Experimental 3D plug track - up to version 244c

[Martin Wynne]

I have not figured out how to get the second slider visible just yet, but I think I will get there

@Phil G

Hi Phil,

Explained here:

https://85a.uk/templot/club/index.php?threads/more-plug-track-software-developments.792/post-10165

cheers,

Martin.

 

[Phil G]

Thanks Martin,
That saved a lot of trawling, I know you had explained it somewhere just was not sure where.
cheers
Phil,

 

[Martin Wynne]

.
Bug warning

As always, within minutes of releasing an update a bug appears which I swear didn't show up in the weeks and months I have been working on this update.

There is a bug in the timber shoving. If you try to use a mouse action, such as moving a timber along, on a timber before that timber has been previously shoved, Templot can crash. This is much more likely if experimental chairing is in force.

Now fixed in 244b

 

[Terry Downes]

Hi Martin, just wanted to say 244a is working well and everything You promised has been delivered. This program just keeps getting more and more addictive. Well done!

*** The following is just a suggestion and is in not intended to apply pressure or criticism in any way ***

I do have a potential timber shoving enhancement for a future release which we can maybe discuss in the forthcoming Zoom meeting tomorrow?

I'm successfully using SC custom as You previously suggested. This is all working great but, I was wondering whether the 'temporary kludge' could somehow include rails 2 & 3 maybe via another tick box within the heave chairs dialogue box?

If this was added then I assume the options for chair heaving on rails 2 & 3 would become active on my D1 timber shown below.
Another potential enhancement would be chair alignment to either timber or rail which would enable the chairs on rails 1 & 4 below to be changed to standard S1 chairs aligned with the rails rather than the timber.

Terry.

 

[Martin Wynne]

I'm successfully using SC custom as You previously suggested. This is all working great but, I was wondering whether the 'temporary kludge' could somehow include rails 2 & 3 maybe via another tick box within the heave chairs dialogue box?

@Terry Downes

Hi Terry,

I think instead of extending the temporary kludge, my time would probably be better spent getting the proper K-crossing chairs done, so that none of the kludge is needed.

If you want S1 and L1 chairs aligned onto the rails, the way to do that is to overlay partial templates containing them. Also, there is a way of getting an approximation to the K-crossing chairs, by overlaying V-crossing chairs. in fact this was possible in 243a.

Copy the half-diamond template, remove the stock and check rails, extend the wing rails to convert the chairs to parallel-wing, blank off and shorten the template to just those 3 or 4 timbers, then snake them over the K-crossing and align the rails. Here is a very scruffy bit of video clip showing that:

https://flashbackconnect.com/Default.aspx?id=2T9d6vRHd4gUwuZWOI2hSQ2

Now you need to:

Shove the timber containing the XN chair along until it is exactly centred on the knuckle (see my note about the bug above). The chair will recalculate accordingly.

Shove the remaining underlying timbers on the K-crossing to match the overlaid chair positions.

Hide the overlaid timber outlines. Make sure there are no conflicting chairs on the underlying K-crossing timbers.

The overlaid AB chairs will have a location for a missing vee rail. This could be filled and filed back to represent the prototype K-crossing chairs, sort of.

(For the above the 4th chair probably isn't needed at this crossing angle and could be replaced with an L1 and CCL/R overlaid.)

It's all a bit of a kludge until I have the K-crossing chairs ready, but it's doable. The above doesn't make much use of the new 244a functions. However, for a slip you do need them because you would need to make a longer SC socket instead, and carve up 2 chairs to fit in the one socket.

n.b. For 3D diamond-crossings, remember to omit the duplicated K1 timber at the centre from one of the half-diamond templates. This wasn't strictly necessary for the 2D paper templates, but it is important for the 3D exports.

cheers,

Martin.

 

[Martin Wynne]

@Terry Downes

Hi Terry,

After making that video clip I thought I would carry on playing with the templates. What's interesting is that all this could have been done in 243a.

Overlaid L1 and CCL/R chairs:

With the result:

The red bit of splice rail doesn't exist of course, so the AB chair space would need to be filled and filed back to look the part. Also the XN and AA chairs should be slab & bracket, but one thing at a time.

Hiding the timber outlines allows the sockets to be captured by the half-diamond template:

Then some duplicating of a group, mirroring and rotating gets you further forward:

It's all quite a lot of faff, compared with a few clicks when I have got the K-crossing chairs done.

cheers,

Martin.

 

[Terry Downes]

@Terry Downes

Hi Terry,

After making that video clip I thought I would carry on playing with the templates. What's interesting is that all this could have been done in 243a.

Overlaid L1 and CCL/R chairs:

View attachment 8937


With the result:

View attachment 8938


The red bit of splice rail doesn't exist of course, so the AB chair space would need to be filled and filed back to look the part. Also the XN and AA chairs should be slab & bracket, but one thing at a time.

Hiding the timber outlines allows the sockets to be captured by the half-diamond template:


View attachment 8939


Then some duplicating of a group, mirroring and rotating gets you further forward:


View attachment 8936


It's all quite a lot of faff, compared with a few clicks when I have got the K-crossing chairs done.

cheers,

Martin.

Perfect, thanks. I'll give this a go later...

 

[Paul Boyd]

I suspect this is a long-term bug which has always existed. The much greater memory usage needed for the chair heaving has made it more evident. It may explain the occasional reports of Templot crashing over the years which I was unable to replicate at the time.

Hi Martin,

In the decades that I've been using Templot, I don't think I've ever used the Forward and Backward buttons - and I do a lot of timber shoving! I just go for the mouse action straight away. I've never had a crash when doing that, so I suspect that it may not be a long-standing bug.

Cheers,
Paul

 

[Martin Wynne]

I've never had a crash when doing that, so I suspect that it may not be a long-standing bug.

@Paul Boyd

Hi Paul,

Thanks for that. I'm coming to the same conclusion. This bug is proving very elusive to find, and I'm now thinking it must be something I've done in 244a.

I've had enough coding for one day, so I'm now trying the filing jigs on the Neptune 4. The results are proving excellent, and only 45 minutes for a 1:6 crossing-jig part compared with 2.1/2 hours on the Kingroon.

But I think unknowingly I struck lucky with the pricing at £210 that day. On Amazon the base-level Neptune 4 is now £240 for Prime users and £300 for non-Prime. Which is a significant jump from the Kingroons at £160 and £180.

If I was asked today which to recommend it would be a tricky decision. Fortunately no-one has asked recently. But I shall have to rethink some of the Cura profiles I have posted. It's puzzling that the Klipper-based Neptune 4 is able to print PLA+ at 220degC when on all the other (Marlin-based) printers I've tried such a high temperature has caused significant stringing, at least for plug track bases. Also Klipper is making a much better job of the 7-seg text than anything I've achieved with Marlin.

cheers,

Martin.

 

[Martin Wynne]

.
ZOOM meeting tomorrow (Tuesday 26th March) at 8pm UK time.

@Phil O

Phil, we can start a meeting in the afternoon at say 3pm if that would help? We could have a short session, and then restart proper in the evening. Anyone else interested?

Sorry I can't do Wednesday.

ZOOM access link at:

https://85a.uk/templot/club/index.php?threads/zoom-meetings-access-links.620/

cheers,

Martin.

 

[Phil O]

Hi Martin,

I can be about at around 3pm, but as Phil G particularly asked for my participation, he's likely to be still asleep.

I will log on at that time and see where we go from there.

 

[Phil O]

Hi Martin,

I can't make this afternoon, I have just had a doctor's appointment come through for 3pm this afternoon. Sorry.

 

[Martin Wynne]

@Phil O

Hi Phil,

Thanks for letting me know. No-one else has expressed any interest in a 3pm ZOOM meeting, so I shall consider it cancelled. Sorry to miss you, see you at a later date meeting.

ZOOM today will be at 8pm and recording available later.

cheers,

Martin.

 

[Penrhos1920]

.
Bug warning

As always, within minutes of releasing an update a bug appears which I swear didn't show up in the weeks and months I have been working on this update.

There is a bug in the timber shoving. If you try to use a mouse action, such as moving a timber along, on a timber before that timber has been previously shoved, Templot can crash. This is much more likely if experimental chairing is in force.

The workaround is easy, once you know about it. First shove the timber using the buttons at least once. For example in the above case click the forward or backward buttons on a timber before clicking the along mouse action. Similarly for the other mouse actions -- use the appropriate buttons first. If the timber has been previously shoved, i.e. showing blue, this is not needed.

I will get this fixed and release 244b as soon as possible.

I suspect this is a long-term bug which has always existed. The much greater memory usage needed for the chair heaving has made it more evident. It may explain the occasional reports of Templot crashing over the years which I was unable to replicate at the time.

cheers,

Martin.

I think it's been introduced recently as I found that I can get around it by turning off Experimental Chairing before shoving. This wasn't a problem for me as there was no chairs that I'd tinkered with on that template. HTH

 

[Terry Downes]

View attachment 8936
It's all quite a lot of faff, compared with a few clicks when I have got the K-crossing chairs done.
cheers,
Martin.

Hi Martin, I have tried this but, I cant get it work 100% and I think its because my crossing is NOT symmetrical. Rather than You spending a lot of time writing a detailed reply, I will attach the box file and IF you get a moment then maybe we could hook up via a Zoom meeting sometime. I'm working from home this week so any day or evening would suit me.

 

[Phil G]

Hi Martin, I have tried this but, I cant get it work 100% and I think its because my crossing is NOT symmetrical. Rather than You spending a lot of time writing a detailed reply, I will attach the box file and IF you get a moment then maybe we could hook up via a Zoom meeting sometime. I'm working from home this week so any day or evening would suit me.

Hi Martin,
if you do find time to run a zoom with Terry, I would very much appreciate it if you could make it evening. This is exactly what we were talking about at the end of last zoom meeting after you turned the recording part off I think. As such I too am very keen to better understand the workaround you were talking about.
cheers
Phil

 

[Martin Wynne]

Hi Martin, I have tried this but, I cant get it work 100% and I think its because my crossing is NOT symmetrical. Rather than You spending a lot of time writing a detailed reply, I will attach the box file and IF you get a moment then maybe we could hook up via a Zoom meeting sometime. I'm working from home this week so any day or evening would suit me.

@Terry Downes @Phil G

Hi Terry,

Sure. Any time today after about mid-day would be fine. Let me know what time would be convenient.

Phil G in New Zealand would prefer evening, but if you can manage only daytime, it can be recorded.

cheers,

Martin.

 

[Terry Downes]

@Terry Downes @Phil G

Hi Terry,

Sure. Any time today after about mid-day would be fine. Let me know what time would be convenient.

Phil G in New Zealand would prefer evening, but if you can manage only daytime, it can be recorded.

cheers,

Martin.

Hi Martin, evening is good for me. shall we say 7pm UK time? I will prepare my list of other questions....

 

[Martin Wynne]

Hi Martin, evening is good for me. shall we say 7pm UK time? I will prepare my list of other questions....

@Terry Downes

Hi Terry,

OK, that's fine. See you at 7pm today.

cheers,

Martin.

 

[Michael Woods]

Hi Martin - I am having a moment of madness. I have created this catch point. And it is clearly marked as solid jaw. But when I export I get lose jaws. I always assume that it is my incompetence rather than a software problem but I am scratching my head with this. Appologies in advance.

 

[Martin Wynne]

@Michael Woods

Hi Michael,

The "solid jaws" label applies to the control template. Have you remembered to store it? Delete the existing background template first. Only stored background templates can be exported for DXF/STL.

However there is something strange in your screenshot. The background template visible behind the control template appears to be to a slightly different track gauge. Does that make any sense to you?

If not please post your BOX file, so that I can see what's going on. Thanks.

cheers,

Martin.

 

[Michael Woods]

As suspected - my incompetence - sorry.

 

[Martin Wynne]

Hi Martin, I have tried this but, I cant get it work 100% and I think its because my crossing is NOT symmetrical. Rather than You spending a lot of time writing a detailed reply, I will attach the box file and IF you get a moment then maybe we could hook up via a Zoom meeting sometime. I'm working from home this week so any day or evening would suit me.

@Terry Downes

Hi Terry,

Thanks for an enjoyable meeting last night. It was good to go a bit deeper into Templot in a ZOOM meeting, even if I did inevitably get into a few muddles.

The BOX file we ended up with is attached below.

The templates which were mirrored and rotated will need more adjustments to the curving to match the opposite irregular half-diamond properly. When doing that, the partial templates for the K-crossing chairs are from the V-crossing on a 1:4.27 half-diamond template. This means that after aligning the radius in the main road to the underlying curve, you can adjust the radius in the diagonal road separately by changing the K-crossing angle (F10 mouse action) even though the K-crossing part is not visible. Change to a curviform V-crossing first, and put the peg on CRL-4 (FP) before changing the K-crossing angle.

cheers,

Martin.

 

[John Walker]

I have been playing with a crossover in 244a. In the past I would extend timbers between the crossings to the full width of the crossover. With chair heaving I can only fit 4 of the 6 chairs to the long timber. Should I represent the long timber as 2 separate timbers, position the chairs and then join them in my laser cutting file?

Regards
John Walker

 

[James Walters]

Hi John,
That works, or you can:
1. extend and shove the timbers of each such that they overlay each other precisely,
2. Omit 2 of the chairs from one of the timbers with the chair heaving controls.
3. One of the two timber outlines can be hidden by pressing H, or using the button in the shove timbers dialog.
You'll be left with one timber outline and 6 chair sockets.
The benefit being that there is no need to use CAD outside of Templot to create what you are after.

James

 

[Martin Wynne]

I have been playing with a crossover in 244a. In the past I would extend timbers between the crossings to the full width of the crossover. With chair heaving I can only fit 4 of the 6 chairs to the long timber. Should I represent the long timber as 2 separate timbers, position the chairs and then join them in my laser cutting file?

Regards
John Walker

@John Walker @James Walters

Hi John,

No. Extend the timber as before. But on the other template, instead of omitting the timber, click hide outline:

The timber will vanish but the chairs will remain. If they are fully over the extended timber (shove them along* as if the timber was still there) it will capture the sockets and the timber will have all 6 sockets in the export.

*there is a bug in 244a. Before along click the forward or backward buttons once (not both).

cheers,

Martin.

 

[Martin Wynne]

@John Walker @James Walters

p.s. John,

Note when making crossovers that the end flanges and webs will very likely conflict into the opposite sockets:

The solution is to remove the end flanges from the timbers, on both templates -- untick the box:

edit: this can now be done automatically, see: https://85a.uk/templot/club/index.php?threads/more-plug-track-software-developments.792/post-10892

And the result is -- sockets captured and free of conflicts:

Always check the preview for no missing or obstructed sockets before starting the printer.

For 2D laser, untick the snibs boxes similarly to remove any conflicts.

cheers,

Martin.

 

[John Walker]

Ah, I see! Thanks Martin, and James. Now to find time to apply myself.

Regards
John walker

 

[Martin Wynne]

.
Bug warning

As always, within minutes of releasing an update a bug appears which I swear didn't show up in the weeks and months I have been working on this update.

There is a bug in the timber shoving. If you try to use a mouse action, such as moving a timber along, on a timber before that timber has been previously shoved, Templot can crash. This is much more likely if experimental chairing is in force.

The workaround is easy, once you know about it. First shove the timber using the buttons at least once. For example in the above case click the forward or backward buttons on a timber before clicking the along mouse action. Similarly for the other mouse actions -- use the appropriate buttons first. If the timber has been previously shoved, i.e. showing blue, this is not needed.

I will get this fixed and release 244b as soon as possible.

I suspect this is a long-term bug which has always existed. The much greater memory usage needed for the chair heaving has made it more evident. It may explain the occasional reports of Templot crashing over the years which I was unable to replicate at the time.

cheers,

Martin.

I'm pleased to say that I have found and fixed this bug, after two full days searching.

As always, it was blindingly obvious and could have been found by a child of 2. Unfortunately I had only a child of 75 on hand.

No sun today, so I will try to get 244b released later.

Martin.

 

[Steve_Cornford]

Just wait until you have a child of 76 on hand!

Steve

 

[Martin Wynne]

.
Someone asked recently if the option to tighten the loose jaw fit was working properly. Sorry, I can't remember who it was or find the post.

Having now tested it, it seems to be working ok. Here I made the loose jaw fit excessively tight for clarity. The rail web thickness in 4mm/ft scale is usually about 0.4mm. Tightening the fit by -0.3mm has therefore had the expected dramatic effect:

Note that this setting is only for the loose jaws, and is not available on the XN knuckle chairs. Notice that the opposite gauge-side jaw is unaffected. This adjustment doesn't affect the chair base in any way, it changes only the thickness of the key.

If a tighter fit is needed on the XN chairs, try swapping a loose jaw from one of the other chairs.

---

To adjust the rail-fit for the solid jaws, the adjustment is done differently:

In this case it is a % adjustment. i.e. -10% means the space in the chair for the rail web is reduced by 10%. In this case both jaws are affected equally and the rail remains centred in the chair.

This setting also affects the loose jaws, i.e. you can combine the two adjustments to get the desired result, but it is likely to need some trial and error.

If there is a significant problem with the rail-fit, it would be better to measure the rail section dimensions and enter a custom rail section.

cheers,

Martin.

 

[Martin Wynne]

Hi all,

Program update 244a is now on the server.

Your copy of Templot should update automatically if you restart it and follow the instructions. There is no need to uninstall your existing copy of Templot.

Getting the chair heaving to a usable state has been a bit of a battle. Not all of it is yet working, and I'm aware of some quirks which will need further attention. But I have released it in its present state because I know some folks have been waiting for it for a long time.

A brief list of some of the 244a changes, which I will explain in more detail later (or I have explained previously):

2D exports:

modified snibs
soleplate detail
indented socket ends
kerf lines can be drawn directly in the background shapes

3D exports:

chair heaving
temporary wrong chairs on K-crossings (editable in the chair heaving)
multiple switch drive slider ribs
parallel-wing changes
default timber thickness changed to 10.08" (3.36mm in 4mm scale)

general:

changes to inherited brick/marker colours

---

244b is the next work in progress.

cheers,

Martin.

@James Walters


Program update 244b is now on the server.

Your copy of Templot should update automatically if you restart it and follow the instructions. There is no need to uninstall your existing copy of Templot.

Only 2 changes, see: https://85a.uk/templot/club/index.php?threads/new-templot2-update-version-244b-is-now-available.889/

A reminder of the previous 244a changes is in the above quote.

cheers,

Martin.

 

[Martin Wynne]

@John Walker @James Walters

p.s. John,

Note when making crossovers that the end flanges and webs will very likely conflict into the opposite sockets:


View attachment 8995


The solution is to remove the end flanges from the timbers, on both templates -- untick the box:


View attachment 8996


And the result is -- sockets captured and free of conflicts:


View attachment 8997


Always check the preview for no missing or obstructed sockets before starting the printer.

For 2D laser, untick the snibs boxes similarly to remove any conflicts.

cheers,

Martin.

I've mentioned before the problem of the timber flanges and webs conflicting into the sockets on an adjacent template. But I fear that I haven't taken it seriously enough. Here is a typical situation:

With this result if nothing is done about it. A great many sockets will be obstructed and unable to receive the chair plugs to the full depth:

What can be done about it at present is that such conflicting flanges and webs can be omitted by unticking the boxes on the timber shoving dialog.

Which works fine -- but it's a lot of tedious work to go through all the templates, timber by timber, and check carefully for such conflicts. They are easy to miss on the trackpad view, so the 3D preview needs to be checked too.

Some possible options would be:

1. make the plugs 1mm shorter. That would not be very convenient for handling and locating the chairs, and the clip-fit tangs might not work very well. At present they clip into place very nicely.

2. make the timbers 1mm thicker, leaving the plugs at the existing depth. Technically that would work fine, but in practical terms it would be overkill -- large areas of a track plan comprising only plain track and single turnouts would be thicker than they need to be, wasting filament and printing time, and needing more ballast infill material.

3. get Templot to fix it automatically, removing the flanges and/or webs where a conflict is detected.

Having written that it is of course the obvious solution. I think I can see a way of doing it, but it will take quite a lot of work. So that's yet another distraction before I can get back to the missing chairs. And it's now April and the sun is shining. Decisions, decisions.

Martin.

 

[Steve_Cornford]

Hi Martin
I suggest more walks in the sunshine th o recharge your batteries, fresh air refreshes the old grey matter.

Speaking for myself, I am quite happy to live with the work around of unticking boxes to remove flanges rather than distracting you from chairing solutions.


Another couple of suggestions to minimise socket obstructions:-
1 reduce thickness of flanges rather than increase thickness of timbers.(especially on turnout brick templates)


2 reduce width of flanges. (I did this on resin bases with no ill effect)

Steve

 

[Steve_Cornford]

For suggestion 1, reduce flange thickness from 0.96mm to 0.56mm, and plug depth is 2.58mm.
Steve

 

[Martin Wynne]

Having written that it is of course the obvious solution. I think I can see a way of doing it, but it will take quite a lot of work.

@Steve_Cornford

Thanks for the suggestions Steve. I didn't want to reduce the width of the flanges because it would mean losing the dropper wire retaining ridges. I'm quite pleased with those, they work well.

But coding is a funny business. Something which you think should take only an hour or two turns out to take 3 weeks of head scratching.

Then something which you think will take several days turns out to need only an evening.

This is now working, and will save a lot of time. Yet another tickbox:

Box unticked -- many sockets obstructed with conflicting timber flanges and webs:

Box ticked -- all sockets clear:

Admittedly the algorithm is a bit primitive, but I think it will work 99% of the time in this application where everything is rectangular.

But it does remove the entire web or flange, rather than just the offending part of it. Which means that some timbers may end up attached to the webs at one end only -- it would be worth running a splint or two under the timbers between the sockets to restore the integrity of the brick.

Note that although the offending flanges and webs will be removed from the 3D output, they will still show on the trackpad screen.

This will be in 244c shortly.

cheers,

Martin.

 

[Martin Wynne]

.
Program update 244c is now on the server.

see: https://85a.uk/templot/club/index.php?threads/new-templot2-update-version-244c-is-now-available.890/

Martin.

 

[Steve_Cornford]

You wait for a number 244 bus to come along, then blow me down 3 come along in quick succession!
Steve

 

[Martin Wynne]

.
With 244c I think we have reached a staging post / breathing space in this project, so I have done some forum housekeeping.

This topic has been merged with the 2 topics which came before it to create a single topic, so that you can use the search within this topic function to find related content.

But this topic is intended to stop here -- 2,500 posts is surely enough.

I have created a fresh topic to take us forward, see:

https://85a.uk/templot/club/index.php?threads/891/

cheers,

Martin.