I got in some 6-32 knurled nuts in today to put on the adjusting screws for the bed. Tomorrow I hope to do a bunch more wiring. I got tied up today repairing my coffee roaster, and then roasting some coffee. 
The display panel I have has 4 encoder pulses for each “click”, or “detent” or whatever. So I changed the firmware to this:
//
// ENCODER SETTINGS
//
// This option overrides the default number of encoder pulses needed to
// produce one step. Should be increased for high-resolution encoders.
//
#define ENCODER_PULSES_PER_STEP 4 // jeffeb3, was 1
//
// Use this option to override the number of step signals required to
// move between next/prev menu items.
//
#define ENCODER_STEPS_PER_MENU_ITEM 1 // jeffeb3, was commented out, but probably defaulted to 5.
/**
The result is that 4 pulses turns into one step, so that when you go into the 0.1mm move, it moves 0.1mm instead of 0.4mm. Then the single step per menu item makes it so that one step is one menu item.
Before that, it would miss steps (it still misses them sometimes, but I think that’s because it’s overworked, and misses pulses), and the 0.1mm move would go in 0.4mm steps.
I bought this from v1engineering.com, in April, or so. @vicious1 will have to confirm if there are different versions or anything like that. If they are all the same, then someone else should try it out, and probably put that into the RC8 or RC9 Marlin firmware. I’m running the MP3DP RC8, BTW.
I may have to look into what you posted in that last bit of code. My LCD requires me to click twice to move off the first menu item (so goofy). I hope I can find a tweak to fix that (you may have guided me in the right direction thanks).
Neil
There are still some LCD issues with RC8 hopefully they will get them sorted out soon.
My firmware already has the first change, after that I don’t think you need the second change. Are you using your own firmware, you might want to specify that as to not cause confusion.
Your firmware changed the first one to 1, not 4. I am using your firmware (linked here).
If you change the first to 1, and don’t change the second, then you get something close, but you still get 4 steps in the move menu, and 1/4 times, you miss a menu step (because it’s set to 5 steps/menu item).
Some wire management done. I want to build a stand/box for it, and put the ramps board in there, but it makes sense to have all the cables go to this corner of the machine, where Ryan designed in a mount for the ramps.
Are you putting any strain relief on your hot bed wires (bottom picture)?
w.r.t. strain relief: No, how serious is that? Is that why there are those 4 holes forward and right (in that same picture)? I tried zipping them there once, and wires were so big, they were keeping that screw from being able to level the bed. The power wires are sort of short for where I’m going too… Hmmm. I’ll have to think about that. I should have done that before I put that cable cover on it.
I have been playing with my bed arrangement, I think @vicious1 is right (I should have known, you probably all did), the thermal pad is not a good choice. I could feel the areas without the pad, and they were significantly colder than where the pad was, and the pad parts were not terribly warm. I took them out, and have better adhesion. I thought I had pretty good luck on my other printer, but now I’m second guessing that as well.
The middle of the aluminum plate is a bit higher than the front or back. I’m using receipt paper to adjust the corners and if I do nothing, it’s pushing hard on the plate in the middle. I adjusted the slic3r settings to 1.3mm before it adheres the way I want. Which is fine for the middle, slic3r settings are easy. I have an inductive sensor I bought for auto leveling though, I think that is my next highest priority. So after some fun, and some good prints and some bad prints, I’ll be going back to aluminum. I’ve seen some people using metal tape on the glass to give the sensor something to mark against.
The wires will break at the connections after being moved repeatedly. The 4 holes in the bed are there to zip tie the cable. If you put the soldered connections on the left side, the cable will pass over the holes for zip ties.
I have a stupid question. The MK8 extruder has that really long throat, and it’s causing the fan to be really far from the work. The bottom of the fan is about 3cm higher than the nozzle. I can move the fan down, but I was just messing with an inductive sensor mount, and it’s also too high, so I’d need to customize that (not a huge deal, but annoying, to say the least).
The stupid question is, should I try to shorten the throat? The threaded part that connects the heater block to the rest of the extruder? That would be crazy, right? I would have to make sure I didn’t end up with any shavings in the nozzle, and I’d need to make sure the threads were still fine, but I could easily remove 1cm, and have enough room for a gap between the heater block. Is that length a good thing? I really don’t care about the z build area, I don’t have anything I want to print for a week.
If you want to shorten it you have to take it apart, and pull the teflon out. Then there is a bit you can cut off the top with no problem. The heater block end though is a little different. You have to cut it at 90 degrees or the nozzle will leak. I do have short ones in the shop. Those geeteck extruders are nicer but that is an odd throat piece.
As I’ve mentioned, I plan on making a box under the printer. This is the drawing for the top of the box. I will cut this out of some 1/4" MDF first, and just place it under the printer to hold it square in the XY.
I haven’t CAM’d or cut it yet. It might need some tolerances around the printed parts, especially. Currently, it’s their exact STL size.
Let me know if you have any suggestions.
I just had a chance to play with those firmware edits you suggested. Works much better know, Thank you. I will update all the firmware’s shortly. I’m going to check the bugfix branch out first. Maybe they fixed the LCD lagging issue.
Cool. Let me know if there’s something else you want me to check. From my testing, it is laggy, but it doesn’t miss steps unless you do like 10 at once.
Something is wrong with RC8, I have tried everything. I just started fresh and only changed the steps and accel. It still messes up unless you are moving slowly. I have spent hours last night and again today.
We have to stick with RC7 for now.
I had some trouble with that drawing. The components weren’t rotated the right way in estlCAM, but I selected everything, and chose “expode” from one of the menus.
I cut it out of 1/8" hardboard, which I figured would fit in the ~4mm gap that I measured between the bottom of the parts and the MDF frame. I just did a through cut, but I would probably remove about 3mm where the parts are if I was using a thicker material.
I should have added 0.2mm to each printed part, because it was too tight, and 4mm of hardboard is not strong enough for me trying to “fudge” it into place.
The key part is the inside though, and it is keeping it super rigid. I think I’ll beef the outside to about 10mm, instead of 4, and I’ll add the 0.2mm to each printed part.
@vicious1, maybe I can look into it. You took the stock RC8 from here?
https://github.com/MarlinFirmware/Marlin/tree/RC
And then you changed these lines:
#define DEFAULT_AXIS_STEPS_PER_UNIT { 200, 200, 800, 200 } //MPCNC
#define DEFAULT_MAX_ACCELERATION { 3000, 3000, 35, 10000 } //MPCNC
Then what do you do, and what does it do? Is there a bug report in the github issues? Are you having trouble with the MPCNC, or the MP3DP?
Wow that is crazy I bet the bottom makes it super rigid. Maybe add some screw holes and use the little upper brackets to hold the y axis mounts from bending in. Can’t wait to finish the new printer. I sketched out the rest by hand yesterday. CAD ASAP.
As for RC8, I think it is a processor issue. Something they have done seems to be using extra processing power. I tested all changes against my RC7 and then tried to change all kinds of things in RC8. The only way I could get things to work was slow the Z axis moves in the gcode…and another funny thing was the script I use to test causes errors in all of my versions of RC8. Can’t tell if that has anything to do with it.
This caused errors after it finished, I could crash marlin if I used the max z axis speed 8.4mm/s (504mm/m)
G91
G1 X-100 F11000
G1 X100 F11000
G1 Y-100 F11000
G1 Y100 F11000
G1 Z9 F504
G1 Z-9 F504
G1 Z1 F504
G1 Z-1 F504
G1 Z1 F504
G1 Z-1 F504
G90
I have attached a basic marlin and a test gcode that crashes on RC8 and not on RC7 and all the feed rates are very slow relative to the max.
Marlin.zip (1.25 MB)
plastic-test.zip (2.6 KB)