Jump to content


Google Ads are only seen by non-members of RMweb - Create an RMweb account and you'll only receive modelling ads.

Photo
- - - - -

Stay alive capacitors

Im a little confused...




  • Please log in to reply
44 replies to this topic

#1 nimbamoss

nimbamoss

    Member


  • Members
  • PipPip
  • 236 posts
  • LocationCleethorpes

Posted 20 August 2011 - 12:22

Hi,

Can anyone clarify please.

The use of capacitors to stop blips in power supplies to DCC locomtives etc.
Is it only for the newest generation of chips or can they be used on any dcc chips including sound?


Ian

Google Ads are only seen by non-members of RMweb - Create an RMweb account and you'll only receive modelling ads.

#2 noiseboy72

noiseboy72

    Member


  • Members
  • PipPip
  • 430 posts

Posted 20 August 2011 - 13:02

With a bit of work, they can be used on any chip. The newer 21 pin ones are easy, as the connection points (+ & - of the bridge rectifier) are available on the connector. On the 8 pin connectors, you can use the function common line as the + and wire directly onto the component.

Lenz use a small circuit as a dc- dc convertor (essentially a diode based voltage doubler) so that they can use a lower voltage, higher capacity capacitor.

I have a Bachmann pannier tank that would benefit from one of these. It has an inate ability to find every bit of dirt on the layout!

#3 Marty Mc

Marty Mc

    Member


  • Members
  • PipPip
  • 758 posts

Posted 20 August 2011 - 13:14

Hi

Some of the newer decoders like the lok pilot 4.0, Zimo, Loksound 4.0, Lenz, DCC Concepts decoders etc now have solder tabs to add stay alive. But that doesnt mean older decoders cannot be done. She photo this is a loksound 3.5 which I modified to take stay alive. Ignore the 2 black wires on the bottom right that's for a hall sensor nothing to do with stay alive.



Posted Image

Now their is a problem to this you void your warranty the minute you remove the heat shrink and solder on the electronic components and it is not the easiest of jobs as you will be soldering in a very tight space.

It can be done as this is now considered an older decoder. Far easier on 21/22 pin decoders. I would suggest using DCC Concpets Stay alive as it is pre wired with very fine grade of wire plus it has the sub circuit needed to make stay alive work.


Best of luck.

Martin

#4 Oldddudders

Oldddudders

    Member


  • Members
  • PipPip
  • 11,405 posts
  • LocationLe Mans, France

Posted 20 August 2011 - 14:31

This chap has been demonstrating it can be done for years http://www.members.o...north/alive.htm

#5 RedgateModels

RedgateModels

    Member


  • Members
  • PipPip
  • 4,041 posts
  • Location16C - Mansfield

Posted 20 August 2011 - 21:07

oh yes, done it for the Lenz Standard, if you can fit a 4700uF cap in it makes a loco pretty much invincible to dead frogs etc :)

#6 Suzie

Suzie

    Member


  • Members
  • PipPip
  • 2,143 posts

Posted 21 August 2011 - 17:00

I have found that a 35V 4700uF between the blue and black/white wires will make even the worse 4-wheel loco impossible to stall even with dirty wheels and track and at speed step 1, but you need to have room for it.

#7 noiseboy72

noiseboy72

    Member


  • Members
  • PipPip
  • 430 posts

Posted 18 September 2011 - 13:09

A few observations following fitting my Bachmann pannier tank:
This is a newer chassis with the DCC socket. Originally I had just plugged a Lenz silver into the socket and that was it! Even the deocder number was still set to 3.
This loco has never run that well, stalling at the merest hint of dirty track and not liking slow speeds, growling and stuttering along. I decided to fit a stay alive capacitor and also tidy up the installation as well.

I removed the DCC socket and hard wired the chip, thus removing the circuit board with its various chokes etc. This allows the body to sit much better, so great news for a start. Removing the chokes also decreased the starting speed of the loco and smoothing the response from the motor with much less growl. Result!

Fitting the capacitor was fairly easy. I used the blue function wire and also soldered onto the negative side of the bridge rectifier. The wires were run through into the cab, where a 16V 1000uF cap would lay across the floor, totally invisible. I considered the charging of the capacitor, usually controlled by fitting a 100 Ohm resistor and diode. The reason usually given for this is that the in-rush could kill the diodes on the the decoder. However, bearing in mind the diodes fitted are rated at 4A, I did not bother, the capacitor is simply wired across the bridge rectifier as a conventional PSU smoothing cap would be.

I had hoped the 16V capcitor would be OK. My multimeter indicated a voltage of 16.5-17V across the cap, but it still got very hot and the can started to expand. I therefore substituted it for a 470uF 25V and this works fine.

Performance wise, the difference is huge. The stutter is gone and the loco will traverse an insulfrog double slip at speed step 1. I hear what is said about good trackwork, clean pick ups and live frogs, but to me this does not feel like a bodge to overcome poor track, but a genuine improvement to running.

Another thing I found on another site was turning off the analogue control option. The decoder is no longer trying to decide if the incoming voltageis pure DC or a DCC signal, so behaves differently on loss of signal. The Lenz deocder would stop and restart the speed curve on loss of signal, but now carries on where if left off.

Please also note that anumber of decoders with stay alive do not function correctly, if at all if you leave analogue control switched on.

Edited by noiseboy72, 18 September 2011 - 13:19 .

  • Informative/Useful x 3

#8 brossard

brossard

    Member


  • Members
  • PipPip
  • 2,535 posts
  • LocationA Bloke in Quebec

Posted 18 September 2011 - 13:49

It looks like there is some very precise soldering required. Not sure if my equipment/skills are up to that. (I had an NCE chip wire termination fail and asked a guy at work who does fine instrumentation to have a go. The repair didn't hold). I am interested in retrofitting some of my locos (Loksound V3.5) with the device. I emailed Richard about this but he didn't respond (I understand he's been travelling). For my non-sound equipped locos I wouldn't have a problem with switching to the DCC Concepts Decoder/Stay Alive.

John

#9 noiseboy72

noiseboy72

    Member


  • Members
  • PipPip
  • 430 posts

Posted 18 September 2011 - 15:15

If I can solder it, anyone can!
In all seriousness, its no worse than 0.1" Veroboard. The trick is as with all soldering to prepare the joint first. Wet the tip of the iron, the wire and pad with a little solder, trim the wire to the correct length and then with the decoder securely mounted, apply the wire and then the iron. This way, you are not attempting to feed in solder or having to hold the iron on the joint long enough for the track to lift.

Iron wise, I use a 45W temp controlled solder station purchased for £45. It has a standard tip and is set to the standard setting for lead solder.

...And avoid lead free solder unless you are re-working newer commercial equipment. You can still freely buy 60/40 solder and it is much easier to work with due to its lower melt point.

#10 Nigelcliffe

Nigelcliffe

    Member


  • Members
  • PipPip
  • 3,138 posts

Posted 18 September 2011 - 17:04

Minor addition to NoiseBoy's comments on soldering. As said, it's all in careful preparation - fix the item down with something (tape, pegs, cocktail sticks) and work out how you will get wire and iron in cleanly and quickly, then do it.

I agree that lead-based solder is easier to work with, but as you say, for any reasonably recent electronics (including DCC decoders/chips) you have to use lead-free. That means a different tip for the soldering iron as you mustn't mix lead-based and lead-free solders if you expect a decent joint to be made.

#11 Edwin_m

Edwin_m

    Member


  • Members
  • PipPip
  • 6,404 posts

Posted 18 September 2011 - 17:46

As well as protecting the diodes the other reason to limit the inrush current is to reduce the risk that the command station cuts out when it is first powered up. This won't be a problem if there is only one stay-alive on the layout but if there are several then the current starts to add up.

#12 brossard

brossard

    Member


  • Members
  • PipPip
  • 2,535 posts
  • LocationA Bloke in Quebec

Posted 18 September 2011 - 18:06

Thanks for the encouraging comments guys. I'm no newb at soldering just not experienced with fine joints and wouldn't want to damage a 100 pound decoder. I tried lead free solder - what a disaster! Again, I tapped into my work network and got an electronics guy to give me some nice solder along with non corrosive flux which works well.

John

#13 Marty Mc

Marty Mc

    Member


  • Members
  • PipPip
  • 758 posts

Posted 19 September 2011 - 04:26

Hi
We all have different experiences and this is what works for me. I stay away from Lead free and use a 2% silver solder see the review here.

http://www.modelrail...Concepts-Solder

I have never had any issues with it but what goes with that is a good Iron I have a 75W temp controlled unit.

M

#14 gavind

gavind

    Registered Member


  • Members
  • Pip
  • 14 posts

Posted 23 December 2013 - 00:55

Hi

Some of the newer decoders like the lok pilot 4.0, Zimo, Loksound 4.0, Lenz, DCC Concepts decoders etc now have solder tabs to add stay alive. But that doesnt mean older decoders cannot be done. She photo this is a loksound 3.5 which I modified to take stay alive. Ignore the 2 black wires on the bottom right that's for a hall sensor nothing to do with stay alive.



002-2.jpg

Now their is a problem to this you void your warranty the minute you remove the heat shrink and solder on the electronic components and it is not the easiest of jobs as you will be soldering in a very tight space.

It can be done as this is now considered an older decoder. Far easier on 21/22 pin decoders. I would suggest using DCC Concpets Stay alive as it is pre wired with very fine grade of wire plus it has the sub circuit needed to make stay alive work.


Best of luck.

Martin

 

Hi,

      Apologies for opening an old thread… I'd worked out the basics of keep alive for myself some while back, and found it to be extremely successful, but I came across this post as I've acquired a loksound 3.5 decoder secondhand to begin playing with DCC sound, and was struggling to get any response out of the cap under dcc (It's charging up OK). I've tried wiring up as shown in your earlier post, and with the cap between the ground pin and the standard function V+ blue wire, but in either case when I cut the power everything just snaps off (lights, sounds and motion), as if the decoder is interpreting the DC. -I've tried changing the settings in CV29 and CV50 (analogue mode setting for ac/dc/off), as well as the brake mode settings in CV51, but none of these seem to make any difference...

 

Any suggestions very welcome...

 

Many thanks,

Gavin



#15 chaz

chaz

    Member


  • Members
  • PipPip
  • 5,267 posts
  • LocationEastleigh Hampshire

Posted 26 December 2013 - 16:18

Gavin, hello.

 

Coincidentally I am doing exactly the same - in my case a Loksound V3.5 with a 25V 3300uf, a 100R current limiting and a bypass diode for the discharge path. A meter shows me that the capacitor is charging as it should but the arrangement has no effect cutting out instantly when the track to wheel contact is interrupted. The capacitor discharges very slowly, taken ten seconds or so to empty (proving that there is a current path).

 

I spent a frustrating evening trying to get the mod' to work, checking several times that my connections were correct and the soldering sound. I also changed the settings in CV29 & 50 but these had no effect.

 

I know that's not much help but at least you know that you are not the only one with this problem.

 

Nigel Cliffe suggested searching German sites in my thread on this subject....

http://www.rmweb.co....-has-no-effect/

but the language problem is making things difficult.

 

I should have said that the ESU forum has not provided any useful info' yet either.

 EDIT

...have a look at posting 11 on this (German) site...

http://www.s1gf.de/i...&threadID=10234

there is a diagram and a photo that might offer a solution - haven't tried it yet but I will.

 

Chaz


Edited by chaz, 26 December 2013 - 16:48 .


#16 gavind

gavind

    Registered Member


  • Members
  • Pip
  • 14 posts

Posted 27 December 2013 - 13:10

Thanks Chaz,
Yes, it's good to know that I'm not just going insane…!

OK, that's effectively doing the same thing. -According to the pinout specifications for the 21 pin decoders, (see http://www.dccconcep...nctionsdesc.htm) those are the ground, decoder V+ and pin 12 is Vcc for the processor sub-circuit. So Martin's original picture should be completely equivalent to the pins used for the 4700uF cap in the thread you refer to. I've tried adding a smaller cap to keep the decoder powered separately, but it makes no difference (I would expect this anyway, as everything should come directly from the main rectifier bridge.
I did see something on a different thread somewhere (but I can't find it again) about the loksound 3.5 not supporting CV11 (the timeout to stop activity after loss of the DCC signal), but it would make sense for no timeout setting to correspond to an infinite timeout, rather than 0s timeout… -Maybe someone else knows more about that.

Gavin

Edited by gavind, 27 December 2013 - 13:21 .


#17 chaz

chaz

    Member


  • Members
  • PipPip
  • 5,267 posts
  • LocationEastleigh Hampshire

Posted 27 December 2013 - 20:41

Thanks Gavin.

 

As I understand it you are saying that a second smaller capacitor is not going to make any difference? Saves me another frustrating blind alley.

 

Now I know that people have made the keep-alive circuit with V3.5 Loksound work - so the trick must be known. Could I ask if you are reading this and have done so could you please post details of what needs doing?



#18 gavind

gavind

    Registered Member


  • Members
  • Pip
  • 14 posts

Posted 27 December 2013 - 21:24

OK, I did some more searching, and I picked up your other post at http://www.rmweb.co....-has-no-effect/

which sent me looking for the lokpilot v3.0 manual. I found this on the dccconcepts pages, and it clearly has the same information: use two capacitors. So I went back and tried again. I've also now set CV124 to 0 and made sure I have analogue mode turned off (again). 

 

I now find this does seem to make a difference to the running performance, in so far as I've got it to run a very slow route over some of my favourite stalling points with the lights on, but no sound, and it seems better.

 

I'm using a 24V/1000uF cap on the main rectifier and a 10V/220uF cap on the Vcc line.

 

What's really odd, though, is that the standard test of just pressing the E-stop button with the lights on still snaps them straight off, so I can't be 100% sure it's really working as it should. I guess at this point the best test would be to up them to larger caps and see if I can see a more definite effect. -I don't really have a feel for how much charge storage is required for the sound system to keep running. 

 

Cheers,

Gavin



#19 Nigelcliffe

Nigelcliffe

    Member


  • Members
  • PipPip
  • 3,138 posts

Posted 27 December 2013 - 21:55

With any decent system, E-stop is a specific DCC command to the track, and any decent decoder will obey and stop instantly.

Your observations suggest the circuit is working. 5000uF or more should show a more marked effect if you pull the power lead to the track.

Nigel

#20 gavind

gavind

    Registered Member


  • Members
  • Pip
  • 14 posts

Posted 28 December 2013 - 14:16

Thanks Nigel,

             That's a good point - I'd always interpreted the red button to be a simple interruption of power. I checked it running into a purposefully dead section of track and it appears to do the right thing...

All my other decoders are from a different manufacturer, and they tend to all slowly fade away when I hit the E-stop. I'm on 00 scale, so 5000uF is larger than I can fit in --I realise you're 2mm scale, but I'm not in that league…yet! I've upped it to 2200 for the main drive, which is about as far as I can go.

 

Gavin


Edited by gavind, 28 December 2013 - 14:35 .


#21 Nigelcliffe

Nigelcliffe

    Member


  • Members
  • PipPip
  • 3,138 posts

Posted 28 December 2013 - 18:43

Gavin,

as indicated elsewhere in one of the threads on capacitors, there are different capacitor technologies.  With Electrolytics, they are cheap but take up a lot of space.  If you want higher capacity in the same space, then you have to change capacitor technology, and accept that they cost a lot more money.

 

2200uF should help quite a bit, it should certainly kill any crackle/pop noises in the sound and lamp flickering.    Its about the most I was squeezing into 4mm scale locos a few years ago, before changing to other capacitor technologies. 

 

Another way to investigate whether its worth changing capacitor technology is to rig up some capacitors outside the loco for a test (make sure your insulation is decent, shorts can be dangerous to electronics).  Then you can see cheaply how much more capacitance is needed for the performance you would like, and from that decide which, if any, of the alternative technologies are worth pursuing. 

 

- Nigel



#22 gavind

gavind

    Registered Member


  • Members
  • Pip
  • 14 posts

Posted 28 December 2013 - 19:34

Indeed… you've reminded me that I was going to take another look, and I now see that I can get 22mF /5.5V gold caps for 55p each from RS, so 4 of those in series would make a nice 5500uF in more-or-less the same volume as I'm currently using.

 

Thanks,

Gavin



#23 New Haven Neil

New Haven Neil

    Member


  • Members
  • PipPip
  • 2,076 posts
  • LocationIsle of Man

Posted 28 December 2013 - 21:59

Don't you parallel capacitors to increase, er, capacity?  Putting them in series makes for a reciprocal value IIRC????



#24 Nigelcliffe

Nigelcliffe

    Member


  • Members
  • PipPip
  • 3,138 posts

Posted 28 December 2013 - 22:32

Need series to raise the working voltage from 5.5v to either 16.5 or 22 volts.
  • Informative/Useful x 1

#25 RedgateModels

RedgateModels

    Member


  • Members
  • PipPip
  • 4,041 posts
  • Location16C - Mansfield

Posted 29 December 2013 - 13:07

Sorry if I'm missing something, but can someone confirm where on a Loksound 3.5 the connections need to be made?

 

cheers

 

Ian









Google Ads are only seen by non-members of RMweb - Create an RMweb account and you'll only receive modelling ads.