Building 3D track

[Steve_Cornford]

Bunched laser sleepers

 

[Martin Wynne]

(I note that Elegoo seem to recomend 8 seconds for the ABS like grey resin)
I am now in the process of printing another set of S1 chairs and locator plugs, and will then measure these to see if i get any consistancy.

I am thinking that locator plugs are easier to measure than chairs.
I suppose it is the equivalent to etching a 10cm scale
with 1mm increments on an etch design, or for that matter the calibration lines that Templot prints for ensuring our templates are printed to the right size.

Picture of "bunched" sleeper sheet to follow, the design of which was performed by the laser cutting service.

@Steve_Cornford

Hi Steve,

Thanks for the pics, sorry you seem to be having a lot of faff. It's difficult to be sure, but the bunched sleepers seem to be wider than 3.3mm ?

I think the recommended 8 seconds is for the previous Mars printer (not Mars 2) which used a lower-intensity screen. When the Mars2 was introduced they made a big splash about how much faster it is than the previous Mars.

I assumed/intended that anyone printing a batch of locator plugs would reduce the size of the plugs so that they are an easier fit through the timbers. Otherwise I can't quite see how they are usable to locate the timbers over the base layer. Unless I'm missing something?

At the back of my mind I'm sure that when I first introduced the locator plugs I made them automatically smaller than the normal plugs. But I can't now find any such code, so I must have removed it for some reason. I will see if I can get it back.

Your problems with the bunched sleepers illustrate why I so very much prefer to do experiments on my own machines rather than rely on 3rd parties. But it seems a laser-cutter of the required quality costs well into 4-figures, so unless I have some other major use for one (which I haven't at present) it's not going to happen.

cheers,

Martin.

 

[Steve_Cornford]

Hi Martin, forgot to say that the Scalefour nickel silver rail fitted chairs ok. When I say ok, bit of a wiggle, but slight;y easier than sliding onto exactoscale chairs. With Exactoscale chairs i seem to get about a 1 in 40 failure rate, which occurs in the bottom of the chair whicj has a hole in it, and one or other of the bottom side splits as I insert the rail, seems always to be a chair in the same position on the exactoscale sprue.

For me It is not really a lot of faff, just jolly good fun as I progress along the learning curve!

Oh, and I have noticed that on a few of the 80 S1 chairs I printed with plug width adjusted by -0.3mm a corner of the chair baseplate had snapped off, as you predicted, but I suppose that happens in the real world too?

I might try printing a set of 5 locator plug rafts with each raft having a reduction of locator plug width of -0.05mm, then measuring the resultant plugs (easier than chair plugs) as a means of evaluating the shrinkage.

It is going to be interesting to establish that if we use the same parameters and resin on several different Mars 2 Pro machines we get consistent results, or if we dont there is a way of fine tuning an individual setup to achieve the desired results.

Enough waffling for now, back to the resin tank to try out my 2nd build plate that has now arrived.
It is in a box labelled "sandblasted build plate" and is identical in looks to the one supplied with the Elegoo Mars 2 Pro.

I have already thoroughly cleaned it after removing the protective (tacky) clear film!
Wont be making that mistake again, famous last words!

Happy chair design, Steve

 

[Martin Wynne]

Enough waffling for now, back to the resin tank to try out my 2nd build plate that has now arrived.
It is in a box labelled "sandblasted build plate" and is identical in looks to the one supplied with the Elegoo Mars 2 Pro.

@Steve_Cornford

Hi Steve,

I thought mine was identical until I looked closely. On the 2nd one the top clamp screw has a plastic knob with an embossed logo. The original one has a metal knob and no logo.

Another thing I noticed by keeping my eyes peeled. I was surprised to find the Mars tool kit included a pair of genuine Xuron flush snips:

https://www.xuron.com

On closer inspection I have since noticed they are branded Xoron Corp in an otherwise identical logo. How about yours? It's great to have these low-cost machines from the far east, but we are not here to be fooled.

Oh, and I have noticed that on a few of the 80 S1 chairs I printed with plug width adjusted by -0.3mm a corner of the chair baseplate had snapped off, as you predicted, but I suppose that happens in the real world too?

Yes it does, but usually only after some time in use, not when new from the foundry! More worrying is what happened to the broken-off bits? If they are accumulating in the bottom of the tank they could spoil some of the next chairs, or jam against the build plate on the first layer.

Also a few on one print could be a lot on the next print, if conditions are marginal.

You could try increasing this:

which widens this red:

So that the corner overhang of the base from the plug is reduced. That should be ok for a bash fit into plywood, but might make it too difficult for FDM.

Or see if 1.8mm or 1.9mm plugs are feasible instead of 1.7mm?

cheers,

Martin.

 

[Steve_Cornford]

Hi Martin,
Spare build plate
Yes mine also has the plastic knob with logo rather than plain metal knob.
But when I said the same I meant the sandblasted finish, rather than CD turned
Flush snips
Yes min also have embossed Xoron but looks like Xuron at first glance!

Broken bits
These occured after curing and a bit more brittle.

Ply Sleepers
The "bunched" ones are are about 3.39mm wide, whereas the non-bunched sheets seem to have 3.3mm wide sleepers

Steve

 

[Martin Wynne]

Recent news item:

https://knowledge.autodesk.com/supp...-Version-1803-older-Effective-March-2023.html

After March 2023, Fusion 360 will no longer support

macOS 10.15 (Catalina) or lower,​

Windows 8.1,​

Windows 10 Version 1803 & older.​

​

Martin.

 

[Steve_Cornford]

Hi Martin,
Thanks for the pointer to the text depth parameter, I will give it a try tonight.

The top of the pin support pyramid is too small -- they all printed ok, but before UV-curing they were very delicate and several were lost in the washing process.

Was this in your paint brush process or in the Mercury wash phase?

 

[Martin Wynne]

Was this in your paint brush process or in the Mercury wash phase?

@Steve_Cornford

Hi Steve,

A few of the loose jaws parted company in the first "dirty dunk" brush wash to clean the back of the build plate, etc. That was before I realised just how delicate they were.

They all survived the Mercury wash (or at least, there are none in the bottom of the tub). I leave prints on the build plate hanging in the washer.

Then a few more were despatched into the next county when I got a bit too close with the air blaster for drying after washing.

And some more got dislodged in handling the rafts to remove them from the build plate and put them on the turntable for curing. At this stage they are still very soft from printing. I collected them up and cured them and they are perfectly usable.

No good hanging the build plate in the Mercury for curing of course because they wouldn't rotate into the light beam. But I have since discovered that if you completely remove the top clamp screw from the build plate housing, you can lay the plate down sideways on the turntable. There is just one position where it can rotate and clear the column by a few mm. So that would be a way to cure them without needing to handle them. It would need longer because they will be out of the light beam for 50% of the time.

Once cured they are much stronger, but still easily removed from the supports with fingers if desired.

What I don't know is how difficult it is to remove cured rafts from the build plate. Only one way to find out.

Having said all that, out of 60 loose jaws printed, there were still a good 50 still on the rafts at the finish. For 40 chairs.

I made a couple of mistakes:

The top of the support pyramid was too small to have any strength. I have now increased it a little. But there is not much scope for that because it needs to be smaller than the pin.

I also made the support pyramids too tall. Being the same length as for the chairs made the loose-jaws stand above the chairs when the rafts are side-by-side on the build plate. Making them much more vulnerable to damage. I have now lowered them with yet another setting:

2mm is the minimum support height, because that is the thickness of the Xuron cutter jaws.

cheers,

Martin.

 

[Martin Wynne]

What I don't know is how difficult it is to remove cured rafts from the build plate. Only one way to find out.

@Steve_Cornford

The good news: The answer is not at all difficult:

After curing the rafts can be "flicked" off the plate by getting the scraper under a corner - done over a soft rubber mat to catch them. Unlike being "peeled" off when they are still soft from printing.

Blu-Tack to keep the plate in the one position on the turntable where it can rotate without catching -- switch the machine off to try it spinning free (stepper motor off). I have made a pencil mark for next time.

I gave them 12 minutes curing to compensate for not being in the UV beam for a lot of the time.

100% yield and untouched by human hand until cured. Apart from the one loose jaw I bumped while fiddling with the Blu-Tack (which you noticed immediately!).

The bad news: Today's loose-jaw pins won't fit the slots. They are about 0.02mm (less than a thou) bigger than Saturday's pins, and the slots are the same amount smaller:

Saturday pins:

Fit Saturday slots easily.​

Fit Monday slots with care.​

Monday pins:

Fit Saturday slots with care.​

Don't fit Monday slots.​

What's the reason for the difference? Everything was exactly the same -- except I rotated the rafts 90 degrees on the build plate. I've been wondering for some time if this makes any difference. Now I know. The screen resolving power is obviously not the same in both directions. The dot spacing is the same, but the dot definition must be different, the dot being better focused or contained in one direction than the other. Light overspill from one dot to the next reduces the effective resolution.

Which means we need to decide on one direction or the other and stick to it. And then make the STL settings accordingly. I wonder if changing from 2.2 seconds exposure to 2.5 seconds would reduce the difference or increase it? I wonder if other makes of printer are the same? Above results with Elegoo Mars2 Pro.

The only other change which might account for the difference -- Saturday was a sunny afternoon. Today it's a wet night here and it was raining while printing. So greater humidity in the air. I don't know any reason why that would make a difference to 3D printing -- but I do know that it makes a difference when baking bread and developing photographic prints.

Back to the good news -- I can actually see the difference when fitting today's modified loose jaws (in Saturday's chairs), the jaw now sits down flush on the chair base. Using the old and new loose jaws in side-by-side chairs makes it obvious (although maybe not to a man on a galloping horse).

cheers,

Martin.

 

[Steve_Cornford]

Hi Martin,
Very interesting, and in a way I have experienced something similar.
I am still using 2.5 seconds exposure.
During the process of attempting to produce some S1 Locator plugs that give the best fit in my ply timbers (the unbunched ones) I have been printing some rafts of S1 locators with various "plug width adjustement" settings.

As ones with -0.15 plug width adjustment were still too wide, several days ago I produced some with -0.17 and -0.18 adjsutments respectively, each on rafts of 20 chairs, and if I call the ones you have pictured above as landscape rafts, I was arranging them portrait 5 rafts to the build plate (as you had originally).
I then measured the plug widths actually produced with a micrometer screw gauge and was surpris ed to find the following distribution of sizes:

-0.17mm adjustment gave 1 @ 1.74mm; 10 @ 1.75mm; 2 @ 1.76mm; 4 @ 1.77mm; 3 @ 1.72mm (20 in total)
-0.18mm adjustment gave 4 @ 1.74mm; 10 @ 1.75mm; 4 @ 1.76mm; 2 @ 1.78mm (20 in total)

in measuring these i did not note the postion on raft, however I surmised it might be something to do with UV emission from the overhead light in the room I was using (at night) when performing the operation of a) first switching the build plate onto the plastic 45degree holder whilst red cover off, then b) similar cause when transferring build plate from that bracket to the metal Mercury wash bracket, again with cover off.
My reasoning being that during a) there is a coating of resin on the printed chairs that could be exposed to UV emitted by the room light whilst red cover off causing a variation in plug width on those chairs most exposed.

The next day I produced some S1 locator plugs with -0.20 plug width adjustment, and some ordinary S1 chairs with -0.00 plug width adjustment (ie no adjustment) but with 6.87mm overall plug depth (as I needed them a bit longer to plug into the trackbed, but also this made it easier to measure the resultant plug width)

This time I used a much dimmer bedside table lamp (not an led) to illuminate the room whilst performing the two bracket change operations and here are the results for all the rafts I printed:-

S1 locators
-0.20mm adjustment gave 12 @ 1.70mm; 8 @ 1.71mm (20)
-0.20mm adjustemnt gave 1 @ 1.69mm; 7 @ 1.70mm; 12 @ 1.71mm (20)
-0.20mm adjustment gave 4 @ 1.70mm; 16 @ 1.71mm (20)
-0.20mm adjustment gave 1 @ 1.69mm; 3 @ 1.70mm; 3 @ 1.71mm; 13 @ 1.72mm (20)
S1 chairs
-0.00mm adjustment gave 9 @ 2.09mm; 10 @ 2.10mm; 1 @ 2.12mm (20)
-0.00mm adjustment gave 18 @ 2.10mm; 2 @ 2.11mm (2)

Assuming S1 chair plugs are designed to be 2.10mm, I suppose an accuracy of plus/minus 0.01mm is acceptable?
So my locators with adjustment of -0.20 (per side should give a target plug width of 2.10 - 2 * 0.20 = 1.70mm

Am I right in thinking that if the printers resolution is 50 microns (0.05mm), then objects produced will have nominal dimensions prouced in increments of 0.05mm (in the X and Y plane), but 0.01mm in the Z plane?

I am wondering whether it would be usefula to actually get a dark room bulb (one of those red ones)
You earlier posed question about incandescent bulbs but t'internet seems to imply that even these can emit UV

129mm divided by 2560 pixels gives 0.050
80mm divided by 1620 pixels gives 0.049

which re-inforces your resolution argument.

Regards Steve

ps I had only my second print failure yesterday, which was completely self inflicted. I have been happily printing away and completely ignored the fact that the resin tank level was getting low, which of course I noted after the print had finished and the build plate rose to the drip level, and the FEP was not completely covered in resin. it only affected a couple of chairs fortunately. Perhaps I need some some sort of dip stick device that when laid acroos the top of the tank has a pointer that should not be proud of the resin surface otherwise more resin is needed!

 

[Martin Wynne]

Assuming S1 chair plugs are designed to be 2.10mm, I suppose an accuracy of plus/minus 0.01mm is acceptable?
So my locators with adjustment of -0.20 (per side should give a target plug width of 2.10 - 2 * 0.20 = 1.70mm

Am I right in thinking that if the printers resolution is 50 microns (0.05mm), then objects produced will have nominal dimensions prouced in increments of 0.05mm (in the X and Y plane), but 0.01mm in the Z plane?

I am wondering whether it would be usefula to actually get a dark room bulb (one of those red ones)
You earlier posed question about incandescent bulbs but t'internet seems to imply that even these can emit UV

@Steve_Cornford

Hi Steve,

Thanks for the info, but something is not right there.

The design size for the S1 plugs (top part under chair base) in 4mm/ft scale is 2.0mm x 4.17mm (6" x 12.5" scaled).

I'm getting typically 1.99 - 2.01 mm. That's because I set the shrinkage to achieve exactly that:

which in my case is 1.5% on all 3 axes. This therefore is the default in the program (it has to be something).

If you are using 2.5 second exposures it is likely to be printing slightly larger than my 2.2 seconds, so you would need to adjust the shrinkage accordingly to get 2.0mm plugs.

We don't know the X-Y resolution of the sliced images sent to the printer in the CTB file. But I suspect they are resampled to 2560 x 1620 and a bit depth of 1 (black/white, screen dot either off or full on). In which case achieving +/- 0.01mm is an excellent result and it's amazing we can get close to that.

Although the Z-axis resolution on the printer is apparently 0.01mm, we are slicing the model into 0.05mm layers. So all Z-dimensions will be rounded to 0.05mm increments. We don't know enough about the internal workings of Chitubox to know how intermediate Z-dimensions are rounded to the layers. I suspect on the free version it is a simple truncation at the slicing section.

I have a single 60 Watt incandescent ceiling light in my work room. It is on a dimmer, and turned well down when I'm resin printing. I believe if dimmed to less than about 75%, the UV emitted is effectively nil. I do have a darkroom safelight, but it's years since I used it and I never found it very pleasant to work under -- one of these:

When I first started photographic printing I had a 1-bar electric fire in the darkroom which served as a combined heater and safelight. Saying that now sounds like I was on a different planet at the time.

Sorry to hear you ran the tank empty! You have obviously been busy.

---

p.s. to see chair dimensions on the control template, user the ruler tool:

Switch off timber infill (trackpad menu, above) so that you can see the ruler. Ruler at utils > show ruler.

CTRL+FULL STOP = mouse cross hairs

ALT+LEFT CLICK sets one end of ruler
ALT+RIGHT CLICK sets other end of ruler

cheers,

Martin.

 

[Steve_Cornford]

I will try some prints at 2.2 secs exposure so that I am matching your settings and report back. I am using elegoo abs like grey resin.
I think that it was all the locator plugs that used the resin. About 1 hour per print.
Steve

 

[Martin Wynne]

The bad news: Today's loose-jaw pins won't fit the slots.

@Steve_Cornford

Hi Steve,

I have now re-done those rafts, laid in "portrait" mode (running the short way across the build plate) as I did before.

The good news is that the loose-jaw pins now fit the slots again.

There isn't any bad news, although it meant fitting one fewer raft on there. They could be made a little longer.

It occurred to me that the deep vertical slots in the chairs are very tiny to wash out properly with a horizontal swirl. So I plunged the plate up and down in the wash bucket several times before starting the powered wash. I don't know if it made any difference, but the pins are fitting fine. Any residual resin left in the slots would be bad news for that.

I will post some pics later.

cheers,

Martin.

 

[Steve_Cornford]

Thats good news, and a useful pin slot wash tip.

I did wonder whether an ultrasonic clean would help until I researched the dangers of using IPA in an ultrasonic cleaner

My chitubox settings

Print in progress!

 

[Martin Wynne]

Print in progress!

@Steve_Cornford

Hi Steve,

Woah!

Several settings there in your version of Chitubox which I don't have.

I don't know what Image Blur does, but it sounds the very last thing we need if we are trying to get accurate engineering parts. It might be used on artistic models to soften any sharp corners.

With pixels 0.05mm apart, smearing 2 of them together is going to reduce the precision of parts significantly. On my printer it is switched off, and the Grey Level is set to zero to get maximum contrast at edges:

Some Googling needed to find out what's going on there. Is there any reference to this stuff in your printer manual? Mine doesn't refer to any of the advanced settings.

I can't remember if the above were the defaults on my printer, or whether I took one look at them and changed them.

edit: detailed article here: https://www.chitubox.com/en/article/support/indepth/technology/52

cheers,

Martin.

 

[Steve_Cornford]

Just reading the homework you have set

 

[Steve_Cornford]

Welcome to the world of voxels and hemivoxels!
Am i any the wiser? Not yet, probably because I had my eggs scrambled this morning!

That article explains the theory and shows the results in a clear and concise way, but what does it mean for our application?

Probably means upgrade to a Saturn to get text onto the top of your bolt/screw heads

PLug track users to use same parameters that you use, then have a standard Templot exported test piece that a new user can produce and measure to check for consistent results.
I chose S1 locator plugs as with a 12mm plug they are easier to measure than just the default S1 chair.
I suspect the pegs and peg slots are the more critica, and harder to measurel.

Steve

 

[Martin Wynne]

Am I any the wiser?

Hi Steve,

The problem with all this tech stuff is that it's likely to put folks off from trying Plug Track.

I'm hoping to make it as easy as possible with videos saying get one of these, click this, do that. But I can't do that until at least I have got as far as a complete turnout. And it's difficult with the internet awash with videos shouting at you, and everyone seemingly printing ugly fantasy miniatures which do nothing at all for me -- and certainly don't involve measuring things or fitting functional parts together.

But here's a nice article about starting from scratch with the Mars2 Pro and the Elegoo "ABS-like" resin to make usable parts:

https://thetechnologyman.com/elegoo-mars-2-pro-review-resin-3d-printing-introduction/

From which this is interesting:

"I found I could go a little lower than the default 2.5s – 2s gave slightly better results. And interestingly Lychee slicer uses 2s for this printer. Lychee slicer is an alternative to Chitubox and can still export .ctb files that can be printed off the Mars 2 Pro. I found it easier to use compared to Chitubox and it has more features in its free version."

I also discovered rather late in the day that the yellow plastic scraper is for cleaning the FEP film, not removing parts from the build plate. That solves the mystery of why they give you two scrapers.

I like the idea of spraying the FEP with PTFE dry lubricant. Something to try.

Here is the same guy covering the same subject in a video. He's talking too fast for a 74-year-old brain to follow, but I solve that by clicking the Settings cog and changing the playback speed to 0.75 :

Off to find the Lychee slicer to try.

Scrambled eggs.

cheers,

Martin.

 

[Steve_Cornford]

So that i can get away with only 1 litre in the mercury wash tank, I have mad a trial modification to the build plate bracket so that the build plate hangs lower in the wash bucket.

A couple of strips of 20mm x 2mm aluminium and some brass m4 nuts and bolts.
Steve

 

[Steve_Cornford]

After testing my modification, I found that with just 1 litre of IPA there was a navigation effect and air bubbles pulled down to below build plate in the middle, adding another 250 mils cured the problem.