Hello my fellow nerds.
I was curious what everyone thinks of the vision encoder board? Seems like a great idea. I got one. I did the whole thing. I wonder if it was actually needed or not. though, haha. I’m not sure I have any test to even see how much it may or may not help, and I kind of realized, I already got fairly good prints and things seemed pretty accurate as it were.
Anyone else been playing with theirs? have bubbling opinions?
I suspect this is of most benefit to
(A) the laser cutter and blade module
(B) when you print a multi-part object spread across 2 printers and need them to fit together, this requires identical calibration between the machines (EDIT May also. benefit print farms who need to print X million identical items)
The accuacry of the system is already greater than the error caused by filament so I am not sure it benefits a user with a single 3d printer only.
I’m guessing the machines are pre-calibrated at the factory and so you won’t see much results with the vision encoder. I think where you’ll probably see the most change is over time as the printer wears from use and starts to lose its accuracy. Then you’ll probably see a difference.
I don’t really know though. It’s so new and hasn’t had time for enough people to use it and see how useful it is yet. It’ll be interesting to see if it lives up to the hype.
I have seen X1 with vertical skew. Unsure if factory calibration didn’t, or it got banged in shipping.
Is there a back to back test that can show a difference?
If so how often should the plate be used? Is the scan only good for the one print? and how is that measured? 1 10 hour print? 10 1 hour prints?
The skeptic side of me is just seeing this as a $100 bottle of snake oil. 
I had to look up “snake oil”. Basically a scam medicine. While this plate probably won’t make a difference for a single small print, I can see that it maybe might be of use to the laser cutter, and print farms who want to identically calibrate multiple machines, so I do not think it is snake oil.
I was partly under the impression that it’d benefit the laser/cutter crowd the most.
This is what I was feeling. Even across multiple printers, the Bambu printers are accurate enough and all of that, that it doesn’t feel needed so much.
I believe they are. It’s noted somewhere in the material on the encoder board.
I’m hoping we see this eventually. I’m not sure how well it’ll show off though with the printers being so new still and being calibrated from the factory.
Their marketing says “5-Minute calibration for weeks-long accuracy”, so I assume every few weeks or something. 
I forgot to snap a picture when I ran the encoder board on my machine. It tells you how much it improved the accuracy, what it was before, and all that jazz.
I believe it’s beneficial. It’s important to have a reference like this for a machine that will change over time as things settle and wear. My first print is with the H2D improved the fitment of my assembly’s tolerance significantly. Going from an average of 0.3mm+ on all mating surfaces to 0.05+ without any issues.
Those figures are along the lines of what I got, from what I remember. I, admittedly, don’t know well enough though to know what those numbers mean in for me in reality.
Well in my printers case if I printed a part prior to calibration it could have been 0.169-0.327mm out of expected design spec. I call that a significant improvement for what I do.
I think it could be for me too. I’ve been getting tighter tolerances going on overall with my Bambu printers, more so than previously. It’d be nice to get an even tighter fit.
Although, I can’t help but think about Makerworld and all of the non-H2D-vision-encoder users. Since I design a lot for Makerworld, I’m not sure how full advantage I can take of the H2D and some of it’s specific features.
mine just arrived as we speak 
i still need to go through the process, but the printer is running right now (as it has been pretty much since it arrived)
i have to say the quality is already great, if it improves even slightly it’s already a win, and i think it will be especially useful for the laser version
that being said, the precision really is impressive, i saw a video that the guy said that he adjusted his tolerances from 0.2 to 0.1 and i thought it was BS, but turns out, it’s actually the case!
So do we get similar numbers running these tests back to back (i.e mechanical slop unfix-able by the encoder plate)?
I’m not sure about that; the screen message says that’s compared to no calibration, but they’re calibrated from the factory, right?
0.169-0.327mm errors would give terrible prints.
That range is over the entire build plate. So a 325mm wide print with a variance of +-0.169mm is not a horrible print. My old X1C would be out ~1.0mm on average with a 150mm+ print.
The largest benefit is being able to continue and maintain the calibration after any maintenance or changes to the printer.
I realy want that plate, but bambu lab does not want to sell it to me :-(.
That “out of stock” is realy mean.
One thing I find slightly disturbing, which may account for the “slop” is that the vision encoder does not align snugly. There is some variance side to side.
Printing now, but can take a picture later.
I am not sure why that would matter, the alignment will be done between points on the plate. If it was. not put in the same place, all that would change would be where on the plate the center of the model was.
I mean I can try doing the alignment with it shifted to the right, then do it again shifted to the left and see if it says there is a difference. May be adjusting where it thinks center is.
So encoder plate is aligning things dynamically? I would’ve assumed the plate would need to be squared perfectly for that level of accuracy. I guess I really don’t understand what it is actually doing
