Using a DCC controller to power a DC layout

[21C123]

Hi all

 

I was wondering if it is possible / advisable to try using a DCC controller with a chip fitted to its outlet to power a DC layout.

I usually model in OO but have a small O gauge layout that I am converting to DCC, using a radio NCE controller. I would like to be able to use the radio controller to also run one of the cabs on the DC layout and thought if I wired a chips track input to the DCC bus and it’s output to the DC cab input the controller would “see” the layout as a loco.

Would that work or place any strain on the controller. .?

 

Any thoughts appreciated.

 

Roger

[Nigelcliffe]

Will work, with various limitations, up to the point where you have a short circuit on the DC layout then stuff might go "pfff". 

 

The limitations - you won't have a small feedback loop from DCC decoder to motor, instead its a variable motor (different locos) at different distances and wiring quality over the DC layout.  So the BEMF features in the decoder may be unpredictable, depends on decoder and luck.  You can turn those off, or turn them down down with CV changes.

 

A short circuits on the DC layout is the big problem - if the decoder has adequate output protection then it might survive, you're doing things to a decoder its not supposed to support, but some decoders do have short circuit protection on the output stage (Zimo, ESU, Lenz and perhaps others, read the manuals carefully).  If it doesn't (and a lot of US brand decoders have no output protection), it goes "pfff" immediately on the short circuit and the decoder is dead.   

If you can limit the current on the DC output of the decoder to a fair bit lower than the decoder's rated current limit, then things might be a bit safer on short circuits.  Can either limit things on the input side (use a smaller low powered DCC system so the current cannot exceed the decoder output, use a low-current DCC short circuit breaker device,  or use a car light bulb in series with one DCC wire to limit the max current at the decoder), or can limit on the output (car light bulb in series).  As to which car light bulb, depends on current limit required, a side-light (5W) is around 0.5Amp, an indicator (22W) is around 1.8Amp.  All numbers with a large margin of error on them, so don't expect protection at any exact number. 

 

But its all well outside how things were intended to work, so if stuff goes "bang" then you're somewhat on your own.   There is negligible risk to the DCC command station, the risk is destroying the decoder. 

[John K]

An ingenious plan if you like the smell of burning electronics.

I agree with Nigel - it is theoretically feasible, but I suggest you will be safer to spend a tenner on an H&M Duette from eBay to run your DC stuff.

John K

[RedgateModels]

I tried it years ago with a Marklin Z Gauge loco and track using a Hornby decoder, it worked for a bit until the decoder let go ......

[Marcyg]

I used to be part of a club that used Hornby zero one in that way. The chips powered the track. It actually worked very well and the layout was a rather large 28ft by 12ft. I think a few chips were lost a long the way, but not many, and was certainly reliable. How modern chips would cope with something like that I have no idea

[Oldddudders]

An ingenious plan if you like the smell of burning electronics.

I agree with Nigel - it is theoretically feasible, but I suggest you will be safer to spend a tenner on an H&M Duette from eBay to run your DC stuff.

John K

I am less confident that a member in NewZealand will get a Duette for a tenner.

[Mike Buckner]

Another potential failure mode is if (probably accidentally)  there are 2 or more locos on the track section, this may exceed the maximum current draw of the decoder.

[kernowtim]

Ztc 511 controller lets you switch between dcc and dc operation, also has a zero1 mode. Does have various quirks and access to only 8 decoder functions..

[GoingUnderground]

Just fit decoders in your locos. It's a lot easier in the long run.

[Phil S]

Busch make a dedicated controller for exactly this purpose for their H0f mine/field railway system which uses analogue 3V locos.... it uses 4 functions to enable any combination of 4 track sections - the dcc speed control then providing the +3V - 0 - -3V variable drive. It allows us to include (2×4) sections of controlled H0f on our totally-dcc controlled skandi layout appearing as (2) locos numbers (9998 and 9999 = easily accessed on a Multimaus by number, cursor or name.

On the computer control side RR+Co has a mode where the block sections are each an individual dcc decoder output.... perhaps looking there may offer further guidance on decoders which have suitable output protection against a derailed train as mentioned earlier.

A crude option may be to include a series high wattage resistor which would limit the current drawn from the decoder both normally and during a track short. Eg drop 50% in the resistor of >10 ohms ...but include a warning LED so that a short is not left unnoticed.

 

In a similar way, we also control 2 Magnorail bicycle systems as 'locos' by driving their motor (below the baseboard)

...but in this case there is no risk of a short as the decoder is hard wired to the drive motor ( locos 9996 and 9997 ... with 2 more systems planned ).

Edited August 14, 2018 by Phil S

[21C123]

Thanks for the suggestions, I might give it a go with an old decoder I have.

The OO layout is 8 by 2 m plus 15 m of branch line, with 30 or so locos on at any one time I’m not rushing to chip the lot at once, I thought this may be an easy way to get wirelesss control.

The duette will have to continue in use for its 4th decade....

 

Roger

[Phil S]

There is always a risk if both analogue and digital systems are combined in close proximity:

Avoid using the Duette (or similar generation controllers) with DCC fitted locos - or you may get strange results (in the short term)

1/ The PEAK voltage out of the 12V output can be well above the decoder rated  maximum. (there may be no 'long term': after the smoke)

2/ The unsmoothed 'dc' from the controller is anything but smooth - it is 100Hz 0-peak-0 cycles ofrectified sinewave

 If nothing worse, this cases the decoder to reset 100x a second - resulting in it getting nowhere fast.

 

ALTHOUGH this period of controller may have '12Vdc' (or sometimes 14 or 16Vdc) quoted on a nice metal plate; they have unregulated and unsmoothed outputs which can peak at 28V or so 100x a second .... they are anything but '12Vdc'  ... a 9V batery is better for testing locos - fitted or otherwise - [provided dc running is enabled]

 

An additional piece of info about the Busch H0f:  they also provide protective cutouts which might be PTC resistors or self-resetting fuses (I can't remember which - they are not to hand)   I can imagine people using '12V bulbs' in series as an indicative protective resistor - but would advocate a more specific protection and indication with an LED and resistance/Zenor.

 

Also note that if/when the s/w such as [RR+Co] enables adjacent sections to allow the loco to progress around the layout, I assume (not having tried it myself) that the system checks that adjacent sections are given matching polarity - or are otherwise OFF.  ... so ensure wire phasing when installing as carefully as if normal dc sections -   If only isolated areas are planned, this is not a problem - for example I plan to use it for semi-automated shuttle sections - using the standard diodes at each end .... the train will run when the whole section is enabled and stop at the end until the controller is reversed.   .... Many such little sections can be run this way from the 1 dcc controller (8 in the case of our H0f sections)  The Busch Module uses relays to enable/isolate each of its 4 output sections ( 0-4 on at any 1 time )

Edited August 16, 2018 by Phil S

[Theakerr]

Roger, I have done a basic proof of concept set up and it worked very well.  I have not taken it any further because I have had no need to.  Essentially I took a DCC power unit (Digitracks  because I had access to a loaner) and used the DCC control unit connected by wire directly to the decoder about 12" away.  Output went from the decoder to my track which is about a 100'  broken into 5 blocks.  I run 18 trains all on their own controller in an automated system using, for the most part, the old KPC control units.  These are superb units but are not available and am continuously looking for a suitable replacement system Small, fits into my panel, switchable feedback).   DCC was not a viable alternate because I run three time periods resulting in having to put decoders in at least 30/33 locos, some of which will be a bit of a challenge.  With the system I was considering I could simply slot one decoder in as required.  I used an HO decoder for my proof and this had a power rating of 1.5 amps which I figured would be more than enough for all but a few of my locos that use some rather old open frame motors.  If I did go this route I would use some of the 0 gauge decoders that have up to 5 amp rating.  My decoder did not have back emf so I cannot comment but a friend who 'knows these things' believes it will work although some experimentation due to the length of line feedback might be required with the CV(?).  His position was that the length of wire and track feedback since they are essentially a constant should be capable of being 'calibrated'  I do not understand all the comments about short circuits because as I understand it the problems associated with DCC shorts, and I know they are there, are all associated with the 16VAC on the track and the signal going down the track to the decoder.  With the system I used, and I think you are proposing, the 16VAC is well isolated from the control unit.  If the DC side were to be a problem it would be very easy to protect using two diodes as we used to do with the Portescap motors.

Jim

[Phil S]

"I do not understand all the comments about short circuits because as I understand it the problems associated with DCC shorts, and I know they are there, are all associated with the 16VAC on the track and the signal going down the track to the decoder.  With the system I used, and I think you are proposing, the 16VAC is well isolated from the control unit.  If the DC side were to be a problem it would be very easy to protect using two diodes as we used to do with the Portescap motors."

 

Phil: The problem of a short circuit occuring with this method

[ where the dcc bus supplies a loco decoder which is then fed to the analogue track section(s) [ 1 per decoder usually - perhaps with extra selection by functions as with the Busch module ] ... the analogue track effectively being the orange/grey wiring to the loco motor ]  of  powering/control is very real:

NO ADDITIONAL 'Analogue' controllers are required - the whole point being to use the same dcc controller as everything else. 

 

The exposed TRACK -  which might have anything fall  on it, or the loco derail causing a short across the rails - this now being the output stage of the beleaguered decoder .... and if the analogue track wiring is good or short (not long), then it will have 'zero' resistance and overload the loco decoder output immediately.   [There may be a remoter risk of the 'analogue track' shorting to the dcc power bus/track eleswhere - which will also destroy many decoders]

 

NOT ALL DECODERS are designed to protect themselves against a 'motor short' - and these will then blow their power stage (look for a pinhole or worse in one of the black components - do not send for a warranty repair 8-)    However Lenz and some other decoders SHOULD protect themselves and offer some form of diagnostic response.

 

Therefore some form of current limiting in the 'motor' (analogue track) path is desirable - such as with a high power rating resistor  (it should be capable of taking the full decoder motor current at full voltage  eg 12V, 1 Amp  P=IV  =12Watts  but in normal use drop maybe 1/4- 1/2 the nominal motor voltage (3/4 for Busch).   If half, then it needs the same resistance as the motor (10-20 ohms?)   A combination of (linear) resistance and non-linear diode-based voltage drop may be best, and more power efficient, especially if the choice of motor is wide ranging.  eg 5V zeners (1 for each direction) or stacked diodes, and a smaller resistance in series.

 

There is an option to drop some of the 'excess volts' at the input side of the decoder by using inverse-parallel diode stacks - which lower the overall decoder voltage from the  dcc-track voltage, - this would avoid inhibiting any back-emf measurements whilst permanently dropping the voltage range for the motor.    With hard-wiring, there should be no risk of the input voltage falling to the NMRAminimum of 7V (at which all decoders should work)

Using CV5 is available to limit the maximum speed (effective motor voltage) is always at risk of being reset at the decoder ... but is more tweakable operationally 8-)

Some additional protection on the motor side would still  be beneficial.

[Theakerr]

Phil, I think I understand where you are coming from.  Going back to my original thought using the Portescap system where two diodes in opposite direction were connected to protect the coreless motor, I had always understood they were there to protect the coreless motor from incoming surges.  My logic was to use them so that they prevented outgoing surges.  So is my understanding in-correct and are they there for another reason.  Not being sarcastic or clever, I am interested in learning more.  

On an associated subject re control, on the advise of a gaugemaster technical Guy, I have just replaced the 14VAC input with 12VAC and what a difference that has made to controllability.  Almost as good as the KPC on the test Bachmann K3 with a lot less voltage fluctuation.

Edited August 17, 2018 by Theakerr

[peach james]

If I was going to do this with DCC decoders, I would go for something like 4 amp decoders, and the trick of a car headlamp as the short protection. At 4a, 12 v, you need to drop below 3 ohms before decoder damage should happen. The 21w headlight gives an illuminated resistance of about 2x that, you may find that 2amp decoders will survive being treated like that, but it would be close. There isn't much decoders available between the 1.3(2) amp decoders and the 4 amp ones, I don't think. There is a large price difference because you move from chip to discrete components for the H bridges.

Again, it may work for you, but I think that you will fairly quickly find that full on dcc is easier. There is a break point, and I think it comes when a layout is designed to have more than two locos running at once, where dcc becomes easier for loco control. For switch machine operation, I think that the only time that dcc makes a lot of,sense is once you want to automate panels in any style. Diode matrix or otherwise, at that point, CDC becomes easier than wiring. Before that, probe and stud is vastly simpler if using solenoid, switches if using servo or torti.

[Phil S]

Phil, I think I understand where you are coming from.  Going back to my original thought using the Portescap system where two diodes in opposite direction were connected to protect the coreless motor, I had always understood they were there to protect the coreless motor from incoming surges.  My logic was to use them so that they prevented outgoing surges.  So is my understanding in-correct and are they there for another reason.  Not being sarcastic or clever, I am interested in learning more.  

On an associated subject re control, on the advise of a gaugemaster technical Guy, I have just replaced the 14VAC input with 12VAC and what a difference that has made to controllability.  Almost as good as the KPC on the test Bachmann K3 with a lot less voltage fluctuation.

The potential/possibility for damage is not really to the motor - because the short circuit is removing any possiblility of voltage across the motor ! **  I do not know of the background to the diodes for the Portescap motors - but the basis of the coreless motors is they have very low inertia - not having a heavy core - and hence want a smooth (or high frequency = effectively smooth drive voltage) to avoid constant accelerating and decellerating forces on them (cogging style) from pulsed controllers of the past. **Although shorting the terminals of a coreless motor WILL stop it instantly - an example of rheostatic braking !!

 

In referring to voltages; to avoid confusion, I add 'dcc' to dcc track or bus voltages, avoiding possible confusion with dc or ac (sinusoidal) supply voltages not carrying data.  I assume the recommendation was therefore to reduce the Trackdcc voltage from 14Vdcc to 12Vdcc ( is this is displayed value or meter measurement - as the latter may be inaccurate due to meters designed for sinusoidal measurement - this is good for a dedicated output used for this application only - but may have adverse effects if reducing the entire layout's track voltage (lower voltages mean more risk of voltage brownout, and higher currents for the same speed). I use 16Vdcc - regulated with my Roco controllers by using a SMPS of 18Vdc in to the Amplifier - this gives about 12V nominal max voltage to a motor after the bridge rectifiers and drive stages.   For narrow gauge (and N?) locos, some users may use a lower voltage (Roco don't)

 

If using the dc-sections-from decoders,as I am - to avoid the need for any '2nd' controller (ie non dcc handsets) - where a lower voltage is desirable JUST for those analogue parts - possibly different for each - then the local inverse-parallel-pair diodes method produces known and consistant voltage drops (about 1.2V per pair) - regardless of the current taken by the motors  ...Obviously, the Busch Module does a more specific version down to its +/- 3V variable outputs.  CV5, as I mentioned can also be used - in LGB we use it to limit locos on our shunting puzzle to 1/4 - 1/2 the maximum possible speed. We also run the shunting puzzles with the Roco controllers, and so the track volts are only 16Vdcc and not the higher 22Vdcc of normal LGB digital garden railways - the same reasoning as Gaugemaster's suggestion to limit the speed range for maximum control.

 

I see that Bulbs  - a popular USA method ?? of current limiting have been suggested - but, in the same way that a PSX or similar intelligent circuit breaker is better - the bulbs have a slow response, and may not protect enough - hence the 'need' for 4A decoders ??

I am assuming that the analogue sections might, as in my cases, be for small narrow gauge locos or efficient oo/Ho which are too small to convert - taking upto 1/2 Amp ....so there is no point in spending money on 4A decoders to control them !!

{One has to be careful in making assumptions about scale / gauge / currents etc unless they have been clearly stated in each message}

 

On our 00 Loft layout  (Barnstaple - North Devon) I was going to have the L&B analogue, as it had been when I was using Zero-1 for the main layout ... but tempus fugit... and dcc in H0e etc became available, and Heljan have released dcc Manning Wardle locos  - this leaves only some old Egger/H0e industrial narrow gauge to run in analogue, from the dcc, rather than retain normal analogue power+control just for them .... Since then, we have used the Busch System our H0 layout - via their dedicated module x2 - and it just keeps thngs so simple 8-) ...everything on out layout(s) can be controlled from any Multimaus (Locos and Accessories) - or via touch screen or tablet /phone with Z21.  We have the flexibility to stand behind the visitors and operate what they are looking at - without them seeing how.   [ We also have Roco joystick controls for them to operate a crane - as a separate dcc system ]

 

In a related way, I also use Uhlenbrock dcc modules for 1 or 2 relay changeover switches  - they are quite small, easily programmed, and allow 'non dcc' devices such as a Faller Carousel/Roundabout with a synchronius AC motor to be controlled from the UNIVERSAL dcc handset - no fixed control panel.

Edited August 17, 2018 by Phil S

[Nigelcliffe]

Phil, I think I understand where you are coming from.  Going back to my original thought using the Portescap system where two diodes in opposite direction were connected to protect the coreless motor, I had always understood they were there to protect the coreless motor from incoming surges.  My logic was to use them so that they prevented outgoing surges.  So is my understanding in-correct and are they there for another reason.  Not being sarcastic or clever, I am interested in learning more.  

On an associated subject re control, on the advise of a gaugemaster technical Guy, I have just replaced the 14VAC input with 12VAC and what a difference that has made to controllability.  Almost as good as the KPC on the test Bachmann K3 with a lot less voltage fluctuation.

 

I can't work out what your two diodes are doing in terms of "protection".  I think its a bit of popular mythology, rather than actual protection.    Two (normal common "rectifier") diodes in opposite directions in a wire will drop the voltage in the wire by about 0.7v and very little else.  That could help with some simple analogue transistor controllers which don't actually go down to "zero", but end up stopping at about 0.7 to 1v minimum.  With that control putting out a small voltage, a very efficient motor (including Portescap types) would turn slowly, so the diodes deal with that shortcoming in the controller design.  The diodes could be fitted at the controller end of things rather than in the loco. 

 

Phil has covered a lot of the issues around shorting the motor outputs of a decoder.  Possibly giving more detail than needed at this stage.    As well as the risk of a direct short (derailment, loco driven into incorrect turnout, screwdriver over the rails), there is one other risk which depends on the layout's wiring and use.  Phil mentioned it briefly:  if there is any means for the normal DC from the KPC controllers to get to the DCC decoder outputs, that also risks blowing the decoder.  The worst would be the full DC (both wires), but depending how things are wired, even one rail could pose a risk.  If the layout is wired with a "common return" (a very common way of doing DC layouts, saves a lot of wire and allows cheaper/simpler single-pole switches), then things may get complicated. 

[RAF96]

I think you will struggle to find any DCC decoder that will power a trackload of up to 30 locos, even if gated within their own power sections.

 

Rob

[Phil S]

I think you will struggle to find any DCC decoder that will power a trackload of up to 30 locos, even if gated within their own power sections.

 

Rob

Phil: I believe the method supported by software applications is to use 1 decoder-per-section - so that 1- 1.5A decoders (or smaller if a known and limited range of suitable locos)  - the Busch Decoder Module for H0f is a specialised but useful example where that is expanded by the use of 4 functions to select between 4 sections -  and allowing more than 1 on at a time facilitates moving a train along a multisection area:  In our case (1): it is used for a 'mine working/ forestry industrial area - with 3 basic routes and a common loco shed/home area to which a loco can be sent ( 1 at a time ) ....the points are operated manually at the moment ( a fragile part of the system !)  The 2nd instance of the H0f is for a 'miniature tourist railway' or 2 separate military shuttle tracks by a castle - all of which can now be controlled from  any dcc controller  - and without having to have an analogue control system and power supply - the locos are too small to fit a decoder in individually  (which used to be the case with N/H0e.... )

 

In 'full size life' - wasn't Prof Eric Laithwaites Magnetic Levitation Linear Motor Train an example of such sequentially switched power provision ???

[Theakerr]

Thks for all the feedback and info.  Fair bit to digest.  As a bit more background I do not run a common return and when in operation each controller is completely isolated from the others.  I run an analog automated system that allows me to run a sequence of 8 trains, each with its own storage track such that when a storage track is activated a specific controller tied to that track is also activated so that the train completes a loop at a realistic speed for what it is (goods, passenger).  This train then activates the next train in sequence.  Basically my system uses ladder logic.  Each track is activated by a reed switch that in turn activates an auto relay and other track are isolated by the points change.   Going DCC is not really an option due primarily to cost (its not just the decoders), but also due to the difficulty of fitting decoders to some of my locos and at the end of the day my system does what I want it too, i.e. have a beer and watch trains go by.  The highest voltage I have seen is 10V but most trains run about 6 to 7V and max amps are about 1.4 with most being in the 0.7/8 range.  

Interesting about the diodes on a Portescap motor.  I will see if I can find some of the original literature and see what it says.  

[47137]

Going off at a bit of a tangent, could you modify the analogue controller to use a motor-driven potentiometer, in the style of a hi-fi amplifier? Configure two function outputs of the DCC decoder to operate so they are 'on' only when their function button is pressed on the radio handset, and use these to drive the motor. Put a changeover relay near the output of the analogue cab to change direction, and use a third function output to drive this.

 

In functional terms this keeps all the circuit protection of the analogue controller, and uses the radio handset to control the analogue controller so to speak. In electrical terms, you might want a couple of transistors to beef up the two function outputs to drive the potentiometer motor. Emergency stops will be at the mercy of the motor turning the potentiometer fast enough. Driving will be in the method of the old-fashioned 'shunting' style controller like the ECM 'Rambler'.

 

This would be an experimental project to begin with but it would protect the chosen decoder from all the things which can go wrong on the analogue layout.

 

- Richard.

[davetheroad]

This concept is doable using radio control receivers. BlueRail trains have a video showing such a setup. The big issue is providing short circuit protection for the receiver chip. The Deltang system I use has receivers which have outputs handling up to 6A with a typical small receiver having 1.3A output. Rectifying would be needed for the power supply, unless you used a big battery, readily available. No other power like DCC on the track either

 

I think such a system would be best suited to a single loco in steam where there are a large number of analog locos the owner does not want to chip. You do get the possible advantage of a radio throttle built in and the ability to control various accessories. 

[steam-driven boy]

Hi,

I continue trying to work on my N gauge layout and this topic reminded me of having seen the SMALL SCALE, BIG SOUND article in Model Rail 168/Spring 2012, the notion of adding this form of sound more than the other aspects of DCC to the layout appealed given my perceived limitations in the on-board quality of sound in the scale even with more recent improved products, the space available just doesn't allow the sound I hope for.

I had a look at the website which includes the description and a video, then emailed a couple of queries to Paul Lawton who very kindly responded in timely fashion with reassurances regarding those uncertainties.

I've only run a small test track with a single loco at a time, further testing will continue when I am able.

 

Regards, Gerry.

[34theletterbetweenB&D]

Thanks for the suggestions, I might give it a go with an old decoder I have....

 Have you given it a whirl?

 

I have done this twice.

 

Once as an expedient, away from home and had DC trains to test with a small layout and it turned out the available DC controller was fritzed; but had the DCC kit with me and a few decoder fitted locos. I didn't even remove the decoder used from the loco, just disconnected from the motor and connected to track. This worked perfectly well

 

Second time to pick off the sound output for input to a static replay system to demonstrate to an unbeliever that the brain maps sound to the 'source' - even a moving source - if the soundstage is reasonably congruent with the putative source location, because sound is multipath in the real world and the grey goo has years of training in matching that heard sound with that visually observed object.

 

If doing this for any length of time I would definitely arrange the current limiting protection suggested above, but short term there was no need