I finally installed JMRI and got it working, fired up decoderpro 3 to try my first simple programming task - changing loco from default address of 003 to cab number (3567). Stuck the loco on the programming track, waited forever for decoderpro to read everything (won't do that again) changed from 2-digit to 4-digit, changed address from 003 to 3567, wrote it out to the decoder. Now I can't control the loco. Sound starts, but headlight doesn't come on. Doesn't respond to the throttle. Worked fine before on address 003. Put it back on the programming track and read back the address - it says it's 3567, just like I told it.
I'm using Digitrax, DT402 throttle, ESU Loksound decoder.
Any ideas what I screwed up?
I have the right to remain silent. By posting here I have given up that right and accept that anything I say can and will be used as evidence to critique me.
Carl,
I assume you have JMRI/Decoder Pro completely installed, Right? You fire up Decoder Pro 3, and get the 'Roster Menu'. You should be able to turn track power on/off clicking the power button on the top of the screen?
Put your engine on the Program Track, and set the configuration in the lower left of the screen to Program Track.
Click on New Loco and let Decoder Pro read the values in the engine. It sound like you select 4 digit addressing, and the cab # - Did you let Decoder Pro write out the entire set of CV's?
What brand of engine is this? I suspect the engine will still answer to address 3; if so, the 4 digit bit did not get set. Try re-writing the CV's to the loco again.
Jim
Modeling BNSF and Milwaukee Road in SW Wisconsin
You're using a Digitrax what - DCS100, DCS50 or 51, DB150, PR3, or ???
Some of the command station programming outputs don't provide enough "juice" for power-hungry sound decoders, but we can't begin to offer any suggestions until we know what you're using.
Command station model and which computer to LocoNet interface you're using, please.
Also, which version of JMRI? For ESU decoders, it should be AT LEAST 3.8, and the latest Test release (3.9.7) would be better still.
JMRI is 3.8, command station is DCS200 connected with Locobuffer USB. Loco is an Intermountain HO SD40-2.
Tried a couple more experiments. From the "basic" page in programmer, write address 1111 - reads back as 1111, write address 3567 - reads back as 3561 or 3551.
Try a non-sound decoder and see what results you get.
The DCS200 is one that is known to have a programming output that is too weak for many of the power-hungry sound decoders.
Your other options are to get a programming track booster, or get a PR3 and use it's stand-alone programming outputs (the PR3 can also replace your LocoBuffer).
I personally went the PR3 route - My PR3 with an 18 VDC regulated power supply has no problems reading from or writing to any sound decoder I've tried so far: ESU, QSI, Soundtraxx (both LC and Tsunami), BLI Blueline, and of course Digitrax...
When you changed the address, did you also change the value in CV 29? If not, the decoder does not recognize the long address. I set CV 29 to a value of 34 to also turn off DC operation.
LokSound decoders require you to press the number 8 on your controller to turn on the sound.
Which version of JMRI are you using? You need to be very careful with Loksound decoders, support in JMRI has only been very recent.
Do not do read all sheets, there are THOUSANDS of CVs in a Loksound Select/V4, and even with the proper version of JMRI, it will take a half hour or more to read the entire decoder. That's assuming you are using Direct mode for programming. If you use Paged, it will probably take hours. And you can easily get misreads or miswrites if there is a slight spec of dust on the program track and it misses one ack pulse but continues on just fine after that.
Program track boosters are NOT needed on any ssyestem to program Loksound.
It's all so that the complete flexibility to make any F key control any function wire and/or play any sound slot is available for any DCC system (thus lots of CVs to cover every possible combination) instead of it being accessible only to the proprietary programming device. However - with the Lokprogrammer interface and software, it only takes a few minutes to read or write the entire decoder, because it uses a proprietary non-NMRA DCC method of reading and writing the decoder memory.
Also, now you know why I rarely use JMRI. All I ever set is the address, and adjust volumes. Nothing against JMRI, but there is just too much going on under the covers and it's just faster to dial this up on a DT402. And since I only use Loksound for my sound decoders, I invested in the Lokprogrammer so if I need to do anything more complex, I can do it through there. My non-sound decoders, literally all I do is set the address, I don't have any flashing or blinking lights to configure, and my era is long before ditch lights, so really all there is is the address. And it's always the cab number, so the benefits of having a JMRI 'database' recording all the settings is also wasted on me. Even if the computer is already on, it takes longer to get to the part to set an address than it does to just set the address with my DT402.
--Randy
Modeling the Reading Railroad in the 1950's
Visit my web site at www.readingeastpenn.com for construction updates, DCC Info, and more.
rrinker Also, now you know why I rarely use JMRI. All I ever set is the address, and adjust volumes. Nothing against JMRI, but there is just too much going on under the covers and it's just faster to dial this up on a DT402. I don't have any flashing or blinking lights to configure, and my era is long before ditch lights, so really all there is is the address. And it's always the cab number, so the benefits of having a JMRI 'database' recording all the settings is also wasted on me.
Also, now you know why I rarely use JMRI. All I ever set is the address, and adjust volumes. Nothing against JMRI, but there is just too much going on under the covers and it's just faster to dial this up on a DT402.
I don't have any flashing or blinking lights to configure, and my era is long before ditch lights, so really all there is is the address. And it's always the cab number, so the benefits of having a JMRI 'database' recording all the settings is also wasted on me.
Not long ago, I installed Decoder Pro and filled out the roster for my 50 or so locos. Took me untold hours. And, I haven't used it since. It is just easier to program CVs without Decoder Pro.
Rich
Alton Junction
Now for CTC panels... there is at least one other (free) option I would like to use, but it only has European style signals and controls, and I'm just not ready to invest the time to develop US&S style controls to add in to it, so JMRI is my go-to for signalling and control. It doesn't help that as a programmer I absolutely despise the Java language and the huge limitations the "run anywhere" concept brings in to the picture.
The only non-sound decoders I have are in some Atlas SD35's that I picked up on Ebay. Tried 2 of them and they don't respond.
Further experimentation with the SD40-2, I was able to write it back to 3 and control it as before. A rewrite to 3567 and read back reports a 5 digit address.
What if you program the address using the DT204 instead of JMRI? Does it work properly then? All JMRI really is doing is emulating a throttle and pressing the 'buttons' faster than you can, so if it doesn't work from a throttle, it won't work with JMRI.
It behaves the same way when programming is attempted from the DT402.
Just tried something else - it seems that I can successfully program to any short address.
So setting the address to 3567 with the DT402, then reading back, you get either 3551 or 3561? Always the same, or one of the two?
This translates into a bit error in CV18 (don't ask how I figured this out, but I can explain it if you like), Bits 2, 3, and 5 are failing to read consistently.
First thing to try, unless you have a lot of saved routes and consists in the DCS200, is do a CV39 reset of it. Resetting the command station is usually a good idea when wierd things are happening, it IS a computer and sometimes the brains can get scrambled.
Second thing, is do a reset of the decoder. Set CV8=8. Then try the address programming again.
If it still keeps reading back an address other than what you programmed it to, it's probably a defect in the decoder, but it could be a command station issue which is why you need to find somethign that will read besides this IM loco.
You can also try setting the loco to 3552. The same bit defect could result in it reading as 3520 or 3558.
Randy's suggestions are good suggestions, but the FIRST thing I would do is clean the program track and see if that makes any difference.
carl425 ...changed from 2-digit to 4-digit, changed address from 003 to 3567, wrote it out to the decoder. Now I can't control the loco. Sound starts, but headlight doesn't come on. Doesn't respond to the throttle. Worked fine before on address 003. Put it back on the programming track and read back the address - it says it's 3567, just like I told it...
One thing you do not mention here is whether or no the loco would still respond to address '3'. If so, then it is just a matter of CV29 not getting set to enable the extended ("4-digit") address.
This is your problem right here. The DCS100 (and DCS200 in your case) do not have enough oomph in there programming tracks to read/write sound decoders.
The lack of power leads to packet loss when programming the decoder which is why you are getting wierd numbers when you read back the CV and hence the lack of control.
Contrary to what RRinker states a programming booster is also necessary for LokSound decoders for the setup you are working with. Unless you use their proprietary programmer you will need a booster or PR3 to program those decoders.
I have a very similar setup to you (DCS100 with Locobuffer USB) and had the same issues until I got a PowerPax Programming Booster. The newer ones have an on-off switch on them. For non sound I turn it off and it programs the units just fine.
There are some cautions to take into account when using a programming booster. I always test locos first with the booster off. this will tell you if the decoder is wired correctly without causing damage is something is wrong. I then turn the booster on to program it after the verification test is complete.
It doesn't really matter what brand of Sound decoder you use with the setup you have. It has always bothered me that the DCS series of boosters have too low of a setting for their programming track.
I hope this helps.
Colorado Front Range Railroad: http://www.coloradofrontrangerr.com/
I thought it strange that I could set the address to 1111 and it would work, so I decided to see how high of an address would work. 2222 worked, 3333 worked so I tried 3567 again and this time it worked. I can address and operate the loco as 3567 with no problem.
So... if the definition of insanity is trying the same thing over and over expecting different results, you must be insane to run DCC.
Tried one of the Atlas SD35's and it worked first time. Tried another and it looks dead - JMRI reports that it failed to respond.
Thanks all for the responses.
Renegade1c...It has always bothered me that the DCS series of boosters have too low of a setting for their programming track...
Well, the DCS series is designed to NMRA standards. What bothers me is the fact that many sound decoders require more current than the NMRA DCC standard limits service mode programming to.
carl425 I thought it strange that I could set the address to 1111 and it would work, so I decided to see how high of an address would work. 2222 worked, 3333 worked so I tried 3567 again and this time it worked. I can address and operate the loco as 3567 with no problem. So... if the definition of insanity is trying the same thing over and over expecting different results, you must be insane to run DCC. Tried one of the Atlas SD35's and it worked first time. Tried another and it looks dead - JMRI reports that it failed to respond. Thanks all for the responses.
Again, I would try cleaing the programming track (and the loco wheels). Anytime I have issues on the porgramming track, that's one of the first things I do and it will often clear things up.
davidmbedardLearn how to manually set address by writing cv 17 then cv 18 then cv 29
Yeah, did that. Just because I know how doesn't mean I want to do it - like mowing the lawn. I'd rather just let the software do it for me.
Spoke too soon on the SD35. It took what I gave it and read it back to me but won't run.
CSX RobertAgain, I would try cleaing the programming track (and the loco wheels). Anytime I have issues on the porgramming track, that's one of the first things I do and it will often clear things up.
Sorry for not responding the first time. I try to do all the basic stuff before asking questions. Cleaning the track was near the top of the list.
LOKSOUND DOES NOT NEED A BOOSTER ON DIGITRAX
Do I need to post a video of programming a Loksound loco, or even my Atlas QSI loco, on the program track, with no booster (and I'll do an OpSw 39 reset first so you can see I am not using Blast Mode or any other such nonsense) before anyone will believe it? Loksound very specifically shuts down the amplifier portion of their decoder when on the program track so it needs no more power than any simple motor decoder. QSI is questionable, but mine happens to work.
There, in caps so it's clear. Loksounds don't need extra current on the program track to work.
If the readback of CV17 and 18 had the actual address being assigned, I'd be inclined to believe that it is just a CV29 issue - however the fact that it reads an address other than that being set, and the difference is compeltely accounted for in CV18, tells me there is some problem reading this CV, either the decoder is giving the wrong response, or the command station is reading it wrong. Yes, dirty track might cause an issue like this, but it would be extremely unlikely that the missed read would be one and only one CV - and the same CV every single time. Not to mention it also causing problems writing with no acks.
Do try OPs Mode, if the loco runs on 3, use Ops Mode and set CV17 to 205 and CV18 to 239. This equates to your cab number. Set CV29 to 34 (or 38 if you want DC enabled). See if it works on that address now. If not, try the address it reads back as - the 3561 and 3551.
Interesting that 1111 seems to work OK. some of the same failed bits that would cause the misreads of 3567 would also cause 1111 to read a different addresses as well, which would make it seem to not be the loco. But corrupted slot info in the command station could cause wierd things around a specific address, which is why I suggested an OpSw 39 reset.
rrinkerLoksound very specifically shuts down the amplifier portion of their decoder when on the program track
That didn't seem to be the case here. It responded to programming commands with engine sounds for both read and write.
Well that's weird, my Bowsers do not make any sound effects on the program track. Just the usual motor jerk like most any decoder. My older (Loksound 3.5) steam locos also do not make sounds on the program track.
rrinker LOKSOUND DOES NOT NEED A BOOSTER ON DIGITRAX Do I need to post a video of programming a Loksound loco, or even my Atlas QSI loco, on the program track, with no booster (and I'll do an OpSw 39 reset first so you can see I am not using Blast Mode or any other such nonsense) before anyone will believe it? Loksound very specifically shuts down the amplifier portion of their decoder when on the program track so it needs no more power than any simple motor decoder. QSI is questionable, but mine happens to work. There, in caps so it's clear. Loksounds don't need extra current on the program track to work.
The DCS 100/200's do not have a Blast mode for programming. This was feature that was added to the Zephyr line. You must be using a DSC 50 or 51 (Zephyr or Zepyr Xtra). The DCS 100/200 is different circuitry than the zephyrs. Digitrax realized at some point that the DCS100/200 had issues programming sound decoders and fixed it in the Zephyr series. Unless they have fixed it recently, the issue still exists in the DCS100/200. I haven't upgraded my command station since it was purchased in 1995 so that could also be the difference.
Please note the OP is using a DCS200. They are different animals. You are correct that a Zephyr will program a loksound decoder without a booster but that is not true with a DCS 100/200.
We have a DCS 50 at the club to program locomotives and it doesn't need booster but the DCS200 command station on the layout requires a booster to program the same locomotive.
The programming circuitry on the Zephyr's is a vast improvement over the DCS100/200 series.
I have a set of PCM PA's which I had nothing but trouble programming (they have LokSound decoders) with a DCS100 and a Locobuffer USB. Once I added the Powerpax unit, I had zero issues. Same goes for soundtraxx and QSI.
I will agree to disagree with you on this one.
After some more research, I ordered a PTB-100. I'll report back when it arrives and I get it installed.
Thanks for the help.
Well, the PTB-100 finally showed up last week in the mail and I got it installed yesterday. Based on my testing so far with a Loksound select that I could read but only sometimes write and a Tsunami that I could neither read nor write, the PTB-100 fixed my problem. I am able to consistently read and write to both of these decoders without issue.
So if you want to accept my 2 decoder sample as statistically significant, this seems to validate the suggestion that the DCS200 programming track output is too weak.
carl425 Well, the PTB-100 finally showed up last week in the mail and I got it installed yesterday. Based on my testing so far with a Loksound select that I could read but only sometimes write and a Tsunami that I could neither read nor write, the PTB-100 fixed my problem. I am able to consistently read and write to both of these decoders without issue. So if you want to accept my 2 decoder sample as statistically significant, this seems to validate the suggestion that the DCS200 programming track output is too weak.