Welcome to the forum   
Welcome Guest! To enable all features please Login or Register.

Notification

Icon
Error

2 Pages<12
Share
Options
View
Go to last post in this topic Go to first unread post in this topic
Offline JohnjeanB  
#51 Posted : 26 May 2020 13:56:35(UTC)
JohnjeanB

France   
Joined: 04/02/2011(UTC)
Posts: 1,115
Location: Paris, France
Hi
Quote KiwiAlan:
I don't believe you need three sensors on the bridge itself. I think the it should be possible to use the radial tracks of the turntable as the entry and exit sensors (positions 1 and 3) with the rotating turntable being the central sensor ( sensor 2). After all a loco will (should) not be entering the turntable at more than 'idle' speed so once whichever radial sensor is no longer energised by the loco then speed step 0 should be commanded, and whatever momentum the loco has will be enough to stop it in the middle of the rotating section. This may require some slight fine tuning if time delays in software for different sizes of loco with different amounts of momentum set, but for the most part should be correct.
End quote
If the goal is to handle small locos (15 cm) on the bridge yes one sensor is OK but with very long ones (29cm) 3 sensors are very much needed.

So your idea of using the bridge rail side contact to have the detection on each connecting track is possible by doing the following:
- each connecting rail would need its two extremities connected to other connecting rails and to S88 entries (all right rails to one -Bridge sensor 1 and all left rails to the other-Bridge sensor 3)
- the middle section of the bridge being connected by the existing rotating contacts - Bridge sensor 2
- of course each access track would need also a regular occupancy detection
- to cut the extremity contact of the bridge regular rail so it does not trigger a false occupancy signal
- to prevent the small section of rail used to "deport" the bridge contact from being shorted by passing axles (which would otherwise create "out-of-sequence detections). One easy way to do this is to cover the access track small detection section with clear cote (transparent enamel)

By doing all this it becomes feasible I think.
Cheers
Jean
My lay-out videos
latest vid
humping yard
Offline mcs51  
#52 Posted : 28 May 2020 11:19:05(UTC)
mcs51

Belgium   
Joined: 25/05/2020(UTC)
Posts: 5
Location: West-Vlaanderen, Kortrijk
Dear,
Thank you for the answers !
I have a TT-DEC decoder and I will stick to that, no DSD 2010 for me thank you Jean.

For the moment I’m looking at Per his solution.
Cut the detection rail into 3 parts and with diodes bring 2 sensors over 1 wire.
But these 2 sensors are not what Rocail needs.
Rocrail needs “Enter” and “In” sensors.

One needs to process these 2 inputs further.
I’m thinking of following:

Call the detection rails as shown in Per pictures D1 & D2 (left or right doesn’t care)
On positive edge of D1 Flag 1 is set
On positive edge of D2 Flag 2 is set

The “Enter” sensor of Rocrail is then: Flag 1 OR Flag 2

The “In” sensor of Rocrail is then given with following equation:
(Flag 1 AND positive edge of D2) OR (Flag 2 AND positive edge of D1)

Flags should be rest with following equation:
(Flag 1 AND negative edge of D2) OR (Flag 2 AND negative edge of D1)

With this method the locomotive will always runs toward the furthest section opposite to entrance side.

I will try this logic with gates or use an Arduino Nano for this… … … any other suggestions?

So far I tested Per setup with 3 short rails and in Rocrail both sensors K1 (D1) and K2 (D2) always came up together whatever side I put some wheels on ?!!!!!!!!!?

I have to say that I don’t use relays but an optocoupler to bring the sensors into an Arduino Mega DCC++ BaseStation.
The diodes “next to the turntable” in Per drawings are the input LED’s of the optocoupler.

I mentioned spikes on the signal after the diodes on the bridge, so I had to put a LP filter to remove them from the signal.
Now I have one or the other detection in the Arduino, no problem, but I need the logic to “translate” these 2 detections into the correct Rocrail signals.


Last question: How to dismantle the TT without breaking anything so that it can be reassembled correctly after modification?

Best regards
Alain
Offline JohnjeanB  
#53 Posted : 28 May 2020 11:59:24(UTC)
JohnjeanB

France   
Joined: 04/02/2011(UTC)
Posts: 1,115
Location: Paris, France
Hi Alain
Originally Posted by: mcs51 Go to Quoted Post


For the moment I’m looking at Per his solution.
Cut the detection rail into 3 parts and with diodes bring 2 sensors over 1 wire.

Not sure what you mean here
Originally Posted by: mcs51 Go to Quoted Post

But these 2 sensors are not what Rocail needs.
Rocrail needs “Enter” and “In” sensors.

With 3 sensors Rocrail deals with ENTER, PR2IN and IN.
Indeed you could deal with only two sensors (ENTER and IN) but still by cutting the detection rail in 2 places so that each has a length of approx 6 cm to the end of the bridge, to allow a loco to brake at the end of the bridge. In this disposition, the center detection part of the bridge would not be used.

To open the TT without breaking anything:
- remove the access rails or side walkways so that nothing will prevent the bridge to go up (4 on one side of bridge, 6 on the other side)
- remove the guard rails
- remove the center screw and remove the metal plates between the bridge rails
- pull up vertically to remove the bridge

When putting the bridge back I recommend using a toothpick inserted in the center hole of the bridge to guide it vertically down to the master axle (Königsstuhl as they say in Germany). This is key to void bending the contacts sideways and create problems

Cheers
Jean
My lay-out videos
latest vid
humping yard
Offline mcs51  
#54 Posted : 28 May 2020 18:52:11(UTC)
mcs51

Belgium   
Joined: 25/05/2020(UTC)
Posts: 5
Location: West-Vlaanderen, Kortrijk
Hello Jean,
Please check the schematic of Per in post #12 on page 1
Divided the detection rail of the TT bridge in 3 pieces, I think to leave 1 cm at each end.
That way the locs will always run toward that last cm.
Ok, it’s not in the middle as, I think, it was in reality but lots of other things aren’t ok too in model railway and we have to live with it.
I tested this setup with 3 pieces of rail first and have working detection in Rocrail.
BUT, like I said in previous post, these signals K1 and K2 from Per drawings aren’t correct for “Enter” & “In” sensors in Rocrail
So I ordered an Arduino nano to process these 2 signals so that Rocrail can work with them.
Thank you for the procedure to “open” the TT.
To be continued… … …
Amicalement
Alain



The resistor of the LP filter is not 1k5 but 1k2!
Offline JohnjeanB  
#55 Posted : 28 May 2020 22:21:00(UTC)
JohnjeanB

France   
Joined: 04/02/2011(UTC)
Posts: 1,115
Location: Paris, France
Hi Alain
One thing I know is 1 cm is way too little even if you have no inertia
IMO at least 5 cm would be better
It is up to you,
Cheers
Jean
My lay-out videos
latest vid
humping yard
thanks 1 user liked this useful post by JohnjeanB
Offline DaleSchultz  
#56 Posted : 28 May 2020 22:37:01(UTC)
DaleSchultz

United States   
Joined: 10/02/2006(UTC)
Posts: 3,470
I would use the short (1cm) length on the way in, if that is free and the track coming onto the TT is also free, then stop.
(Instead of waiting to get to the far side short piece.)

In fact, I would electrically connect the two short pieces together since you only have to know one edge in any operation.

The sequence of a loco travelling onto the TT could be:

Start with loco on feeder track. feeder track detection ON, TT detection off
loco moves onto TT #1 feeder track detection ON, TT detection ON
loco moves onto TT #2 feeder track detection off, TT detection ON
loco arrives in 'center' of TT. feeder track detection off, TT detection off
STOP loco

This approach lets one combine a single electrical connection on the bridge with the feeder tracks.

Dale
Intellibox + own software, K-Track
My current layout: https://cabin-layout.mixmox.com
Arrival and Departure signs: https://remotesign.mixmox.com
thanks 1 user liked this useful post by DaleSchultz
Offline mcs51  
#57 Posted : 10 June 2020 19:33:00(UTC)
mcs51

Belgium   
Joined: 25/05/2020(UTC)
Posts: 5
Location: West-Vlaanderen, Kortrijk
Hi there,
I tried something else to have the loc in the middle of the bridge Cool

Using an Arduino Nano I take the millis() at which:
a) The loc makes contact with the 75mm feeder track
b) The loc makes contact with the bridge track
c) The loc is fully off the 75mm feeder track

Using those time intervals, the length of the bridge and the length of the feeder track one can calculate the time needed to have the loc (almost) in the middle of the bridge.
I’ll review my math later on because now the loc always stops at the same location but not the middle of the bridge, see picture.

I did the test 5 times, always that same location!

An Arduino Nano output connected to the Arduino Mega running DCCpp code activates the “In” sensor of the TT to stop the loc.

But I have another problem, dragging the loc from the Turntable block where it is located to for example “blk-001”, upper left location, the turn table rotates to the correct location but the loc starts in the wrong direction, that is toward “blk-007”.


Any idea why??
Thanks a lor for some inputs!
Best regards
Alain
Offline DaleSchultz  
#58 Posted : 10 June 2020 21:39:33(UTC)
DaleSchultz

United States   
Joined: 10/02/2006(UTC)
Posts: 3,470
sounds like nice progress.

The real test is with different locos that have different lengths and different speeds!

I am weary of using a time calculation as it is not a positive feedback. If the loco does not travel at a constant speed, or gets stuck, or the layout loses power (Eg because of short elsewhere) then your time calculation will be inaccurate.

Dale
Intellibox + own software, K-Track
My current layout: https://cabin-layout.mixmox.com
Arrival and Departure signs: https://remotesign.mixmox.com
Offline mcs51  
#59 Posted : 11 June 2020 06:50:34(UTC)
mcs51

Belgium   
Joined: 25/05/2020(UTC)
Posts: 5
Location: West-Vlaanderen, Kortrijk
I’m not worry about loc speed & length because calculation takes care of it.
But the other things you name will indeed be a problem but also with any other detection method.
Just drive the loc manually back from the TT and restart the process.
Or drive the loc manually on the bridge and set the “In” sensor in Rocrail.
Users browsing this topic
OceanSpiders 2.0
2 Pages<12
Forum Jump  
You cannot post new topics in this forum.
You cannot reply to topics in this forum.
You cannot delete your posts in this forum.
You cannot edit your posts in this forum.
You cannot create polls in this forum.
You cannot vote in polls in this forum.

| Powered by YAF.NET | YAF.NET © 2003-2020, Yet Another Forum.NET
This page was generated in 0.406 seconds.