DCC Loco Conversions - 26 - Dublo & Wrenn 2-6-4 Tank

Discussion in 'DCC Control' started by Jim Freight, Jun 30, 2022.

  1. Jim Freight

    Jim Freight Full Member

    Messages:
    1,103
    Likes Received:
    882
    Joined:
    Sep 9, 2019
    Installation of a decoder is straightforward, modifying the chassis to insulate the original brush holder takes most of the time.

    Very Important - the chassis was modified as in articles 22 and 23, failure to do so can destroy a decoder.

    https://platform1mrc.com/p1mrc/inde...rsions-22-dublo-brush-holder-extraction.6017/

    https://platform1mrc.com/p1mrc/inde...sions-23-dublo-brush-holder-replacement.6025/

    Issues

    Just skip over this section to the solutions if you wish to avoid the gory details.

    This proved to be quite a marathon compared with the other Dublo locos converted to date.

    I originally had three to convert which also required repairs to get back to running order, however after initially writing one off as beyond economic repair and buying some more over the last few years for breaking for spares I ended up fixing them all so I ended up with six!

    Initially conversions used the RoS-6D decoder which all ran well on the rolling road (DC and DCC) and the test track (DCC only) so I converted all six to use that decoder, mistake!

    When I started to track test them on the layout most did not behave as well as I anticipated which then led to some long and drawn out CV adjustments and on layout testing.

    They are a mixture of ages, manufacture and wheel tread condition from as new wheel tread plating to 50% worn through to the alloy base material of the wheel, some with Alnico magnets and some with NEO magnets.

    This is summarised in the following table.

    Key Loco Parameters

    Key Loco Parameters Table.jpg

    As can be seen 5 out of 6 locos preferred a Gaugemaster DCC93 decoder.

    BEMF

    Previous experience with the RoS-6D was positive with worn wheels but another issue manifested itself in the form of erratic running above crawling pace, yes, the RoS decoder works well with these locos when starting but as soon as the they were running nominally above a scale 30 mph various degrees of hunting was exhibited, from annoying to the eye to the totally unsusable loco 5.

    I use BEMF for two purposes, good low speed control, especially for shunters and goods train locos and for leaving mainline trains running unattended on the continuous mainline with 1 in 40 grades.

    Without BEMF enabled none of these locos would start very easily, they would suddenly leap into life and need the speed backing off again, BEMF control reacts much quicker than a human hand on the controller, so turning off BEMF completely was not an option.

    BEMF Cut Out

    I tried using the BEMF Cut Out option where above a certain speed BEMF is disabled. These tank locos are powerful and heavy with good traction and barely notice a rake of six British Trix coaches fitted with 14mm Hornby wheel sets so this was worth a try. Although they are more likely to be used with 4 suburban coaches or a mixed frieght in practice.

    However this can be a little tricky to set up well, I use the JMRI Decoder-Pro tool and the cut out is specified as a number between 1 and 128, which is all very well if you are using 128 speed steps but I use 28 steps. In which case I found that you need to set it according to the 1 to 128 which maps to 1 to 28 something like this.

    BEMF Cut Out Value Translation

    BEMF Cut Out Table.jpg

    A gotcha that can arise here is that when the cut out threshold is reached, depending on the loco, it's load and the grade it is on it may not be a smooth transition especially at low speeds.

    A way around this is to use the custom speed table option as opposed to using the default speed table with the start,middle and max speed CVs. Typically at the cut out threshold a step up in the speed table value is required to maintain a smooth transition, this is tricky to achieve and trial and error was required while programming on the main. Even so there was still some hunting occuring at scale speeds of 40 mph plus, I typically limit these type of locos to a maximum scale speed of 60 mph.

    This was confirmed by running with BEMF turned off completely, so possibly an effect due to the PWM applied to the locos or was it a dynamic related to the locos getting near the current limit of the decoders?

    For whatever reason this problem seems to be confined to these early design low revving open frame Dublo/Wrenn motors. However loco 4 ran perfectly with the RoS-6D!

    Another Decoder

    My fallback decoder for difficult locos is the Gaugemaster DCC93 decoder which I tried initially by fitting a socket to the worst performer, the unusable loco 5, although it was not as smooth to control when starting as with the RoS-6D it was acceptable and ran smoothly at all other speeds.

    Apart from loco 4, the RoS-6D decoders were replaced with the GM DCC93 decoder.

    Despite using BEMF they needed start voltage CV settings from a minimum of 12, most 20 and one 30, supposedly a non-zero should not be needed with BEMF enabled but it does make a difference.

    So as these locos will spend most of their time running at 40-50mph on the main line this was an acceptable solution, it must be remembered that these relics were designed for NASCAR speeds on a train set oval some 60 years ago!


    Solutions

    First two images show the brush holder replacement and wiring common to all of these locos.

    1 DSCF0529.JPG

    2 DSCF0527.JPG

    1) Rails RoS-6D Decoder

    Decoder mounted on an offcut of laserboard, held in place with double sided sticky pads.

    Pins 5 and 6 are cropped off as not required.

    Wires lap jointed onto the decoder pins as before, loose sleeved with unshrunk heat shrink sleeving. Multi-stranded 10/0.1 equipment wire with PVC insulation is a good choice for these locos.

    Note that I do not connect the wire from the collector plate directly to the decoder but only have a short length attached to the decoder, an easy to access joint then connects it to the wire from the collector.

    The reason is that when the pickup collector wears out in maybe 5 to 10 years time I do not need to get my soldering iron near the decoder for the replacement when my eyesight and steadiness of hand may be somewhat diminished. It is important to consider future potential maintenance and repair difficulties when age is not on your side.

    3 DSCF0537.JPG

    4 DSCF0539.JPG

    2) Gaugemaster DCC93


    Mounting of the DCC93 is similar however the pins of this decoder are seperate pieces of crudely cut wire soldered to PCB (Printed Circuit Board) pads, whereas those on the Rails decoder are part of a 6 way inline plug and stay in place during wire attachment.

    Covering the rest of the PCB with masking tape avoids any solder splatter going where it shouldn't on these naked boards.

    The pins de-solder very easily from the naked PCB of the DCC93 so it is better IMO to remove all 6 completely, just touching them with a solder wetted bit is generally enough to lift them off with the iron tip.

    Multi-stranded 10/0.1 equipment wire with PVC insulation is soldered directly to the PCB pads, a steady hand and a temperature controlled iron is required, preferably with static electricity grounding, this is not for the faint hearted

    Alternatively make up a 4 pin socket and solder to that.

    The decoder is mounted on an offcut of laser board as for the Rails decoder.

    In addition Duck tape is used to insulate the inside of the body shell in case the decoder gets loose, although the DCC93 does have automatic shut down if track power connects to the decoder motor drive outputs, a certain law stipulates some other part of the PCB will contact track voltage from the chassis or body with destructive effect.

    5 DSCF0567.JPG

    6 DSCF0545.JPG

    In Operation


    Generally the current consumption of this open frame motor with a re-magnetised on re-assembly Alnico magnet, or a NEO magnet on the rolling road without the body on DC was 0.25A at 6V .

    Some were slightly higher at 0.375A (3/8ths amp, my old meter unit was calibrated in eighths), all within the continuous capabilities of both decoders. Locos with Alnico magnets and a higher current draw were NEO fitted to lower their current consumption.

    No issues pointed to any preference for a NEO or a healthy re-magnetised Alnico magnet, these are strong locos with excellent tractive effort, probably the best power to weight ratio locos of my Dublo fleet and of course no heavy tender to drag behind them either.

    Could I have done this better ?

    Yes, three aspects come to mind.

    1) A full on layout track test should have been conducted before bulk installation, running on the layout is different to the test track for reasons unknown, mind you if I had chosen the wrong one i.e. loco 4 I would still have been mainly wrong.

    2) Should not have stocked up on so many decoders to start with.

    3) If I could only solve trying to design parts for my 3D printer then I could dispense with sticky pads and make cradles that could be screwed in place, particularly for the naked PCB decoders.

    Summary

    This turned out to be a rather long exercise due to the number of locos I inadvertently fixed prior to conversion and my initial choice of decoder, hopefully you will have better luck with your choice, but this really does illustrate how different some of these motorised relics can be.

    The Rails Connect family of decoders are not that good with low revving, low geared motors with worn gears, it will be interesting (I'm mad!) to see how the Dublo N2 and A4 locos which share the same basic motor arrangement convert, luckily I only have two of each of these.

    I still consider the RoS-6D decoder to be the preferred first choice for good slow speed running without requiring the expense and large physical size of a Stay Alive system, additionally these decoders are usefully small.

    Discussion always :welcome:

    Jim
     
    Last edited: Jul 2, 2022
  2. Jim Freight

    Jim Freight Full Member

    Messages:
    1,103
    Likes Received:
    882
    Joined:
    Sep 9, 2019
    The RoS-6D decoders that were unsuccessfully used in 5 of the 2-6-4 tank loco conversions are now fitted to, (these conversions will be posted shortly) :-

    Dublo N2
    Wrenn N2
    Dublo A4
    Wrenn A4
    Dublo Co-Co (Deltic)

    Although another Dublo N2 and a Deltic did not like them, be prepared for some serious fun when converting these old locos for DCC operation, no two are the same!

    Jim :)

    Return to Index
     
    Last edited: Feb 24, 2024

Share This Page