Templot Club forums powered for Martin Wynne by XenForo :

TEMPLOT 3D PLUG TRACK - To get up to speed with this experimental project click here.   To watch an introductory video click here.   See the User Guide at Bexhill West.

  • The Plug Track functions are experimental and still being developed. Some of the earlier pages of this topic are now out-of-date.

    For an updated overview of this project see this topic.   For some practical modelling aspects of using Plug Track see Building 3D Track.

    The assumption is that you have your own machines on which to experiment, or helpful friends with machines. Please do not send Templot files to commercial laser cutting or 3D printing firms while this project is still experimental, because the results are unpredictable and possibly wasteful.

    Some pages of this and other topics include contributions from members who are creating and posting their own CAD designs for 3D printing and laser-cutting. Do not confuse them with Templot's own exported CAD files. All files derived from Templot are © Martin Wynne.
  • The Plug Track functions are experimental and still being developed.

    For an updated overview of this project see this topic.   For some practical modelling aspects of using Plug Track see Building 3D Track.

    The assumption is that you have your own machines on which to experiment, or helpful friends with machines. Please do not send Templot files to commercial laser cutting or 3D printing firms while this project is still experimental, because the results are unpredictable and possibly wasteful.

    Some pages of this and other topics include contributions from members who are creating and posting their own CAD designs for 3D printing and laser-cutting. Do not confuse them with Templot's own exported CAD files. All files derived from Templot are © Martin Wynne.

3D printers - fun with FDM printers (Marlin-based)

Quick reply >
@Paul Boyd

Hi Paul,

I've been doing battle with Cura 5 on my new Neptune 2S.

The problem for a long time has been the width of the sockets in plain sleepers in 4mm scale. The plugs are 2mm wide, plain track sleepers are 3.33mm wide. That means the wall each side of a socket needs to be about 0.6mm thick, or a bit more. That's a problem using a 0.4mm nozzle.

Using a narrower plug on the chair isn't an option -- it wouldn't be wide enough to contain the slot for the loose jaw pins, and the delicate thin chair base surrounding it would be wider than is good for reliable resin printing. It would also make CNC-milling the sockets more difficult.

So the choice is to do a single wall and over-extrude from 0.4mm in the hope of creating a thicker wall; or do two walls and under-extrude in the hope of creating a thinner combined wall.

Neither was working -- the most I could get with a single wall was 0.5mm, making the wall unnecessarily fragile, and the chair was able to twist out-of-square in the too-wide socket.

With two walls the thinnest I could get was 0.7mm, making the chair too much of a bash fit in the too-narrow socket.

Then hidden among the 10,000 Cura 5 settings I found Flow Equalization Ratio:

cura_flow_eq.png


Which applies where a wall width needs to be modified to fit the model.

That's normally set in the range 0-100%. So I set 200% . :)

And that seems to have worked! A single wall but now 0.6mm thick. What that setting does is to cause the forward nozzle movement to slow right down, while continuing to extrude filament at the normal rate, so creating sufficient over-extrusion to create a 0.6mm wall from a 0.4mm nozzle. Cura was already doing that at 100%, but not enough and creating only a 0.5mm wall. Setting 200% increased it to 0.6mm. Why Cura couldn't have done that by itself I have no idea.

The Prusa gcode viewer is showing the effect of the over extrusion.

What that setting is % of only Cura knows, and it seems odd that 100% isn't the usable maximum. But it's working, and by changing this % I can now fine-tune the wall thickness.

Every day is a schoolday with Cura. The next mystery is why the same settings which produce nicely-fitting brick connector clips on the BIBO, are now producing a tommy bar which won't fit in the claws on the Neptune. :confused:

How are you getting on with the Anycubic Mega? I see it is now discontinued, just like my BIBO. It's difficult to keep up.

cheers,

Martin.
 
_______________
message ref: 5362
@Paul Boyd

Hi Paul,

I've been doing battle with Cura 5 on my new Neptune 2S.

The problem for a long time has been the width of the sockets in plain sleepers in 4mm scale. The plugs are 2mm wide, plain track sleepers are 3.33mm wide. That means the wall each side of a socket needs to be about 0.6mm thick, or a bit more. That's a problem using a 0.4mm nozzle.

Using a narrower plug on the chair isn't an option -- it wouldn't be wide enough to contain the slot for the loose jaw pins, and the delicate thin chair base surrounding it would be wider than is good for reliable resin printing. It would also make CNC-milling the sockets more difficult.

So the choice is to do a single wall and over-extrude from 0.4mm in the hope of creating a thicker wall; or do two walls and under-extrude in the hope of creating a thinner combined wall.

Neither was working -- the most I could get with a single wall was 0.5mm, making the wall unnecessarily fragile, and the chair was able to twist out-of-square in the too-wide socket.

With two walls the thinnest I could get was 0.7mm, making the chair too much of a bash fit in the too-narrow socket.

Then hidden among the 10,000 Cura 5 settings I found Flow Equalization Ratio:

View attachment 4614

Which applies where a wall width needs to be modified to fit the model.

That's normally set in the range 0-100%. So I set 200% . :)

And that seems to have worked! A single wall but now 0.6mm thick. What that setting does is to cause the forward nozzle movement to slow right down, while continuing to extrude filament at the normal rate, so creating sufficient over-extrusion to create a 0.6mm wall from a 0.4mm nozzle. Cura was already doing that at 100%, but not enough and creating only a 0.5mm wall. Setting 200% increased it to 0.6mm. Why Cura couldn't have done that by itself I have no idea.

The Prusa gcode viewer is showing the effect of the over extrusion.

What that setting is % of only Cura knows, and it seems odd that 100% isn't the usable maximum. But it's working, and by changing this % I can now fine-tune the wall thickness.

Every day is a schoolday with Cura. The next mystery is why the same settings which produce nicely-fitting brick connector clips on the BIBO, are now producing a tommy bar which won't fit in the claws on the Neptune. :confused:

How are you getting on with the Anycubic Mega? I see it is now discontinued, just like my BIBO. It's difficult to keep up.

cheers,

Martin.
Hi Martin

That's a useful tip to know, thank you. Somewhere in the back of my Covid-befuddled brain I seem to remember trying to adjust a wall thickness that wasn't a multiple of the nozzle width - can't remember what it was now - but that tip may work there as well.

The Mega is doing fine! I tend to use it mostly for printing functional things such as circuit board mounts or adapters, and will shortly be printing some alternative spacers for the DCC Concepts rolling road to suit my narrow gauge stuff. I've even printed prototype parts for work before committing to expensive machining. I've learnt to wait until it's nicely warmed up before levelling the bed, and then cleaning the bed with IPA immediately before printing - I've put a 3-second warning beep into the code so I can keep cleaning almost right up to the point at which the head starts to move for printing. It's now essentially a process of uploading the sliced model to OctoPrint (the trouble I had finding a Raspberry Pi!!) and setting it off.

I do still have a few timber bricks to print for my narrow gauge layout when I get back around to that. No sockets though, just the rail head guide indentation.

Cheers,
Paul
 
_______________
message ref: 5364
@Paul Boyd

Hi Paul,

Sorry to hear your brain is Covid-befuddled. I hope you are soon feeling better.

My brain is just ordinary-befuddled. :(

cheers,

Martin.
 
_______________
message ref: 5365
.
New slicer option for FDM:


Not sure if it adds anything to Cura for timbering bricks, but worth a try. I will have a go and see.

I do wish 3D printing reviews would print something actually useful of known dimensions, instead of always these toy figures. Do folks actually print these things? What do they do with them?

cheers,

Martin.
 
_______________
message ref: 5611
.
New slicer option for FDM:


Not sure if it adds anything to Cura for timbering bricks, but worth a try. I will have a go and see.

I do wish 3D printing reviews would print something actually useful of known dimensions, instead of always these toy figures. Do folks actually print these things? What do they do with them?

cheers,

Martin.
Hi Martin,

Sounds worth a try, although I gave up watching the silly video!

I’ve often wondered what people do with all these weird figures! I’ve come to the conclusion that many people buy printers to either make parts to break their printer (and then ask “I’ve changed the hotend and the extruder and the motherboard and the firmware and now my printer doesn’t work - what have I done wrong?”), or to print silly figures. Each to their own, I guess!

Cheers,
Paul
 
_______________
message ref: 5612
Hi Martin,

Sounds worth a try, although I gave up watching the silly video!

I’ve often wondered what people do with all these weird figures! I’ve come to the conclusion that many people buy printers to either make parts to break their printer (and then ask “I’ve changed the hotend and the extruder and the motherboard and the firmware and now my printer doesn’t work - what have I done wrong?”), or to print silly figures. Each to their own, I guess!

Cheers,
Paul
Paul,

You forgot about the stepper motors and the belt drive cogs. Oh, and there's also the Z-axis limit zero micro-switch :)

(The last one was very simple on my printer and did improve the repeatability quite a lot.)

Cheers!
Andy
 
_______________
message ref: 5616
@Steve_Cornford

Hi Steve,

You mentioned on Zoom this week that you were still dithering carefully considering the options about FDM printers.

Now 180mm x 180mm is definitely much smaller than ideal for the timbering bricks, but we do have the connector clips working well and I might be able to add some functions to speed up adding them to the bricks.

With 180mm any typical 4mm/ft turnout should be possible in only two sections. For example the Finetrax EM B-7 turnout base is comprised of two 160mm sections.

The trade-off is that this 180mm printer is sub-£200 and getting excellent reviews. It's direct-drive with manual bed levelling:






So I have ordered one:

https://www.amazon.co.uk/Aluminum-Function-Intelligent-Temperature-180x180x180mm/dp/B08CZD7ZLC

and can report the results. :) It's due on Monday.

This will be my 4th FDM printer. Obviously I don't actually need it, but I want to find one which I can say in the plug track instructions "get one of these and use these settings for it" for newcomers to FDM printing who want to try plug track. I think being sub-£200 will be important for such folks, and 180mm is plenty big enough for general FDM modelling jobs. The Neptunes are now more expensive, and apparently never in stock anyway.

It will probably need a glass build plate, but being a cantilever design the exact size of the plate doesn't matter too much, so I should be able to use the one off the BIBO for testing.

It's interesting that the details the reviewer complains about -- separate power supply, no integral roll holder, manual levelling, are all things that I prefer anyway.

It's also interesting that my Neptune 2S is also now effectively a cantilever design after I slackened the right-hand guide wheels right off. With a single-sided Z-lift I found it was the only way to get consistent Z dimensions.

But given the price of eggs now, one of the 4 is likely to end up on ebay. :)

cheers,

Martin.
 
_______________
message ref: 6378
@Steve_Cornford

We can all agree that 180mm x 180mm is small for our purposes. But how small?

Here are 4 180mm bricks -- allowing some overlap for the clip connectors, it is still sufficient to print 4 quite large 4mm/ft scale items. Three B-8 turnouts and a 1:8 diamond-crossing in P4:


four_bricks_180mm.png



So that's one item per brick in effect. You would need to go quite a lot larger to do the same in fewer bricks.

A larger build area would give you greater flexibility to arrange the bricks for more complex formations, but it's all doable on 180mm with a bit of re-arranging on the screen. Once you have split up your track plan into 180mm sections and added the connector clips, individual items can be shifted around on the screen to get the best fit to the brick areas for printing.

At present keeping the connectors (in the background shapes) aligned with the brick templates as you move them around is a bit tedious. I have in mind to look at adding the connectors to the symbols function, so that they can be integral with the template and saved in the BOX file.

If you are happy to build individual items and lay them on a paper template in the traditional way, with most items only 2 connectors are needed to link together the two sections.

I'm looking forward to seeing what the Kingroon can do. It's only a fraction more expensive than I paid for the Neptune 2S. That came as a full kit of parts which took several hours to assemble getting everything accurately aligned, and needed a lot of adjustment to the slider wheels to get it running backlash-free but without any binding. With aggressive retraction settings the Neptune bowden tube has worked better than expected, but the tube fittings are the weak part and I have now got through 2 sets of them. If the tube fittings work loose, the full retraction effect is lost, leading to excessive stringing on the printed parts. A direct-drive extruder, as on my BIBO printer, is definitely preferable, and I'm hoping the Kingroon will confirm that.

cheers,

Martin.
 
_______________
message ref: 6381
@Steve_Cornford

Kingroon report

I'm very pleased with this little FDM printer. It is nicely made and assembled, all metal, and supplied well packed. It sits firm and square on 4 rubber feet (unlike the rocking Neptune, packed up with expired credit cards). The linear X and Y rails are excellent and glide smoothly without any play. It would have been nice to have the same on the Z column instead of the rubber V-wheels. The direct-drive extruder is a simple design, but works very well and the results are excellent.

The printer itself is quiet and smooth. What lets it down is the external power supply. It's an unpainted tin box with sharp edges, and a loud cooling fan. In fact it's the same one as supplied with other FDM printers, including the Neptune. But in that case it is painted and fixed to the printer. It is obviously intended to be built into other equipment rather than supplied loose. I understand the reason for supplying it separately from the small printer, but Kingroon should have put it in a simple case -- and supplied a longer connecting cable. I can see this one getting an extension to the 24-volt cable and living in a box under the table. The cable has a ferrite chunk on it, so there may be problems with a longer one. We shall see. Electrically it is fine, and the RCD trip is happy with it.

The assembly instructions are brief to the point of omission. How hard can it be to insert and tighten 4 screws? Well hard enough to need 3 hands, or an assistant. Preferably one who doesn't mind being shouted at. :) Or getting greasy fingers from the linear rails. It would have been so much easier if Kingroon had supplied ball-spring T-nuts instead of the plain ones. They are available from Amazon, and if ever I have to assemble another one, I shall be ordering some! Also a plug and socket connection in the wiring loom would have been welcome instead of the hard-wired link between the two parts. Just getting both of them out of the packaging undamaged was tricky.

I started by taking the bottom plate off. There were a couple of points mentioned on YouTube videos which I wanted to check:

kingroon_pics1.jpg


The Y-belt runs inside the case. No cables were actually rubbing on it, but there was little to prevent them from doing so, so I added a few extra cable ties to be sure. All the connector screws were tight. The idler pulley for the belt runs on a simple stud bearing, best described as an "optimistic" design. But it seemed firm enough, and probably adequate. It would not be too difficult to replace it with something better if the need arises.

On both X and Y there is no means to adjust the belt tension, which is a shame. As supplied the belts were just about tight enough, but I would have set them a little tighter given the chance. It should be possible to fit spring tensioners if they do get slack, so I have ordered some anyway. They will come in handy on the BIBO too:

https://www.amazon.co.uk/DollaTek-Timing-Tensioner-Torsion-Printer/dp/B07DK41ZNS

The V-rollers had been set too tight on the column, with the result that they had developed flats in storage. This produced detectable notching on the Z movement. I slacked them off on the eccentric nut, and they seemed to recover and run freely. In practice the Z-movement never sees reversals during printing*, so there is no need to worry about Z-backlash. Not much point in worrying about it anyway, given that the Z-screw has a plain brass nut.

The direct-drive extruder is a bit basic to look at, but works really well. It is reassuring to see the retraction working very positively on the gear wheel, unlike the iffy results with a bowden tube on the Neptune. Also, a cocktail stick works great as a wedge to take the pressure off the filament if the printer is being left for any length of time. Otherwise there is a risk of the filament breaking, especially with older filament. The pressure spring is adjustable. They are reports on YouTube of the extruder getting clogged because of the internal design, but so far it is working fine. It's a good idea not to leave the nozzle heated for long periods if not extruding. Turn the temperature down until you are ready to go.

The printer is small enough to have a permanent home in the computer den, and within range of the computer for a USB cable to provide direct control if preferred instead of the Micro SD card:

kingroon_pics3.jpg


kingroon_pics2.jpg


I borrowed the glass plate from the BIBO for testing. My 3 foldback clips were just long enough to reach. They do mean losing about 10mm of Y-space in total, but there is no restriction on the full 180mm on X. With an exact-size piece of glass smaller clips could be used. Proper glass bed clips are available which use the minimum amount of bed space.

The plate is 4mm thick, so I added a 4mm spacer on top of the Z depth adjuster screw. In theory it just needs raising 4mm on the screw adjustment, but I forgot to check how much free thread is available on the screw -- and I'm not taking it all apart again just to find out!

I started off using the standard Kingroon settings supplied in Cura, for a test bit of timbering. I didn't expect the results to be very good, but in fact they were awful and unusable for plug track, with mis-shaped sockets.

Cura finally agreed to use the same settings I had been using on the Neptune. The results are excellent and easily the equal of those from the Neptune, and possibly better. It just needs some tweaking to the shrinkage and flow settings to get accurate sizes, no doubt resulting from the change to a direct-drive extruder.

It's still early days, but if you can live with the 180mm x 180mm build size, this looks to be an excellent little FDM printer. That's a big IF of course. For a few individual turnouts for a small 4mm layout it is plenty sufficient -- ideal even if you are short of working space for a printer. For 7mm or a spread of complex pointwork on a large 4mm layout you would likely soon wish for something larger.

I will write again when I have done more with it.

* unless you use Z-hopping with retraction, which I have never found to achieve anything, apart from unnecessary wear of the brass Z nut.

cheers,

Martin.
 
_______________
message ref: 6398
It should be possible to fit spring tensioners if they do get slack, so I have ordered some anyway.

Always amazed by Amazon. Ordered at 10pm last night. Delivered 10:15 am this morning.

Very fiddly things to fit, and must be in exactly the right place on the belt. They need 20mm space on the belt, so you need a belt run clear of the pulleys at least 20mm longer than the bed travel. But useful if there is no other means of tensioning. I shall keep them as spares for now.

Martin.
 
_______________
message ref: 6399
* unless you use Z-hopping with retraction, which I have never found to achieve anything, apart from unnecessary wear of the brass Z nut.
Hi Martin,

I've just picked up on that last line - I've idly noticed that both Z lead screws rotate backwards and forwards ever so slightly during retraction but never thought much of it. Is that Z-hopping? I might have a play with turning that off, if only to save a bit of wear. It wasn't something I've consciously turned on.

Regarding Bowden tube, I was having retraction issues with the stock Anycubic tube. I replaced that with tube and fittings from Capricorn, who make all sort of claims, and actually found that the claims seemed to be justified! I certainly found retraction far more reliable, although clearly direct drive is better. I just don't like the idea of a kilo of filament wobbling about on top of the printer!

Cheers,
Paul
 
_______________
message ref: 6400
Hi Martin,

I've just picked up on that last line - I've idly noticed that both Z lead screws rotate backwards and forwards ever so slightly during retraction but never thought much of it. Is that Z-hopping? I might have a play with turning that off, if only to save a bit of wear. It wasn't something I've consciously turned on.

Regarding Bowden tube, I was having retraction issues with the stock Anycubic tube. I replaced that with tube and fittings from Capricorn, who make all sort of claims, and actually found that the claims seemed to be justified! I certainly found retraction far more reliable, although clearly direct drive is better. I just don't like the idea of a kilo of filament wobbling about on top of the printer!

Cheers,
Paul
@Paul Boyd

Hi Paul,

Yes that's Z-hopping. The nozzle is physically lifted off the work while the filament is being retracted from the nozzle. I tried it years ago on the BIBO, and found that it did nothing to improve the results -- just wasted a bit of time and caused extra wear and tear on the Z nut. If anything it seemed to make stringing worse, in the sense that the nozzle is lifted into fresh air and able to ooze more freely than if it remains in contact with the work. It seems to be enabled by default in many standard printer profiles, so someone must find it helps somewhere. Easily switched off in the Travel settings in Cura:

z_hop_cura.png


One point to note is that if Z-hopping is switched on, all printing is done with the Z-drive falling.

Whereas if Z-hopping is switched off, all printing is done with the Z-drive lifting. With a conventional open gantry printer, that means the printing is done with the Z-screw pushing up against gravity, which is more likely to provide accurate positioning then when relying on gravity to take up the backlash in the nut when the Z-drive is falling.

On cased printers such as the BIBO however, it is the other way round because the Z-screw lifts the build plate towards the nozzle, and drops it away from the nozzle as printing progresses.

Does your printer have a sprung anti-backlash nut on the Z-screw? None of my FDM printers has, although the MINIBO uses a Z-belt instead. With a Z-screw there is also a small amount of backlash in the stepper motor to consider.

There is so much less to think about for resin printers! Only one moving part, for a start. :)

Not sure of your point re the filament reel? I also don't like having the weight of a full Kg reel hanging on the printer itself, regardless of the type of extruder. Here is my hi-tech patent wall-mounted reel holder:

filament_holder.jpg


1. upside-down shelf bracket, fixed with single top screw.

2. self-tapped jacking screws to ensure the reel is running level.

3. plastic screw plug on outer end to stop the reel falling off.

4. Pledge furniture polish applied inside the reel to permit smooth rotation (other brands are available).

5. cheapo kitchen sponge (dry), folded round the filament to remove any dust or other rubbish that you don't want in the extruder.

6. garden-twist tie wire to retain it and tie it back to the holder as the filament pulls through it.

This design has worked flawlessly for many hours of printing. I'm open to receiving design awards and general acclamation. :)

cheers,

Martin.
 
_______________
message ref: 6401
Yes that's Z-hopping.
I see Cura has a selection of z-hop settings - definitely worth a perusal next time I print something.

zhop.png


A design award is definitely due for your reel holder! I hadn't given it much thought but for direct extrusion especially, I suppose it doesn't really matter where the reel is, within reason. I can just imagine my landlord's reaction to your design hanging off the wall though :) I might have another look into direct extrusion.

Cheers,
Paul
 
_______________
message ref: 6404
@Steve_Cornford

Kingroon report

I'm very pleased with this little FDM printer. It is nicely made and assembled, all metal, and supplied well packed. It sits firm and square on 4 rubber feet (unlike the rocking Neptune, packed up with expired credit cards). The linear X and Y rails are excellent and glide smoothly without any play. It would have been nice to have the same on the Z column instead of the rubber V-wheels. The direct-drive extruder is a simple design, but works very well and the results are excellent.

The printer itself is quiet and smooth. What lets it down is the external power supply. It's an unpainted tin box with sharp edges, and a loud cooling fan. In fact it's the same one as supplied with other FDM printers, including the Neptune. But in that case it is painted and fixed to the printer. It is obviously intended to be built into other equipment rather than supplied loose. I understand the reason for supplying it separately from the small printer, but Kingroon should have put it in a simple case -- and supplied a longer connecting cable. I can see this one getting an extension to the 24-volt cable and living in a box under the table. The cable has a ferrite chunk on it, so there may be problems with a longer one. We shall see. Electrically it is fine, and the RCD trip is happy with it.

The assembly instructions are brief to the point of omission. How hard can it be to insert and tighten 4 screws? Well hard enough to need 3 hands, or an assistant. Preferably one who doesn't mind being shouted at. :) Or getting greasy fingers from the linear rails. It would have been so much easier if Kingroon had supplied ball-spring T-nuts instead of the plain ones. They are available from Amazon, and if ever I have to assemble another one, I shall be ordering some! Also a plug and socket connection in the wiring loom would have been welcome instead of the hard-wired link between the two parts. Just getting both of them out of the packaging undamaged was tricky.

I started by taking the bottom plate off. There were a couple of points mentioned on YouTube videos which I wanted to check:

View attachment 5468

The Y-belt runs inside the case. No cables were actually rubbing on it, but there was little to prevent them from doing so, so I added a few extra cable ties to be sure. All the connector screws were tight. The idler pulley for the belt runs on a simple stud bearing, best described as an "optimistic" design. But it seemed firm enough, and probably adequate. It would not be too difficult to replace it with something better if the need arises.

On both X and Y there is no means to adjust the belt tension, which is a shame. As supplied the belts were just about tight enough, but I would have set them a little tighter given the chance. It should be possible to fit spring tensioners if they do get slack, so I have ordered some anyway. They will come in handy on the BIBO too:

https://www.amazon.co.uk/DollaTek-Timing-Tensioner-Torsion-Printer/dp/B07DK41ZNS

The V-rollers had been set too tight on the column, with the result that they had developed flats in storage. This produced detectable notching on the Z movement. I slacked them off on the eccentric nut, and they seemed to recover and run freely. In practice the Z-movement never sees reversals during printing*, so there is no need to worry about Z-backlash. Not much point in worrying about it anyway, given that the Z-screw has a plain brass nut.

The direct-drive extruder is a bit basic to look at, but works really well. It is reassuring to see the retraction working very positively on the gear wheel, unlike the iffy results with a bowden tube on the Neptune. Also, a cocktail stick works great as a wedge to take the pressure off the filament if the printer is being left for any length of time. Otherwise there is a risk of the filament breaking, especially with older filament. The pressure spring is adjustable. They are reports on YouTube of the extruder getting clogged because of the internal design, but so far it is working fine. It's a good idea not to leave the nozzle heated for long periods if not extruding. Turn the temperature down until you are ready to go.

The printer is small enough to have a permanent home in the computer den, and within range of the computer for a USB cable to provide direct control if preferred instead of the Micro SD card:

View attachment 5466

View attachment 5467

I borrowed the glass plate from the BIBO for testing. My 3 foldback clips were just long enough to reach. They do mean losing about 10mm of Y-space in total, but there is no restriction on the full 180mm on X.

The plate is 4mm thick, so I added a 4mm spacer on top of the Z depth adjuster screw. In theory it just needs raising 4mm on the screw adjustment, but I forgot to check how much free thread is available on the screw -- and I'm not taking it all apart again just to find out!

I started off using the standard Kingroon settings supplied in Cura, for a test bit of timbering. I didn't expect the results to be very good, but in fact they were awful and unusable for plug track, with mis-shaped sockets.

After a battle of wits with Cura -- in effect telling it that this isn't a Kingroon printer, it finally agreed to use the same settings I had been using on the Neptune. The results are excellent and easily the equal of those from the Neptune, and possibly better. It just needs some tweaking to the shrinkage and flow settings to get accurate sizes, no doubt resulting from the change to a direct-drive extruder.

It's still early days, but if you can live with the 180mm x 180mm build size, this looks to be an excellent little FDM printer. That's a big IF of course. For a few individual turnouts for a small 4mm layout it is plenty sufficient -- ideal even if you are short of working space for a printer. For 7mm or a spread of complex pointwork on a large 4mm layout you would likely soon wish for something larger.

I will write again when I have done more with it.

* unless you use Z-hopping with retraction, which I have never found to achieve anything, apart from unnecessary wear of the brass Z nut.

cheers,

Martin.

Returning to this to say that the above printer seems to have been replaced with an improved version:

https://www.ebay.co.uk/itm/155563902302

The photo says "US Seller" but the actual delivery appears to be from the UK (Lichfield) in 3 days.

Significantly updated, and now maybe the ideal low-cost FDM printer to get started with plug track:

Size now 200mm x 200mm
Glass bed instead of bendy tinplate (manual levelling). Not clear if it is removable.
Belt tensioners
Power supply now integrated
Z rail changed to linear rail (good), but Y rails* reverted to V-rollers (?).
There's a filament run-out detector, but not in a convenient place for overhead feed. I wouldn't bother with it -- put an oddment of filament through it to fool the system.

*This later version has double linear rails for the Y-axis, but more expensive:

https://www.ebay.co.uk/itm/155551205026

cheers,

Martin.
 
_______________
message ref: 7074
.
Auto Bed Levelling

As so often over the years when developing Templot and plug track I feel that I must be on a different planet and missing something. :confused:

Today's mystery is the automatic bed levelling on the Elegoo Neptune 3 FDM printers. I recently obtained a Neptune 3 Plus, and I've been tinkering with it.

I have now solved the problem and it is working great. I've got the timbering thickness to +/-0.02 mm across the entire 13" x 13" bed. That's +/- less than 1 thou. That's likely to be better than any plywood or injection-moulded track. :) I shall be posting the relevant files for this.

But what is everyone else doing -- because as supplied the auto bed levelling is disabled. ???

I followed the User Manual to the letter. I started the auto levelling. I watched as the head moved to 49 positions across the bed, measuring the undulations in the bed at each position. I looked at the table of results, showing variations in the range +0.17mm to -0.08mm. I watched numerous Youtube videos and web sites explaining how it works and how to set it up.

The printer measures the undulations in the printer bed, and then as it prints it bobs the extruder up and down on the Z-axis as it moves across the bed to compensate for the undulations and ensure the first layer sticks to the bed. Right?

Wrong. Or at least, as supplied Wrong. It does nothing at all.

I suspected it was doing nothing and attached a pointer to the top of the Z-screw with Blu-Tack, to see the slightest rotation of the Z-screw. Nothing. It remained rock solid and not moving a fraction as the extruder moved all across the bed.

Cue Google. And this page:

https://marlinfw.org/docs/gcode/M420.html

The Marlin Gcode interpreter requires the M420 command after a G28 homing code to switch bed levelling on:

M420 S1

The Neptune 3 printer is supplied bundled with a custom version of Cura 4.8.0 which includes a printer profile for the Neptune 3 Plus.

There is no M420 code in the startup Gcode script.

As a result it doesn't perform any bed levelling. It measures the information but does nothing with it. Am I the only one to have noticed this? Did all those folks making Youtube videos not discover this? Or am I missing something?

In fact I'm not using that version of Cura, I'm using the later 5.4.0 which comes with its own profile for the Neptune 3 Plus, which includes this line:

;M420 S1 Z10 ;Uncomment to enable progressive compensation height of 10mm

So it's not working there either -- the relevant command is commented out with the leading semi-colon.

But I'm not using that profile either. In order to use my standard brick-printing print profile it was necessary to create a custom printer. To which I have now added the line

M420 S1 Z0

And (y) -- it's working. Great. I can see the Z-screw rotating a little to and fro as the extruder moves across the bed.

The Z0 setting means the fade-out effect is disabled. Meaning that the printer will continue to bob up and down the same amount on every layer of the print. The result is that the timbering brick is the same thickness all over, even if it's not dead flat, but following the undulations in the bed. It will be flat, when it is fixed down to a flat baseboard.

So we are on our way, even without a rigid glass bed, and I shall be posting the required profile files. But how did it get to be such a battle of wits -- or is it just me?

Martin.
 
_______________
message ref: 8494
Hi Martin,
With reference to your mods to the Neptune 3 plus, what effect do they have when printing jigs?
Would you need to switch on the face option?
Steve
@Steve_Cornford

Hi Steve,

I haven't made any physical changes to the Neptune 3 Plus. I have abandoned any idea of using a rigid glass bed, it's not usable with the inductive sensor on the Neptune.

Had I known that I'm not sure I would have gone for it, although the bendy tin bed is growing on me, now that I have tamed the auto-levelling function and know what I'm doing with it. I like the way the print sticks firmly to it when hot without needing any glue stick, and simply breaks free by differential contraction as it cools. I haven't needed to remove and bend it to release prints from it, which means that the levelling is unchanged for subsequent prints and produces excellent repeatable results.

The only changes I have made are in the settings in the Cura slicer program, so obviously you can make your own changes as you wish. For the filing jigs you need the top surface to be dead flat, rather than following the bottom surface at constant thickness, so that means a simple change in the Cura settings when doing the jigs. I will post all the settings and profiles I'm using.

I think you meant the "fade" option rather than "face"? Yes, that's what I'm referring to above. :)

Are you thinking of getting one? I received this email yesterday:

"Save 10% on your order with our "SHOPNOW-23" voucher code! Sale ends on 24.10.2023 at midnight (CET)"

from: https://www.3djake.uk/elegoo/neptune-3-plus

which is a lot less expensive than on Amazon. Mine arrived from Austria just fine and post free.

cheers,

Martin.
 
_______________
message ref: 8525
Just tried your discount code but it gave 0.0% so suspect it might hve been personal to your self.
Yesterday price was £296.00, today it is £303.00
Steve
@Steve_Cornford

Hi Steve,

Looking at that page I think the 10% has already been applied, with £340 crossed out. Now showing as £420-80 = £340 on Amazon:

https://www.amazon.co.uk/dp/B0C2HWS5YK

Might not be wise to wait?

This looks to be a new delivery in Europe from China at the higher price, it wasn't available on Amazon when I ordered mine (£296). Perhaps yesterday was the last of the old stock?

Is the extra cost of laser cutting down to the cutting, or the cost of plywood?

cheers,

Martin.
 
_______________
message ref: 8529
Not the plywood as this was £2.50 plus VAT for 1.6mm 304mm x 304mm.
The cutting was £68.00 plus VAT to produce 7 strips of 80, 3.33mm wide sleepers, even though I had bunched them & removed duplicate lines.
No nibs or snibs.
Steve
Hi Steve,
Where are you sourcing the plywood from and what is the quality like ?

I always use a new blade in my cutter each time. I am usually only using 0.8mm though :)
Knife 1.jpg


Rob
 
_______________
message ref: 8531
Not the plywood as this was £2.50 plus VAT for 1.6mm 304mm x 304mm.
The cutting was £68.00 plus VAT to produce 7 strips of 80, 3.33mm wide sleepers, even though I had bunched them & removed duplicate lines.
No nibs or snibs.
Steve
I would suggest an enquiry to Tim Horn if you want timber strip cut, much cheaper I think. He does the strip for Scalefour but he does the scale thickness as well if you ask i.e. 1.6mm.
 
_______________
message ref: 8540
It is not just the cost of the sleeper/timber laser cutting of the ply, but also the extra cost of laser cutting the sockets in the 3mm cork trackbed that is a factor in tipping the balance between my planned method of using the sockets in the cork trackbed to locate the ply timbers using the chair plugs versus FDM printing the bases.
With almost every new tick box that Martin adds tipping the scales!
Steve
 
_______________
message ref: 8543
_______________
message ref: 8546
@Phil G

Hi Phil,

You mentioned the difficulty of levelling the Ender printer. I've been pondering this, and I think fitting stronger springs might be making matters worse if the printer creates a lot of vibration. If the levelling goes out of adjustment the only possible cause is that the adjuster wheels have turned by themselves, and they are more likely to do that with a strong spring pushing against them (and possibly some unwanted lubrication on the screw?), especially if being vibrated.

The 4 springs need only to be strong enough to support the weight of the bed, and a bit more for luck.

These machines rely on gravity to take up any backlash in the brass nut on the Z-screw, and they can't do that if the Z-rollers are adjusted too tight. Try easing the rollers so that you can feel the X-bar drop back a fraction if you lift it manually. Also make sure that Z-hopping is switched off in Cura. Not only does that wear the brass nut rapidly, it changes the backlash direction while extruding. If the backlash freedom is restricted the Z levels will be all over the place.

In my priming run I have the extruder going to zero before lifting to begin extruding. This ensures that the entire print is done with Z driving in the same direction with no reversals or backlash problems. Measure the thickness of the priming run and adjust the levels accordingly. I don't know what the Ender default priming does, but if there is a Z-reversal between levelling and printing, you will find it very difficult to get accurate Z dimensions.

Having got the bed levelled, try pressing a blob of Blu-Tack between the wheel rim and the Y-carriage to help prevent the wheel turning under vibration.

It's probably time I posted my Cura profile for the timbering bricks, and a detailed explanation of how I do the levelling. What is the size of the Ender print bed?

cheers,

Martin.
 
_______________
message ref: 8622
Hi Martin,
Are there any videos on making adjustments to the filing jigs to ensure the rails are a good sliding fit in the jigs?
if not maybe a good one as a topic for next Monday's zoom?
cheers
Phil
 
_______________
message ref: 9383
Hi Martin,
Are there any videos on making adjustments to the filing jigs to ensure the rails are a good sliding fit in the jigs?
if not maybe a good one as a topic for next Monday's zoom?
cheers
Phil
@Phil G

Hi Phil,

It hardly needs a video -- just change this dimension until you get the result you want:


jig_rail_width.png



A starting point is to measure your rail and add maybe 0.05mm clearance.

Then trial and error tweaking of the setting to get the desired result on your printer.

If the problem is a Z-seam on the corner of the slot, carve it away with a craft knife.

cheers,

Martin.
 
_______________
message ref: 9386
Hi Martin,
Thanks, can you explain why a 1in 8 filing jig works fine with std settings, a 1in 10 is tighter but does work and a 1in12 simply will not fit? how can that have anything to do with the rail epically as its the same rail being used ?
cheers
Phil,
 
_______________
message ref: 9387
Thanks, can you explain why a 1in 8 filing jig works fine with std settings, a 1in 10 is tighter but does work and a 1in12 simply will not fit? how can that have anything to do with the rail epically as its the same rail being used ?
@Phil G

Hi Phil,

It's caused by the resolution of the printer, and rounding effects in Cura.

Find a setting which works ok for the worst case, and use it for all the others. The angle difference caused by a slightly loose groove won't significantly affect the filed angle over the length of the groove.

You might find that increasing the printer belt tension on the Y-axis helps to get more consistent results.

You might find that rotating the jigs 90 degrees on the printer bed helps to get more consistent results.

cheers,

Martin.
 
_______________
message ref: 9389
Hi Martin,
I am pleased to report, I took your advice, increasing the rail width from 0.92 mm up to 0.94 mm (actual rail was measuring 0.88 mm)
and turning the jig though 90 degrees, has resulted in a nice fit for the 1 in 12, very similar to the one I was getting with a 1in 8 jig.

Next step will be to try the switch blade jigs with these changes. Unfortunately because i did both at once. I don't know at this stage which of the two changes had the biggest impact.
thanks for your help once again.
cheers
Phil
 
_______________
message ref: 9402
Hi Martin,
On measuring both the old Jig rail slot (0.83) clearly the reason it would not fit, and the new Jig slot now exactly 1 mm.
I think the biggest change has been the turning the job though 90 degrees, its either that, or the software has round down the first time from 0.92 and up and up the second time from 0.94.
I think this variation is something that needs to be investigated, as either I have a poor FDM machine (still on my old ender 3 as wife says no room for two printers! at least not yet. So that's very possible) or the level of accuracy of FDM printers is just not as good as it needs to be for simple download the jig file and print scenario.

I will do more investigation at my end. All things do seem possible, but they also appear at this stage at least to be a bit hit and miss.

On a side note that 1in12 using you latest jig profile setting which you posted, took a suggested 5 hours and 2 min in Cura and almost 6 hours in reality, again putting my ender 3 machine under quite a bit of suspicion. Do you find your machines and Cura suggested print times match better?
cheers
Phil.
 
_______________
message ref: 9406
Hi Martin,
On measuring both the old Jig rail slot (0.83) clearly the reason it would not fit, and the new Jig slot now exactly 1 mm.
I think the biggest change has been the turning the job though 90 degrees, its either that, or the software has round down the first time from 0.92 and up and up the second time from 0.94.
I think this variation is something that needs to be investigated, as either I have a poor FDM machine (still on my old ender 3 as wife says no room for two printers! at least not yet. So that's very possible) or the level of accuracy of FDM printers is just not as good as it needs to be for simple download the jig file and print scenario.

I will do more investigation at my end. All things do seem possible, but they also appear at this stage at least to be a bit hit and miss.

On a side note that 1in12 using you latest jig profile setting which you posted, took a suggested 5 hours and 2 min in Cura and almost 6 hours in reality, again putting my ender 3 machine under quite a bit of suspicion. Do you find your machines and Cura suggested print times match better?
cheers
Phil.
@Phil G

Hi Phil,

Most FDM printers quote 0.05mm resolution, and I've found that when adjusted nicely they do achieve that on X and Y. If you have a double-column machine with a single-sided Z-screw the Z dimensions will be all over the place, but mostly Z dimensions are not so critical.

Generally measuring timber widths and lengths are within 0.05mm of each other, and correct to size if the shrinkage settings are good. Even the little child's Minibo printer managed it when used with the backlash correction utility which I provided.

I know nothing about the Ender printer, but if you are not achieving such results perhaps you have the roller wheels set too tight on the eccentric adjusters? Adjust them until you can turn the wheels against friction in your fingers without the carriage moving. If you can't turn them without the carriage moving they are too tight. If you can turn them freely without feeling any friction they are too loose.

Check also that the pulleys are fixed tight on the stepper motor shafts, and the belts are tight enough to "twang" when flicked.

All this stuff should be in the printer manual.

On my machines Cura's initial time estimate is usually quite close. If you run the machine via the USB cable from the computer, Cura re-adjusts the estimated time as the job proceeds.

cheers,

Martin.
 
_______________
message ref: 9407
Thanks for the info, I will recheck my ender again, I currently one exasperated step away from binning the dammed thing to be honest.
The reason for the Zoom topic was simply because this stared when I asked you about a possible zoom to cover off the jigs and why there not working so great.
Maybe we should have a few new topics such as possible issues and fixes for
1, FDM printers,
2, resin printers,
3 laser cutting (know you have less direct input on this one, because you don't have a laser) but it should still have its one topic
at least that way its not a mine field to try and find the key bits of info
 
_______________
message ref: 9408
Thanks for the info, I will recheck my ender again, I currently one exasperated step away from binning the dammed thing to be honest.
The reason for the Zoom topic was simply because this stared when I asked you about a possible zoom to cover off the jigs and why there not working so great.
Maybe we should have a few new topics such as possible issues and fixes for
1, FDM printers,
2, resin printers,
3 laser cutting (know you have less direct input on this one, because you don't have a laser) but it should still have its one topic
at least that way its not a mine field to try and find the key bits of info
@Phil G

Hi Phil,

Those topics already exist:

https://85a.uk/templot/club/index.php?threads/3d-printers-fun-with-fdm.277/

https://85a.uk/templot/club/index.php?threads/3d-printers-messin-with-resin.103/

https://85a.uk/templot/club/index.php?threads/3d-track-fun-with-laser-cutters.803/

cheers,

Martin.
 
_______________
message ref: 9409
Hi Martin,
Ok however, I feel I have to point out if I was an absolute beginner the topic names don't exactly spell out there content, and like all threads they meander along ( given I am partly responsible for this I hold my hand up), its not meant to be critical, just a comment I have no issue with keeping these thread topics.
My point was really the topic name should be quite clear,
I believe a name such as "possible issues and suggested fixes" spells out the intent, but it should be limited to only possible issues and fixes.

For example I would have no issue posting a photo of progress on work such as laser on the 3d-track-fun-with laser-cutters.
but for possible issues I would be looking to post things like laser power, cutting speed material and maybe even a burn intensity grid.
likewise the current discussion with FDM is very very helpful, and should be in a topic which clearly spells out its intent.
That was the point I was trying to raise.
cheers
Phil,
 
_______________
message ref: 9410
Back
Top