DCC Loco Conversions - 25 - DCC Gotchas

Discussion in 'DCC Control' started by Jim Freight, Jun 7, 2022.

  1. Jim Freight

    Jim Freight Full Member

    Messages:
    1,101
    Likes Received:
    878
    Joined:
    Sep 9, 2019
    During my conversions and use of a variety of elderly locos two curious effects have been found which I thought I would pass on, to which I have added further gotchas.

    1) Spurious Short Circuits

    My layout is divided into 15 separate areas each covered by a solid state breaker to aid fault finding when running multiple locos, and to prevent shunting mishaps suddenly stopping a fast freight on the mainline causing it to concertina.

    Typically they are set to limit the current to any one area according to anticipated maximum load, full rated for the mainline feeds, minimal for yards.

    When fitting a fuse to a plug in the UK you may have a choice of 1A, 2A, 5A or 13A to suit the appliance and protect the cable, so I apply the same reasoning in protecting loco pickups in particular, there is no need for a 6A rated electronic breaker such as in the command station or booster when it feeds a yard with generally only one loco running.

    However, if I have a breaker set below about 2A I can get spurious short circuit tripping when a loco has below parr pickups, or when using a wheel cleaning brush on older loco wheels.

    It would seem that surges due to inefficient pickups can be reacted to as short circuits, solution, keep wheels clean and when conductive wheel cleaning brushes such as the Peco or Tri-ang types are used do not power them from a track with a breaker set to a low value.

    Similarly some locos with plain alloy wheels, or the cheap Hornby 0-4-0 plastic chassis should also not be run on track with a low breaker trip current.

    2) It programs but does not run

    Two options here :-

    2.1 For some mysterious reason the data store on the decoder has been corrupted, usually resetting the decoder, typically setting CV8 to 8 cures this. Word of warning - if it is a sound decoder check the decoder instructions carefully to avoid deleting the sound data.

    2.2 Programming on the programming track works with very little current, deliberately so, if the power pickup by the loco is just the right resistance, programming will work but loco will not run on the main as the voltage drop between track and decoder when the motor is powered can be sufficient to prevent the decoder from functioning.

    This is difficult to prove but seemed to be the case during one Dublo 2-6-4T conversion.


    I would be interested to know if any one else has experienced similar playful responses from DCC operation.

    Jim :)

    Return to Index
     
    Last edited: Feb 24, 2024
    Rob Pulham and jakesdad13 like this.
  2. paul_l

    paul_l Staff Member Administrator

    Messages:
    9,848
    Likes Received:
    5,902
    Joined:
    Dec 5, 2015
    Hi Jim

    I have had several decoders over the years loose thier settings, often after a short somewhere on the layout. These were generally cheaper decoders. My normal practice is to setup the decoder on my programming track (rolling road), and back up to JMRI Decoder-Pro. A simple re-write of the settings fixes the loss of memory - wish it worked for me :whatever:

    I am fortunate in that I use a NCE PowerCab for my programming track, which can provide 1.5A, which is also very helpful for programming Sound decoders, my PowerPro system is not so reliable for programming sound decoders as the programming track current is less.

    For programming up Bachmann DMU's which require a decoder per coach - I use the same decoder model for each coach - mainly the Gaugemaster 8pin as it fits.
    The decoders normally pulse the motor to give feedback to the controller that a read or write has taken place. This gives an issue when no motor is fitted (i.e. a dummy car). My way around this is to fit the first decoder in the motor car. Save the config to JMRI. Remove the decoder and fit into the dummy car. Install another decoder into the motor car and rewrite the config into the decoder. Jobs a goodun (just repeat if you have a third / fourth car). All have the same decoder address, and because the dummy car and motor car point in opposite directions the the directional lighting works correctly.

    Paul
     
    Rob Pulham and Jim Freight like this.
  3. Jim Freight

    Jim Freight Full Member

    Messages:
    1,101
    Likes Received:
    878
    Joined:
    Sep 9, 2019
    Hi Paul

    Yep, had that one.

    I wonder why they do that, but I really think that is the motor responding to command signals as blips of motor power, however if I try and program or use a standard decoder designed for motor driving and the motor is out of the circuit for what ever reason the decoder is deemed to be not present.

    So whether it is due to the motors presence being required to satisfy the design of the electronics or the firmware I do not know, decoders for accessories must have an alternative arrangement :scratchchin:

    Jim :)
     
  4. Jim Freight

    Jim Freight Full Member

    Messages:
    1,101
    Likes Received:
    878
    Joined:
    Sep 9, 2019
    Finding more to include here, mainly gotcha's, so I have incorporated this thread into my DCC Conversions series.

    3) Runs well on DC, programs, but trips out DCC controllers when on the main.

    This type of fault does not always get detected or show up on DC or when the loco is static on a programming road, these are often dynamic.

    Early (1960s) DC controllers I have used either had electromagnetic trips or thermal trips, the essential difference is that the electromagnetic device would operate relatively quickly in those olden days, whereas a thermal trip would tolerate momentary short circuits, only a prolonged short would trip them.

    The electromagnetic trip could be a little temperamental and could be forced to not trip by an aggressive users finger, whereas the thermal trips were not user accessible, so could not be overridden and were self resetting. However problems can be masked when checking out a loco prior to implementing a conversion to DCC running, e.g.

    3.1 Wire and oil debris - A strand of wire caught up in an over oiled loco stuck to the chassis side making occasional contact with a metal insulated wheel, easily masked with older trips, noticed instantly on DCC.

    3.2 Excessive end float on a driving axle - 1 - Dublo shunter, the centre axle has excessive end float from worn bush allowing an insulated wheel to make contact with the adjacent outside frame.

    3.3 Excessive end float on a driving axle - 2 - Wrenn 2-6-4T, front insulated driver making contact with metal bracket support for the piston slides due to back to back opening out caused by an aging wheel bush.

    3.4 Flaking plating on a pole piece - Wrenn 2-6-4T, why they were plated I do not know but it appears to have been done prior to forming judging by fracture lines in the plating, this did not catch me out but I have replaced the pole pieces with a couple of salvaged unplated parts. The plating appears to be ferrous and may be magnetised, just looking for a place to cause mischief.

    3.5 Drooping metal coupling - Wrenn 2-6-4T, the Tri-ang Mk3 tension lock coupling is not a happy coupling on the front pony truck, it can be just low enough such that the dropper used with an uncoupling ramp can momentary contact point blades if the truck bounces just right. Seemingly at random making the fault tricky to isolate, I suspect this is why on another 2-6-4T the previous owner had cut the dropper off.


    Jim :)
     
    Last edited: Jul 6, 2022
  5. paul_l

    paul_l Staff Member Administrator

    Messages:
    9,848
    Likes Received:
    5,902
    Joined:
    Dec 5, 2015
    Hi Jim

    The pulseing or notching of the motor during decoder reads and writes is part of the feedback mechanisim that the NMRA standard that decoder manufacturers and DCC system makers have adopted to provide simple feedback. On the Programming track, no decoder address is sent as the command is sent direct to the decoder connected. The decoder acknowledges the instruction by an increased current draw of at least 60mA for 6ms - usually by pulsing the motor.

    The good decoder manufacturers - like ESU pulse in alternative directions, so the loco on the programming track effectively stays still, my early Hornby and I think the Hatton ones would require pushing back the loco every few min before it ran out of track. A rolling road alleviates this issue.


    Another advantage of using the programming track is its very reduced current and voltage. Thus if a decoder is wired up incorrectly it wont be distroyed, whereas being connected to the higher voltage of the Track bus would fry it. For this reason it is advisable to try all new decoder installs on the programming track first.

    Most modern systems support Programming On the Main (POM) mode, this allows users to send Decoder CV changes when the loco is on the Main Layout. My NCE systems support it, but it's not something I ever do. I prefer a dedicated program track, less chance of mistakes - I make enough as it is without increasing the risk :whatever: ;)

    However, most programming tracks cannot provide enough current to verify the Motor and function operations, so these will need verifying on the Main.

    Paul
     
    jakesdad13 and Jim Freight like this.
  6. Jim Freight

    Jim Freight Full Member

    Messages:
    1,101
    Likes Received:
    878
    Joined:
    Sep 9, 2019
    Hi Paul, thanks for that info.

    Lenz supports POM, but I only use POM when fine tuning a loco set up, such as the mid and max speed settings once I know the installation is sound.

    I like the sound of alternative directions, very useful, most I have used only go in reverse, mind you there is so much backlash in some old locos a 60mA pulse does not move the loco!

    Jim :)
     
    Last edited: Jun 9, 2022
  7. Jim Freight

    Jim Freight Full Member

    Messages:
    1,101
    Likes Received:
    878
    Joined:
    Sep 9, 2019
    4) Hunting

    A familiar term for those of us that experienced this with our old school gas guzzlers of yore, this is when the speed of a vehicle engine cyclically speeds up, corrects itself and slows down again when all you demanded was a constant speed.

    I have just been having some tedious 'pleasure' with a Wrenn 2-6-4T which had multiple issues that makes locos with integrated motors a real pain to fault find on, finally having solved most issues, smooth running is something that only really occurs when it has a load behind it.

    This is certainly an issue with low gear ratio motors such as in the Dublo 2-6-4T and Duchess locos where backlash of the gears can be significant.

    However this can also be problematic for a loco with a high gear ratio motor descending a slope with a train, especially if thrust bearings are not fitted at each end of the armature shaft causing varying frictional load.

    Back to our Dublo motors with a low gear ratio worm, generally with a steep helix that can be back driven, i.e. the drive axle gear can rotate the armature, on the bench this is easily demonstrated when the magnet and brushes are not present on a Dublo open frame motor.

    This can also occur to a limited extent when the motor is running and the train behind the loco starts to push it downhill or bounces off the loco due to a speed surge followed by a slowing or hesitating due to a momentary power pickup failure.

    Also the backlash in the gears means that the load on the motor is not constant, which means that like a high gear ratio loco descending a slope and being pushed, sometimes the loco motor is pulling other times it's being pushed.

    An interesting example can be found here :-

    https://www.rmweb.co.uk/topic/157352-back-emf-and-gradients/

    It would seem that BEMF firmware does not handle backdriving well or at all, if someone can confirm or deny this authoritatively I would love to hear from you!

    This hunting effect will also occur if two locos are coupled together working as a pair and both have BEMF enabled, they will quickly start to fight each resulting in wheel scrub on the rails and the train behind them being jerked causing the extra fun as described above due to backlash or downhill running. The way around this is to disable the BEMF on one of them.


    A Trix Warship was particularly troublesome, details to follow in a future article, from my notes :-

    "First attempt used a Hattons DCR-8PIN-HarnessMini but erratic pickup issues, 2nd a Rails RoS-6D, solved pickup issues but severe hunting when BEMF enabled, when disabled difficult to control at low speeds. 3rd attempt using a Gaugemaster DCC93, loco runs well with BEMF."

    I think I will fit a socket to the Wrenn 2-6-4T to see if it runs better with a DCC93, the wheel plating of this loco has little wear so advanced brown out protection may not be required.

    Oh, and by the way, BEMF issues can probably be had with DC electronic controllers too.

    Jim :)
     
    gormo likes this.
  8. Jim Freight

    Jim Freight Full Member

    Messages:
    1,101
    Likes Received:
    878
    Joined:
    Sep 9, 2019

    5) One size (or rather family) does not fit all.


    5.1) Fitted a Gaugemaster DCC93 into the Dublo/Wrenn 2-6-4T that was unusable, it now runs well.

    I Now have one 2-6-4T that runs well with the RoS-6D, four not really good enough and one running well on a Gaugemaster DCC93.

    So four more to try with the GM DCC93.

    5.2) The RoS-6D out of the 2-6-4T that was unusable runs a Dublo Deltic well :scratchchin:

    5.3) I tried a RoS-218 in the Deltic (plenty of space for the wider package), supposedly the same core, and it seems to waver at higher speeds, and the CV speed settings had to be higher to reach the same top speed as with the RoS-6D. :facepalm:

    Otherwise very good at low speeds on BEMF.

    Jim :hammer:
     
    Last edited: Jun 18, 2022
  9. Jim Freight

    Jim Freight Full Member

    Messages:
    1,101
    Likes Received:
    878
    Joined:
    Sep 9, 2019
    Update 06/07/22

    Added "3.5 Drooping metal coupling"

    Jim :)
     

Share This Page