Setup is a DR4018 connected to a Pi-Sprog running JMRI 4.16. Point motor is Fleischmann.

I have to send each command to throw or close a switch twice for the DR4018 to respond. The red light illuminates on both commands but the motor only switches on the second.

From what I can see with a meter across the output, nothing happens after the first command, I only see a voltage change after the second one, so it's not the point motor lagging.

Interestingly, the first command is "remembered" by the DR4018 if I send other commands... so for instance to throw switch 56 I have to press "throw" twice. If I "throw" 56 once, then "throw" 55 or any other outputs, then "throw" 56 again it responds.

Pretty much a noob with DCC, I think I've managed to factory reset it once and reprogram, but not really played around with setting or reading CVs otherwise.?

Interesting, I've had trouble operating signals from DR4018 decoders.
They often miss commands that are sent quickly one after another. I have JMRI send multiple and delayed commands. Making sure there was a delay between commands improved the response, repeating the commands is far better but not 100%.
I haven't had problems sending a single command. One DR4018 operates switches and signals. The switches always work, after they have moved the signals are set where maybe 6 commands are sent then it may fail.

Mike Ruby

Thanks Mike, that's interesting/. How did you go about setting up JMRI to send multiple and delayed commands? Is it in the configuration somewhere or does it require scripting?

Martin

My word, I think he's got it!

I had a closer look in the turnout tables and set up the turnout in there. I then went to the Automation menu and set NoFeedback - times to try 2, Interval 100 (ms presumably, the minimum). I then set the turnout automation from "Off" to "NoFeedback" and it now sends the command twice and the DR4018 works as expected.

I would like to know if the problem is in the Pi-Sprog setup or JMRI or all Command stations.

Have done a lot of testing of the DR4018 with no problems, using all Digikeijs products.

One think i still need to do is try to blast the DR4018 with a lot of data.

Have one DR4018 setup with three different setups.

This is a interesting topic and if there is an issue we need to documents and post it here.

Dennis
Owner of the Forum

I also raised this question on the Sprog group and there's some interesting information there -?/g/sprog-dcc/topic/pi_sprog_with_dr4018_only/34303883?p=,,,20,0,0,0::recentpostdate%2Fsticky,,,20,2,0,34303883

It seems that it is normal practice with train decoders to send commands several times because it's quite possible commands will be lost on the track. However the JMRI/Sprog setup is to only send accessory decoder commands once since the command should never be lost.

It could be that the Dijikeis command station (and possibly others) automatically send the command several times and therefore the problem isn't seen.

However this doesn't answer the question as to why the DR4018 doesn't react to the first command. Maybe it's saying "Are you sure?" :-)

I've raised this with support, I'll wait to see their response.

开云体育

Yes Martin you found how to do it.

As to systems used, I've tried Lenz and the DR5000 with the same results. I know JMRI with Lenz sends several off commands, but I think one on command.

I use the same delays for each DR4018 as the problem is just in each decoder. I set a different delay for each output, I think with repeat twice. If I send a single command from the JMRI turnout table it always works. The other thing with JMRI is I set Logix outputs to always send even if JMRI thinks it already in the correct state, then if it does miss a command a rerun of the Logix normally corrects it.

Iain do you operate all turnouts at the same time or one after another?

Mike

开云体育

That is why you don't have a problem, I find putting delays between commands makes the decoder work. The problem when operating signals is that too much delay can make the signals momentarily show the wrong aspect.

I did contact Digikeijs and they said not to send commands too fast, but as signals are set by a code to three addresses per signal head they need to accept at least three rapidly.

Mike

开云体育

开云体育

Just thinking outloud, so this may be total nonsense Mike, but if you have more than one DR4018 could you split the three addresses per signal across the two (or more) units and speed the commands up that way?

The decoders are on different boards and modules so not practical. I did do a similar thing in our storage yards with a different make of decoders to give them time to charge before firing the next motor. I use the same make for the yard signals and they have no problems I can change all outputs as fast as the commands go out, but they can't do the nice signal effects that the DR4018 can. Which is a thought, I wonder if they work better if I remove the dimming action. I'll try that at the next meeting.

Mike

It's quite likely that the decoder is tied up for the time it takes for the light to dim, and is ignoring further commands, so that might be worth trying.

Martin