Benny,
Some interesting challenges with this layout.
I was able to re-create the station scenario running from C to 1b. According to the route table, running through track 10 is the best route since the straight route has an additional hop and therefore a higher metric. By setting S35th to 500, it uses the straight route. If a block is occupied on the straight route it goes through track 10. Note: When an alternate route is used, each segment has to be cleared individually.
By making 40 -> U2 into 2 NX segments and deleting the one segment 40 -> U2 route, you no longer need to use metic values since there is only one route.
If you use tvsHs40 and tvsX21aa directly with an occupied block, a Stack dialog request is displayed. If you use tvvHs40 and tvvHsU2 somehow the occupied block problem is ignored or bypassed. If you use the green buttons, the NX process waits on the dialog response which never comes since the process was initiated by Logix.
Dave Sand
toggle quoted message
Show quoted text
On Sep 2, 2018, at 7:29 AM, Benny Tj?der <etxbct@...> wrote:
Hi,
I have changed the layout close to 38-U2, 40-U2 and named it Bilbox.xml in the zip file.
In that file I have added a sensor for the hidden signal X21aa and added EntryExit 38-X21aa, 40-X21aa and X21aa-U2. I also removed the existing 38-U2 and 40-U2 EntryExit routes.
Now it looks it's working from 40 out on U2 or 38 out on U2. 40 out on U2 uses turnout Vx25. If I put turnout Vx25 occupied and select 40-U2 EntryExit everything freezes. I need to kill it from task manager.
I added a picture JMRI-nx4.PNG which shows the area where Vx25 (Vx25a side) is occupied.
/Benny
Den s?n 2 sep. 2018 kl 12:05 skrev Benny Tj?der <etxbct@...>:
The station case is set if you make the invisible green button close to C visible. Than push C and next push the green button by the station on track 10. Or select a route on track 2 over turnout V29 and then select a route from A to Track 1 by using the green buttons.
The hidden signals was put there to allow the closest signals to show correct aspect (sh3) during local movments on the station when the turnouts can be local operated (not from the panel). Before adding the hidden signals the visible ones showed stop (sh0) if the turnout was in wrong direction. By adding hidden signals fixed the problem.
In the case 40-U2 I have put a higher metric on turnout Vx21 block S21ab located between 21a and 21b. Could that cause NX and SML to choose different route?
The higher metric is to use Vx25 more than Vx21.
If I set it back to default (200) and put higher value (2000) on block S25ab it looks better. SML and NX select same route.
But if block S21b is occupied I get same problem but SML do not set signal 40 to green since the block is occupied. Should it not select Vx25 in this case?
/Benny
Den 2 sep 2018 07:28 skrev "Dave Sand" <ds@...>:
Benny,
As I mentioned before, I am not able to re-create the ¡°station¡± scenario, but I do know what is happening with the other 2 scenarios.
As a general rule, Signal Mast Logic (SML) and Entry/Exit (NX) both make extensive use of layout block routing which is part of Layout Editor. This applies to SML when the ¡°User Layout Editor Paths¡± option is selected (also the default). NX always uses this mode.
Another general rule is that SML signal mast pairs match NX sensor pairs. They share the same anchor points, turnout connections, etc. This way the path from A to B includes the same turnouts and blocks for both SML and NX. This is important because NX uses the turnout list from SML.
Both 40 -> U2 and 31 -> 33 have a hidden signal mast. In both case, the NX route goes through the hidden mast to the destination sensor. SML created an alternate route that skips the hidden mast. Since the turnouts are based on the SML alternate route, the turnouts on the NX route are not set correctly.
The easy fix is to delete the alternate SML route which forces SML to follow the same path as NX.
Another solution is to add NX sensors to the block boundary that holds the hidden mast and create two NX pairs. Since the current setup is Logix driven, the Logix could be changed to combine two NX routes to make it look like one route. I have not tested this theory.
Dave Sand
On Sep 1, 2018, at 4:36 PM, Benny Tj?der <etxbct@...> wrote:
There is a memory field in the lower left wher you type BCT in the memory field below and push the sensor button above the memory field.
That makes the panel to be in local mode and the panel sensors can be operated.
/Benny
Den l?r 1 sep 2018 23:31Dave Sand <ds@...> skrev:
Benny,
Thanks for the added files. The panel xml file now loads without any errors.
? The 40 to U2 appears to be a NX problem.
? The 31 to 33 turnout issue applies to both NX and signal mast logic. This implies an issue with block routing.
? I am not able to re-create the ¡°station¡± problem because I cannot find a sensor pair that goes straight through with no intermediate sensors.
What is the trick for getting the sensor icons to work in local mode? I am currently manually removing the Disable from the icon.
The NX implementation is quite unique. Now that I can set and clear routes, I hope to identify the issues.
Dave Sand
On Sep 1, 2018, at 1:11 PM, Benny Tj?der <etxbct@...> wrote:
I have ziped it now.
/Benny
Den l?r 1 sep. 2018 kl 18:57 skrev Dave Sand <ds@...>:
Benny,
I am also going to need the custom components located in the resources folder, and maybe the jython folder. These are located in the same directory as the panel xml file. It works best if you zip the folders and their contents and upload the zip file.
Dave Sand
On Sep 1, 2018, at 11:41 AM, Benny Tj?der <etxbct@...> wrote:
The xml file and the picture are uploaded now.
/Benny
Den l?r 1 sep. 2018 kl 18:13 skrev Dave Sand <ds@...>:
Benny,
Attachments are removed from postings. You will need to upload your picture to the photos section. It would also be helpful if you could upload the panel xml file to the ProblemsBeingWorkedOn folder.
Dave Sand
On Sep 1, 2018, at 10:53 AM, Benny Tj?der <etxbct@...> wrote:
Hi,
I have downloaded the RPi-JMRI image and installed on a new SD card.
No problem to get it working. Downloaded my layout made in 4.12.
It started up and have connection to my railroad.
Raspberry Pi is running JMRI 4.13.2
The problem is if I select a route over two turnouts the turnouts are both thrown but the route shows thru both turnout in closed position, which should be the best route.
The route table shows thrown path as active, but closed path as reserved.
I attached a picture showing the route table and the layout with the route reserved.
Turnout 27 is two right handed turnouts (27a and 27b) faced each other.
Turnout 26 is similar as 27 but two left handed turnouts.
The route table is from turnout 27a (block 27a).
Block 27as is between signal 31 and turnout 27.
Block 27acd is the block after a closed turnout (towards turnout 26 in closed position).
Block 27ab is the block in between 27a and 27b turnout.
In turnout table both 27a, 27b, 26a and 26b are thrown.
The main path thru all these turnouts are set to closed (default).
I have seen other strange behavior in EntryExit routing. Sometime it selects a route from one track down and then one track up instead of going the shortest way on same track.
/BR
Benny
