Solved Need for fixing the accessory addressing fault

Solved Need for fixing the accessory addressing fault

Postby ron&bram » 07.07.2008, 22:54

After some testing with a) a serial port monitor program to send hex codes and b) the published Xpressnet protocol specification from Lenz I can confirm that what a lot of people have rpeorted in this forum that for the Elite the accessory address (in Rocrail definition) 1, port 1, does not exist.
Hence, the turnout connected to the first port of my first accessory decoder has to be adressed as address 1, port 2, the second as 1,3 the third as 1,4 and the fourth (now it bocomes irritating) 2,1. The first turnout on my second decoder is 2,2 and so on.
I find this very irritating. At first, the workaround looks very easy, but when problems arise over one year and your turnout, as reported on the screen as decoder 3, port 1, is not working, where do you start looking? Do you then remember the nasty fault in the Elite before or after you have thrown one perfect working decoder in the dustbin (having tested 15 times the wrong wiring hanging upside down at the back of your layout).

After achieving some success together with Jean-Michel in reading and writing cv's with the Elite, Jean-Michel has asked me to have a go at the addressing problem. I have written a fix and tested with normal addrressing, FADA and PADA, and my Elite it works.
I have sent the changed code per email to Jean-Michel and suggest that, based on the above arguments over the fault coming over a year, it is implemented, but since a) the lenz lib is Jean-Michels work and b) I am not the only Rocrail user with an Elite I invite other Elite users to comment on my proposal.

Regards to all from a slowly getting familiar with C++ Ronald
Last edited by ron&bram on 06.09.2009, 17:11, edited 1 time in total.
Best regards, Viele Grüsse, vriendelijke groeten, Ronald :rr_cap:
GCA85/50/93/145/146/153 Loconet, Twincenter+booster for locs, seperate booster for accessory decoders, dedicated laptop with XP and Ubuntu for Rocrail, AndRoc, LGB indoor OTF layout
ron&bram
 

Postby RAFHAAA96 » 08.07.2008, 08:11

Ron
We Elite users had similar offset problems in the early days - please see thread http://forum.rocrail.net/viewtopic.php? ... highlight=
This was mostly cured in the end by introduction of FADA for the Elite and Hornby Acc Decoders, although I found it easier to use turnout number *2 instead of the formula suggested to determine the interface values.
It worked OK for me until the laptop died and now I have to load up an old pc with RR and start all over again.
On the problem you have seen - that off RR not firing point motors hard enough. I have also thought that was a problem as they fired OK direct from Elite and also from a trial version of a voice activated control system, but not always reliably in RR.
Myself and I'm sure other Elite users appreciate the work you and J-M have done sorting out problems associated with the Elite. Its a pity Hornby hasnt given support by issuing details of the protocol they use.
I don't have the pc coding skills to carry out the investigations you two have done or I would be more pro-active.
Robbie
Regards Rob.
Halton Brat
OO gauge - Hornby Elite & eLink CS & Accessory Decoders - Win 8 - 64 bit.
RAFHAAA96
 

Postby ron&bram » 08.07.2008, 10:14

Hi Rob,

I hope that you had a good time in France and did not return as overheated as your laptop was before you went on holiday.

I am quite aware of the content of the thread you refer to, but since I was happely debugging the Elite anyway I decided to give it a thorough test and try to work out a fix for the reasons mentioned in my post and below.

The Elite situation is as follows.

After 5 years of modelrailroading silence (reason: see my avatar) I started again and after two days my Fleischann twincenter kindly blinked all its leds at me and said bye bye. So, after looking around a little bit I bought an Elite. The controls are not as sophisticated as those on the twincenter, so I had a go at Rocrail. This worked up to a certain level. I started posting here and through the replies came in contact with Jean-Michel, being the "father" of the lenz library (also the base for using the Elite with Rocrail), who supported me in finding all the bugs in the Elite's firmware (version 1.3).

For every investigation I do on the Elite I start of with the Lenz Xpressnet protocol specification. With a serial communication test program I send the command as defined in the Xpressnet specification as a series of hexadecimal numbers to the Elite and observe what is happening on the layout and what kind of response I get back from the Elite to the pc.
If the response on the layout or the answer from the Elite is not what it should be according to the Xpressnet specification I see it as a bug. Then I check with Jean-Michel how the lenz lib handles this situation (or I find out myself, everytime I learn more about how Jean-Michel composed the lenz lib and how C++ works) and try to develop an Elite fix for the lenz lib. These I send to Jean-Michel to incorporate or not in the sources on sourceforge and from there on into the snapshots.

At this moment the bug overview is:
- turnout operation reliable, Jean-Michel turned off the deactivation of the other gate on the same port when using an Elite.
- cv read/write on the pt working, the Elite does not send a broadcast message when entering service mode, fixed by adding special timing and checks when reading/writing cv's with an Elite.
- cv read/write on the main track: working, like above.
- accessory decoder offset fixed but not -yet- released. This point has been present in this forum for a long time and the general accepted work around is to use FADA and add 2. This works but I find this a wrong solution. WHen building a layout and documenting the wiring and so on, it should be that the turnout attached to port 1 on decoder 1 is also known in the software as address 1, port 1. The worst deviations occur at port 4, for the Elite you must make an address increase to get it working, so turnout on decoder 1 port 4 must with the Elite be addressed as address 2, port 1. When you build a layout you must take all precautions to make sure that any defects problems that may occur in the future can be traced easily and solved. One of the keys to that is to avoid exceptions.
- feedback when manually controlling a loco with the Elite. This is not handled by Rocrail, so no issue (yet?), but there is bug there also, when you look in the Xpressnet specification the message from a command station when controlling a loco should be starting with 0xE4, the one from the Elite starts with 0xE5 (which is related to double header).
- no broadcast message when pressing emergency stop. Can not be fixed in Rocrail, only in the Elite firmware.
More can follow, I have by no means finished reading the Lenz manual on Xpressnet protocol, but for the time being, everything I want to do with Rocrail and the Elite I can do, without having to make special provisions for using an Elite (so I can also switch from command station without having to correct al my decoders adresses).

In general, one could say that Hornby has done a half job implementing the Xpress protocol into their Elite (or maybe they will finally admit that Elite stands for Economic Lenz Impersonation, Too many Exceptions).

Regards,
Ronald
Best regards, Viele Grüsse, vriendelijke groeten, Ronald :rr_cap:
GCA85/50/93/145/146/153 Loconet, Twincenter+booster for locs, seperate booster for accessory decoders, dedicated laptop with XP and Ubuntu for Rocrail, AndRoc, LGB indoor OTF layout
ron&bram
 

To all Elite users

Postby ron&bram » 08.07.2008, 15:37

I also posted a list of deviations between the Elite's implementation and the Xpressnet standard on the Hornby DCC forum of Hornby's website and contacted Hornby's customer support on this matter. I hope that the technical guys at Hornby are as quick with fixing the flaws as their forum administrator was with closing the topic.
But maybe it is time for other Elite users also to get into contact with Hornby's customer support and complain about their Xpressnet implementation (Robbie?). When nobody complains, nothing changes.

Regards,
Ronald
Best regards, Viele Grüsse, vriendelijke groeten, Ronald :rr_cap:
GCA85/50/93/145/146/153 Loconet, Twincenter+booster for locs, seperate booster for accessory decoders, dedicated laptop with XP and Ubuntu for Rocrail, AndRoc, LGB indoor OTF layout
ron&bram
 

Postby rjversluis » 08.07.2008, 22:23

Hi Ronald,

you are a good P.I. in case of the Elite and Xpresnet.

You are right to call all other users to contact Hornby to push them to check their compatibility with XPressNet or spit out the diffs.

The Elite issue did caused Jean-Michel to get some long gray hears, but I will buy him some paint to repair it. ;-)

But let's not complane too hard on Hornby because they choose an existing protocol instead of inventing the wheel over again...

"Let's re-inventing the wheel because the old one is not round enough..."

This statement is found in internet...
Best Regards, Rob.
:!: PS: Do not forget to attach the usual files.
:!: PS: Nicht vergessen die übliche Dateien an zu hängen.
[ macOS - Linux] - [ N: CBus - CAN-GCA ] - [ 0: RocNetNode - GCA-Pi ]
rjversluis
Site Admin
 

Postby ron&bram » 08.07.2008, 23:14

Hi Rob,

If it will be of any comfort to Jean-Michel I will exchange my avatar from a picture of my son for a picture of me (I turned grey at the age of 25 and having now almost doubled that number exchanged grey for near white).
I'm glad my work is appreciated, even it is only a small contribution compared to the work done on Rocrail. I see it as a sort of challenge and good fun to iron out the bugs in the Elite.

Best regards,
Ronald

P.S. in three days there will be a period of radio silence on my side, being on holiday without internet access and out of reach for any form of communication from my work.
Best regards, Viele Grüsse, vriendelijke groeten, Ronald :rr_cap:
GCA85/50/93/145/146/153 Loconet, Twincenter+booster for locs, seperate booster for accessory decoders, dedicated laptop with XP and Ubuntu for Rocrail, AndRoc, LGB indoor OTF layout
ron&bram
 


Return to Hornby Elite