TCS FL4 Reverse Direction Issue

trptman Feb 27, 2019

  1. trptman

    trptman TrainBoard Member

    10
    2
    5
    I'm attempting to program the FL4 decoder before mounting it in a loco. I'm using DecoderPro and Sprog3 and am able to set the lighting effect on each function without issue. However, the functions only work when the function is set to both directions or forward direction. When I put a function into reverse direction, it will not light up at all regardless of the direction set on the throttle. Separately, I'm testing a Loksound Micro Select and don't have any issues with direction. Any suggestions on things I can try for the FL4? Thanks!
     
  2. Onizukachan

    Onizukachan TrainBoard Supporter

    752
    1,198
    24
    Trpt, are you programming on the main? I did an fl4 for lighting in 4449 a few weeks ago, and didn’t have any issues programming my reverse light, nor for Mars to turn off in reverse.

    I just “did it” in jmri. Probably because I didn’t know any better, lol.

    Are you writing changes, or sheets?
    I found that because of the lack of feedback /ack , I had to write sheet to ensure the setting took, then I’d flip the track power off and on, and it would be working as expected.
    Doing the resistor on one of the light outputs as suggested in the manual did NOT work at all for me, thus the blind on the main as write sheet approach.
     
  3. trptman

    trptman TrainBoard Member

    10
    2
    5
    I've been writing in various combinations - single CV, changes on sheet and whole sheet. I do get feedback as the light flashes for each CV written and can alter the light effect without difficulty. I agree that, like you, I have been unable to read the decoder and have tried various combinations of resistors without results. I've also tried changing to programming mode and putting the JMRI connection to the Sprog in sprog mode - no difference.

    The other thing I'll note is that when I put a function in forward direction only, the throttle direction doesn't make a difference - that is the light will remain on when putting the throttle in reverse.

    I think tonight, I may try to reset the FL4 and start over - not a big deal. I've tried so many changes to CVs that I'm not sure what I messed up, so it may be better to start over and test each change carefully.
     
  4. Onizukachan

    Onizukachan TrainBoard Supporter

    752
    1,198
    24
    Sounds like the address isn’t taking, to me.
    Have you considered putting the loco this will be paired with on the track so it can ack for you?
    Or a completely different one?
    I put my 6104 f3a on the track when I did the fl4 address change, then removed the fl4 and wrote all sheets for the correct entry/ the correct address etc back to the f3a.

    Also, you do have the fl4 as it’s own roster entry in jmri, correct? Mine for example is named “z lights 4449” so it goes to the bottom of the roster, jmri can’t apparently handle multiple decoders in one roster entry.
     
  5. trptman

    trptman TrainBoard Member

    10
    2
    5
    No difficulty changing address and using either the short or long address. I've been programming the decoder using a separate roster entry and with nothing else connected to the track except the FL4. The lights respond find to the throttle buttons as long as the function has been programmed to both directions.
     
  6. Onizukachan

    Onizukachan TrainBoard Supporter

    752
    1,198
    24
    I apologize for asking but, are you sure you aren’t seeing the default DC override on the lights?

    I think you’ve got the right idea, reset it fully.
    But you might consider putt8ng a loco on the track so something responds to the acks. Can always go back and write all sheets on the loco to put it back the way it was after.
     
  7. trptman

    trptman TrainBoard Member

    10
    2
    5
    I reset the decoder tonight and proceeded to make each change individually and test each one before proceeding to the next change. I got it working the way I intended. Onizukachan - thanks for your responses.

    For what it's worth, I have a few notes that may help the next person working on one of these decoders. I'm using JMRI version 4.12. First, attach an led with resistor to one of the outputs - every time a CV is written successfully, it will flash - very helpful feedback. Second, when changing the Function Map, make sure to write the entire sheet. I made the mistake of initially writing just the changes and it left some old settings in some of the CVs and got some weird results as I tested using the function keys. You can confirm the setting are right by comparing the CV settings after changing the Function Map to the TCS manual and/or pressing all the function keys that shouldn't respond to make sure it doesn't light the output.

    Next, on the Lights sheet, some of the wire effect drop-downs aren't matched to the correct CV setting. For example, when the purple wire is set to constant dim for both directions (CV52=44), Auto Mars shows in the drop-down. I recommend setting the effects by using the CVs directly.

    In order to lock the decoder with multi-decoder installs on the same loco using the same address, a recommended practice is to set CVs 15 and 16 which this decoder supports. When they match, you can make changes. When they differ, the decoder is locked. Each decoder in the loco will have a different value for CV16. In order to set CV15, you need to use the Simple Programmer to change this CV. The key here is to set the Mode to Ops Byte.

    Lastly, I struggled to figure out how to change the decoder address. Every time I attempted to make the change, the led would light indicating the CV was being written, but the throttle would only respond when the address was the default of 3. What finally worked was changing the preference to Sprog from Sprog Command, using the programming track option with the Direct Bit. I may not have needed all three of those, but at least, it is working now.

    Next up, is to finish the preliminary set up on the Loksound decoder (almost done), and test the two together. After that, I'll be ready to install the decoders and leds in the loco!
     
  8. Onizukachan

    Onizukachan TrainBoard Supporter

    752
    1,198
    24
    Man, I’m glad to hear it all worked out. I didn’t realize you didn’t have anything connected!
    I can tell you One thing it does NOT do well with LEDs is dimming, recommend dimming via additional resistor, or via tinted canopy glue.

    Cab interior is dimmed via canopy glue method and numberboards via extra 220 ohm resistor.

    If you will be using mars light, I put the settings that worked for me (603 LED) in my home thread below, along with a video.
    Once it is installed, it’s pretty easy to tweak those. All the numbers interplay though, so as you change brightness, the cycle speeds up or slows down a partial step also.
     
  9. Onizukachan

    Onizukachan TrainBoard Supporter

    752
    1,198
    24
    FYI, for those who might find this topic in the future, JMRI has some bad lighting definitions for fl4.
    For example green doesn’t show automars at all, and the other three colors show it, but puts in 44 instead of 45 for automars. Several of the definitions put in the wrong number into the cv.

    So if you are using an fl4 set it all up, then go modify the lighting CVs manually.
     
  10. RCMan

    RCMan TrainBoard Member

    271
    132
    12

    Have you let the JMRI programmers know about this issue?
     
  11. Onizukachan

    Onizukachan TrainBoard Supporter

    752
    1,198
    24
    They have been aware since sept of last year. Supposedly fixed it but I tried updating definitions and nothing changed.
     
  12. RCMan

    RCMan TrainBoard Member

    271
    132
    12
    I have a couple FL-4's bought 4 years ago, will check them using the latest release of Decoder Pro.

    Thanks
     
  13. trptman

    trptman TrainBoard Member

    10
    2
    5
    Here is a follow-up on my FL4 saga. After my response above on 2/28 that all was well, a few days later it started acting up again - functions would not respond to reverse throttle and worse, function 4 couldn't be shut off - light was on continuously. I reset the decoder several times and tried changing functions little-by-little, but it didn't fix the issues. Several weeks ago, I sent the decoder to TCS for warranty service. I received the same decoder back again yesterday. TCS's only note was "repaired".

    I tested the decoder last night - all my changes to the functions worked as desired including proper response to reverse throttle. Now here is where it gets interesting. The last change I was going to make was to change the address of the decoder. I put JMRI into SPROG mode and used DecoderPro in Programming Track mode and successfully changed the address from the default of 03. Now, the decoder no longer responds to reverse throttle setting, just like before. I tried resetting, but that didn't fix the reverse throttle issue. Does anyone have any suggestions on why using the Programming Track mode or changing the address would cause this issue? I really don't know if it is the FL4 or some issue with DecoderPro and/or SPROG that causes this in the FL4.

    I'm not going to send it back to TCS, as the cost of postage (twice) is about the cost of the decoder. I plan to live with it and use the functions that don't rely on throttle direction.

    The other decoder in this loco was to be a Loksound Select Micro that I bought on eBay. I made a careless mistake in testing where I let the alligator clip from the blue positive return wire get too close to the speaker magnet. The magnet attracted the clip and fried the decoder. ESU was gracious to allow me to get a replacement at no charge, but a day after receiving it, they said they have used up their spare inventory of Select Micros and I'd have to wait on one of the new V5s. I like the specs for the V5 compared to the Select Micro (I can use some of the additional functions to replace the FL4 functions that aren't working), but it will be a challenge to program as JMRI hasn't yet released updates for this new decoder.

    I'm brand-new to model railroading and DCC, and my wife keeps asking me "are you sure you want to deal with all these challenges?" I'm beginning to wonder myself!
     
  14. Onizukachan

    Onizukachan TrainBoard Supporter

    752
    1,198
    24
    What happens when you change it back to 03? Does the function return?
    If so it has to be the JMRI /sprog version causing the issue.
     
  15. RCMan

    RCMan TrainBoard Member

    271
    132
    12
    JMRI programmers are working on the ESU V5, give them time as they are all volunteers and have a family life also.

    Watch for the test versions changes and it will show up there first.
     
  16. trptman

    trptman TrainBoard Member

    10
    2
    5
    Richard - yup, setting it back to 03 was the first thing I tried. Also tried a reset. No luck. It seems that changing the address the first time did something more or less permanent to cause reverse throttle to not be recognized. Also, the opposite setting was affected as well - that is putting the function in forward only would light regardless if the throttle was in forward or reverse.

    Dennis - thanks for that info. I wasn't complaining about the JMRI updates. I recognize that the coding is handled by volunteers - just remarking that it is going to be a challenge to program the decoder. May not be an issue in the near future anyway as ESU doesn't have an ETA for the micro versions of the V5.
     
  17. RCMan

    RCMan TrainBoard Member

    271
    132
    12
    I was just letting others know about how JMRI gets changes. JMRI has always been volunteer only since the inception.
     
  18. trptman

    trptman TrainBoard Member

    10
    2
    5
    Another update - I emailed TCS and they emailed me back today. Their suggestion didn't fix the issue completely, but it put me on the right track. They recommend using the JMRI Reset function in the menu (I had been using the single CV programmer to set CV8=2). This Reset function did reset the decoder completely and I was able to modify the functions and get them working as intended with the reverse throttle.

    After that, I tried changing the default address using the Basic sheet and using the "Write changes on sheet" button. This caused the reverse functions to fail as before. I then Reset the decoder and got it working as intended, but instead of using the Basic sheet, I went to the CV sheet and wrote only the address change for CV1. Bingo! Address has been changed and the functions didn't get messed up!

    Thank you all for your suggestions and to TCS for responding back to me on a Sunday!
     
    chadbag and Onizukachan like this.
  19. Keith Ledbetter

    Keith Ledbetter TrainBoard Member

    279
    195
    12
    Not to put you through anymore pain but can you pull the CV sheet up as you write the changes and see what it impacts? It's like jmri might have a bad CV mapping in there or something worth investigating further. I have found some funkiness like that very occasional and they will correct in jmri fairly quickly if a reproducible issue.

    Glad you got it working! And super sorry your newbie experience has been so much work. I promise most of the time it's much more straightforward.
     
    chadbag likes this.
  20. trptman

    trptman TrainBoard Member

    10
    2
    5
    Keith - As much as I'd like to help improve JRMI, I'm a little reluctant to reset and reconfigure this decoder.One issue is I haven't been able to successfully read the decoder's CVs, so I'm not sure if I could be certain which CVs were changed with the actions I took as described above.

    I've looked briefly at the JMRI website, and I don't readily see how to report issues or make suggestions for specific decoder definitions - does anybody know?
     

Share This Page