, then little, if any, current is being discharged into the decoder. That suggests something more fundamental is wrong.

 

I've never tried the circuit onto a Loksound V3.5, so can't comment on whether there are other settings or considerations. I'd never been able to find enough information to feel confident about connecting into that decoder.

[chaz]

3000uF would be a few tenths of a second at best. Whether you can see it on lifting a loco is uncertain, but it should improve running. If you expect a second or more of running, you need well over 20,000uF. I have 5000uF in a 7mm 0-4-0, and that is fine. ( space isn't an issue if willing to use expensive capacitors. Cheap electrolytics are not very space efficient. ).

 

Diode should be arranged so that current flows from capacitor positive, via diode, to decoder when there is no power from track.

 

But if your meter is showing the voltage holding up on the capacitor for more than a second, then little, if any, current is being discharged into the decoder. That suggests something more fundamental is wrong.

 

I've never tried the circuit onto a Loksound V3.5, so can't comment on whether there are other settings or considerations. I'd never been able to find enough information to feel confident about connecting into that decoder.

 

Thanks for the advice. Having been on line to look at capacitor suppliers I must admit to being rather baffled by the huge range of designs and spec's on offer. If you could suggest a suitable type, value and supplier I would be grateful. Cost is not an issue but size most definitely is.

In the meantime I will have another look at one I have done and see if I can work out if it matches what I should have done.

 

Chaz

[chaz]

Each diode has a ring marking the cathode. Current only flows from anode to cathode. You can easily make a test circuit: disconnect the capacitor (and associated parts) from the decoder. On the + side of the capacitor you connect the diode and resistor as per the way it was in the loco. Now, use a 2nd diode and connect it to the original diode, ring to ring. From that junction connect an LED (with it's own resistor) to the - of the capacitor. Next, connect this - to the power terminal and the other wire from the power pack to the open end of the 2nd diode. The LED should now light up. Disconnect the power, the LED should stay lit for a few seconds (long enough to notice). If that works, remove the 2nd diode and connect the remaining parts to the decoder. As said, 3300uF is not very much (the u is micro: 0.000001) but you can wire multiple capacitors in parallel should you have the space for more, smaller capacitors. The total capacity is then the sum of all capacitors.

 

Disconnecting will be easy as the capacitor etc are connected via a 2 pin plug and socket (polarity marked, of course!) - so setting up these tests should be simple. Unfortunately space is very tight so more capacitors will not fit in. It seems my best option is to find a smaller type of capacitor.

[RedgateModels]

I had no joy with the conventional arrangement of a home grown Stay Alive on a Loksound 3.5. Whilst the capacitor did charge up when power was removed it did not discharge - indicating that the circuit was not taking the stored energy. I gave up at that point not wishing to risk potential damage to the decoder looking for a suitable point to attach the capacitor.

 

 

The above shows where I made the attachment.

[Nigelcliffe]

Thanks for the advice. Having been on line to look at capacitor suppliers I must admit to being rather baffled by the huge range of designs and spec's on offer. If you could suggest a suitable type, value and supplier I would be grateful. Cost is not an issue but size most definitely is.

In the meantime I will have another look at one I have done and see if I can work out if it matches what I should have done.

 

Chaz

 

I have recently written a two-part article which is in Scalefour News.  Part 1 appeared in the most recent issue, part 2 should be somewhere in the new year. That has lots of detail.   There is also a shorter article in the 2mm Magazine on my 0-4-0 diesel shunter.

 

 

Capacitors come in several types, and at different voltages.   Broadly speaking:

 

Electrolytics - cheap, but quite large.  By trading off voltage the energy store per unit volume can be increased, 16v rated will store more energy than 25v rated.  But, you need to know your track voltage to determine whether 16v is safe to use.  Some authors suggest you need a significant safety margin on the voltage rating, others will say that the rating is the rating.  Quite a few commercial suppliers of DCC parts use 16v rated components and say it is for systems up to 16v at the track.

 

Tantalum - fairly expensive, small.  Typically surface mount devices, and usually the largest storage is around 220uF at 16v.   So, for more storage you have to build a bank of them.  Higher voltage rating means connecting in series which drops the total capacitance.

 

Ceramic - fairly expensive, smaller than tantalums.   Again surface mount devices.  The largest I know of is 100uF at 16v.  Again, higher voltage means connecting in series with loss of total capacitance.  Cost can be well over £1 per capacitor, price falls if hundreds are ordered.   I have used those in banks; 900uF in a 2mm scale 0-4-0 (on my blog, link in signature line), and 2200uF in a 4mm scale GER tram loco, fitting easily between the frames, between the two axles of the loco, with the chip above the capacitors. I've been sizing up a GWR 14xx in 2mm scale, and there is room for at least 1600uF, though will probably tackle the "easy" 800uF first, see how it runs, and then decide whether the other 800 are really needed.  

 

Gold/SuperCap - expensive, smallish.  Massive energy capacity, 1F or more is possible.   But, voltage rating is typically 2.7v.  So, need to put several in series to achieve suitable operating voltage (as done by TCS in their KA series keep alive modules), or fit a voltage multiplying circuit (as done by ESU and Lenz on their power modules).    

 

 

Off the shelf, and 16v rated, the TCS KA1 and KA2 pack a massive energy punch for modest price.   If you know the track voltage is safe on 16v, then consider those. Using a TCS KA2 can give a run time of 30 seconds or more than a yard of track is possible in an O gauge loco.  Long run times are not a universally good thing; jam a mechanism and derail leaves the motor buzzing on full power stall for over half a minute, or an out of control loco can wreck scenery, other stock and then drive onto the floor well before the power has died away. If you can't get a DCC-Stop signal to the decoder, it keeps trying to run the loco.  I think a second or two is the sensible maximum.   And, with massive energy stores, I think a discharge resistor across the capacitors is a sensible addition;  that does sap a tiny bit of power, but if you know the discharge is fitted, then the capacitors will be empty after a certain number of minutes (perhaps 30 minutes) and thus no electrical risk after that time if dismantling the loco.  An accidental short over a charged up gold-cap could be fairly explosive and dangerous.

 

But, for experiments, may as well stick with the electrolytics - if you can't get an effect with electrolytics, you won't get anything with the use of more expensive components. 

 

Sources;  Farnell/CPS for off the shelf components, the range is huge, post costs and price for small volumes can make them expensive.  Ebay sellers might have some.    Zimo dealers should have access to some super-caps and tantalums listed in the Zimo price list.    Reading various German forums might lead to suppliers in Germany for the model hobbyist. 

 

 

 

I note from RedgateModels picture that his connections to the LokSound V3.5 decoder are different to those used by Martin in the other thread.  Martin put two wires onto the diode bank, not one.

 

 

 

- Nigel

[chaz]

I had no joy with the conventional arrangement of a home grown Stay Alive on a Loksound 3.5. Whilst the capacitor did charge up when power was removed it did not discharge - indicating that the circuit was not taking the stored energy. I gave up at that point not wishing to risk potential damage to the decoder looking for a suitable point to attach the capacitor.

 

DSC01405.jpg

 

The above shows where I made the attachment.

 

Thanks for that - exactly my experience - however I don't want to give up, if I destroy the chip - c'est la vie - if the keep-alive can't be made to work then I will have to replace it anyway.

 

I wonder if there is someone reading this who has successfully "done" a Loksound V3.5 and can tell me the secret?

[Nigelcliffe]

I suggest searching in German, rather than English.  There is more likelihood of finding someone who has done it.   One term to search for is "Loksound v3.5 Schaltung Elektrolyt Kondensator".

 

That threw up several possible pages, I found the following two had some interesting comments:

http://www.stummiforum.de/viewtopic.php?f=5&t=44318

http://www.s1gf.de/index.php?page=Thread&threadID=10234

 

They indicate that two capacitors may be required, to different places on the decoder.  One to provide track voltage over the diode pins, and another at 5 or 6v over part of the processor.  

 

I make no claim as to whether the information found in the pages is accurate, or relevant, you'll have to judge that for yourself.   If you do make it work, please report back on the solution, I'm sure others will be interested no matter how complicated it gets. 

 

 

- Nigel

[chaz]

I have recently written a two-part article which is in Scalefour News.  Part 1 appeared in the most recent issue, part 2 should be somewhere in the new year. That has lots of detail.   There is also a shorter article in the 2mm Magazine on my 0-4-0 diesel shunter.

 

 

Capacitors come in several types, and at different voltages.   Broadly speaking:

 

Electrolytics - cheap, but quite large.  By trading off voltage the energy store per unit volume can be increased, 16v rated will store more energy than 25v rated.  But, you need to know your track voltage to determine whether 16v is safe to use.  Some authors suggest you need a significant safety margin on the voltage rating, others will say that the rating is the rating.  Quite a few commercial suppliers of DCC parts use 16v rated components and say it is for systems up to 16v at the track.

 

Tantalum - fairly expensive, small.  Typically surface mount devices, and usually the largest storage is around 220uF at 16v.   So, for more storage you have to build a bank of them.  Higher voltage rating means connecting in series which drops the total capacitance.

 

Ceramic - fairly expensive, smaller than tantalums.   Again surface mount devices.  The largest I know of is 100uF at 16v.  Again, higher voltage means connecting in series with loss of total capacitance.  Cost can be well over £1 per capacitor, price falls if hundreds are ordered.   I have used those in banks; 900uF in a 2mm scale 0-4-0 (on my blog, link in signature line), and 2200uF in a 4mm scale GER tram loco, fitting easily between the frames, between the two axles of the loco, with the chip above the capacitors. I've been sizing up a GWR 14xx in 2mm scale, and there is room for at least 1600uF, though will probably tackle the "easy" 800uF first, see how it runs, and then decide whether the other 800 are really needed.  

 

Gold/SuperCap - expensive, smallish.  Massive energy capacity, 1F or more is possible.   But, voltage rating is typically 2.7v.  So, need to put several in series to achieve suitable operating voltage (as done by TCS in their KA series keep alive modules), or fit a voltage multiplying circuit (as done by ESU and Lenz on their power modules).    

 

 

Off the shelf, and 16v rated, the TCS KA1 and KA2 pack a massive energy punch for modest price.   If you know the track voltage is safe on 16v, then consider those. Using a TCS KA2 can give a run time of 30 seconds or more than a yard of track is possible in an O gauge loco.  Long run times are not a universally good thing; jam a mechanism and derail leaves the motor buzzing on full power stall for over half a minute, or an out of control loco can wreck scenery, other stock and then drive onto the floor well before the power has died away. If you can't get a DCC-Stop signal to the decoder, it keeps trying to run the loco.  I think a second or two is the sensible maximum.   And, with massive energy stores, I think a discharge resistor across the capacitors is a sensible addition;  that does sap a tiny bit of power, but if you know the discharge is fitted, then the capacitors will be empty after a certain number of minutes (perhaps 30 minutes) and thus no electrical risk after that time if dismantling the loco.  An accidental short over a charged up gold-cap could be fairly explosive and dangerous.

 

But, for experiments, may as well stick with the electrolytics - if you can't get an effect with electrolytics, you won't get anything with the use of more expensive components. 

 

Sources;  Farnell/CPS for off the shelf components, the range is huge, post costs and price for small volumes can make them expensive.  Ebay sellers might have some.    Zimo dealers should have access to some super-caps and tantalums listed in the Zimo price list.    Reading various German forums might lead to suppliers in Germany for the model hobbyist. 

 

 

 

I note from RedgateModels picture that his connections to the LokSound V3.5 decoder are different to those used by Martin in the other thread.  Martin put two wires onto the diode bank, not one.

 

 

 

- Nigel

 

Thanks for all that useful info' Nigel.

 

"Martin put two wires onto the diode bank, not one."   That's how I connected mine, but I suspect that RedgateModels use of the function "common" may be electrically identical.

 

Chaz

[chaz]

I suggest searching in German, rather than English.  There is more likelihood of finding someone who has done it.   One term to search for is "Loksound v3.5 Schaltung Elektrolyt Kondensator".

 

That threw up several possible pages, I found the following two had some interesting comments:

http://www.stummiforum.de/viewtopic.php?f=5&t=44318

http://www.s1gf.de/index.php?page=Thread&threadID=10234

 

They indicate that two capacitors may be required, to different places on the decoder.  One to provide track voltage over the diode pins, and another at 5 or 6v over part of the processor.  

 

I make no claim as to whether the information found in the pages is accurate, or relevant, you'll have to judge that for yourself.   If you do make it work, please report back on the solution, I'm sure others will be interested no matter how complicated it gets. 

 

 

- Nigel

 

Thanks for the additional info' Nigel. I will follow up the German websites but this will probably have to wait until after Christmas....

 

If I get a successful outcome I will certainly post in this topic with details.

 

Chaz

[tender]

Might be worth taking a look here http://www.rmweb.co.uk/community/index.php?/topic/43324-stay-alive-capacitors/?p=468314

I think your problem may be using the blue common wire as a connection point. If I remember rightly from a circuit I once saw but cannot for the life of find, this is on the voltage regulator side of the decoder so your circuit will only be keeping the decoder circuit alive but not providing anything to the motor power circuit.

You might want to try leaving DC enabled in CV29 as when you loose power you also loose DCC commands so the decoder will be running on DC power. Not sure how long the decoder will function without DCC commands in DCC only mode.

Ray.

[GoingUnderground]

ESU have published a circuit for connecting stay alives in their older decoders, but it was in the LokPilot V3 manual, page 19. ESU clearly state that 2 capacitors are needed, one to buffer the motor output, and the smaller capacitor to keep the decoder alive. The wiring diagram shown is for a 21pin MTC decoder, so the wiring might be the same for a LokSound 21pin MTC V3.5. AFAIK they have not published such a circuit for LokSound V3.5s.

 

Your observed slow discharge may be the clue, you could be keeping the decoder alive, but due to the design of the LokSound there's no current going to the motor as you only have one capacitor.

[chaz]

Might be worth taking a look here http://www.rmweb.co.uk/community/index.php?/topic/43324-stay-alive-capacitors/?p=468314

I think your problem may be using the blue common wire as a connection point. If I remember rightly from a circuit I once saw but cannot for the life of find, this is on the voltage regulator side of the decoder so your circuit will only be keeping the decoder circuit alive but not providing anything to the motor power circuit.

You might want to try leaving DC enabled in CV29 as when you loose power you also loose DCC commands so the decoder will be running on DC power. Not sure how long the decoder will function without DCC commands in DCC only mode.

Ray.

 

Thanks Ray. My connections are both to the DC side of the rectifier diodes - I'm not using the blue common wire. I will try resetting CV29.

[chaz]

ESU have published a circuit for connecting stay alives in their older decoders, but it was in the LokPilot V3 manual, page 19. ESU clearly state that 2 capacitors are needed, one to buffer the motor output, and the smaller capacitor to keep the decoder alive. The wiring diagram shown is for a 21pin MTC decoder, so the wiring might be the same for a LokSound 21pin MTC V3.5. AFAIK they have not published such a circuit for LokSound V3.5s.

 

Your observed slow discharge may be the clue, you could be keeping the decoder alive, but due to the design of the LokSound there's no current going to the motor as you only have one capacitor.

 

Thanks Keith, I will download the V3 manual and investigate.

[chaz]

ESU have published a circuit for connecting stay alives in their older decoders, but it was in the LokPilot V3 manual, page 19. ESU clearly state that 2 capacitors are needed, one to buffer the motor output, and the smaller capacitor to keep the decoder alive. The wiring diagram shown is for a 21pin MTC decoder, so the wiring might be the same for a LokSound 21pin MTC V3.5. AFAIK they have not published such a circuit for LokSound V3.5s.

 

Your observed slow discharge may be the clue, you could be keeping the decoder alive, but due to the design of the LokSound there's no current going to the motor as you only have one capacitor.

 

Sorry Keith. that may be so but there is no reference to keep-alive wiring in the V3 manual that I recently downloaded. This is dated 2005 and only has 8 pages. Maybe ESU have shied away from providing the info' as they don't trust our soldering?

[gavind]

Hi Chaz,

      Here's the link for the lokpilot manual with the wiring diagrams… 

 

http://www.dccconcepts.com/PDF_Downloads/esulokpilotv3.pdf

 

Gavin

[GoingUnderground]

I think you must have downloaded the LokPilot 2 manual. You need the LokPilot 3 manual.

 

The latest edition of the LokPilot V3.0 manual is the 7th edition produced in November 2009 and has 72 pages. For some reason it is not currently on the ESU web site.

 

There is an 8th edition, but that is only available in German, and you can download it from the ESU web site from here

 

The wiring diagram is on page 19. It is identical to the diagram in the 7th edition.

 

I have attached a copy of the 7th edition manual.

 

ESU do trust our soldering as if you buy a V4 series decoder and their stay alive power pack the user has to solder leads from the power pack to the decoder. The power pack is apparently not suitable for the V3/V3.5 series of decoders. If you do attach a stay alive to a V3/V3.5 decoder I believe that you will have to detach it when programming. That limitation doesn't apply to ESU's own power pack when used with the V4 decoders. 

51978_LokPilot_V30_Familie_ESUKG_EN_User_Manual_Edition_7_eBook.pdf

[chaz]

OK. I have rewired using two capacitors and the connections as detailed in the manual linked to in the previous post. I have also tried all combinations of settings in CV29, 50 and 124. A frustrating afternoon with no success at all. Lifting the loco from the track, analagous to it hitting a patch of dirt that breaks contact (yes?) has the motor and the sound cutting instantly with no detectable carry on at all.

 

I think I will have to bin the decoder and fit something better. I like the sound project (an edited download from the ESU website) that is currently in the loco very much so I think I will stick with ESU Loksound - a V4 might well perform as I want.

 

Chaz

[chaz]

Just to finish the story...

 

I replaced the Loksound V3.5 decoder with V4 and added ESU's own very expensive, but small, keep alive. No problem fitting this combination into the model. It ran for the whole weekend at Guildex 2014 in Telford quite faultlessly - transformed from a Prima Donna into a paragon.

 

I will continue experiments with home-brew keep alives - but if you need a very small solution the ESU unit might well answer.

 

Snap shows the decoder and KA installed in the frames of the 7mm 0-4-0ST.

 

 

Chaz

[Cantabrigian]

Wondering if anyone can help me. I've been struggling for months to get any effect with a stay alive on a Loksound V4 installed in a Hornby Class 8 with DCC sound (Hornby insulfrogs are brutal for this 0-6-0 despite careful leveling).

I started with the 2200uF shown in the V4 documentation and then went to 5 2200uF in parallel. Loss of sound and power still seems to be instantaneous.

Based on the back of an envelope calculation, I realized the this set up would still deliver much less than a second of stay alive, so I convinced myself that this was the issue.  I then purchased a KA1

The KA1 has 10X the capacity of the capacitors I was using, so I expected to see 2-5 seconds of stay alive.

But although I can see the KA1 charging to 13V, and I can see it slowly discharging on power down, I'm still seeing instantaneous loss of sound and motion as soon as the loco loses contact with the track.

 

I've continued my web search for any quirks of the Loksound V4, but to no avail.  There are numerous references to CV settings, but I have now tried every combination.

 

Both ESU and TCS have not provided much help

Somenone posted that the Loksound V4 needs the 3-wire ESU powerpack and does not work with 2 wires, but the ESU manual clearly shows a 2 wire solution.

 

ESU commented the my decoder looks like an early V4.0, and that the images of my wiring look good. So they could not explain it.

 

I’ve also attached a few images of the wiring, and my original attempt using electrolytics, which I have now replaced with the TCS KA1 (sorry no picture of that)

 

I set CV113 to 255 to try to max out the stay alive time, but to no avail.

Any ideas anyone?

 

Many thanks

[Nigelcliffe]

Whilst not mentioned in the manuals, have you disabled DC running in CV29 ?  There is also another DC running settings in CV50 which I'd be tempted to disable.    I would also check if brake-on-DC is having any impact (CV27, bit 3 and 4).

 

When using the TCS KA1/KA2, a in-rush resistor/diode is not strictly needed as there is one within the KA packaging.  The KA1/KA2 is only good to about 16v track voltage.  Your measured 13v sounds fine on this score, and as you indicate you should see somewhere in the region of 5 seconds of running with a KA1. 

 

 

- Nigel

[chaz]

Are you allowing enough time for the stay-alive to charge up? When I installed an ESU V4 with their power pack in an 0-4-0ST the combination was initially disappointing with no discernable effect. I was testing it on a friend's layout which may have been rather dusty. After a lot of swearing and threats ("You'll get stamped on, you *$%^@&!") and some head scratching and standing around feeling cheated the loco' was transformed.  Don't forget if the loco' has dirty wheels and the track is dirty the stay-alive may not be able to charge. However once it has charged it will cope with a surprising amount of dirt.

It's not unusual for my 0-4-0ST to work throughout a two-day show with no stalls and no need to have its wheels cleaned.

 

Chaz