2018 Atlas Nscale SD35 ESU sound decoder JMRI Pane Selection

[conrad]

I posted this on jmriusers@groups.io and thought maybe someone here knows.

I have just placed an order for an Atlas 2018 issued N scale SD35 sound loco (ESU decoder) and need to know the installed decoder model. 

My DCC system consists of JMRI, Digitrax DB150 command station and Digitrax PR3 programmer. This setup does not reliably recognize/read/write sound decoders on the programming track and, by design, cannot read decoders on the mainline.

Thus to program on the main with JMRI I need to know the decoder model.  My research and guess is that it is ESU's ESU 73199 Select Direct Micro OEM sound decoder which I believe is listed in JMRI as "LokSound Select OEM" under ESU LokSound Select. Can anyone verify this?

Afterthoughts: One may ask why I would buy a loco (Conrail by the way) issued 2 years ago.  Well, I do the home finances and noticed that during the pandemic the wife (54 years of bliss) has been heavily into online shopping.  I feel she is really keeping the national economy going.  So, why not me!  Well for $150 I get my first ESU sound loco and learn to deal with ESU's 200 or 2,000 CV's. For the price it's not worth doing an installation and the loco has good reviews (see the posts back in 2018 and Spookshow).  I really like ESU decoders and have installed a half dozen LokPilot Nanos in the past.


Conrad

[conrad]

No need to reply.  I got the info direct from ESU!

"So far Atlas has only used LokSound Select decoders in their N scale locomotives.
This decoder is the same as the 73199. LokSound Select Direct Micro OEM.
Hope this helps!
Best Regards,
Matthew  Herman
General Manager
ESU LLC"

Now to work the JMRI file

Conrad

[RBrodzinsky]

As long as you have a current version of JMRI, the best method to get the exact decoder is to let DecoderPro identify it.  Dave Heap has done an excellent job of getting all the different ESU identifications into DecoderPro, and it will select the correct model automatically.  Do New Loco|Read Decoder

[conrad]

Rick,

"As long as you have a current version of JMRI, the best method to get the exact decoder is to let DecoderPro identify it.  Dave Heap has done an excellent job of getting all the different ESU identifications into DecoderPro, and it will select the correct model automatically.  Do New Loco|Read Decoder"

I think you missed my note that for sound decoders my system cannot recognize, read nor write on the programming track and for all decoders cannot recognize/read on the main. 

Conrad

[peteski]

I think you missed my note that for sound decoders my system cannot recognize, read nor write on the programming track and for all decoders cannot recognize/read on the main. 

Conrad

The fact that in standard DCC mainline programming is "write-only" is quite normal and expected.  Only the few Railcomm capable command stations and decoders are capable of CV read-back on mainline.

[conrad]

Gee Peteski, I always thought programming on the main was both read and write except with cheapo "command" stations like the Digi DB150 (aka really a booster).

Anyway, for using the programming track with sound decoders, the advice from jmrikusers@groups.io seems to be to increase the voltage supplying the PR3 to maybe 16 - 18vdc.  I'll have to check what my existing wall wart is putting out.

Conrad

[reinhardtjh]

Anyway, for using the programming track with sound decoders, the advice from jmrikusers@groups.io seems to be to increase the voltage supplying the PR3 to maybe 16 - 18vdc.  I'll have to check what my existing wall wart is putting out.

I see in the JMRI posting you state you're using a Digitrax wall wart.  That's probably the PS14 so it would put out about 14V.  Do you know if you have the original PR3 or the PR3 Xtra?  There may be no difference, but with the PR3 Digitrax originally specified up to 20VDC as the input, when the PR3 Xtra was released they lowered that to 15VDC.  They never specified if the internal components changed. The Xtra is the PR3 with the addition of a PS14 wall wart which may be the reason their specifications changed.  I have an original PR3 and I ended up switching to an 18VDC which helped on everything except BLI Paragon 2 decoders. I ended up eventually going to a SPROG 3 which then successfully programmed the BLI Paragon decoders.

Original PR3 Instructions: https://www.digitrax.com/media/apps/products/computer-control/pr3/documents/PR3_1.pdf
     PR3 Xtra Instructions: https://www.digitrax.com/media/apps/products/computer-control/pr3xtra/documents/PR3Xtra_4.pdf

[conrad]

I have the PR3 with the digi wall wart, PS14-03.  It's output is 14.9vdc, 300ma.  I should have investigated this 11 years ago when I first got sound decoders.  I have many spare wall warts so I'll look for a higher voltage one.  Wow, years wasted with programming sound decoders on the main.  It works but is a pain as I can only indirectly tell if a particular CV has changed by verifying with a bell or horn CV change and listening! 

Conrad

[RBrodzinsky]

I have absolutely no issues reading any ESU decoder with either my DCS50 (original Zephyr) or PR3.  I do need to use a programming track booster for older Tsunami decoders.

As stated, one can only read a decoder on the programming track, unless using Railcomm

[peteski]

I have the PR3 with the digi wall wart, PS14-03.  It's output is 14.9vdc, 300ma.  I should have investigated this 11 years ago when I first got sound decoders.  I have many spare wall warts so I'll look for a higher voltage one.  Wow, years wasted with programming sound decoders on the main.  It works but is a pain as I can only indirectly tell if a particular CV has changed by verifying with a bell or horn CV change and listening! 

Conrad

Just as an interesting tidbit of information, QSI sound decoders, when programmed on main, actually speak (in male voice) the CV and its new value. I always thought that was cool. Too bad QSI is not really a viable manufacturer (and no N size decoders).  They made excellent decoders (Revolution and Titan).

[conrad]

The following is my post on jmriusers@group.io.  Note, Dave Heap, and I quote, is "the developer who has done the special JMRI code and definitions for the ESU V4/Select/5 decoders and endeavour to keep them up-to-date."

Update on 2018 Atlas Conrail SD35 Sound Decoder Pane Selection

Just received the loco.  It’s a beauty.  Very detailed in comparison  to my last one purchased over 10 years ago.   Lots of progress in manufacturing and, of course, sound.

Using my original setup (Digi PR3, Digi PR14-03 p/s, and JMRI 4.20) I began the loco setup.   Taking Dave Heap’s advice, I used JMRI’s New Loco→Read Type to ident the loco’s decoder.  It worked: ESU LokSound Select Direct Micro OEM.  I then opened the Programmer and entered/saved the Roster Entry.  Again, per Dave Heap, I then read the full decoder, being careful to only use the “Read Full Sheet” on each CV pane instead of “Read All Sheets”.  Must be hundreds of CV’s as it took forever, especially during the Function Map pane read.  Luckily, I only have to do this once to get to “factory defaults”.

After the full read I then was ready to make changes.  My first, and so far only changes, are to drop the two digit 03 address, enter the loco’s four digit road number, 6020, and enter my user ID.   I followed Dave’s instruction and entered the changes on the Basic sheet but did the actual CV changes on the CV pane, writing each CV individually.   All of this to avoid the “dreaded?” rare JMRI condition.

One concern with making changes on the basic sheet but actually changing the CV on the CV pane is the indexed CV’s.  The LokSound Select Direct Micro OEM user guide states “Make sure that Index CV 31 is set to 16 and Index CV 32 is set to 1 before changing a volume CV”.  However, in JMRI, when changing the sound level value in the Sound pane only the actual CV is changed but not the index  registers CV 31 & 32 (as far as I can tell).  This could be a problem if you forget.  Additionally,  the user guide states “CV 32=0 if accessing CV’s 1- 255, CV 31 (sic, appears to be a typo) =1,2,3 if accessing CV’s 257-511”   I don’t know when to use 1, 2, or 3.  Will have to read the comprehensive manual.

A word about my inability to read my older sound decoders (MRC, Tsunami):  I did get an 18vdc wall wart  for the PR3.   It works for my older sound decoders although sometimes CV’s come up as unread.  I noticed that JMRI does do a second and maybe third try.    In addition,  I discovered that my other “test” layout, run with the Arduino  based JMRI DCC++, will read and write the sound decoders.  It’s a little flaky at times but generally works.   If I get an unread CV I just try again and usually it’s good.

Finally, I see that ESU decoders are complex.  My plan is to learn with this OEM decoder loco then retrofit my old locos with appropriate ESU sound decoders.

Conrad

[reinhardtjh]

JMRI automatically handles the indexing for you.  That's CV31, CV32 and so on.  You don't have to worry about it.  That's part of the magic that it does.  And yes, reading every CV on an ESU decoder (Both sound and non-sound) takes a long time.  Before the issue with the "Read all sheets" you owuld start it and go off for between 30 minutes to an hour.  V4 decoders were a little quicker due to fewer CV's to read.  With the V5 sound decoders there are a possibility of nearly 2,000 CV's to read.