Bed Leveling and Adhesion Struggles

Two questions here and I think they may somehow be interrelated so I thought I’d combine them in one thread. I’m having big time struggles with adhesion but only in certain spots on my bed. I tried telling Slic3r that by bed was only 125X125 so it would shift the “middle” over and the next print stuck like a dream. However a repeat of the same print (in the same offset location) seemed to have struggle sticking on the back corner areas. I had tried the same print to absolutely no success in the machine’s natiral middle (wit hthe bed set to 200X200).

Then last night after Googling for answers I saw a thread on RepRap that said not to use 70% Isopropyl - which I had been using. So in hopes that was my failure, went and got 95% Iso today and immediately saw the same skirt line not sticking in the same manner - no change. Darn.

I was originally trying to print a second LCD body (Saturday night) in the machine’s middle but no matter what I did it wouldn’t stick on the end sections or the loop around the screw hole. I bet I had at least ten aborted prints before finally giving up and hitting Google. I had already printed two more universal spool holder things but in both of them the north east corner lifted. I just let it print on - the bearing middle had stuck good and the top shape is all that REALLY matters on that print.

Now the bed leveling bit.

I’ve been using the piece of paper method thus far but today it occurred to me that I have feeler gauges in the garage I could use. Would this work for bed leveling? And if so, what gap should I be giving the nozzle at “0”? And do I do it hot or cold?

I am short on my time so i will give more details tomorrow. Many many factors play into bed adhesion. Bed material, surface coating, filament type and brand, bed temperature, bed level (including bed warping), ambient temperature, drafts (air blowing on your printer) and a few others. I will explain more details tomorrow. If you can please answer the following.

what is the surface material of your print bed? (Glass, metal etc.) What coatings tapes or glues are you using if any?

What type of filament are you using? (Type and And brand)

what temperatures are you using? Bed temp and extruded temp.


Thanks for digging in. It’s a PEI bed 3m’d to the one Ryan ships. I’ve been using Hatchbox PLA 1.75mm (white). Last night I used all kinds of combinations of temps but never lower than 60 on the bed - I took the bed up to 70 but then thought “PLA doesn’t even need a heated bed so why am I doing this?” And extrusion temps - 196 up to 210. I’m sure it’s a combination of bed leveling nozzle clearance (it’s level) and just bad spots here and there on the PEI. All attempts were fully preheated. My next attempt was going to be to hit the PEI with 1500 grit sand paper according to the RepRap posts. But I thought I’d ask here first.

I’m interested to see what Aaron has to say about it. Kelly’s bed is the aluminum mk3 (not mk53 from the mk3).

PEI needs heat, or at least it loves it. At 60C, it’s sticky AF, at 25C, it mostly lets go. Uneven hearing can also cause issues on the outside of the print bed, but I don’t know how to fix that.

A piece of paper is about 0.1mm thick, but the number doesn’t really matter. It’s a dial for you to turn to get it just right. You can go up or down to get it tuned just right. There are some settings you can enable for babystepping in configuration adv. They let you adjust the total Z height with fine adjustments after the prints start. You can watch the skirt at adjust it up or down to get it just right. It won’t fix unflat or untrammed beds though.

Have you tested with the paper near these trouble spots? Are you measuring just at the screws or in the middle as well?

Is there anything mechanically lifting the nozzle in these spots (filament spool pulling?)

That feels wrong to me. 1500 won’t do much damage, but IDK how that would help. I would worry it would end up even less flat. I haven’t tried that myself. I wouldn’t mind having a rougher surface finish though.

I wouldn’t hesitate to add a brim to fickle parts. Some designs just don’t have much footprint and it’s not worth fiddling with. 5mm is enough for the trouble prints. You add it in the slic3r settings. It can help with adhesion and warping.

The speed and acceleration have an impact as well. As the filament comes out in a corner, if you whip around it fast, it will pull it back up. Not much to do besides slowing down. AFAIK, there aren’t acceleration settings that only apply to the first layer, unfortunately.

If you have a good machine, and it’s printing well in the middle, it’s probably just not completely flat bed. To make it worse, the aluminum bed changes shape with temperature. There are two more complicated ways to fix that.

  1. Auto bed levelling. Add a Z probe, and map the height of the bed automatically before each print. You can add a sensor, wire it up, change the firmware, and then adjust the hell out of it. It’s still fiddly work, but theoretically, it can compensate for problems that otherwise couldn’t.

  2. Manual bed leveling. You can adjust a mesh manually, and the software will do the same compensation, but without the probe. You basically make the printer go to each X,Y in the grid, and manually adjust the height in Z with your feeler gauges at each point, which it saves in eeprom. I haven’t ever tried this myself.

In either of these cases, if your back 1/4 of the plate was 0.05mm lower, then whenever the nozzle was over the back 1/4, the Z would move down by 0.05mm. The compensation is in both dimensions, and it adjusts the offset in between the measured points to something reasonable. You should end up with something that is much flatter, and I can say that I’ve printed much larger items with leveling on. I can also say I’ve spent more time on leveling than anything else in the printer. I would wait until you’ve gotten through the other troubleshooting before adding these, but if you want to jump in. I’ve got a few dozen tips for you.

If start with the brim. And I’d leave the temps around 205C/60C for the first layer. Slic3r can adjust the nozzle temp after the first layer to 195C or whatever. I would also post one of your great vids with the trouble.

1 Like

Where to begin? I have been 3D printing for about 6 years. I have printed over 1500 parts on FDM type printers alone and another 700 on SLA printers. I have owned and worked with about a dozen different printer models. So most of what I am sharing here is from personal experience and what I have gleaned from research online.

In summary I suggest you try the following: Make sure once again the bed is level while it is warmed up. (The paper trick works well enough for this to get it level.) Like Jeffeb3 said use 60 for the bed temp. (I might even go down to 50 or 55 but never above 60). Adjust your Z homing height to be a little higher off the bed after the paper tick. Make sure your print bed doesn’t have cold spots. Make sure there are no drafts of air blowing onto your print bed. Or you could try my solution I explain at the end.

Now for some details on WHY bed adhesion fails or succeeds.

PEI is some really good stuff. I have had pretty good luck with it. However, I had better luck with it when I was using ABS. PLA I always had more problems with the first layer. Like Jeffeb3 mentioned PEI gets more sticky as it gets warmer. Problem with PLA is it gets more loose and melty as it gets warmer. At about 70 c it is soft enough that it can change shape if pressure is applied. So even it is “Sticks” some pressure could still pull or push it off the bed. So I don’t let the bed get hotter than 60 with PLA. As a contrast ABS doesn’t get soft until about 170 c. From my experience ABS stuck better on PEI for the first layer. ABS suffers from shrinkage so it could pry back up later in the print job. So bad news either way. I have tried dozens of other solutions like hairspray, glue sticks, tapes, slurries etc… I could speak to the pros and cons of each but I won’t waste our time. PEI is the second best solution I have found. The best is explained near the end.

Level bed is huge problem for many people. Use the paper trick to make sure the bed is all the distance from the nozzle. But then you need to set the nozzle distance from the bed as a different step. Here is my reasoning. There is no one size fits all for correct nozzle height when you switch filament types. PLA and ABS need to have the starting Z height just a little different to get best results. I don’t have a science reason for this but I know from experience that I needed to reset the z homing height every time I switch filament type for best results. So the paper trick is great at getting a level bed but not for setting the nozzle height.

Aluminum is more likely than glass to warp as it gets hot so the center of your bed may be higher up when it is hot. 200 mm x 200 mm at 60 c this shouldn’t be a big problem. But still I would do the paper trick while it is warm. When you are doing the paper bed leveling make sure you level the corners in the correct order.

Correct Order? If your bed is leveled with 4 screws (one in each corner) then do this order. Front left then back right. repeat this several times until they are both level. Then start working on Front Right and Back left. When they are level do the Front left and Back right again. keep repeating this until all 4 are at the same level. it can take time.

If your bed is held in place with only three screws this is a lot easier. Start with the edge where two screws are closest. For example if you have two screws in the front two corners of the bed and only one screw in the center back then start with the front. level the front left and then the front right. Stay in line with the screws don’t be checking the very front of the bed unless the screws are right there at the very front. If the screws are set back about an inch then check an inch back so you are in line with the screws. Go back and forth until they are equal. Then go to the back center near the rear screw. After that test all four corners. Last you can check the center of the bed to see if it is warped up or not. I primarily use a TAZ 5 with a huge print bed something like 270 x 270 (I don’t remember exactly). It does warp up in a little in the center when I get it up to 90 c or higher. I have a solution for that near the end of this.

Flow rate vs nozzle height. This is a big problem for many people that they just don’t realize. I hope I can explain it well enough to be helpful. If the nozzle is too high off the print bed it won’t physically smash the plastic into the print bed. So it won’t stick well. If the nozzle is too low it will smash it so flat that it oozes to the sides. It will stick a single bead to the bed pretty well but the next bead that tries to lay down next to it will either have problems sticking or it will push the first bead sideways and make it peel off. You can see this happen early with a brim sometimes. When you pull a printed part off your print bed you can also see the beads that made up the first layer. If you see the beads have small gaps between each other then the nozzle was too high. If the beads are all messy and overlapping then the nozzle was too low and too much material caused the beads to run together. But nozzle height isn’t the only factor here. Flow rate is the other side of this same coin. It is possible for your nozzle height to be perfect but your printer could push out too much or too little plastic at any given moment. This is more noticeable when you have a Filament that isn’t a consistent thickness. Some brands are better than others.

Here are some photos that show these problems. This first one shows the messy lines from too much plastic getting smashed down into too small of a space. You can also see how the bed wasn’t level because the plastic is even more smashed together on one side than the other.

[attachment file=74653]

This one was about perfect. Printed on PEI with some damaged spots perfect flow rate and homing height. Almost all perfect lines with a few spots that you can see the filament diameter was inconsistent so there was more or less plastic extruded. This is about as good as you can get without a raft.

[attachment file=74654]


Briefly I will say that air moving across your print bed can cause problems. Cold drafts especially. Do what you can to minimize these. It is possible but not likely that your print bed can have cold spots. Aluminum usually normalizes this so you probably don’t need to worry. I have had some heat beds that are hotter in the back than the front or there is a dead spot near a corner.


Moving onto my best suggestion for solving print bed adhesion. I have been doing this for over a year and 99.9 % of the time my prints stick to the bed first time and stay stuck through out the print. Larger ABS prints still tend to warp a little but most of that is reduced as well.

In short I use these printer bed adhesive pads and I also print using a raft every single time. Before you buy any of these I hope Ryan can do something so he can get a small kickback if you click the link and buy it. I don’t know how that works but I would prefer to have Ryan get a kick back for a product I recommend on amazon. Here is my review of this product on amazon.

So this stuff works great for bed adhesion. But you still have to get your Z homing height correct. Too close you can stick so well that you actually rip these pads. Too far and you still have problems. Again different material types still need a slightly different z homing height because of the flow rate vs nozzle height issues. So this is how I solved the problem. I use a raft for all of my prints now. The raft lays down a few really Thick beads that get smashed down onto the print bed so they stick well. But then it leaves a space between each of them so they don’t knock each other loose. This means my z homing height has more flexibility or tolerance. I could be a little off (too high or too close) and it will still work. My bed could be slightly out of level. My bed could be slightly warped. I haven’t had to level my print bed or adjust my z homing height in over 8 months and I switch between filament materials all the time. Now there are some settings for the raft that you may need to adjust. I am using CURA as my slicer and some settings can make it create a space between the top layer of the raft and the bottom of my printed part. But usually the default settings are good so I don’t have to mess with them much. If you have questions feel free to ask.

Here is a picture of a part that was printed on a raft. The beads are all just about perfect. I used a black marker to make them visible on the light colored plastic. The small circle of plastic is actually the bottom of the raft for another part I just printed a few minutes ago. You can see the thick smashed beads and the space between them.

[attachment file=74655]

There are so many things that I have tried and learned from. This is the best solution that I can recommend. I mostly use ABS, PLA, PETG and HIPS but this has also worked well with Laywood, t-glase, Nylon, N-Vent PVA and a more that I can’t think of at the moment. Nylon did damage these pads a little as can be seen in the pictures attached to my amazon review.

Thank you for your patience. I have tried to write this over about 5 hours but I keep getting distracted with other tasks that require my attention.


Wow. Thanks hardly goes far enough here. I really appreciate you taking the time to write that all out.

I went on a whim before reading this but it seems my intuition was in line with your suspicions. I have the three screw setup and although I was good and “paper trick” levelled it was still a bit off I think. I figured the side with the two screws was maybe just marginally lower so I turned the single screw down a half of a turn and adjusted the z homing accordingly. I started a print (that I probably shouldn’t have) and it seems to have laid down nicely. It’s still printing the base layer and has been going for 30 minutes - it’s halfway through the infill but all appears to be sticking well. I’m hearing some clicks from the extruder every so often and seeing some misses but nothing terrible. I wonder how much of that is inconsistent filament diameter? In any case, that’s a whole other conversation.

I picked up another caliper today - a bit better constructed than the other one and with a lock screw so I can get reliable readings and lock them down so I don’t bump it taking it to and from the machine. I’ll have a good look at everything when this print is off the bed. It was an MPCNC kinda day here but I’m back in the office now working the other toy.

I am glad to be of help.

Clicks in the extruder could be caused when the extruder is slipping. Meaning it can’t push the filament out fast enough so the teeth slip on the filament. Two possible causes. You might have your speeds set to high or the nozzle temp might actually be too low. Keep in mind that PLA can only get so hot before it turns into a glue like substance. Most brands start doing that around 225 or 230 from what I have seen. So don’t let it get that hot. Even when you are changing from PLA to ABS. Set the temp to something like 210 so you can slowly push ABS through to clean out the PLA. When you think it is clean then you can raise the temp for the ABS up. This will help avoid an extruder jam.

I have to say - I had one large footprint print (150mm circle base) adhere beautifully and according tot the above the bottom surface looked magnificent (I’ll get a pic of it later and post it). However as you’ll see in the video here - this is typical of the struggles I’ve been having with adhesion. You’ll see at the end of the clip when it’s filling in around the right tube opening - the tube opening lifts - it’s only partially stuck although it appeared to have laid down fine. Anything obvious jump out at anyone? The bed it level and clean and true (to the best of my knowledge).

The video really helps.

First thing that jumps out at me is the fan. It looks like it is on even though you are on the first layer. I can see some fibers blowing in the wind. Usually it is best to keep the fan off for the first few layers. You will notice in the video all of the spots that are pealed up are facing the fan. In the video at the 39 second mark you can see the fresh bead curl up (starting on the side facing the fan) from the fan blowing on it. Keep the fan off for the first 2 mm or 5 layers. Try that first.

Next try slowing down the first layer.

That should do it but if not then try moving the Z up just a tad. A tiny bit. Like a quarter of a turn on the screw. I say that because you can see how the beads lay down next to each other. Near the edges you can see that the top surface is made up sharp blade like lines. That is caused by the plastic coming out of the nozzle then oozing out to the sides and back up the side of the nozzle a little bit. Raising the nozzle will make it so the beads lay next to each other with a flat smooth top. But your Z height is really close to where it should be so the first two suggestions should solve the problem.

I’ll try slowing things down. How much do you think? I’m not sure if slowing down is like 10% reduction or 50% reduction. I also assume that Slic3r figures out the extruder rate based on the print speed?

The fan you hear and the blowing you see is the nozzle fan. Is there a way to disable that for the first few layers? The only control I’ve found is the print fan. Oddly, the lifting edge you saw riffling in the breeze was actually then facing away from the fan.

I quite routinely see a bead go down - like in the skirt for example…it all appears to stick but then the next pass around dislodges it. Same with first layer perimeters. Sometimes they stick and it’s the infill that dislodges things, sometimes the final pass lets go. The thing I find odd is how it seems like it’s always releasing in the same sorta spots. The back left corner of the prints or those oddball spots right around the middle of the bed. I suspected it was the isopropyl I had been using (70%) but I’m now using 95% and really rinsing down the bed even between each adhesion failure.

You shouldn’t have to clean it that often.

The heat sink should be oriented so the find are horizontal. That will make the air blow away from the prints.

[attachment file=75377]

My default settings I use have the first layer running at 50% of the speed of the rest of the print. Since most failed prints are caused by the first layer failing it is worth the time to get it right. Yes, your slicer will slow down the extrusion to match the speed of the print.

Slightly confused by which fan is on now. If it is the fan in Jeff’s picture then he is correct. The fins should go horizontal so the air goes out the sides. (Also I prefer to have the fan pull the hot air through the fins and throw it straight out in one stream. I have a few “Scientific” reasons for this but I don’t know if they actually work.) You can’t control that one. It should automatically turn on when the extruder is hot. (Or on some machines it is just on all the time.) If the fan that was blowing in the video was the other fan that points down toward the nozzle then yes you can control that one. It is usually in the “Cooling” section in your slicer software and there should be a way to keep it off for the first few layers.

As a side thought you may want to google to see if you can find some default material setting profiles for your slicer program. I know they are out there but haven’t looked in so long I don’t know where. essentially these are saved profiles that people have found to work with that brand of PLA or something. That should include all the recommended speeds, temperatures, fan settings etc.

A few things can cause that. Nozzle height vs flow rate is the biggest contributor to the problem I have seen. Because the first beads are stuck on the bed then they get dislodged by the next beads that are trying to take up more space than they should. So they get pushed to the side. In your slicer you may want to set the first layer thickness down a little. I have seen %400 work but I used to have good results between %150 and %200. OR you could move your nozzle up but in the video it looked pretty good to me.

The next most likely reason for the dislodge is temperature fluctuations. Any wind on the first few layers can cause it to curl or cause the PEI to become less “sticky”.


I really think the fan is the problem. One test you can do is to let it only make one circle, and just push it with your finger. With a hot bed, and recently extruded plastic, it should stick pretty well. If it just pops off, then something is wrong with the chemistry (probably temperature).

Do you think I can flip the heat sink and secure it with only one screw and still get enough contact to the body for cooling? Or should I try to drill another screw hole through it? It only has two so in order to mount it with both screws the fins are vertical.

[attachment file=75442]
[attachment file=75443]

A piece of tape across the bottom works. I think I have that in the build instructions.

By that you mean mount it vertically with both screws but then seal off the heatsink bottom with a piece of tape?

You can drill another hole or put tape on the lower side of the heat exchanger fins - this will force the vast majority of the air up. It will likely also reduce the efficiency of your heat exchanger, but for testing purposes, it may be easier than drilling the extra hole.

1 Like

Speeds are currently set at 30 mm/s for perimeters and infill and the first layer is set at 50%. External perimeters are set at 80%. I’ll try bringing them down to 25mm/s and get the heatsink fan breeze tended to and report back.

Is it bad practice to clean the bed before each print - like does the ISO gradually degrade the outer surface of the PEI so I’d be better off only cleaning when it’s deadly obvious it needs cleaning?

I clean the bed several times a day before every print. No noticeable changes over the years. Once and a while I use Windex a magic eraser and it seems to go back to normal.

Yes tape off the bottom to force the air up. Most don’t have an issue but some do for some reason.I print with the bed and PEI at 63C.

1 Like

I wonder why mine has different holes?

[attachment file=75466]