The Secret to a flat 3d printer bed 

Not sure if "truly enjoyed" was dripping with sarcasm or not (it should be), but I was very disappointed with the initial outcome and spent a week procrastinating removing the magnetic sheet that I had applied incorrectly. Hopefully others will gain from this. I never thought about large multi zone beds, but this would be useful. In the future, I plan on having the printer wait until the top probe hits the correct temps prior to starting the print. But that is for another day. I originally thought it was a gimmick, but it's turning out to be very useful. Thank you for watching!

Thanks! Believe it or not, I just consider this "ball park". I was targeting < 0.05 variance. So close, but not quite there.

@@KapmansBasementWorkshop to finish at .05, the build plate would have to be much more precise, then each layer brings it's own imprecision: plate, glue, magnet, flex plate. after all .05 is only 50 micron! , I guess you would still need a bed mesh to print on the whole surface and have anice and shinny bottom

Great to hear! Thank you for watching!

I think the biggest issue here was actually me and the way I assembled the bed. I'm pretty sure many others both before and after me have flat bed issues that are related to how the bed was assembled. Hopefully this will help future builders as well as others that need to fix their beds. Thank you and thanks for watching!

It's a beautiful plate. Very well made. The endstops are a game changer! Make sure that magnet is cured on the flattest surface you can find! Thank you for watching!

Thank you! In this case, I actually think the original magnet was fine. The problem was me and how I attached the magnet. Ad long as the adhesive cures under pressure on a very flat surface, this should come out great. This was purely my mistake. Up until this point, I failed to see the need for beds with embedded magnets, but they suddenly seem like they are worth trying. Thank you for watching!

The make beautiful beds. I can attest to their quality. They did not send me that to review, I paid for it.

It's a beautifully machined bed plate! Thank you for watching!

I hope you find some of the others useful as well. Thank you very much!

So that's very interesting. I've wondered multiple times about doing this. Over time, I seem to get changes in my bed mesh and I'm convinced it related to the glue. If I may ask, how big is your print bed? My main concern about doing this has been about if the clips are strong enough on a 250mm bed.

@@KapmansBasementWorkshop 235mm, it’s on an Ender 3 V2. I still have a little remaining glue around the outside which I guess helps and I almost never use the entire print bed. I’m actually thinking of moving to a Voron 0

I hope so. It's actually something I was wondering about for some time. So it was a pain, but I'll always do this correctly next time. And, no more bed mesh. Thank you!

Glad I could help! So here comes a shameless plug...subscribe to my channel and browse around. There is a pretty long list of videos that could provide some help/tips. If you are building from a kit, I think these recommendations in this video will help make that plate better/flatter. That said, Nero and Steve Builds are quite popular.

@@KapmansBasementWorkshop My 2.4R2 300mm is finished and I have a variance < 0.09. My unklicky probe is allowed a 0.003 deviation. Not much worse than you, thanks! [Actually it's only one point that's 0.05 lower than the rest, at X30 Y150.]

@@iniqy Thank you for reporting back. That's great to hear and I'm glad you've now got a flat bed! Most importantly... congratulations on the new printer! Have fun with it!

Funny thing. I actually had a tube of thermal paste sitting on the work bench as I had full intentions of using it. But, yet another mistake, I forgot to use it as I was so busy with everything else, especially shooting video....I would recommend the use of thermal paste. That said, it's working petty well without it. Best of luck with your improvements. The key to this video (and you seem to recognize it, but some are not) is the flattest surface you can find to make sure that magnetic plate is flat. Thank you so much for watching!

Thank you for watching!

I meant to say that I'd be surprised if the thickness of the magnetic sheet is any better than plus or minus 20 microns which is what you're measuring compared to the bare plate.

I really don't think it's the aluminum bed. My guess is that it's the most accurate thing on the printer (hot or cold). I'm pretty sure its more of a factor of the magnet, the quality of the flexible plate and the gantry itself. In another video I switched to a single rail on the x-axis and the bed magically got flatter as well. Thanks for watching!

I really need to try garolite soon. It's a really nice bed. They make great parts. Thanks!

I did not enjoy removing the bedplate :). I do intend to go back to the original, but I think I may build a second printer and use that plate. Or, get the tools to properly measure that plate for flatness and add a new magnet. I'm convinced it's also very flat, but the magnetic sheet was the problem. It's a great idea for a part two, I should have thought of that! Thank you!

Thanks for the kind words!

Ha! If you like suffering. Thank you for watching and the comment!

It's a really nice bed. I would not buy another without it!

@@KapmansBasementWorkshop I couldn’t agree more. I definitely intend to upgrade and I think this one’s it.

Thank you! A really good question.. I don't use SuperSlicer and I could not find documentation that could explain it to me clearly. But I think it's relatively easy in Klipper to get the software to heat the bed until the temperature of the top sensor hits the correct value and ignore the heater sensor as it normally operates.

I'm digging out and restarting an old 3d printer I stopped using once my Voron was running. It have one of those old heat pads. I'm going to have a look in the next few weeks to see how well it does. This is a good thought. Thanks for watching!

Also a feeler gauge has another use in 3D printing and that is to use it for bed leveling rather than a piece a paper.

You are braver than I!

Great point! At some point, I want to make another video about this as some of it could be better, but I think there is more I have been learning over time. Thanks for watching.

That just might work. I actually thought about something similar AFTER I did all this work. I was going to take out the back panel, unscrew the bed and try to flip it out the back onto glass. I think the cables I have are too short to support that. One other thing. Look into gantry backers. The curve you see are actually a combination of things. One of those being the dissimilar metals on the gantry curving in the heat.

Have a look at the 1st tip in the video: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE--DcdVez_v5Q.html Thanks for the comment!

I didn't. I believe any bed sold for a Voron is pre-drilled and tapped. Both these beds were. Thank you!

I fully agree. Hopefully others will gain from seeing my mistakes.

Have you setup QGL and Bed Mesh to avoid those magnetic spots? They are not flat due to the pull of the very strong magnets on the steel sheet.

Thanks a really good question. I would wonder about it's ability to transfer heat, my cast iron pan takes a long time to heat up. And it would need to be cast iron for stability with the temperature changes. I suspect it would work. Interesting idea.

I think it's a common belief, and I've never seen any specifications for the sheets. But I actually believe it's the glue that's used that has the biggest impact. Nothing really wrong with the glue, but making sure the glue sets while pressure is applied on the flattest surface possible is really the trick. I'm convinced with a little more pressure and applying the magnetic sheet before the heater on a very flat surface should make up for much of any inconsistencies that exist in the magnet.

My apologies, but I don't remember. However, have a look at this: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-i8a2-Jd-dIo.html I eliminated one of the MGN9 rails and the bed appeared to get flatter due to less heat expansion in the gantry. The measurements (before and after) are presented in the video. It would appear to me that backer plates could improve that even more. Thank you for watching and the kind words!

Thanks for posting this. It appears to be a great idea. However, I think you will understand when I state that I hope to never do this again! But if I do, I will try PAM!

I think it is the reason. But it's not cheap....thanks for watching!

My third Voron (and first kit Voron) was a 0.1 that included an upgraded Mandala Roseworks bed with embedded magnets. Once I saw it and thought about it, I bought a new Roseworks bed for my 2.4 and I do NOT regret it. It works really well and I’m able to consistently lay down a good first layer without bed mesh. I do wish I had a thermistor on top, but I calibrate temp by pressing my DMM’s heavily-thermogreased thermocouple into the bed and correlating thermistor reading to actual bed-top temp.

If you are using a high quality rail such as a Hiwin, try the single MGN9 rail, take the bottom one off. Have a look at this: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-i8a2-Jd-dIo.html

@@KapmansBasementWorkshop I took it off today and low and behold bedmesh working the way its meant to be and getting a near perfect squish on the first layer across the entire bed. guess I will be ordering that mgn12 after all since my current mgn9s have play in them. variance came down to 0.112 as well.

@@lxc647 I'm glad to hear you have this working well now! In general, high quality rails are not considered important for 3d printers, but for the x-axis, I would recommend a high quality rail like the Hiwin with some preload. They can be ordered with various levels of preload which will tighten things up on that rail and eliminate any chance for play.

The bed, since it's cast aluminum and not rolled or forged, really doesn't change when heating. There are some very minor changes, but I believe it's due to the glue and magnetic sheet as far as I can tell.

I think it depends on your needs. If your bed is 0.08 and you are having no problems. No need to change. Yes, that was terrible! Thank you!

Based on my observations starting with the worst (no actual tolerances): 1. Magnetic sheet adhesive 2. Rough PEI print surface 3. Magnet 4. Gantry

And there was enough glue left over to stick it to the new bed?

@@KapmansBasementWorkshop No, I got a new sticker off of aliexpress. same adhesive.

@@M1America Thats good to know. This may impact what I do with my old bed. Thank for these details and thank you for watching!

It does compensate, but your print will still have it's shape impacted by the shape of the bed. So it's always better for the bed to be as close to perfect as possible. Bed mesh will help you have a successful print, but now your print has not technically "correct". On some high precision (as good as precision can be with heated plastic) prints, this can be a problem depending how far off the bed is as well as the bed shape.

They did have a option for increased accuracy at some point. But honestly, I'm not sure it's needed. The magnet and glue we attach to the bed will most likely negate the benefit of the increased accuracy. Those are already not as accurate as the bed plate. That said, it's a beautiful bed.

@@KapmansBasementWorkshop I really like the idea of a top mounted bed sensor. And you're right about the build plate negating any accuracy gained by surface grinding.

I plan on purchasing all my future beds from them. It's a great bed. I have the heating rate set to 80 vs the usual 60 that is recommended. The bed has not changed shape at all. And the bed temps at the surface is really nice, but the built in bed plate stop is a huge benefit. It makes me want to replace my other beds just for that feature.

Those back stops are fantastic. I have another printer that lacks them. These really make a difference. I never even considered WD40. That probably would have worked great. Thank you for the tip!

Are you referring to the temperature sensor just under the magnet? It's a very narrow slot. Once the steel sheet is sitting on the top, I don't this it will impact anything, but it might just to that small area. But also, I just don't measure that spot. Thank you!

What 3d printer brand is it? I've been thinking about testing that out.

That would probably end up being pretty flat. That's my guess anyway. Thanks!

I don't have a referral code. But I highly recommend the build plate from Precision Printer Parts. The same plate works for both printers. Wait until you open the box and see the plate 1st hand.

No, the trick is how you apply the magnetic sheet and let it cure. It will cure in the shape of the surface used. I used a new bed plate to start with a known quality of flatness. An effort to rule that out of the discussion as I was suspect of my original plate. Turns out it was not the plate.

In hind sight, I wish I tried that. If you do try it, please let us know what happens. I'm 60% sure it will work. One major thing to consider, will the heat cause the glass to shatter? It might make sense to take the bed out and lay it magnet side down cold and then turn on the heater. It will heat up slower and allow the glass to warm up slowly. Then figure out what weight can be put on the back of the plate that's ok with the heat. The glass table in the video is actually very expensive. If I broke it, contending with my significant other would easily have been the most difficult part of the project :)

@@KapmansBasementWorkshop I tried this, and it didn't seem to work. Next, I will try a stone rolling pin at 120.

@@tonyrouse6864 Not a bad idea. I should have given that a try. Depending how badly it is warped and the shape, you might want to try a slice of paper on top of the magnetic sheet, under the PEI to raise a low area....

I was wondering if that could be done. But I did not consider it until after I did this. But thank you very much for verifying this works. I've noticed over time, it seems to change and this technique could get it back to where it needs to be. And I agree. I have noticed not all steel plates are of the same quality. Especially the ones with the rough pattern PEI. Thank you as I'll be using your trick every once in a while.

If it does, I could not smell it. It's a bit of a mystery what's inside Goo-Gone...

@@KapmansBasementWorkshop I actually went to get acetone the other day and saw Goo-Gone. One of them definitely showed it having acetone. The other types didn't really mention it so maybe yours doesn't.

@@jackshett I could not find acetone that week. So Goo-Gone it became. Does the acetone work well? I was told online to use alcohol, and it did nothing. It might have even made the glue harder!

@@KapmansBasementWorkshop Acetone does work, but it is probably about as effective maybe slightly stronger.

@@jackshett Probably stronger. But the fumes.....

I've never been that lucky. But, If the bed isn't level, you run the risk of part of the 1st layer not sticking to the bed. Thank you for watching and the commentts!

Thats a great tip. Thanks!

Possibly. But I have never tried it. I also thought about polishing with the finest sanding paper possible on a very flat surface. But before you go too far, be sure to check your gantry 1st.

That is the one thing I wish I had tried 1st. But the wiring was problematic (way too short). I still have my old bed. Someday soon, I plan on hooking it up to a power source and a micro-controller and giving it a try.

@@KapmansBasementWorkshop You can get stand-alone PID controllers (capable of switching AC power, even) for about $50 on Amazon.

@@benbencom That's good to know. I had no idea. I suppose the $50 is worth it if it works and eliminates the need to remove all that glue.

A bed mesh will certainly help. But have a look at those gantry backers. I don't use them yet, but people I know claim they help alot.

That is a good point and worth reminding people about 100% power. I think the beds that come with most of the known kits are good beds. By using this new one, I was trying to take the aluminum bed out of the equation and limit the issue to something else and make sure people would not point to my bed provided in the kit as the source of the problem. I think the video succeeded there, but I'm second guessing how I was able to communicate that. Thank you for watching!

I've seen it done. You need clips, but it certainly works.

Really good question. The mesh doesn't fix everything and It would be nice to be able to print without a mesh. My original goal was to see if a flatter bed could provide more consistency and more importantly to see what a flatter bed would bring to the printing experience compared to many of the beds that come part of a kit. I can't answer that yet as I don't have enough time with the flatter bed. I would assume for many, 0.16mm vs 0.07mm just isn't important and I can easily understand why. I ended up publishing this video as a help for people building a new bed/printer so they can hopefully avoid the mistakes I made and have a better bed using the parts they purchased. I also hope for those that are having bed problems that mesh isn't fixing, some of these tips can be used to improve the bed existing in a cheap manner.

​@@KapmansBasementWorkshop Well, its very good information and something I wouldn't have thought would impact the bed flatness. Definitely good info for people building Voron's. If I'm ever preparing a bed again ill take care to apply the magnet on as flat a surface as possible. I ask because I'm the type to also go down rabbit holes. I guess most attracted to the Voron are those who like fiddly things. But I also try not to follow my natural tendency to push fiddly types of inquiry or tweaking past the point they actually produce a net benefit. Its a hard line to recognize at times, and even harder to respect. my bed fluctuates between .12 and .25 depending on bed temp. But I find the bed mesh takes care of it well enough. I probe with one reading (dip down) as my PN-08 is surprisingly accurate and repeatable. That combined with high speed for movement leads to fast probing. Ohhh, also, you MUST setup adaptive bed probing. that is brilliant and saves time for small prints and allows you to use a higher resolution mesh. check that out for sure. Enjoyed watching you spiral ;) Thanks for the info !!! github.com/Frix-x/klipper-voron-V2/blob/main/macros/probing/bed_mesh.cfg

@@ThePhilbox Thank you. This was definitely a rabbit hole. This was supposed to be quick and easy, but when I did not get the results I was expecting, I just had to get to the bottom of it. Adaptive bed mesh looks like a fantastic idea. I considered it. But now, with this new setup, I've been printing for days and getting great prints, even the large ones over 300mm wide. All with no bed mesh. So it was a pain up front, but no more bed mesh at all! In the end I'm pleased. Thank you for watching!

My original variance with the same kit was 0.31-0.38 depending on temperature and how long it sat. Even with the mesh I had print issues and according to the Voron team going over 250mm leads to more complications printing which I believe is a result of the issues covered in this video. I originally went to the Voron discord for help with my variance after a few prints failed to stick well. The general conclusion was the X rail was bowed due to the dual rails. I had used a 0.0001" graduated dial test indicator to ensure parallelism between them and as a result had extremely smooth movement indicating the Hiwin rails were well made so I rejected their conclusions and went with the same bed in the video. Needless to say my X rails were fine.

@@ChrisHarmon1 Fair enough....I have a magic studio plate with .15-.3 variance. Bed mesh seems to work fine for me. But have still had to chase issues like thermal drift of the probe, and thermal expansion of the frame/gantry. Once I started heating with the probe over the build plate (recent change) all my problems were pretty much solved. Now if I'm doing back to back long prints just have to remember to drop it down a tad to fight thermal expansion.

I would argue that a bed mesh tries to compensate for a bed that isn't flat and give you reasonable prints. It cannot make a bed or your print flat. If you print a large object, even with bed mesh, it will still have a variance of 0.5215 on the side that was on the bed. For most objects, this may not matter so you should be fine. But if you need the print to be better than that, you need a flatter bed. That said, a flatter bed plate actually didn't fix the situation. It was the application of the magnetic sheet that was the issue. Thanks for the comment!

@@KapmansBasementWorkshop ahh I see exactly. If you print a 250x250x250 cube, the bottom will be wavy, and if those cubes need to be perfectly fit together, they wouldn't. Great explanation and great video!

Exactly. One other thing, it also just bothered me that I switched plates and my results were worse. I needed to understand why. Hopefully this will help people when building their next printer. Thanks!

I ran measurements both hot and cold. I wanted to see the difference between hot and cold and how the magnet sheet impacted that flatness.

Next time, I'll use it. At the time of the video, I could not get acetone locally, so I used Goo Gone. Thank you for the tip and thank you for watching!

You should. Thats good!

About $15 for 400x400

That is a good idea. That would soften the entire surface in one go. I should have thought of that! Thank you for watching and posting!

You bed should not be changing that much from heating if it's a good cast aluminum bed. I resorted to this after fixing other things, especially the gantry. Each fix seemed to improve the bed mesh. There really wasn't anything left to fix/adjust. So this bed when hot, differed from the cold by < 0.024 with the final magnet attached and a 1 hour heat soak. Now that still might not represent the final results because the printer's side panels were removed, so the interior of the printer was cooler than usual. I probably should post a follow up at some time, but I'm currently making some additional changes. I'm glad the single mgn9 rail worked for you as it did for me. Finally, Thank you for the great comments and watching. I really like your sense of humor!

Thanks for the comments. I'm not convinced the cardboard did much other than distributing weight evenly But, you got me thinking....

Believe me, I know the feeling! That was the start of this video. How many rails on the x-axis? MNG9 or MGN12 or something else altogether?

2 Rails mgn9 .. i think i build a new gantry with one mgn 9 and the stealthburner.. or i go direct to a carbon fiber x axis with mgn 9 , lgx lite and stealthburner…

@@rsmaster5637 In case you didn't see it: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-i8a2-Jd-dIo.html and the stl for single rail stealthburner: github.com/Kapman1/StealthBurner-Single-MGN9-Carriage You can see the before and after of the single mgn9 rail and it's impact on the bed flatness (it improved). Also, if you go with a single mgn9, make sure it's a high quality rail such as Hiwin. I've been looking at carbon fiber square tubing. I've received 3. None of them are as straight and flat as the aluminum extrusion. 2 of them were very far off. So I'm still on aluminum. Finally, others have told me what a big difference backer plates have made to the flatness of their bed. So you might want to look into that for the x & y on your gantry..

Yes the backer plates are also in my list.. someone used old linear rails as backers on the gantry … if i go for a single mgn9 hiwin would be the only choice.. i dont understand why some of the new single rail revisions of the voron are based on an mgn12. Maybe if the gantry moves fast from the front to the back you get more wobble on the nozzle due to less stability.. if the center of gravity of the printhead is to far away from the rail then the small rail could get an issue.

@@rsmaster5637 I suppose so. But so far, I have not had any issues.

If you are referring to a difference in your prints, I think it depends on how much the bed is off. Especially on large pieces. But the way it was installed did make a measurable difference. There is a aspect of OCD here, but to some it matters. Thanks for the comment and watching!

The smell should go away in a month or so of use. But in my case it was very strong. And then the odor disappeared. I ended up building nevermore and it seemed to help during that period.

@@KapmansBasementWorkshop , Thank you very much for the answer, I am thinking of removing the magentism for that reason and for not being as flat as aluminum alone. But in addition another reason is that I have found the Garolite (FR-4) plaque much better than the PEI, in addition the steel plate can also be used without magnetism, that is to say, hold it with tweezers.

@@valdolandar534 My next bed probably will not have a magnetic sheet any more. I've been planning to try out Garolite for some time now. Does it withstand the heat for ABS printing?

@@KapmansBasementWorkshop, The Garolite that I use comes from Ali-Expres, Named as FR4, and works perfectly, the maximum temperature I have tried is 130 degrees Celsius at first Layer. The adhesion between the Garolite and the ABS is much more than with the PEI. If in the end one does not want to fold the garolite or force to take off it is necessary to wait until lower than 50 degrees Celsius. Another advantage of the Garolite is that it can be sanded with fine grain sandpaper and thus restores possible scratches. The sanding fine grain adds power of adhesion too.

Have you tried this?

@@KapmansBasementWorkshop Not yet, but I'll be doing it on my son's CR-10 when I'm done modding this Ender 7. I've been trying for months to think of a reason that this might be a bad idea, and I can't. As long as the printer is on a level surface and it is applied super-thin ("squeegee" it with a razor blade), it should work. Some of those resins are good for 275-300 Celsius, so bed temps won't be an issue.

@@ApexNick I'd love to hear back about how this worked. It sounds reasonable and I can't imagine why it won't work. But the permanence is something I need to mentally get over. Thanks for watching!

@@KapmansBasementWorkshop Agreed, though I can't imagine much actual damage coming from it as either way it will be covered (spring steel, pei, glass). An idea I need to think on is some sort of strong magnets (but not strong enough to affect the electronics) glued on the underside of the platform so we can use those magnetic build sheets without the adhesive backed magnetic sheets.

@@ApexNick There are companies that sell beds with embedded magnets. I wondered why, but in hindsight, they seem like a great idea.

I don't use v wheels on this printer, but this is a great tip. For my other printer, I will keep my eyes open for this. Thank you for the tip!

Thank you for the tip!

The magnetic sheet and the aluminum bed are placed directly on the thick glass surface. I placed the cardboard on the backside to deal with the heater bump where the heater cables come in. That is why it came out so flat afterwards. The cardboard had nothing to do with it. The Glass surface provided a much better surface than the previous wooden bench. The result here is that you need to use the flattest surface possible when installing the magnetic sheet and letting it cure. Your wooden work bench is not good enough for this task, not is simply placing the magnetic sheet on top and letting dry on it's own.

That's a big variance. I would check the gantry and the bed mount and the associated screws. What printer is it?

I would check your gantry. Is it perfectly parallel to the bed? It may not be.

Hey @@KapmansBasementWorkshop , I fixed it, the bed now has a delta of 0.35. The Bed was ok, but my printhead was lose, it was so soft attachted that you can wiggle it. It's not so flat as your printbed, but it's ok. What's the point in getting flatter now? I mean the printer will compensate for the error and now my error is homogeneous spaced. So no more valleys and mountains.

As long as it's printing well, that all that matters. I'm glad you got to the bottom of this.

There is certainly truth in this. That said, moving to a single rail made a big difference without adding any weight and this new plate with a careful application of the magnetic sheet yet another. My goal was to not use a bed mesh and have parts 300 mm and larger be as flat as possible. Hopefully others will gain something from my mistakes. Thank you!

​@@KapmansBasementWorkshop Now i checked my bed variance before print and it is very nice 0.039 mm, after heating it is something like 0,1+ something ... but titan backer are ready to install! :) Looking forward to new videos, I really like how you explain everything.

@@davidbucek My previous bed was not as good a yours. I'm sure the titanium backers will provide a big improvement. Taking off one of the mgn9 rails provided a substantial improvement for me in probed flatness, especially when hot. Let us know what difference you see with the backers on. Thank you for the kind words and thanks for watching!

I initially installed it on my wooden work bench which wasn't as flat as I had hoped. That is the method most videos show when attaching the magnet. The second time, The glass table provided a flat enough surface to create a significant improvement. The key takeaway, your bed will only be as flat as the surface used to allow the glue to dry.

Exactly. I screwed up. Hopefully others will learn from my mistakes. And if they have made the mistake previously, here is a path to fixing it without purchasing a new build plate. I might have gotten carried away when creating the video thumbnail...Thank you for watching!