Great support + mirror cleaning = faith restored! Death to the Giant Flake(s)!

[Seagull]

The problem of contaminated/dirty mirrors has to affect everyone who has a Form1(+) sooner (2 weeks for me) or later (could take months or longer). There are some good threads out there already Attack of the Giant Flake (thanks to @Ante_Vukorepa and other contributors), also a similar one to this IT’S ALIVE! Mirrors Cleaned, Printing Fine! (nice one @Frank_Guthrie), but for those despairing at terrible print quality, particularly after a honeymoon period at the beginning, I thought I’d post a report of my positive FormLabs support experience.

One word of warning right at the start - you’ll find this elsewhere on the forum: DON’T just follow the instructions on how to ‘blow dirt off the main mirror’… you WILL blow it into the back of the printer, sooner or later, and thus contaminate the MUCH more critical smaller fixed mirror and the 2 tiny ‘galvos’ (the movable mirrors that guide the laser spot).

Anyway, the first 2 weeks of printing: blissful… I knew I’d made the right choice for a reasonably priced first 3D printer… way better than any FDM printer I’d seen in the same price class, and I’d seen a lot at the EuroMold exhibitions in 2013 and 2014.

Then suddenly, after 2 weeks, things started going downhill, bit by bit; after another 2 weeks I opened a ticket in despair - that has to be the best thing I did. Support, in the form of @Nicholas_Quinn, was great - he was patient, helpful, responded quickly to mails, didn’t miss questions but took time to answer them in depth.

The problem I had: dirty mirrors. That wasn’t immediately 100% certain, but parts I printed started looking like this (26th may):

Instead of like this (a week before):

or like this:

or like this:

Then things got worse, with practically no area of the parts being good, even small parts:

And then EVEN worse, with numerous failures, and even the ‘non-failures’ being terrible, with a combination of massive flakes firmly attached to the parts, coarse crappy surface even where there were no flakes, jello in the resin after one tiny print… aaargh!

(Thanks to @KevinHolmes for the tall cross file… but beware, it’s not kind to your resin tank!)

The printer was unusable - support ticket opened (if you’re having a problem, use the forum by all means, there’re lots of clever folks here with good ideas, but nevertheless, get a ticket - use both resources parallel; the forum also can’t give you permission to open the printer without voiding your warranty ).

To cut a long story at least a bit shorter than it usually is with me: bit by bit we ruled out things like the tank, the resin, the large mirror (free replacement mirror and PEC-PADs were sent, easy to change); I took a picture of the laser spot, support said it looked like dirty optics, not a degraded laser (and was right on the money):

(what a ‘laser dot’… no wonder that I was getting flakes and jello a-plenty in the tank after even the smallest of print jobs… all thanks to dirty mirrors)

So, I got the ‘ok’ from Nicholas from support to take the back off the printer and ‘go for the galvos’ (DO NOT DO THIS WITHOUT GETTING A TICKET AND BEING TOLD TO DO SO BY SUPPORT!!!).

Once the galvo block was off, I could easily light and photograph the dust in the machine, the dirty small mirror:

The galvos were also obviously dirty - most of it you can’t see… I’m just not good enough with the camera:

The actual cleaning was fairly nervewracking… I was sure I was going to break or scratch the galvos, despite following the instructions to the letter… and indeed the first few wipes made the mirrors look worse than before I started. I was really worried it was going nowhere. But, I persevered - you have to be REALLY thorough not just with the cleaning, but also with examining the mirrors… otherwise you will think they’re clean when they’re not - shine that light on the mirrors from ALL directions, like, a thousand times, particularly with the galvos.

In total I needed about 90 minutes from starting to dismantle the printer to getting it back together and running again, and after the first print, it was obvious the problem was fixed… faith restored, yippee! (Although, there might be room for minor improvement still… plus I didn’t rinse well enough to open up the smaller holes)

(I found the minimum feature test file from @JoshK / @Steve_Johnstone to be a great, quick way to test the printer, better than the butterflies - you can spread the feature tests around on the platform, printing is done pretty quickly, post-printing is easy; it also isn’t harsh on the resin tank, in comparison to the tall cross. It’s not the ‘last word’ in quality testing, I guess, but it’s a good start!)

Most of that time was spent cleaning the mirrors, examining, re-cleaning, re-examining, etc. - dismantling the printer is easy, it’s a simple process and the printer is well-designed.

Anyway, hope this helps at least one person in despair! Even if your problem is not dust or dirt on the mirrors, a ticket for support is the way to go, with parallel help, advice and ideas (and a shoulder to cry on ) in the form of the forum.

Sol

P.S. That was also the absolute end to jello in the resin… now there is really no need to filter the resin after a print at all - just search for bits, of course, and don’t forget to ‘stroke’ the silicone base of the tank with love to keep it happy and you’re good to go.

[Kara]

Thanks for this encouraging post!
We are always happy to help in support! I will let Nicholas know about your positive experience with us!
Never hesitate to contact us at the first site of trouble or even if you would just like some tips on printing your files for the greatest success.

[Alex_Vermeer]

Have you printed anything like your first two prints (handle + crim-track) to see how it prints when going over the area multiple times? Though the mimimum feature test is a nice way to test prints, small prints often come out fine, even if there are a few spots still on the mirrors. tall flat parts (like yours or Kevin’s) are going to be the challenge.

[Seagull]

Well, I haven’t printed that part again… but I DID try something fairly ambitious, more so than that part, I think:

There were a few areas that failed:

1. The yellow marked one was a hollow part that I hadn’t realised was hollow… PreForm did highlight it red, but I thought it was an error (because that part HAS to be solid… right ;)?)
2. The red marked part I also overlooked… should have been supported
3. The blue marked part… well, a chunk’s missing, for whatever reason.
   Other than that, it was very good; the only things that bothered me were a slightly curved bottom edge and it being 99.3mm wide instead of 100.0mm… the length was within a half a mm, which for 170mm I thought was good enough. All in all, for my first part, which was so big I had to and could only tilt it one direction, I was pretty happy.

Even the top part of the print came out really nicely:

A close-up of the blue area:

A close-up of the green area:

This was the third attempt at printing this part - the first two attempts looked like this:

After two fails, and scouring the forum, I decided to try raising the part to 10mm (instead of the standard 5mm), as @MarcusKnorr suggested in the surface quality variation thread, as I thought it was probably due to resin starving after peeling, what with both prints failing in the same manner at almost the same place.

That worked wonders… although I did also change a) the resin tank, b) the material and c) the orientation (rotated through 180 degrees about the z axis on the platform)… don’t normally change so many parameters at once when I’m ‘debugging’ anything, but I was a bit conscious of what each failed attempt was costing. May go back and try the successful .FORM file in grey in a fresh tank in the original orientation but still raised 10 instead of 5mm over the base… but for now, I’m pretty happy with the results for such a large part.

Now, if I could just get a) dimensional precision generally to +/- 0.1 (max. 0.2mm), and b) eliminate warpage as far as possible for such panel-like parts , things’d be pretty much perfect ! Oh… and the back surface was not great… not critical for me, in this case, but I do see what everybody else sees, a sloppy support side. But you can’t have everything, right?!

[JasonSpiller]

I’d be interested in how long that print remains straight - every circuit board mock up I’ve printed has warped within a few days of printing.

[Matthew_Bohrer]

Are you post-curing them with UV?

[JasonSpiller]

I’ve done some with and some without. They have all warped within 3-5 days of print - I just don’t do that sort of print now.

[Seagull]

I’ve read a lot of posts on here about UV curing, and have yet to invest in a ‘proper’ controlled method of doing so. Initially I left them in (direct) sunlight , and AFTER I removed the supports - doubly bad idea (bear in mind we were getting up to nearly 40 degrees C in the shade here near Munich in Germany last week)… unless you want SUPER bananas, that is, or are in the market to make miniature see-saws for model railways. On the ‘glass half-full’ side, they weren’t sticky any more

I also tried the ‘glass bowl of water’ method FormLabs recommends for flexible parts: this seemed to work well, but then I haven’t printed anything in flexible that would warp… it was only to ‘cure’ the stickiness problem, for that it works fine.

Then I decided to leave the supports on and make sure the parts stayed in the shade, turning them every 20-30 minutes… this worked fine, stickiness gone in 1-2 hours, parts not warped. However, like @JasonSpiller says, sooner or later, they start to warp, even when not exposed to further UV (left in a drawer, to avoid even the minimum UV that passes normal double-glazing):

Warping is not sooo severe, from my point of view - particularly when we are using the parts for immediate verification, warpage at a later point in time is unimportant. However, it’d be nice to be able to keep models to show customers… and the ‘banana’ effect is a big hindrance with regards to that.

I’m still wondering where the warpage comes from…

I’m pretty sure that one factor is the dripping of non-cured resin onto the back side (and therefore edges lf layers still being exposed) of partially built models, where it gets ‘half-cured’ by the laser on the next pass. This has to be partly to blame, and I’m not sure what could be done about that… it’s probably simply an inherent disadvantage of upside down resin printing.

If that’s the main factor, then gradual curing of a model over a few days, especially without supports but probably even with, or anything else to clamp the model in it’s form is always going to create warpage, perhaps even with clamping it’ll happen. One thing I want to try (if noone else has done this already…?) is post-curing heat relaxation… not sure what temp to try, but I’ll start with something like 60 degrees C and gradually raise it ‘to destruction’.

Could another possible factor be moisture absorption? That the part absorbs moisture more one side than the other? In which case a finish immediately after curing and separation from the support structure ought to cure (sorry !) the problem… will do a forum search on that subject when I have time.

[Matthew_Bohrer]

For my part, I use a cure tank I built; it’s about the size of a pint of milk with UV leds wrapped around it. A five minute cure in there takes any printed part from tacky to fully hardened, with a nice firm solidity. It cost me about $20. Considering your investment in the printer is a minimum of $3k, plus resins, it’s worth $20 (or hell, spend $100 and build a really cool one) to improve your parts.

Pragmatically, I’d suggest that some (maybe all?) of the warpage is due to uneven curing over a longer period of time. If it’s curing in ambient sunlight (instead of wrap-around UV from a cure box), you’re going to have different parts of resin (at a molecular level) curing at very different rates. These molecules are not perfectly rigid and so this means the plastic can deform quite a bit. Then, as they cure, they cure into that deformed position.

By curing everything with a nice high blast of UV all at once from all directions, the whole surface cures uniformly. There will be an extremely minor contraction (after all, this is part of what the curing is: the molecules contracting into harder, rigid polymer chains), but it will be uniform across the surface, and occur at a uniform rate. This should dramatically reduce warpage.

[Seagull]

@JasonSpiller: did you try this before giving up printing these kind of parts?

@Matthew_Bohrer: yeah, I can see I’m going to invest in something like this… what I’m doing at the moment is way too hit ‘n’ miss. I’m drawing up a list of things I’ve realised I’ll need/want since I got the printer a couple of months ago, then I’ll splurge, and for sure some kind of UV curing item will be on that list.

[JasonSpiller]

Yes, I use a uv cure oven. The problem, as I find it, was the thin cross section, coupled with the stress built into the print during the peel process. I only ever have this sort of warping problem with thin flat prints, so I stopped!

[KenCitron]

I have been using a water based cleaner as a pre-wash in an ultrasonic cleaner to remove the majority of excess resin. I then rinse the part off under a faucet then with my airbrush compressor I dry it off. Lastly I toss the part into a quick ipa dunk and some agitation. The part should not feel slippery. Remove and use the air compressor to dry the ipa so it doesn’t soak into the surface.

I haven’t had any tackiness after this and haven’t had the typical warping or part distortion issues.
Yellow Magic from Bradley Systems is one cleaner used in the Flexo industry and works well on Formlabs Resin here.

I have found the tackiness is raw resin on the part. Water based cleaners won’t typically cross contaminate parts with color, ie: black resin shouldn’t cause a clear part to be hazy. You should be able to recover majority of the cleaner back with a uv light since the resin is coagulated and not dissolved in the solution.

I originally used heated baths but found that room temperature seems to be enough and since the final cleans bath as others suggested works well and is less damaging than a long soak in IPA.

I recently got my printer back because I had some issues with clarity and sharpness of prints along with various failures. They had worked on the galvos and the printer is amazing. i do have a small flare on the laser but it seems to be non issue. Formlabs has a really good support team despite some of the horrors you may read on some of the posts. A real pleasure working with them.

[Seagull]

Thanks for that input @KenCitron - we have 3 ultrasonic cleaners here, for various purposes… I will investigate. Failing suitability of those 3, I guess that’s another thing to put on my ‘splurge’ list !

This would probably explain why I am getting ‘thick’ walls (see my post Calibration Part III)… I definitely notice slippery surfaces and the need to always post-cure the parts to avoid tackiness; in addition, I find the surfaces often need a gentle rubbing with gloved finger to make sure there are no tiny particles on the surface - a soft toothbrush leaves scars (perhaps a very soft artist’s brush is the way to go), and even strong agitation in the IPA bath doesn’t get these all off, so I guess the bath is simply not removing all raw resin either.

@KenCitron : do you print large, thin parts, where warping seems to be most common? As soon as I get to complex parts with 2-3mm or thicker walls, I don’t have a problem… but print a panel at under 3mm and bingo, I get warpage, even if I keep the IPA bath short, like 3-4 minutes of shaking and then that’s it.

[JonathanBaker]

Looks like Yellow Magic is only available in quantities of 4gal+. How’s their pricing?

[KenCitron]

They shipped me a sample quart for testing, been using that. They are supposed to be putting the product on Amazon. Not sure the pricing but certain it is less than IPA.

You can email Bradley Systems and find out more.

[Seagull]

Looks like my printer is beginning to suffer from dust, again… ticket opened, let’s see what the result is.