Block detection and automation test/practice layout

[inobu]

34 minutes ago, chadbag said:

 

Maybe better with dremel cutting wheel or something 🙂   And wear safety glasses.

I know, its such a trivial task you don't pay attention to it. 

[gavino200]

1 hour ago, chadbag said:

 

Maybe better with dremel cutting wheel or something 🙂   And wear safety glasses.

 

It's a fairly isolated little project. You should make a thread dedicated to it, with a title that's searchable. No one's ever going to find this in this thread of mine.

 

You don't need to use a dremel. You just need to hold the pin with something. Either use a small hobby pliers and hold the piece by the part you're cutting. Or better still, hold the component by the plastic part and place the "legs" in some blu tac or adhesive putty and cut the leg with a nipper. It won't go anywhere. 

[chadbag]

40 minutes ago, gavino200 said:

 

It's a fairly isolated little project. You should make a thread dedicated to it, with a title that's searchable. No one's ever going to find this in this thread of mine.

 

You don't need to use a dremel. You just need to hold the pin with something. Either use a small hobby pliers and hold the piece by the part you're cutting. Or better still, hold the component by the plastic part and place the "legs" in some blu tac or adhesive putty and cut the leg with a nipper. It won't go anywhere. 

 

When I first read Inobu's remarks I misunderstood what happened and that is where the dremel remark came from.  I er-read it afterwards and saw that the action of the snip caused the snipped off part to go sailing.  My solution is to wear safety glasses and aim it in a safe direction 🤣 .   You don't want to watch me working with a dremel and cutting wheels and stuff like that.  I am sure OHSA would put me in jail if it were at a work place 

[cteno4]

22 hours ago, chadbag said:

When I first read Inobu's remarks I misunderstood what happened and that is where the dremel remark came from.  I er-read it afterwards and saw that the action of the snip caused the snipped off part to go sailing.  My solution is to wear safety glasses and aim it in a safe direction 🤣 .

I was once working away on the computer and then noticed something sticking out of the side of the monitor (old days with tube monitors) and it was a pin stuck into the plastic bezel of the monitor. I had been making a bunch of connectors the day before and nipping pins here and there. With some more looking I found a few more stuck into things. I always realized they went ping quite well and always directed them into the wall somewhere but this drove home that it would be quite bad in the eye if it could stick into the plastic bezel...

 

my solution was to use a balled up old wash cloth on the bench and nip with the cutter right over rag and let the rag snag the bits. No longer little metal bits all over the place as well. Have been doing now for like 30+ years that way. If I do t have a rag handy I put the cutter up against my shirt when I go snip.

 

jeff

[gavino200]

While starting to learn about routing on my iTrain tutorial, I realized I need to connect my loops, to increase my potential for different routes. So I added in my double crossover. It took a bit of thinking to work out how to make a double crossover in iTrain, as I haven't covered this in the tutorial, but I was able to do it by placing 4 single junctions and giving them the same address. 

 

But then I ran into an interesting problem. My Kato double crossover sometimes works perfectly - all four switches. But about 4 times out of 5, only three of the switches move. It's always the same switch that doesn't move. When it moves, it usually moves slower. 

 

I tried a different double crossover. I functioned worse. Only one switch moved. Both crossovers worked perfectly last time the old layout was up and running. I'll test them again tomorrow with my kato switch, but AFAIK they both work fine. 

 

I wonder if adjustments need to be made to change the voltage output for the double crossover. But I'm not about to experiment with this myself, at this stage. I'll pose the question to the Digikeijs dealer and see if they've encountered the problem. 

 

For now I'll swap out the double crossover for a single crossover, and hope that works for the practice layout. I'll use the last of my 8 switch outputs connected to an unconnected double crossover so I can troubleshoot this problem. I'll need to solve it at some stage for the permanent layout. 

 

Incidentally, I tried Chad's LocoNet trick to see if it would make a difference, though there's no reason it would make a difference in this case. No luck. 

 

Double crossover in place. 

 

 

Current state of the switchboard

 

 

The top right switch is the sluggish one. 

 

 

The Chad LocoNet trick. 

 

[gavino200]

Small update. I tested both the double crossovers with one of my old Kato switches. They work fine. In fact, a single Kato switch can operate both Double crossovers at once with no problems. This must be a voltage output/pulse duration issue. I'll send off an email to ironplanethobbies tomorrow. 

 

[inobu]

What is the rating on the DC power supply. This goes back to the discussion yesterday. The DR4018 uses the external DC power supply to throw the switch.

If this power supply cannot provide the current needed to throw the switch it will twitch instead of snapping.

 

Looking at the images check the connector to the  DR4018 power side first. If it is good start looking at your DR4101.

 

What you can do is plug the Cross over directly to the DR$101. If it works then you know its wire loss to the switch.

 

If it fails then its low output from the DR4018

 

Inobu

 

 

 

 

 

[gavino200]

39 minutes ago, inobu said:

What is the rating on the DC power supply.

 

Inobu

 

 

 

 

 

 

That was it, inobu. I could have sworn it was a 12V. It was only a 9. Swapped it for a 12 and the crossover works fine. 

 

Possibly also the root of the reset problems. Who knows?

 

 

Edited February 3, 2021 by gavino200

[inobu]

Good.......

 

Inobu

[gavino200]

I simplified the practice layout. It's closer to the tutorial layout now and easier to follow. I'll add more complex elements in later. I'm on tutorial 12 now and trying to set up track routes. The concepts are easy to follow, but I have an error somewhere in my block setup. I'll have to troubleshoot it tomorrow. I'm currently not able to drag and drop a train onto my layout. When I try, I get a cursor/arrow flickering with a No enter sign (circle with a line) whereas the tutorial just has an arrow and then the train appears in the block. Other than that, physically everything works and the DR5000 seems to be detecting my blocks. 

 

One likely error is that the block detector at the east side of the platform shows up differently than the others. It says 52 for some reason, but I don't see anything different about how it's set up. Hopefully fresh eyes in the morning will spot something. 

 

 

Edited February 6, 2021 by gavino200

[gavino200]

Some success today. Last-night's problem was just a missed step. I had set up the feedbacks but not the blocks. So I could run trains in demo this morning when I fixed it. After fixing a miswired section of track and fiddling with a troublesome junction I was finally able to run my first automatic train. It was quite satisfying. 

 

The trains don't stop properly as the  system has no real speed table information yet. That'll probably be the next step. But I think I'll take a little pause at this stage and switch to something else, so as not to get burned out.

 

Edited February 6, 2021 by gavino200

[Martijn Meerts]

It's real easy to mess up some wiring for automation, especially with many blocks all close to each other. Some labelling of wires definitely helps in that case. Also, considering your test layout is flat, you should be able to do speed measurements using occupancy detection. I do believe there's a video about that in the tutorial series at some point. It takes a while doing the measurements, but it's a quite critical step, and at least you can just start the measurement and then go do something else.

 

Looks like you're getting the hang of iTrain, those tutorials are really quite nice.

[gavino200]

5 hours ago, Martijn Meerts said:

Looks like you're getting the hang of iTrain, those tutorials are really quite nice.

 

Yes, they're great. Fuller really understands how to teach. I used the manual too when I was trying to troubleshoot. It's not bad either. I was a bit afraid of the computer part of automation. I was imagining a lot of DOS type commands. But iTunes is quite intuitive. I like it.  

[inobu]

Actually this has taught me something. After setting up iTrains myself i learned its usage. Its ideal for people who want's to get to

automation directly and don't mind spending the money. Where as JMRI is for people with a programming background that want's 

to create their own interfaces. Rocrail sits in the middle of the two.

 

The modeling style also has it's input where iTrain and Rocrail both fits the European and Japanese layout style with passenger trains

where as JMRI caters more to the American Club style freight operations.

 

Gavino, I'm always peeking over your shoulders to see what you are doing as I learn a lot as well. You are always doing something.

 

Inobu

 

[Madsing]

I also follow this interesting thread. I have played a lot with Rocrail a couple of years ago and this inspires me to start again 😀.

Rocrail looks a lot like iTrain but the menus and dialog boxes are much less sleek and user friendly. There is one thing I have struggled with in Rocrail, it’s the signals. They don’t react in a Japanese way and I had to resort to “complex” programming to make them work. How does iTrain manage signals?

[gavino200]

1 hour ago, Madsing said:

There is one thing I have struggled with in Rocrail, it’s the signals. They don’t react in a Japanese way and I had to resort to “complex” programming to make them work. How does iTrain manage signals?

 

I have no idea yet, but I'm going to be adding signals to this practice layout. It's something I want on my layout. I have zero knowledge at this stage of either prototypical Japanese signalling or how iTrain will handle it. I haven't even given any thought yet to model signal manufacturers. I have a lot to learn. It'll be a steep learning curve. 

[gavino200]

Getting back to this project. It's been about two months. I didn't intend to leave it that long, but I got caught up with some modeling. It took me a while to remember and re-learn what I've done so far. As it is I have some basic automation working. I can give a train a destination and itrain will take it there automatically, but it won't stop at the right point. I'm on about tutorial 11 of the Bob Fuller series. I'm going to plan on roughly one tutorial per day. 

[Martijn Meerts]

He's uploaded quite a lot of new ones recently. Around Christmas is was very quiet for a while, and I thought maybe he quit doing them, but guess not 🙂

 

[gavino200]

I had sort of a break-through with this today. I've been (sporadically) following the Bob Fuller iTrain tutorial. A problem had been that I built my practice layout before finding the tutorial. Bob's practice layout is very different. It's just two yards connected to each other. As I began having problems carrying out Bob's examples, I couldn't tell if it was because of the difference in the layouts or because the commands/parameters weren't correct. I really didn't want to rebuild the practice layout. I was hoping that Bob would move on to different layout patters. Today it just struct me that simply removing the west loop from the circuit and adding three terminals turned my layout into almost a copy of Bob's layout without relaying any track or touching any wires. 

 

My main problem was that the trains wouldn't stop on my "platforms" even when I set their lengths to ridiculously low values. The trains just kept going until they ran over the second junction and triggered an error. This seriously limited my "automation". I didn't know if this was because different commands were needed for platform sidings rather than yard lines. 

 

Now on the two platforms (and the north main line) are yard terminals. My trains still don't stop on time in the new yard terminals, but they are stopped physically by buffers after spinning wheels for about 5 seconds. I think the problem must be related to track lengths - perhaps the east loop that is now the central connecting segment. I measured actual track length for this and have now arbitrarily shortened it. But the values for the train speed tables are all wrong. They are just guesses that Bob had me enter. I think this problem will be solved when I enter actual speed data and track length data. 

 

But for the time being I can work alongside the tutorial again, so I'm happy enough. 

 

[Martijn Meerts]

For stopping at certain positions, having the correct values for the train's speed tables is essential, since iTrain uses those to determine where the train should stop. So if your speed table says that at a certain speed step the train is going 25km/h, but the train is actually going 50km/h, it'll be running much faster than iTrain thinks, so it won't stop where iTrain thinks it's stopping.

 

I think Bob has a video on speed measurements, so you might want to watch that one if you haven't already done so. It's fairly easy to set up a speed measuring section just using sensors already on your test layout. I'd definitely recommend doing a speed measurement for at least 1 locomotive, since it does take a while to do.

 

You might also be surprised at how fast trains actually are, I have steam locomotives that can easily hit 300 scale km/h 😄

 

[chadbag]

5 hours ago, Martijn Meerts said:

You might also be surprised at how fast trains actually are, I have steam locomotives that can easily hit 300 scale km/h 😄

 

 

 

Yeah, N scale models are much faster than their true scale prototypes.  I did some simple speed tests by measuring time to cover a known distance (not using sensors or anything -- just stopwatch) and a US freight train looked rather "slow" when going a nice scale clip loaded with container well cars.  Then I try to put myself in a small N scale persons shoes and I see things differently...

[cteno4]

We had a time table for the shinkansen loops on the last JRM layout to help us judge speeds, really helped but was a slow process. Was great but tedious, then we had a photo sensor speedometer donated to the club (the guy who made them sold them at the train shows we were at and the Shinkansens were a fun way to show off the system for him) and that was great. The shinkansens were long enough that you could fine adjust the speed while it traveled over the sensors. Really takes awhile to get into the n scale driver’s perspective! 
 

another nice arduino project.

 

jeff

[gavino200]

5 hours ago, chadbag said:

Yeah, N scale models are much faster than their true scale prototypes.  I did some simple speed tests by measuring time to cover a known distance (not using sensors or anything -- just stopwatch) and a US freight train looked rather "slow" when going a nice scale clip loaded with container well cars.  Then I try to put myself in a small N scale persons shoes and I see things differently...

 

Sometimes it's nice to get right down at layout level to see the trains from a scale point of view. I think it would be cool to have a few embedded tiny cameras to see the trains whizzing by from an N-scaler's perspective!

 

[gavino200]

3 hours ago, cteno4 said:

We had a time table for the shinkansen loops on the last JRM layout to help us judge speeds, really helped but was a slow process. Was great but tedious, then we had a photo sensor speedometer donated to the club (the guy who made them sold them at the train shows we were at and the Shinkansens were a fun way to show off the system for him) and that was great. The shinkansens were long enough that you could fine adjust the speed while it traveled over the sensors. Really takes awhile to get into the n scale driver’s perspective! 
 

another nice arduino project.

 

 

Definitely. The ready made ones are a little pricey IIRC. There's probably some pre-written code for this out there too. It might be a simple assembly job. 

[Martijn Meerts]

The ready made ones are more expensive than an Arduino yes, but they come with lots of features. The one I'll be using can also control motors and servos, and has inputs as well.

 

Pretty sure everything can be done with an Arduino (or, multiple Arduinos) as well, but if you want to control them using DCC you'll also be needed a DCC shield thing, and no idea if they're compatible with the smaller Arduinos, and what the cost is going to be like. I've not looked into them much, apart from having a look at controlling a custom turntable using one.