Jump to content

3D printer furstration...


Recommended Posts

I recently joined the gang and bought my first printer. I went with what is arguably the most detail-capable printer of the new budget class lot, the Phrozen Sonic Mini 4K.

I didn't expect it to be an easy ride, like a one button push and then *magic* the perfect parts appear. I know from working with print contractors for a number of years that it can be difficult to get a particular part or geometry print perfectly and getting the orientation, support and parameters right.

 

However, so far, after a couple of weeks of testing, there's mainly frustration, neither calibration prints nor actual model parts print anywhere near to my liking. I can't seem to be able to dial in the exposure correctly, I seem to get over and under-exposure at the same time... The dimensions of prints aren't true to the original designs (oversized, but to different degrees in X, Y and Z direction) and any detail on the sides (i.e. in X and Y directions) is always totally washed out, no sharpness in the depth at all.

The exposure is also uneven across the build plate, one side always shows a lower exposure than the other.

The prints also show some weird angle in the vertical, i.e. they don't sit completely parallel to the build plate but at an angle and vertical bits sit slightly slanted.

I did re-level my build plate many times over and did all sorts of wizzardry for improving the printer I learned from user groups and youtube etc.

I'm on a couple of Phrozen user groups on FB but the vast majority of people print those miniature figures where dimensional accuracy and super fine detail is perhaps of lesser concern.

Therefore I'd like to reach out to the expert printers on here.

I print at 20µ layer height, but maybe this is too thin for certain resins to work properly? I went with this figure as the parts I had printed externally before were never printed thicker than that, usually 15-16µ...

Stuff I may need to learn about:

 - grey scale masking to even out the exposure across the whole build area

 - understanding resin properties better and understanding vertical bleed

 - pixel dimming

 - slow vs. fast resins etc.

 

Many thanks in advance,

Jeffrey

 

Link to post
Share on other sites

A couple of thoughts...

 

1. Have you checked the *actual* screen size/pixel count?  I recall seeing a video on Youtube that mentioned some printer specs weren't 100% accurate, which leads to wonky dimensions.  Simple test would be to print a few different sized cubes/squares and check the dimensions vs. the expected dimensions; if they're a consistent percentage off in any particular dimension, you can adjust your screen size by that value in your slicer.  If they're consistently the same finite measurement off, it's overcure.

(found it: https://youtu.be/-rEUyR9yoBI and hey, it's a Phrozen printer, too.  So that might be one of your problems)

 

 

2. The Para-LED set-up *should* mean fairly accurate exposure across the screen.  Probably worth checking to see if something has been knocked out of whack inside?  Especially if it's an unusual complaint in the user groups.  Otherwise...  at least the latest Chitu release makes it easy to deal with masking.  :(

 

3. Washed out X/Y detail sounds odd.  It's common to have issues with the side facing the build plate, but the rest should be okay (at worst).  Assuming it's not a slicer setting (do you have anti-aliasing cranked way up?  Masking or greyscale?), is your build platform fully tightened down?  Some seem to have issues where they're a little wobbly even after tightening, so if it's not rock solid, that could be worth looking at.  And it might also explain your Z-axis slant.  Otherwise, when the print bed is installed, try wiggling the whole assembly, as you might need to tighten down something there.  If that's okay, try printing a 2 - 3" tall (hollow) block to check for Z-wobble - a regular, repeating pattern along the Z-axis, caused by a bent leadscrew.  This *shouldn't* be a concern, thanks to the linear rails, but still...  Only other thing I can think of is to throw an engineer's square on the Z-axis to make sure it's not out of whack.

 

4. Levelling is a pain, but it's, like, the one critical input you have with the mechanics of the printer (vs. dozens with an FDM printer).  With such thin layer height, if you're out by a sheet of paper across the build plate, it could be 5 or 10 print layers.  So that hassle does pay off (and once it's locked in, it's relatively permanent).  If you're using the sheet of paper method, make sure you check the tension all the way around the build plate; check each corner individually, to make sure you don't have a slight twist.  You can check your levelling by printing some ~1mm squares in a grid pattern; if they all measure 1mm, your build plate is level, if one corner or side is too thin/thick, it's not level, and if the middle or ends are thicker, your build plate is warped.

 

5. Personally, I'd work on the mechanics first.  Nail the levelling.  Even out the exposure.  Then do gross calibration.  Then dial in your exposure settings.  Then adjust per individual prints.  It's hard to adjust everything at the same time, so start with the major stuff, then work your way down.  Resign yourself to wasting a bottle or two of resin and stick to coarse calibration prints until you work through the troubles.

  • Like 2
Link to post
Share on other sites

Some great advice, thank you! The video is also very helpful.

 

I had printed calibration cubes already to check the true sizes. I printed 5 and 7mm cubes and the results were oversize in all directions, but not equally. The largest growth was in Z direction, 2-3% (depending on position on the plate). X was nearly true to size and Y about 1% out.

Two ideas, apart from levelling once again:

 - the resins I've tested (Siraya Tech Hard Grey and Phrozen Aqua Grey 4K) may not be suitable for 20µ? They state 30µ-100µ layers on the bottles... Maybe at 20µ there is too much bleed through causing "over growth"?

 - I wash the parts and plate in an Elegoo wash station, usually 3-4 minutes. The resins say "don't immerse in alcohol for longer times" - perhaps I wash too long??? Does it "wash off" the detail or cause some other reactions?

 

Yes, I definitely have some Z bending going on - when the build plate reaches the print position (i.e. specified layer distance above the FEP) the jack tries to pull down further and the whole plate assembly starts to tilt back slightly. Phrozen support told me there may be a issue with my Z nut and auxiliary palte (the one the build plate mounts to) and they will send me a replacement. But I have my doubts that this is THE or the only issue...

 

How would I create a mask to even out the exposure though? I started using Formware as my slicer. I also have UVtools where you can upload a greyscale mask and include it into the slice file, but not sure how to do that.

 

Anyway, many thanks so far!

J

Link to post
Share on other sites

How frustrating. I can't offer truly expert advice, just sympathy as it is so gutting to be within reach of 3d nirvana but the machine is playing up. 

 

Did you mean to say the prints were 1% out at 5mm? Ie 0.05mm out? That's not a degree of inaccuracy I would be too upset by...I typically get slightly different performance from different resins, eg Monocure clear tends to come out a bit bigger than Elegoo - I know this because I've made prototypes in both and the Elegoo part fits perfectly and the Monocure always needs a tiny bit of adjustment. 

 

Re the washing, you will not damage the parts no matter how long you soak them in IPA - I've left them for days (by mistake mind). 

 

Re the build plate coming out of alignment, that would really ruin you day. Hope you get your replacement soon. 

 

Also just to add, cold weather is a big issue at the moment, depending on your set up might be a problem. 

Link to post
Share on other sites

Thanks for your help!

It may not sound much, but I intend to print masters for casting and if the 1st stage master is already out and not true to design silicone moulding and repeated casting will grow and amplify the problem until I get to the finished product. Parts need to fit into certain places.

There's also the problem with inconsistency. On the same test print I had two 7mm cubes in different locations. One cube came out 7 x 7.11 x 7.19mm, the other came out as 6.97 x 7.02 x 7.19.

Apart from that, my concern is the current lack of definition in the detail on my prints. The printer is certainly capable to do what I want to achieve, but I'm nowhere near that.

 

I'm trying to take precautions against the cold, the printer stands on an electric warmer plate for pets and has the shipping cardboard box over the top. I also pre-heat the resin and build plate and try to keep the room as close to 20°C as I can, difficult though as I also need to ventilate.

I really hope the mechanical issues can be solved. Anything else I think can be rectified by learning more about the printer, the resins and the ways to fine tune things with software, hopefully. 

Cheers,

J

Link to post
Share on other sites

Hi Jeffrey, what are you using as your slicer?  I just have the basic Photon printer, and I initially tried the bundled software.  I now use Z-suite, as it gives me better results.  I have to scale the X&Y dimensions by 95% when using it, while the Photon software works at 100% in all three dimensions.  With it factored, I can get close enough to use it for bearing housings and gears (it;'s not just for modelling).  The Z-suite fails at some complex 3D models (so I have to use the Photon 'ware) but works well on the simpler ones.   Trying different slicers may tell you what's hardware and what's software problems.

 

I have wrapped my printer in Thermawrap which is basically bubble wrap with aluminium foil.  With a terrarium heater plate inside, and the original box on top, I can print at 24C with the ambient at near zero in my workshop.  I will add an extractor to outside next year. 

 

Hope you get it working.  It has taken me 3 months to get my first full size model completed, but I've already used successfully it for a lot of smaller bits.

 

Good Luck   

  • Like 2
Link to post
Share on other sites

Testing at 5mm isn't going to give you a big enough margin, IMO, since at that size, overcure will be as much, if not more of an issue.  I'd go with 25mm (or bigger - 50mm would be good).  Really, the bigger the better.  Doesn't have to be a solid block of resin, either.

 

Z *shouldn't* be a problem.  If you think about it, each layer hardens enough to stick to the model (hopefully...), then the build plate rises up to peel it off, and lowers down to 0.02 mm above the previous layer.  Every layer, the build plate is (or should be) 0.02 mm above the last layer.  So, first, there's only a 0.02 mm film of liquid resin between FEP and the print.  And second, the printer doesn't only ever raises a finite amount - it's not raising 0.02 mm above where the resin actually *is*, it's raising 0.02 mm above where it *assumes* the resin is.  At a practical, functional level, lift distances, lift speed, layer heights etc. get translated via GCode into "send X-many pulses of electricity to the stepper motor ay Y-rate."  So even if your resin expands by 50% and the finished layer is actually 0.03 mm, the printer still only raises a net 0.02 mm...  meaning you'd have a net 0.01 mm gap (.02 - the excess .01 from the previous layer), so it would average out, rather than accumulate.

 

So.  If you're getting a consistent sized overage on your Z-axis (if it's always, say, .05 mm too tall), that indicates your bed isn't levelled properly, and if it's always higher on one side, your bed isn't level.  Don't expect it to be exactly right, but it *should* be accurate to +/- half of whatever the smallest adjustment increment is when you're setting Z=0 (usually 0.1mm)  Which is why you want to use the largest calibration print you can, to get a more accurate measurement.  If you're getting a persistent *percentage* deviation - that would be... odd... and likely suggests something mechanical/hardware related.  (I'd be wondering if the leadscrew is wrong, or something)

 

Anyway...

 

The resins shouldn't have any problem curing at your level heights, although I guess you could try contacting Siraya for their thoughts.  (as an aside, why does everyone call it 'Sierra Tech', just because one Youtuber doesn't know how to pronounce words?!?)  Light bleed, overcure and 'overgrowth' will generally be around the edges as the light bleeds around the edges of pixels on the LCD.  It's not (much of) a 'depth' thing.  That same Youtube channel from my first reply has a really good video explaining how/why overcure happens, along with a utility to reduce/eliminate the elephant's foot it causes in the base layers.

 

Washing 3 - 4 minutes isn't going to be a problem.  You *could* see issues if you leave resin in alcohol for days or weeks, theoretically, but seconds or minutes will not be a problem.

 

Z-banding is... different.  THIS is Z-banding - basically, the leadscrew isn't properly aligned, so it shifts the build plate around horizontally as the screw turns (pic is from an FDM printer, but it can happen on resin printers, too), resulting in offset bands along the Z axis.  What you're talking about sounds more like a calibration issue.  My hunch is you've got the plate too high (too far from the bed, too close to the release screw), which would have the 'jack' hitting the side of the vat on the base layers, and continuing that offset throughout the print.  Essentially, your build plate could be trammed to the LCD, but thrown off because it's hitting the vat, causing all sorts of potential troubles. (especially if you're only doing short-ish prints).  It could even make the detail seem mushy, since the result would be somewhat random.  A simple check would be to move the bed to Z=0 and check to see if there's any clearance between the 'jack' mechanism and the vat - if the arm that your print bed attaches to is hitting the vat, that's a problem, and you'll need to re-level.

 

I know, I keep harping about levelling, and it *sucks* having to do it over and over, but it really is critical, and one of the biggest causes of error for most users, so unfortunately it usually pays to suck it up and work on dialing it in.  Reddit is drowning in "my printer is levelled, how do I fix this totally ~~WeIrD~~ problem" posts that are a.) really common and b.) clearly caused by levelling errors.

 

Lastly, as for cold...  I'm not convinced it's a massive issue.  The LED adds a lot of heat to the printer, so after a short printing session, the resin will be warm and fluid no matter what.  I do find I'm more likely to have adhesion issues with "cold" (say, 18`C?) resin vs. "warm" (from a previous print) resin... but it's not a huge difference, and it definitely won't cause the kinds of size/detail issues you're getting.

  • Like 1
  • Thanks 1
Link to post
Share on other sites

Thank you very much for your advice guys!

 

As for the slicer: Phrozen has a Chitu motherboard and comes with Chitubox slicer software so I started out with that. I then tried Lychee (free) as many recommend it as being a better slicer and now I'm trying out Formware, which is the most powerful and option-rich of the three. Drawback is it's not free, costs around €130 after your 30 day trial. I also did some tests with UVtools and its pixel dimming function. There is also an option to load a grayscale filter mask to even out the light output of your LED array, Haven't tried that yet.

As I like the Formware workflow and options I may end up buying it, we'll see...

 

Thanks for the tip with trying larger test pieces, I will try that.

I have made a small test array with a square block and a few thin vertical pillars and made a build plate setup with this test copied all over the plate space. That way I hope I can get a picture of the evenness of exposure.

 

I start to think that there may be a resign flaw in that the Z beam/screw and build plate attachments aren't strong enough to displace the resin in the VAT. I can clearly see the whole thing flexing back when the plate reaches the print position wich of course is only 0.02mm above the bottom so all that resin needs to be pushed to the sides.

I'll keep testing, re-levelling etc.

I don't notice any jack wobble, but will have a close look for that as well.

 

Thanks again,

J

 

 

Link to post
Share on other sites
  • 2 weeks later...

... a small update:

I'm now getting much better results. The most important change I made was my wash procedure. I didn't realise that just immersing the parts in a wash station (with IPA) and running the cycle doesn't actually remove the small layers of uncured resin around the detail etc.

What I do now: quickly rinse the build plate in an IPA bath to remove the most of the excess resin, and wipe the plate with paper towel. Then I put the plate and prints in the wash station (now with 85% anhydrous Ethanol rather than IPA), give it a 30sec "spin" and blow off the alcohol and resin with an air brush. This is followed by another wash cycle in the wash station and another blow-off by air brush. Additionally, I load some IPA into the air brush and blast the parts with it.

 

As for the printer itself, for now I'm printing at 30µ instead of 20µ and the results are a bit more consistent. I may revert back to 20µ for tests though.

The printer still has a problem with the build plate flexing although I have improved it a little by making some mechanical adjustments. Next, I will install nFEP to improve or rather reduce the stickiness and I'm looking into an upgrade part that may make the build plate assembly a bit sturdier.

In my slicer software calibrated/corrected the resin for shrink rates in X,Y and Z direction according to test prints of 15mm cubes. Printing some actual parts have so far shown a pretty much spot on dimensional accuracy...

 

Cheers,

J

 

Link to post
Share on other sites
4 hours ago, JeffreyK said:

... a small update:

I'm now getting much better results. The most important change I made was my wash procedure. I didn't realise that just immersing the parts in a wash station (with IPA) and running the cycle doesn't actually remove the small layers of uncured resin around the detail etc.

What I do now: quickly rinse the build plate in an IPA bath to remove the most of the excess resin, and wipe the plate with paper towel. Then I put the plate and prints in the wash station (now with 85% anhydrous Ethanol rather than IPA), give it a 30sec "spin" and blow off the alcohol and resin with an air brush. This is followed by another wash cycle in the wash station and another blow-off by air brush. Additionally, I load some IPA into the air brush and blast the parts with it.

 

As for the printer itself, for now I'm printing at 30µ instead of 20µ and the results are a bit more consistent. I may revert back to 20µ for tests though.

The printer still has a problem with the build plate flexing although I have improved it a little by making some mechanical adjustments. Next, I will install nFEP to improve or rather reduce the stickiness and I'm looking into an upgrade part that may make the build plate assembly a bit sturdier.

In my slicer software calibrated/corrected the resin for shrink rates in X,Y and Z direction according to test prints of 15mm cubes. Printing some actual parts have so far shown a pretty much spot on dimensional accuracy...

 

Cheers,

J

 

 

- which IPA do you use? 

You'd better use 99% IPA.

 

And the most efficient process is called 2 bath process.

- 1rst wash with already used 99% IPA, to wash most of the uncured resin.

- 2nd wash with brand new or nearly brand new 99% IPA, to wash the most intricate uncured resin.

Link to post
Share on other sites
18 hours ago, plasticguy said:

 

- which IPA do you use? 

You'd better use 99% IPA.

 

And the most efficient process is called 2 bath process.

- 1rst wash with already used 99% IPA, to wash most of the uncured resin.

- 2nd wash with brand new or nearly brand new 99% IPA, to wash the most intricate uncured resin.

Yes, I use 99% IPA. But I find the results are a bit better with denatured alcohol, i.e. ethanol. It's something like 85-90% ethanol, IPA and 10% methanol

J

Link to post
Share on other sites

You might investigate an ultrasonic cleaner. I found a 2L one on ebay for ~$80. So far I have used denatured alcohol as the cleaning solvent with fantastic results. I've only used this setup outside as I'm slightly skeptical about the safety of that much open alcohol inside the parts washer inside a structure. 

Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...