Roco z21 with LGB "serial" system via Massoth eMotion L

LGB 28002 interface is functionally dual-edge-triggered. Connected to a standard on/off aux interface, it detects transitions from on to off (and vice versa).

Caveats:
  • Includes some sort of "glitch suppression". Rapid [off]-on-off (or [on]-off-on) transitions are not counted as a signal. In this cases, "rapid" is a couple of hundred milliseconds.
  • Counts transitions and detects a pause before acting on the signal.
  • Have not yet tested whether a rapid triple pulse ([off]-on-off-on) is treated as a single [off]-on or ignored.
LGB 55020/55021 and Massoth eMotion L both have modes that are (as far as I can tell) supposed to emulate this behaviour out pin F1/A1 based on DCC commands F1-F8.
  • 55020/55021 enable this with CV51=1
  • eMotion L enables this with CV49=[8+1]
However, I can't get it to work with either decoder, and I don't know if it's a failure of the decoder implementation or the 28002 interface. Pretending that I have serial mode handset by manually pulsing F1 (not too fast) works as expected, which implies the 28002 is at least partially working.


Aside: the Massoth implementation is a little odd.
  • Can enable "serial mode", which enables toggles of command F1 to drive commands F2-F8 (e.g. [off]-on-off or [on]-off-on => F2, [on]-off-on-off-on-off => F5).
    • This is useful if you have a serial mode only handset and want to let it drive multiple outputs bound to F2..F8 on a newer engine
  • Can enable "fast pulse" mode, which (according to the documentation) lets commands F1-F8 toggle A1 output an appropriate number of times.
    • This is useful if you have a parallel mode handset and want to drive a legacy LGB output that uses serial mode for command multiplexing (one pin, multiple commands).
  • Note that if you want a legacy serial handset to drive F1, you don't need to do anything special. The handset toggles F1, which toggles A1, and it all just works without the decoder needing to emulate anything.
However, enabling fast pulse mode on eMotion also requires enabling serial mode. As far as I can tell, both modes work simultaneously. So you are simultaneously mapping multiple F1 commands to F2..F8 (to support legacy serial handsets), and mapping F2..F8 to multiple pulses out A1 (to support legacy serial locomotive features). Which means that you're simultaneously supporting:
  • a legacy serial handset driving serial pin F1/A1 (legacy LGB interface) and converting those same signals into F2-F8 to drive some additional feature
  • a parallel handset driving serial pin F1/A1 (legacy LGB interface) and whatever additional features you added on F2-F8
In my opinion, it would be better if mapping F2-F8 as serial toggles of pin A1 were mutually exclusive with mapping serial F1 to standard F2-F8.

Note: it may be that this is actually what the eMotion decoder does. However:
  • The behaviour I see on my decoder is that enabling CV49=8 does not suppress CV49=1 behaviour.
  • The documentation (and practice) requires CV49=1 to enable CV49=8.
  • In other words, it does not seem possible to enable serial emulation out A1 without also enabling serial control from the handset.
(For comparison, this is not an issue with the 55020/55021, as there is only a single output. Compatibility mode (the default) passes F1 straight through and maps F2..F8 to pin F1. There's no need to map serial F1 from the handset to anything else.)
 
LGB, Massoth and Zimo all offer a LGB serial pulse compatibility mode.
  • LGB 55020/55021: set CV51=1
  • Massoth eMotion: set CV49 bits 0,3 [1+8]
  • Zimo: set CV112 bit 7 [128]
My understanding of all of these is that they translate F1..F8 into toggles of FO1 (function out 1, to use the Zimo naming), emulating serial mode to the sound controller from parallel mode commands.

So how does this mode actually work?

Actual serial mode is straightforward. The cab sends chains of F1 toggles to represent keys F1-F8. For example, "F1-on, F1-off, F1-on, pause" or "F1-off, F1-on, F1-off, pause" would trigger function 3.

But if we are turning F-keys into toggles of FO1, how does that interact with other F-keys? Especially since F1-F8 are modelled to the cab as "states" but processed out FO1 as events / actions.

For example, consider this sequence sent from the cab, with sufficient pauses in between:
  1. F1=on
  2. F2=on
  3. F1=off
  4. F2=off
  5. F1=on
  6. F3=on

What is the behaviour of FO1 under this sequence (assume it starts off)?
  • Does it simply toggle some number of times every time any F command is received?
  • What is the specified toggle period?
  • What happens if another F command is received while the first is still executing?
  • What happens if a redundant F command is received (e.g. F2=on when F2 is already 'on')?
  • Do the "states" of the various F command flags interact with each other? Are they even tracked as states?
    • Obviously, the state of FO1 needs to be tracked, since it toggles "on" and "off". But are the states of F1..F8 tracked?

I haven't found a specification of this behaviour anywhere, but maybe my Google-Fu is weak.
 
Post #14
CV 7 = 155, CV 8 = 123, CV 255 = 130
Firmware version 1.5 is dated from 02.07.08
Serial Pulse Mode was not introduced until Firmware version 2.4 dated 20.12.10

Referring to the LGB manual prior to 2014, when Massoth no longer issued "P" updates, some seven years ago, and unless the equipment used can access the original LGB Serial commands, seems a fairly forlorn hope of success.
 
Last edited:
The Shed said:
CV 7 = 155, CV 8 = 123, CV 255 = 130
Firmware version 1.5 is dated from 02.07.08
Serial Pulse Mode was not introduced until Firmware version 2.4 dated 20.12.10
Click to expand...
Ok, that totally explains the Massoth not working. Is it possible to update the firmware?

The Shed said:
CV 7=28 CV8=123 CV 106=192
Looking at the LGB 55021 interface, I found the serial / parallel mode setting.
On the 55021, CV 51=1 (default) enables a mode where F1-F8 trigger the "F1" output that many times.


Tested with both PIKO and Massoth systems, does exactly what it says on the tin.
LGB 55021
CV 51
= 9 F1 Key serial
10 F2
11 F3
12 F4
13 F5
14 F6
15 F7
= 16 F8 Key serial
Click to expand...
Sorry, but I don't understand what you're getting at here. Setting CV51 to 9..16 causes the F1 output to be driven normally by one of F1 to F8. Have tested it, it works, but not what I'm wanting.

The mode I think I'm looking for is CV51=1, which claims to drive F1 out by F1. And by F2..F8. Wording is slightly different in 55020 vs 55021 manual but both call out that the output is driven by "button 1" and "buttons 2-8". And I haven't been able to get it working.

However, I obtained this 55021 by grabbing it off a 90200.1 (70255 digital starter kit diesel), so it's possible that this function isn't enabled on it.
 
Checked and rechecked the notes you referenced with a PIKO Navigator with Serial Pulse activated, It appears your equipment is unable to use Serial Pulse Mode.

As far as the Massoth L decoder goes with reference to upgrading,
a) would need to invest in a Massoth Programmer.
b) if the Firmware Upgrades are still in existence.
 
Update: I spoke with Massoth tech support, and they've been quite helpful.
  • Firmware is version 2.2, which supports P-update. CV7=155 seems to be a mistake on my part.
  • The update mentioned above is for SUSI. A1 fast pulse behaviour has been in for a long while
Tested by hooking one of the headlights up to a jumper off A1. The light does indeed pulse. However, it pulses very rapidly - fast enough to see it turn on & off, but faster than I can count. This is way quicker than the quarter second or so required by the 28002 interface. I don't know if this is a failing in the 28002 interface or not.

I plan on running the same test with the 52021, just to check.

To use my existing equipment, I'll need to slow down the pulse rate on the eMotion L, and I don't know if I can do that. Otherwise, I might just need to pull off all the electronics and replace them with a Zimo or Massoth loco decoder with sound. This is somewhat annoying as I'll need to re-jig a whole bunch of wiring, but I don't see another option.
 
HI Andrew, I'm confused why you are so keen on retaining MTS 1/early MTS 2 era serial operation. Is it just to operate a couple of outdated loco decoders? If so, seems kind of silly to me. Just replace them. Decoders are not that expensive. If it's for the learning experience, then it kind of makes sense. But it's an early, long outdated, digital system in any case. Just curious.
 
phils2um said:
HI Andrew, I'm confused why you are so keen on retaining MTS 1/early MTS 2 era serial operation. Is it just to operate a couple of outdated loco decoders? If so, seems kind of silly to me. Just replace them. Decoders are not that expensive. If it's for the learning experience, then it kind of makes sense. But it's an early, long outdated, digital system in any case. Just curious.
Click to expand...
That's an easy one to answer.
  • I'm a software guy at heart. I'm fine with modifying software, or even popping out one component to put in a new one. Replacing all the existing electronics means I need to take responsibility for getting it all wired up (and probably soldered up) correctly, and also dealing with issues like keep-alive circuitry for the sound system and voltage levels for the lights and smoke generator. I prefer my hardware to "just work" and deal with the issues in software.
  • I also cringe at the thought of throwing away not-broken existing circuitry, even if it is almost 20 years old.
  • I also feel isolated out here in Australia. I'd feel a lot more comfortable with someone to walk me through the process rather than try to figure it out myself.
I do have some friends who are involved in HO DCC nearby. I suspect if I asked around I'd find someone to walk me through replacing the whole decoder kit - it can't be that much different between doing it for HO and for G.
 
Like you mention, the "Fast Pulse String" provided by the decoder is too fast for the Sound board to process. This cannot be slowed and it was why Massoth originally provided "P Updates" to locos with sound boards that I mentioned back in post 15.
 
muns said:
Like you mention, the "Fast Pulse String" provided by the decoder is too fast for the Sound board to process. This cannot be slowed and it was why Massoth originally provided "P Updates" to locos with sound boards that I mentioned back in post 15.
Click to expand...
Ah, that makes more sense to me now.

I can't find anything online about updating the firmware in the 28002. It might not exist, or be lost to the depths of time. I'll try the same trick with a 55021 and see if it behaves the same way.

If all else fails, I've now found a Zimo distributor here in Sydney, so I might end up just stripping the existing electronics in-toto and replacing them.
 
Finally got around to testing the 55021 from the LGB starter set diesel.

Set CV#51=1. Triggers F1 as normal, but nothing special from other F keys.

DecoderPro lists "buttons 1-8 with fast pulse string (for 55021 with P-Update only)". I try this and it sets CV#51=33, which is undocumented in the decoder manual. I now try F2, F4, etc and they pulse rapidly, exactly like fast pulse mode on the eMotion L.

So I try some of the other options.
  • CV#51=10 is documented as "control via button 2". CV#51=9 is documented as "control via button 1. Buttons 2-8 will be ignored". DecoderPro documents CV#52
    • with CV#51=10, I can double-toggle F1 to trigger output once. Single toggle F2 also works
    • with CV#51=9, double-toggle of F1 does nothing (presumably triggers F2 which is not programmed). Single toggle works.
  • CV#51=2 is undocumented. CV#51=1 is documented as "Control via button 1 with buttons 2-8, the command will be sent the respective number of times."
    • with CV#51=2, F2 causes a single whistle, and sometimes a very brief light blink. Slow double-toggle activates the bell. F1 does nothing.
    • with CV#51=1, F1 causes a single whistle. Slow double-toggle activates the bell. F2 does nothing.
(And I think I've burned out my light by putting a full 18V across it - ugh)

Despite the documentation, it seems that CV#51=1..8 is simple parallel control for output, CV#51=9..16 is parallel+serial control, and CV#51=33 is fast pulse. DecoderPro has it right, and the 55021 manual has it wrong.

But still got the problem that "fast pulse" is way too fast for the 28002 decoder. Not sure how they managed to mess that one up, since these features should all have come out within a couple of years of each other.

FYI: CV#7=28, CV#8=123
 
Last edited:
Andrew_au said:
That's an easy one to answer.
Click to expand...
HI again Andrew,

OK, I can see your desire to not have to throw everything out and especially the attendant re-wiring needed with completely new sound decoders. But, how about just getting rid of the old serial sound board? You can replace serial only sound boards with new parallel eMotion S sound decoders and keep the driving decoders as they are. Only requires two wires for power from the track and a pair of wires to the speaker (the speaker plug is probably the same too.) Of course, you need to have the correct sound project loaded. There are other manufacturers of sound only decoder boards out there too. In fact, you could probably use a smaller scale sound decoder and just ignore hooking up anything but the sound if you want to do it on the cheap. Again, only requires two wires for track power and a couple to the speaker.
 
Last edited:
Hi Phil,

As far as I can tell for the 28002, the sound board is built into the mainboard. Of course, I could just unplug the speaker from the mainboard and connect it to a sound decoder or sound/loco decoder. The other locos don't have sound built in, so that's not a concern.

Would be really convenient if I could get an adaptor between 3rd party decoders and the old 55020 pin interface. As far as I can tell, the only decoder (or adaptor) on the market that fits that interface is the Massoth eMotion L, and that's both long in the tooth and doesn't do sound. Also, I don't know if the mainboards are providing anything other than power regulation and pin outs for the lights / smoke / built-in sound module (where fitted).

I'm also keen to (over time) replace the ancient 55021 and slightly less ancient Massoth eMotion L with something more modern, like a Zimo MX or MS series. If I can't find an adaptor, one option is to keep the existing mainboards and wire up some crimp sockets to connect to the existing pin-outs. A thin insulating layer and some zip ties (or similar) should keep the decoder in place. And either run wires directly to the existing speaker (bypassing the existing sound electronics) or add a speaker and run patch wires.

The alternative is to completely strip the stock electronics. This frees up a bit of space, but I also need to re-do all the wiring connections and figure how to mount the decoder to the loco frame.
 
You must log in or register to reply here.
Back
Top Bottom