Best posts made by Gerrard

I thought I'd add my experiences on this subject. I've built a very large printer. The heated bed and build surface have been two of the biggest concerns, but so far both seem under control. The build area is about 870 x 800, and I'm using 6mm aluminium as the structural element, with 20+ adjustable screws on the underside supported by a 40x40x4.0mm EA aluminium frame. Heating is done with a custom bed heater from China with holes for the screws - it's about 2300W on 240VAC. I looked at getting Borosilicate glass for the build surface, but I was quoted about $1200 + freight from the states (probably another $1200), and PEI sheet was pretty similar. I ended up using a large mirror from IKEA ($60 for 2!), which I cut to size. Before the heated bed went in, I had all kinds of surface adhesion issues, but since then it's been great.
Mesh bed levelling hasn't been perfect, I get about 1-2mm max error across the whole area, but given the size I'm not concerned with that.
So far I havne't printed anything particularly large, I'm still debugging the design, but I have a project coming up that will start to use more of the area, so I'm keen to see how it goes.

Hi All,

I've had a quick look around, and I can't find what I'm looking for.
I'm wondering if there's a way to connect a digitizing touch probe to the Duet 3 board for reverse engineering purposes?

Something like this: https://www.youtube.com/watch?v=FJ8aKo_qqos

Cheers

right... There's a gap in my understanding there. Thanks for clearing that up. I'll give that a shot this weekend.

@Phaedrux, I've run the pressure advance script (here) several times, and I've picked a value that gives the best results (0.025 in my case), but it hasn't helped. Based on the explanation here it seems that pressure advance is simply a multiplier applied to the acceleration - a higher number means more extrusion during acceleration/deceleration.
In my observations however, As I increased the pressure advance value, I noticed that the mid-line regions (constant speed regions) became under extruded - for higher values the lines wouldn't touch. For low values I noticed that the ends became shorter - and for really low values not quite touching the perimeter lines.

To be honest, I've only tested this "on the fly", and I haven't dedicated any real time to it, so it's possible that other effects are at play.

The main intent of exploring this avenue was to explain my observation that the motor went backwards slightly - job done.

Hi All,

I've skimmed over this topic a little, but having just figured out this problem I thought I would add my 2c.

I have a very large bed - roughly 800x800, with 16 manual adjustment screws (4x4 arrangement), and 4 independent Z motors.

In my case, I have (at some point) needed to manually level both the lead screws, and the manual adjustment screw - I have also used G32 and G29 bed levelling.

I partially agree that there is a hierarchy to bed levelling, which needs to be done after assembly, maintenance or a component failure (like a loose screw on a Z motor coupling for example?). With so many components there's plenty of opportunity for something to go wrong, so the correct procedure helps.

I manually adjust the Z screws with the hot end close to the screws first - this gets the nozzle within a reasonable range for G32 to work its magic. Then I run G29 to do a mesh bed levelling - now that all 4 motors are coplanar this gives me a good idea on the flatness of the bed. If the min/max deviation of the mesh bed levelling is less than 1mm (an arbitrary number I've decided to use), then I'll leave it be - otherwise, I'll adjust the manual levelling screws...

A more thorough method would be to look for a deviation greater than Xmm (some number) between any two neighboring points in the mesh bed levelling, but for now I'm satisfied with maximum deviation...

On smaller printers like your typical 300x300 size, then I think a similar approach is easy enough to do, so it doesn't really matter whether it's necessary - do what ever you feel works best for you.

Thanks, that worked.

So, a quick follow up on this.

Back story - the printer is a huge CoreXY built as a MVP (minimum viable product) - there have been a lot of short cuts taken, so I expect a bunch of problems.

The job I was printing when I posted this was for a casting pattern - it took almost 6 full days to print, and the quality was very average. The blobs and zits were the biggest problem, but there were several horizontal bands that were several layers thick - kind of like the magic numbers problem scaled up a few layers.

Anyway - I made the error of printing the whole job 2% too small :'(, so I'm now re-printing it. I took some feedback on board, tried a few test prints, and turned the settings up. I disabled the Accel and Jerk control in Cura, then tweaked the M5665 and M201 values on the duet while it was printing.
We're now printing at 150mm/s (max), Jerk is 750mm/s, Accel is 700mm/s^2, and layer height has been increased to 0.32mm. All with the intention of speeding the print up firstly - and hopefully to get some improvements in quality...

Well... Here's a before and after shot of the print at roughly the same point in the print.

Before

After

It's not perfect, but it's a huge improvement.

@CaLviNx "Best approach" is subject to perspective. I agree with @Piet here - you build the most light weight frame that you can (because it's cheaper) and you let the software/hardware smarts take care of the rest for you - because it can. It's very normal for machines to require commissioning before being ready for use.

Remember that 3 points is the most stable, only if you assume a rigid body. In real life "rigid" doesn't exist. All bodies are elastic to some extent.