Printable View
Why doesn't the build plate just stay calibrated. What is actually happening?
Is it :
-the build plate springs have memory
-The nuts loosen
-support part slippage
-Gremlins
-some other sort of gravity/time/heat/warp/voodoo
Has anyone tried using nyloc plate leveling nuts?
2 things happen, bed can be warped... which leads to alot of levelling, alot of the time and its never quite right... my advice is print in the center :)Quote:
Originally Posted by avocadobaby
OR... when you take something off the buildplate and nudge it a little, there goes some calibration. Any Force on the plate anywhere causes pressure on the springs and eventually the thumbscrews loosen slightly over time. If printing a small piece I rarely level the plate, but if it's going to be wider than say 7cm, it's a must.
A lot I have seen throw themselves way off by not leveling at the same general temp range. One time they will level at 90degrees, then they will get lazy and decide to check the leveling with the machine cold. When they level cold it throws the leveling off. Then when they start a print they heat it up and it is all out of whack.
There are several nyloc style leveling nuts out there. Here is mine. http://www.thingiverse.com/thing:353353
That's probably one of the best tips on here. I used to do my levelling cold all the time :rolleyes: - makes a helluva difference doing it hot, saves much pain.Quote:
Originally Posted by jfkansas
Whats odd is with my creatorX and the old plastic build platform the front of the plate would be lower when hot. With the metal platform the back is lower than the front when heated.
That is my issue!!!Quote:
Originally Posted by jfkansas
If I am printing little parts and raftless... I actually push the whole platform up at the front while its printing that first layer (just a smidgen with my finger ) because no matter how I freakin level it, the back ends up being a bit higher than the front, so I literally push the platform up and make sure the nozzle is really close for that first layer. After getting a set of spirit level guide things for 3d printers (like you stick them all over the thing) I found my hotbed was not an 'L' shape it was more a droopy L.
The good thing is the plastic platform droop is relatively predictable. Just Level when hot, around 80-90 then set the gap. I don't remember where I read it but a guy explained the cause of the plastic droop was the build plate expands and there is no play in the leveling bolts. So when the build plate expands it pushes on the front bolts and this flexes the plastic platform down. This is also why it is predictable since the build plate expands at a certain temp then expansion tapers off. I imagine the cause is the opposite with the metal platform. The steel expands more than the aluminum and so the back sags.
Either way the steel platform is totally worth it if you have an older X. The plastic platform turns to jello when heated and isn't very stable.
Rather than starting a new thread I'll throw this here since it's definitely related. I realize that a machine with a perfectly level build plate is bordering on impossible for the average DIY'er building something in his garage. Even a prefab machine at the price point of a flashforge or similar is probably too low to overcome manufacturing variances that would result in non level build plates. That being the case a levelling mechanism is certainly a must in any machine. What I don't get is why those mechanisms seem to consistently be so inadequate to the task. To Geoff's point that any force on the build plate causes some loss of calibration I don't understand why the leveling device isn't more robust. Something as simple as a threaded hole in the carrier plate with a bolt run through it to a spherical or heim joint on the build plate would allow for adjustment just as easy as the wingnut/screw/spring arrangement. A thumb/set screw positioned at 90° to said bolt would allow it to be locked into place once "level" is achieved to prevent vibrations from moving things around.
Now I understand that the rest of the machine would also have to be built such that things don't move around when force is applied to the build plate. The existing setup means that any force is taken up by the bed and springs so it's not transfered (or minimally anyway) to the rest of the machine structure which could be potentially more damaging than just screwing up bed alignment.
I'm saying all this because in my gigantic list of things I want to design/build (most of which will never happen if I'm completely honest with myself) is a new 3d printer to supplement my ffcp. I bought this machine as a primer to 3d printing so I had something that "just worked" to figure things out on. Now that I have it the thought of building my own seems much less daunting. From an engineering standpoint I can see a TON of places where this thing can be improved on. From better motion guides, to better motion control, to a more robust frame, etc. All of which (i think) could be done without adding a huge amount of cost (comparing the purchase cost of the ffcp to the purchase cost of the components for my own build). One of the big things in my mind would be moving to all metal parts for rigidity and therefore repeat-ability along with durability. This would allow for a more rigid fixturing of the build plate as well. Of course that all assumes that I haven't missed some greater truth to this whole thing that would negate all my assumptions. That's where you all come in. Is there a reason other than what I've stated above to explain WHY the leveling setup can't be built as described above and if it is built that way, is there a reason why it wouldn't work as well as I think it will?
It is really a cost/benefit situation. The cost and amount of parts to "auto level" just isn't worth it. You also would need software to control it and the firmware space on the standard boards really can't support that much programming at this time.
You don't have to be perfectly level but you can't be way off either. The first layer is always laid down level since the x/y axis moves in a relatively flat plane. It's at least pretty flat since there might be a little sag in the middle. I read in another forum that it is like .05mm sag due to the weight of the printhead and the x axis rails flexibility. Since the first layer is always laid down flat to the plane of the X/Y then the only other thing that could be a problem is build plate adhesion.
I think people way over think bed leveling. I spend maybe a couple minutes checking the gap in three spots on the build plate after preheating.
I actually designed a part a while back that would help decouple the expansion of the build plate vs the platform. In the mean time though Flashforge changed from the flimsy plastic platform to the metal and 80% of the problems went away.
My current setup and what has worked best so far is:
1. Wingnuts replacing the thumbscrews and some washers to reduce friction (M3x0.50 wing nuts and M3 washers). This makes adjustment pain free, especially when the bed is hot)
2. Metal arm stiffeners (there are several kinds out there, mine don't seem to be available anymore, but I got mine on ebay)
3. Thick glass plate (1/4", Mcmaster part 8476K74, or I believe the Flashforge glass is the same)
4. Give the bed a thorough heat soak at the intended bed temperature before leveling
Yep, I have 2 - 1/4" glass and 2- 3/16" glass. I rarely use the 1/4" on hot stuff, it just takes too long to heat up.