Hey Snapmaker team and @Tracy . I had an idea today. Why not build in a metal or glass plate as a thick substrate into the heated bed? That might be a way to help with insulating the webbed frame as well as adding more rigidity. Just a thought.
a substrate in the PCB would not be the finished surface…
You think a 1/2" aluminum plate mounted under the heated bed would do the trick? In terms of a modification I mean, not a manufacturer solution. This would keep the magnetic bed on top (instead of some hack like embedding magnets in the aluminum) and add some stability to reduce flex … but the aluminum would still conduct heat down to the carriage.
Well, they are using a really thick pcb for the heat bed. I proposed, instead of using organic cores and prepregs, to use a thick non-organic metal as one layer. I’m honestly not sure what would fix all the issues. I think multiple upgrade/Rev changes are required, possibly.
I would be curious to know what substrate the used. PCB substrates have known CTE’s (coefficient of thermal expansion).
Probably the cheapest Fr4. It would be sufficient.
This helped me a lot with the bed’s seesawing.
Tightened the “bracket” bearings in one of the Y axis linear modules as described in the forum in the below link. Check out the before and after videos at the bottom.
(After taking the bed tray off, the wiggle in the bracket was very hard to detect. I could Barely feel it and didn’t think it would matter as much as it did. I sat there trying to wiggle both the brackets for a while making sure I wasn’t making it up in my head 
)
Wobble Before (Turn down volume!)
After
So far the prints have become acceptable with the stock bed, not sure how long this will last though. I also have the tray without the square border if it matters.
1 Like
Similar to this, support reached out to me recently with detailed instructions. Having calipers, I measured a whole 1.25 mm below theoretical true values at the bearings!?! Needless to say, after adjusting to true in both Y axis linear modules, it’s printing much better already. There is no visible wobble. I will do some smallish but tall prints (in x-y, some vases for my wife), and then try something big to see how much better the whole bed is.
I have machine 389. I do hope they have added some extra QC in this regard. it’s a pretty easy fix… but it’s also something pretty easy to get right out of the factory.
These threads are why it took me 6mo to set my printer up. It arrived in Dec was setup 13 days ago. I am mystified at the issues folks are having based on my anecdotal experience of 1. But I have been thinking.
-
Do they have QC issues where items are sent out to customers outside of tolerances. I saw they sent out z axis brackets that didn’t fit in the indent in the base plate, if they missed this, what else did they miss?
-
Aluminum isn’t as hard as steel, are folks who make mistakes in builds warping objects accidentally. For example if you put the build plate frame upside down and tighten it does that permanently bend the bed plate or some such.
-
Are folks over tightening things by using power drivers? And conversely under tightening as I have see power drivers ratchets kick in when things are stuck but not actually fully tight.
-
Similar to 3 - I used approach to tightening bolts and screws using what I learnt 35 years ago tightening car engine head bolts - pre loading all bolts screws in a non linear pattern to avoid tightening in a way that pulls objects off center. For example one should never attach the linear bars by doing all the screws/bolts on one side to full tightness followed by the other side, one should use a kitty-corner like approach to sequence tightening including one pass to preload tightening followed by final pass of hand tightening.
I know I am new so may have no clue what I am talking about, but the amount of folks having issues is disturbing. Hoping I am adding to the conversation with these musings. This topic also drives a lot of the dysfunction on the Facebook group where folks with no issues are unreasonably dismissive about those folks with issues.
Everyone is welcome on the forum!
I do hope people on Facebook are steered towards the forum for I formation.
1 Like
@NilartPax yes this forum and the discord have been super welcoming 
Another question, when doing the final step of the z-offset with the paper i assume we are looking for a gap of specific physical size between nozzle and heated bed.
Has any one measured that distance with more accurate feeler gauges - would using a feel gauge be better (assume that my gauges would have no oil).
I assume a hotend with different diameter extruder nozzle would need different z-offset as the extruded layer would be different size? (is my mental model right here?)
If you have access to a proper feeler gauge they are definitely more accurate than using tthe paper will be. The slicer mainly handles the amount of extrusion (though calibrating is important), so actually I’d recommend using the same z offset for various nozzles and adjusting in the slicer. If the only slicer you will use is Luban… Maybe modifying the Z offset would be a way to adjust elephant foots or something. Most slicers have first layer extrusion flow settings though.
Thanks,
Just to give some encouragement, i did this print today after 11 x 11 calibration.
I have not calibrated esteps (though i did adjust to a common number others have used) and i have not calibrated flow, linear advance or anything else. Looks like i may need to adjust retraction to get rid of some of those strings. This is I think all one layer in height. Printed on Fast.
You can see i continue to have the adhesion issue in the lowest left corner - where the skirt failed. I am hoping the new firmware might correct that… also for some reason the little triangle objects infill is incomplete. But its more like it just never tried, there was also a strand of infill missing in two of the squares at the bottom left.
Note i did damage a couple of the squares using the pallet knife - oh one word on that to folks, the pallet knife blade should be 100% parallel to the surface at all times - never use the point to dig, never use it to lever - you will damage the bed easily that way (i am unclear many folks know the correct use of pallet knifes in general, not just 3d printing…) oh and this is stock snapmaker PLA too.
So the right combo of steps and parts that are within tolerances for warping etc does seem to work.
Aluminum may not be as hard as steel, but it’s far more brittle. It cracks and fractures more than it bends. Steel is more “flexible”. It’s more that the webbed frame changes dimensionally when it heats up and they did mill the mounting faces but didn’t do a final milling of the mounting faces on the frame in this regard. I’m not a mechanical engineer, someone who is though had explained this much better than I can.
1 Like
Each hot end needs its own calibration. The heat break and the nozzle meet in the middle of screw threads inside the heat block, but there’s no designated datum surface for where they meet. As a result, the height of the break/nozzle can vary by design. If you swap hot ends, recalibrate height.
Oops, I thought I said that. I was talking specifically in reference about it being cast. Yes forged aluminum is not so brittle. Casting steel would just be stupid so I never mentioned it for the steel. My bad, didn’t mean to confuse anyone!