SimonD’s workbench

[Rob R]

Have you looked at the sliced file layer by layer in the region it is failing?

Sounds stupid, but try another slicer.
I think I had something similar a long time ago (with Chitu?), resliced in Lychee and all was well (might have been the other way round, Lychee then Chitu, also might have resupported as well, try on one component first).
I know it all seems like a bit of straw clutching but that's the way the Dark Arts seem to work.....

 

[simond]

Thanks Rob, that’s an option.

But given that the exact same file (high body) has printed once ok, twice not ok, and twice more ok, I’m inclined to believe it’s something downstream of the slicer. I tend to the view that the sliced file will either work properly, or fail every time, but not both.

Obviously the file I tried to print today (low body) hasn’t yet printed ok, but all the parts have been printed, and I’ve used Chitubox since I started this lark, and the majority of what I’ve done has worked, or I’ve been able to assign a reason for it not working, and then work around the failure.

I know this 3DP is a dark art, but I’ve got a modicum of science (and automotive quality) in my background and I can’t help looking for cause-effect, is-is not, Ishikawa, etc…

4M - Man, Machine, Method, Material: almost nothing has changed that I’m aware of!
The Man’s the same but another day older, I guess I might have made a mistake in slicing, removing islands, shaking the resin, cleaning the tank… but the print file worked, so it should be ok, I definitely shook the resin, I’m religious about cleaning the tank…
The Machine hasn’t been changed (but something might not be working as it did). I’m suspicious because the other day, it didn't turn on first time when I pressed the button, I had to prod it a couple of times. It’s been ok since but it just raised a question in my mind.
The Method is pretty much routine, particularly with a known successful print file, but has changed very slightly - my reptile heat mat USB plug has failed so the resin is slightly cooler than normal, but we are indoors, and the weather is not cold . I noticed this today. I don’t know whether it had failed earlier, but it appeared to be working last week when I had the earlier failures. I’ll hopefully get a plug to repair the lead tomorrow.
The Material is the material, and it’s well-shaken. I’m using Elegoo ABS-like black resin & have done something over a dozen wagon bodies, probably more.

I’ve reviewed the Elegoo site, and had a bit of a Google, there’s nothing that seems to fit. Most reported problems are raft adhesion to the build plate, but the Elegoo site does tell you how to replace big bits of the printer including the screen and the “constant current board”. I’ve checked the screen, there’s no physical damage (nor to the FEP) and by laying a piece of paper over it, and doing the “tank clean” it’s possible to check that there’s even illumination, which I’ve confirmed.

I have a tame electronics engineer at work, I’ll ask him for inspiration tomorrow: I can probably check the voltage of the supply under load, this would confirm or eliminate the brown-out hypothesis. After that, I’m stuck. I’ll drop an email to the Elegoo helpline and see what comes back.

cheers
Simon

 

[Rob R]

Simon,
You are far more methodical than me.
I just twiddle a few random knobs/settings and print again.
It works eventually....

 

[simond]

maybe.

but when it works reliably for weeks and months and then doesn’t work…

 

[JimG]

maybe.

but when it works reliably for weeks and months and then doesn’t work…

I'm wondering if your setup is skating on the edge of working/not working and something has changed recently to cause the print's failure at the ends of the supports - working temperature;change in resin formulation;etc... I've sometimes had problems like this at the support end level and got round it by either increasing the number of supports or upping the diameter of the end of the supports.

Jim.

 

[simond]

Thanks Jim,

Chaos theory - the butterfly that flapped its wings in Hong Kong causing the print failure in Kent

And, yes, of course, this “on the edge” case is possible, and I thought about it in discussion with Madame last night when expressing my frustrations. My provisional conclusion is that I’ve been printing with the machine in the same place through two winters, and most recently, I’ve been printing various iterations of the N6 coal wagons using the same resin and when I have had partial failures (see the early N4 wagons further up thread) this has been down to the supports. The failures tend to be very obvious distortions.

I don’t have a photo of a successful print (I will take one, soon, I hope!) but you can see in the failed shot, there’s the wagon, four sets of brake gear, four axleguard assemblies, four spare buffer stocks, three brake guards, three V hangers, two brake levers, the majority of these are small, light parts, and they’ve all failed more-or-less simultaneously.

and on previous prints I did manage to squeeze in two wagons plus spare bits per print, which meant the build plate was very cramped. I’d decided to simplify it by doing one at a time…

So, overall, I don’t think so. But I can’t prove it either way, apart from adding a load more supports.

 

[Osgood]

Concern over over integrity of power supply, known potential issues with constant current board?
Even when dinosaurs walked the earth (i.e. BE - before electronics), when voltage went down current went up and often as a consequence something went fut.
With fancy electronically regulated voltage and current systems it’s amazing stuff attains the reliability is does these days.

 

[Rob R]

Simon,
Does your printer (Saturn?) have a vertical leadscrew for the build plate? No build up of gunge etc at about the failure height? Regularly greased etc?
It does seem like it is missing a layer or 2 but why there and why not on every file are the pertinent questions.
Will think a bit more whilst walking hounds..

 

[simond]

Well, I don't think so, but I haven't looked....

so that's a job for this evening.


I sorted out the reptile heater this morning. no reptiles were harmed...

 

[JimG]

Well, I don't think so, but I haven't looked....

so that's a job for this evening.


I sorted out the reptile heater this morning. no reptiles were harmed...

Another factor might be if there is any whip in the vertical axis. I had this with my previous Phrozen printer, still there even after I fitted a modification supposed to get rid of it. But I did notice that when there was a change in area in the build from very little to quite a lot, that the build plate would stick to the FEP, then whip off at the end of the rise movement. This used to give me marks on the model and sometimes the finish of the printing process. My present Elegoo Mars 4 does not do this and I have had few printing problems.

The change in area from the top of your supports to the first layer of your models might have enough area change to cause a bit of whip in your build plate.

Jim.

 

[Yorkshire Dave]

Just an outside thought after reading the last dozen or so posts - have there been any software updates (inadvertently?) applied to the printer's software (whether automatically or manually downloaded) which has affected it's operation?

 

[simond]

Rob,

vertical leadscrew has been checked and regreased.
Surprisingly grubby when you consider it works in an essentially clean environment.
There is no detectable play in screw or carriage.

Jim,

comments as above. I guess the fact that nothing (obvious) has changed leads me to think that there’s nothing wrong in the column and leadscrew. It seems very rigid.



The column itself appears to be an aluminium extrusion, 48mm fore-and-aft and 130 wide, a flattened “U” with supports built-in for the guide rails, which are steel. The webs that you can see are 8mm thick. The screwed-on cap is 10mm thick. The carriage is an aluminium casting. We’re into “brick outhouse” territory here. They certainly put some effort and material into preventing flex.

Dave,

thanks for that thought. No, no updates, unless it converses secretly with the robot lawnmower….
Its only connection with the outside world is via the memory stick with the print files on it.
I haven’t downloaded any revised firmware, though I believe there are updates available.
It seems to me that would be adding another layer of uncertainty so I’ll steer well clear for now.


The reptile heater has been on all day and the resin is appreciably warm.
I have not had a reply from Elegoo yet, apart from an acknowledgement.
The power supply remains a question in my mind. I don’t have an easy way to stress it to see if it’s still delivering the rated power.

 

[Rob R]

Simon,
X axis was a very long shot but it needed eliminating.
It appears to be failing at the same layer, or there abouts?
If it is a power supply issue, what is difference between this layer and the one before?
The x axis stepper is doing it's thing as per the previous layers, the lamp is doing it's normal thing - we are well off the base layers with longer exposures - and, if I have got it right, the screen is using a bit less power because there are fewer pixels turned on - supports only not the raft of the previous layers.
The "known good file" you resliced, is that a supported .stl or a Chitu project file (.ctb?).
I would be tempted to start over with the .stl, pain in the butt to re support etc. but I think it is still more likely to be file related especially if you have successfully printed a different good file since.

 

[simond]

thanks Rob,

agree re the lift axis. Here are some screenshots;

Layer 168 is where everything starts - the lower left 4 circles are the spare buffers, the top right is a motor mount, top left is 4 axleguard assemblies and bottom right is four brake assemblies. The bottom of the solebars is obvious, and the other bits are v-hangers and brake levers



The bottom of the actual body begins in layer 343


but of course, if it failed on 168, it'll never get here...

and the finished print should look rather like this



Everything except the motor mount has been printed before, using the same settings, and the same Z=5mm lift off the plate.

I can obviously re-support everything, but as you say, it would be a PITA to do so. Before doing that, I am inclined to increase the exposure time per layer by maybe 10 or 20%.

Resin warm, and shaken, I have just kicked off a new print, so I will go and see what it sounds like as we approach layer 168... Trouble is, you cannot see whether it has printed successfully or not because it is pretty much still submerged within the print tray, if not still within the liquid resin at Z=5mm.

Further report to follow, probably over breakfast tomorrow!!!

thx
Simon

 

[Rob R]

Can't see clearly on my phone but does everything fail at layer 168? Even the smaller items?

 

[simond]

Very difficult to say. The brake lever guards survived at least some layers, as did the spring stops.

It cleared through layer 168 about 10 minutes ago - there was nothing at all to note, visual or auditory, as it did so.

 

[simond]

Hurrah!



which of course raises probably more questions than it answers.

firstly, is this a fluke? That’s not a comfortable prospect, because, even though setting up, and cleaning up are not onerous, they're still a significant waste of time, money and effort if I don’t get a print for my troubles.

Secondly, if it’s not a fluke, what was the problem? Again, a provisional conclusion. If the reptile mat that warms the resin had stopped working some time before I noticed, then that might explain the two failed print runs. As it was an intermittent failure, it may have worked sufficiently well for the subsequent two good runs. And then I noticed it following the bad run. I guess this lends credence to @JimG suggestion that I’m running close to the edge of disaster…

otherwise, I don’t really have any idea.

anyway, I’ll run another print tonight, and see how we go.

thanks for the suggestions
Simon

 

[Kev T]

That's a relief, at least you now know that nothing has failed.
Did you get these problems with ST FNG, and how long have you been using the Elegoo resin? Why did you move away from FNG? I'm using Sunlu with mixed results and am going back to FNG to see if that improves things.
Kev

 

[simond]

Hi Kev

The Elegoo ABS-like is far tougher than the FNG, and almost as good detail-wise

There is no way I’d have trusted FNG axle guard assemblies but the ABS is fine.

At
Simon

 

[simond]

This is curious but I suspect I have used light supports where medium or heavy would have been wiser…



Still, it fits and it won’t be seen…

(It’s a motor torque link & decoder shelf for a Premier Canon 30:1 to go in Tony’s Manor. I might get round to fixing it in a future print!)