Help with top layer

[umbral bane]

I've having issue on a large print, with the top layer in various spots on the print. I just printed a new one a few days ago, and it seems along the edges and there are 2 bridged areas where there is wrinkling. I looked at it more last night and it seems like it's being overextruded as I can see what looks like squish along the edges. I had asked about this a week or so ago and it was brought up that maybe my build plate is lifting from the bed. I have the same look at the middle of the part, so I don't think the bed is lifting. I've done an hour + heatsoaks to make sure the bed doesn't move any more. I've been curious if my bridge settinsg might need help as after the infil and before the top layer, the preview shows a lot of internal bridging.

Also something that's strange, I'm getting something along the bottom of my print. It somewhat looks like shifting, but it's not everywhere on the print. I have one feature that has this, and the next one to it doesn't. The feature is patterned in Fusion, and all the area are to be smooth. I can live with that for now, but it makes me wonder. This is the only model that I've seen this on. Not sure if it's just slicing incorrectly or it's not able to read the model right.

I have several pics to share and I'll try to explain them in groups. First will be the top layer that's the most concerning for me.

Here are the rings that I'm talking about near the bottom of my print. In one of the pics you can see it the rings in one cutout but not the next one to it. There are 4 of the same cutouts and only that one has the rings in it.

My printer is a Voron 2.4 350, LDO rev C kit. The material is Polymaker ABS, printing at 260c with the bed at 110c. I've tuned EM and PA although I can't recall those values off the top of my head. I'll get some pics of my settings when the boss gets home to turn my computer on. I'm using the default speeds from the Voron profile in Orca. During the last print I didn't have any wavying or warping of the first layer. I was actually too far away in 2 spots and adjusted it on the fly by .020 and didn't touch it after.

At a loss on where to look or what can be causing the top layer to be so bad......when there are spots where it's almost as smooth as glass!

[primal brook]

Are you running any kind of gantry backers, and are you doing your QGL and bed mesh after the heat soak?

[umbral bane]

I am running backers, 3 total (one on the X rail and 2 on the y rails). QGL runs after as I have a macro for my heatsoak. Machine will home all axis and go to bed center, heatsoak, and then I manaully start my print when the heatsoad is done. My print_start does QGL and bed mesh before it starts to print. Short answer to your questions, yes and yes.

[wise granite]

It sounds like you are doing the right things, yet it looks like your bed mesh isn't good enough. What are your bed mesh settings? Take a screenshot of your mesh in mainsail and share it here too.

[umbral bane]

I'll take one tonight, but I think this one was .16x range in difference. I'll heatsoak it and get a picture.

[primal brook]

Is your print surface smooth, or textured?

[wise granite]

I'm guessing it's bed mesh because your prints don't look like they are lifting but the top surface definitely looks too close. Only other thing I can think of is if the ironing flow is a bit high.

[umbral bane]

textured

[umbral bane]

Here is my process profile settings.

Material (Polymaker ABS Grey) settings

My printer material setting override. Under printer pretty much everything is stock, except I did change my printable area to X-1.0 (I can't physically move my head past or closer to the 0,0 origin beyond x-1.0 as the linear rail bearing will hit the XY joint)

[umbral bane]

According to my settings, I don't do any ironing. I've tempted to do it on a test piece to see exactly what it is and what it looks like while doing it, but I haven't seen a need for it.

[wise granite]

Oh wow, the main top surface was so smooth I assumed you were ironing it.

[umbral bane]

Here's a test I did a while back ago, tuning EM and PA.....no ironing.

[wise granite]

Interesting how uneven it is. Again, feels like bed mesh.

[primal brook]

Ok, as shown here, your first layer squish/z-offset is a little high, or you are under-extruding the first layer, or your bed mesh is less than ideal. Need to start by getting that dialed in.

[wise granite]

I think it's both, could use a touch more squish, but also looks uneven. Maybe just do a first layer test print, .2mm big square.

[umbral bane]

yup, I saw that while printing and I corrected it while it was printing, went down .020mm and it cleaned it up the rest of the print

On the top layer, something happened here as it is a different height. I can feel it with my fingers and you can visually see it. Don't know if that's a cause or effect of something.

[wise granite]

Getting a perfect bottom layer first will help you get a perfect top layer. And help us rule out bed mesh issues.

[umbral bane]

I have done the flow calibrations and I figured close to 20...I think it was like 19.8, but I put in 17.5 for volumetric flow rate.

[wise granite]

Lower it to 15 will probably fix that.

[umbral bane]

Done.

I'll print a big square tonight with a 1 hour heatsoak, get a pic of the heightmap and post.

[primal brook]

Which hotend?

[umbral bane]

I believe its the E3D Revo that comes with the LDO kit from West3D...... 0.4mm nozzle

[wise granite]

You may need to go even a touch lower, those don't flow very well. https://e3d-online.com/pages/revo-nozzle-maximum-flow-rates?srsltid=AfmBOop2iXqjutfxfO_4f2_n-eJGF_P-OtXTiPZu5c8Ar4_1QtdDIQXp

[umbral bane]

ok, I will get a high flow nozzle on the way. I did change it to 15. I had it at 14 before but then I did the Max Flowrate calibration in Orca and it was around 19.8mm....that's how I came up with 17.5....gave it more headroom.

[primal brook]

Yeah, that seems pretty high for a Revo, unless it is actually their HF version.

When you hit that 20, was that when the extruder starting skipping? Or where you got a flow drop off? The two can be at VERY different points.

For example, in this chart, (d) denotes where the flow dropoff started to exceed 5%, while (s) denotes when the extruder actually started skipping steps:

Note: grams is actually supposed to read "milligrams."

[wise granite]

I like to take my max flow rate tests and pull them apart, and I stop where I can feel the layer bonding dropping, which is usually much earlier than you can see degradation.

[umbral bane]

I believe it was when I hit 20 in the max flowrate calibration. I then measured up on the print and 19.8mm high was about the last good print all the way around. Using the calculations it gave me a number and I went more than 10% lower based upon the recommendations of the video I was watching. Around 20mm high, I stated to have skips in that layer height.

Here you can see some of the "skips". Unfortunately I don't have the print any more, but there were some on other sides at that same height.

[wise granite]

Looks like you did everything properly, sometimes you end up needing a bigger safety margin. If you still have that print, or don't mind making another, try pulling the layers apart and you will likely notice that part strength gets worse near the upper end. If you care about part strength take that into account as well, but for parts printed for looks only you may not care.

[umbral bane]

Noted.....I'll see how much time I have tonight, as I'm leaving town tomorrow for the weekend and won't be back till Sunday afternoon. I'll reprint that test. For now I'll leave my max volumetric flow at 15. (I had it at 14, and my speeds capped out at 195, so to get that exta 5mm is the only reason I ran that test and put in 17.5)

[primal brook]

So that's probably when the extruder started skipping. You were technically under-extruding long before that. It's one reason I don't like this test.

[umbral bane]

How would I test/meausre that?

[wise granite]

If you still have your max flow rate test you can break it apart and measure the width at different heights to see where the width starts falling off.

Some people extrude blobs at various speeds and weigh them. I don't personally use that technique as I fear that without the hotend pushing against anything it would be able to flow more, ruining the test.

[primal brook]

I did it by using a milligram scale and weighing the same length of filament at different flow rates, like Stefan/CNC Kitchen did here, but there may be other ways.
https://www.youtube.com/watch?v=0xRtypDjNvI

[umbral bane]

This is after a 97 minute heatsoak....a 60 and a 37 right after that so I could make it home.

[umbral bane]

So here's a 325x325 one layer square. I didn't touch anything for a while, but I could see lines. I dropped z down .020 for a bit. Seemed better but saw more lines. Dropped speed 10%...better for a bit then lines appeared. Dropped another 10% speed, better but still a few lines. I tried dropping z another .010 and started to see wrinkles in the 350,350 corner so went up .010. Still wrinkles so went up .010 and slowed down another 10%. After that, I didn't see any lines, but all along the top or back of the build plate, wrinkles all across the top. So I ended with 70% speed and -0.010 Z end stop.

The curling started when I removed the build plate from the printer and set it on the kitchen table. That's why the bubbling look.

So I'm thinking 2 things.....I think I need to increase the Z endstop a little more when I need to use those areas of the build plate. Second, the hot end just can't keep up. It catches up when it changes directions, but in the long runs, it creates gaps or lines.

[primal brook]

This almost looks like you are probing the bed mesh, but then it's not actually getting used during the printing. A 0.15 bed delta should be easy for the mesh to overcome.

Verify that the mesh is actually being activated for the print.

[umbral bane]

How do I do that?

[primal brook]

It would be a command somewhere in your G-code, probably in the PRINT_START macro.

Should be something like BED_MESH_PROFILE LOAD=<name>

Before that you need to make sure that the mesh is being stored too.

You can read here about the bed mesh commands that Klipper uses: https://www.klipper3d.org/G-Codes.html#bed_mesh

[umbral bane]

I use the Better Print_Start macro.....

I don't know how to paste the code to view in here

[primal brook]

Ah, yeah, here is where it probes the mesh and saves it, but then I don't see a command after that for actually loading it for use:

Add the line:
BED_MESH_PROFILE LOAD=default
right below that highlighted part.

Then try that first layer print again.

[umbral bane]

I know in the dashboard in Mainsail, above the Z axis position, it will say something like adaptiveXXXXXX and there's numbers above the Z axis position that move

[primal brook]

No clue what that stuff means, I have just never used adaptive.

[umbral bane]

Do you do a mesh calibrate before each print? I have been....and if it's a small print, it only probes around the area it's print instead of the whole bed

Will take a bit, the bed was cool when I got back to look at the print, so doing a heat soak again.

[wise granite]

You'll know if you are loading the mesh during the print if you go try to look at it while it's printing, it will only show the loaded mesh, if you don't see a mesh you don't have one loaded.

It sucks because it's slow but with a mesh that bad you may need more points. I mesh before every print now mainly because it only takes a couple of seconds, but back when I had a slower probe I saved meshes at different temps and plates and just reloaded them when using them, like if you have a "heatsoaked_textured_pei_60C" mesh it shouldn't change unless you make a mechanical change to the printer. That allows you to make a slower, finer mesh and not pay the time penalty each print.

[umbral bane]

So in the heightmap, you should see something in there, right?

heightmap in Mainsail I mean

[wise granite]

Yes, if it's loaded.

[umbral bane]

ok, yea, I always have something in there. It does change and it probes before each print.

[wise granite]

You're good then. If it's still not behaving well for a first layer then you most likely just need a finer mesh. You have really big changes over a small area at some points, without taking more measurements klipper doesnt' know enough to properly correct for the bed shape.

Do you know how to increase the number of points measured?

[umbral bane]

Not off the top of my head, no I don't. I've been using adaptive, but I'm sure there's away to still use that but increase the points?

[wise granite]

yeah, look for bed mesh in your printer.cfg, it should say 5 or 5x5, something like that, change it to 9x9 and see if you get a much better bottom layer. If that doesn't help it's not going to, if it does help but takes too long then you can decide whether to lower it to 7x7 or to save some meshes for later use.

[umbral bane]

I don't have anything that says that in my printer.cfg.

https://github.com/Grimmy1427/Voron_24_350configs/blob/main/printer.cfg

[wise granite]

You do.

[umbral bane]

doh....not sure why I didn't catch that

I'll change it to 9,9 after my heat soak macro is done...few more minutes

[wise granite]

It really need more info in places like this, 9,9 should help.

[umbral bane]

Should I take this BED_MESH_PROFILE LOAD=default out of my print_start? I added as MxBrnr mentioned to add it

[wise granite]

I think you already get a bed mesh loaded with or without it as you see a mesh in your mainsail page for it. Likely redundant but makes no different as far as I know.

[umbral bane]

kk, trying it now

I have mesh bounds of 40 and 310.....should I edit that to something like 20 and 330 to get closer to the edges of the build plate, since that's where the higher spots are?

[wise granite]

Get as close as you can, I have no idea where your sensor is but yes, get the probe as close to the full plate coverage as you can.

[umbral bane]

it's straight behind my nozzle

[wise granite]

I see, yeah, cover as much of the plate as you can.

[umbral bane]

it's the Omron square prox sensor

[wise granite]

I've never used one of those but shouldn't matter much.

That sensor might have some distance you are supposed to stay away from the edge of the plate but I wouldn't know if so.

[umbral bane]

I'll change the mesh limits to like x-10 to 340....give it 10mm on both edges, and Y to like 25 and 310 (I'll check if I have more room on Y when it does it, but to be safe for now)

[primal brook]

It's here:

Whoops, totally didn't see that you all had continued the conversation past that. Sorry, haha.

[umbral bane]

You're good man...it's getting late for me, but want to get this tested. I deleted all saved mesh's to make sure I'm not loading something I shouldn't be.....9x9 mesh takes FOREVER!!! 😢

[umbral bane]

Since I deleted the "default" mesh, when it called for it in the print start macro, after it meshed the whole bed, it faulted, so I commented that line out and rerunning

[wise granite]

Yeah, they suck. When I used to have to use fine meshes with slow probes I saved them by temp and plate and reused them. Now that I have a beacon probe I just scan on demand in a couple of seconds.

[umbral bane]

Does beacon require canbus?

I would like to do something like that, but I want to get my big project done first, and then I'll look into doing a mod like that

ok, finally! A 9x9 heightmap!

That's a lot more of a range than I've seen before

[primal brook]

Oh yeah, that's because the bigger mesh grid is getting closer to the edges where the big changes are at.

I bet if you did 11x11 it would be even more...but that will take like twice as long as the 9x9, haha. It will feel that way anyhow.

[umbral bane]

and I might do that for large prints.....how often will that happen, probably not very

Just an on the fly update.....I've had to lower everything .030 to get rid of the gaps, and I've upped my temp 5c and it seems to be pretty good.....haven't gotten to the corners yet

I'm going to assume the moving of the Z numbers that are in brackets indicated that the Z is moving up and down according to the mesh.

[wise granite]

It requires usb.

That plate is ok. I mean it's not great, but for a plate that big .2 isn't terrible. Basically, it's one layer, so should be compensated for, but you don't need to run out and buy a new bed.

[umbral bane]

It might get better, as I ran into a few issues where I had to restart, and the heaters turn off, and I turn them back on right away, but I still counted that as heat soaked.

[wise granite]

I mean, even if you buy a expensive bed from mandala roseworks or something like that I think .2 is all they guarantee for flatness. To get closer than that you'd need to do some work, either sanding the high points down or shimming the low point up say with aluminum foil. My bed is pretty similar and I just print with it like that.

[umbral bane]

so I don't want to jump the gun, but it looks like I'm getting wrinkling in the same spots, in the back of the build plate. In the console, the probing for the best mesh only went back to Y-284.959......the print is going back to Y-337.xxx so there's over 50mm that's not been probed.

[wise granite]

So what if you print in the best spot on the plate? Is that possible, I'm not sure how big your print is.

[umbral bane]

For the most part, yes. The large print is 346 in X and 245 in Y. I can move that around in Y but on X, I'm right there with just over 1mm to play with, as I can't move X in the minus direction back X-1.0 or the head hits the XY joint. Found that out the hard way.

I'm also thinking of upping my initial layer to .3mm from .25mm. I think my mesh is compensating (much better now due to expanding the scanning area) but a little more to help compensate.

The print is finished. I think it's better that the previous. Had to adjust squish down -.030 to get rid of the lines, but after that i put the speed back to 100% (105mm/s). I might be able to gondown a little more, but if I need to print way in the back, unless it probes back there, I think I'd get wrinkling from being too low.

I did get a few lines after I upped the speed, that's why I think I can add more squish....and maybe .3 layer heigh would help with that.

[umbral bane]

I've updated my bed_mesh settings to cover more, almost all of the bed that I can for felt comfortable with, doing it remotely and not physically being able to be in front of my printer. I'm hoping that should clean up the back portion of that 1st layer print. I won't be able to actually print it till Sunday night as I'll be out of town for the weekend. I'll also do the max flow and do the pull apart test on it to see where it lands and report back. I do have a high flow nozzle with a 60w heater on the way for next week to make sure.

[umbral bane]

It was a cold test, but got Y back to 340mm......happy with that!

[umbral bane]

Max volumetric test done

Noticed this on the front left corner

[umbral bane]

The height of the skips measures above 29mm. In the formula that's about 19.5 max flow, so the 17.5 I had in there was 10% lower.....I have since changed my Max Volumetric Speed to 15 (that does allow me to reach the 200mm infil speeds).

[umbral bane]

I checked this and I read .668 to .625mm and I stayed under where I saw some extruder skips in the print (under 29mm high). So not a whole lot of varience in line width....about .043mm from those measurements

[wise granite]

If you start pulling that apart where do the layers start getting weaker. I can't be sure but from the picture it almost looks like the layers stop bonding as well closer to 20.

[umbral bane]

I tried pulling it apart with just my hands and I couldn't. I can get it to break, but straight pulling on the top and bottom I couldn't.

[umbral bane]

Just tried another 1 layer square and across the back was wrinkly. I adjusted my z height within the first 25%.......went -.080. Going to try a .3 first layer.

[umbral bane]

Same problem corner but at .3 1st layer and considerably better!! Overall this print is the best of them all! Maybe the outside edges are just a bit much for the adaptive mesh to compensate for? I think the height map on this print was .28x.....the same corner is where it was the highest.

[wise granite]

Yeah, likely the mesh is struggling to compensate for that. If you are aiming for near perfection you could try to level your bed. Some people use aluminum foil or kaptop tape under the sheet to level it.

[umbral bane]

I think knowing what I do now, what to tweak for a print that goes back that far. I'll measure up the prints tonight for an approximate location to avoid/change layer height if I need to print there. I'm tempted to try my large print again to see if I have the wrinkling on the edges again. My high flow nozzle and 60w heater will be here tomorrow some time, which can only help.

[umbral bane]

printing now, putting down the first layer at .3 first layer height. Looking good so far, but not to the back part of the print yet.

This is the heightmap generated before the print started

[umbral bane]

This is the problem corner or the corner with the most issues.

Most of the whole print, 1st layer.

13.5 hours to go.

[wise granite]

That's a good start! Prints like that are tough, not only because of bed warping but sometimes I have to raise the bed temp 5 degrees because the edges of the bed don't usually reach the same temperature as the center 80% or so.

[umbral bane]

O know ow it doesn't really show anything but this is almost 13 hour in, 1 hour to go. Pretty much top surface. So hard to look at and see anything. It all looks good as of now, but 4-5 more layers to go. Makes me want to stay up and watch bit my body says no! LoL

I've never been able to watch the last layers so I going to try to stay up. I have noticed a shift in lower layers as I've mentioned in my first posts, bit I can live with those results. I'm curious why they show up though. Will pic and post results of this printer when I can.

[wise granite]

Video it if you can, might help you see what is going on if it repeats.

[umbral bane]

Tried to make a finish video. Everything is acceptable, but I do wonder why.

I took other videos, but they didn't post for some reason. I even subbed to Nitro to be able to post.

Here are some anomalies that I can't explain. It's ok for my project but I wonder why they are there when it should be smooth.

SO........I think my top surface issue is solved. I'm going to say it was the lack of probing the entire surface that I'm printing. I think increasing the 1st layer height to .3 from .25 greatly helped, but I think the adaptive mesh in that area helped even more. I will say I'm more than pleasant in the results. There are minor issues that I still ponder about, but in my project they are acceptable. I would like to hear your thoughts on the minor imperfections so possible corrections could be made.

[wise granite]

I'm stumped on that one. I wonder if someone else has ideas but it's not an issue I have run into before. It almost looks like underextrusion in some areas, and over in others, but that doesn't make a lot of sense. Just wildly guessing here, one thing that could cause that is varying height not just through your bed mesh, but maybe through your infill and top layers? Like if bridges were sagging in some areas and not in others. That seems like a longshot guess to me, but if so then adding another top layer or two might help. Hopefully someone else has more insight.

[umbral bane]

So just to update, I found my gantry was racked and the last print i didn't, while it didn't fit like I had hoped it should, it also wasn't square on opposite corners. I deracked it last night and I also installed a 60W coil and high flow nozzle. I'll run some test prints tonight, but I agree with you about the top layer and possible adding some more layers. I have a bridge that goes to the top layer so I'll have to adjust the model to add a little height for that. I think it will fix those issues.

I'm going to call this finished and completed. If I run into anything else, I'll make another post. Thanks for all the help guys! It's greatly appreciated!

[umbral bane]

So I'm back at this a little bit. Most of the wrinkles are gone and I'm 90% happy with it. The piece I'm printing gets covered up by a top piece anyways. It seems when my top layers start to print (I increased from 5 top layers to 7 top layers) I get some bulging at the top. In some small areas, it seems I'm over extruding and the surface is a little rough.

Also in the other picture, there's a line where it looks like there's a change in layer height but on the print just before this, it's more like it didn't adhear to the previous layer. This is visible in the exact same spot on the last 2 prints.

Looking for advice on what to adjust to clean these up. I've though of decreasing my top infill flow down from 1, but not sure how far to adjust. Would .9 be too much? The line, I'm not sure. I thought too much cooling, as I run my fan at 60%, but it's noticeable in a few other spots toward the end of the layer.

[wise granite]

Not sure about the line on the side, but both it and the top could be caused by just a touch of over-extrusion. Like 1-2%

[umbral bane]

Would it be best to drop my top infill flow from 1 to say .99 or .98 (the 1-2% you suggested)? I'm thinking of a way or a print to test this rather then trying the whole print (15 hours) over.

[wise granite]

Drop flow on the hole thing. For a quick test there is this quick test print a few of us have been working on, you can give it a go and post the results and I'll see if anything obvious jumps out at me. #1363262234767261777 message

[umbral bane]

I'll do that tonight! Thanks!

When you say drop on the whole thing.....drop the flow in the filament (I'm at .93xx for my gray ABS) settings? Not the top infill flow. Just want to be correct on this.

[wise granite]

I'd be tempted to try 92 for flow rate. Might be pressure advance issue though.

[umbral bane]

Here's the flow I calculated using the calibration prints.... .9394 flow

Here's .9294 flow.

Here's .9194.