Rostock Max V2 ultra level frustrations

[captainpotatofudge]

Ok so I have this situation. I have searched the forums over and over and over and read and read and read and read. I have implemented better than half the ideas I've read and gotten somewhere? And then nowhere. The short data: Rostock Max V2 w/Repetier 0.91 for Rostock, MatterControl v1.5, Trick Laser cork dampers, Trick Laser 300mm CF arms, E3D v6 hot end, 0.4mm nozzle, injection molded carriages w/Trick Laser end stop thumbscrews, and belt tension by printed cams found on Thingiverse.

I've got a Rostock Max V2 and I really want to love it. It however, does not seem to want to love me back. Please help!!
So let's try to start at the beginning. Sadly I have managed to lose a lot of the notes I've been keeping so I'm going to try to get the timeline and situation as straight as possible.
I received the assembled machine the week of Christmas 2015. My father in law purchased and assembled it with the help of an EE friend. He printed a few things, but didn't dial in filament settings or anything like that.

So I setup the machine and got to calibration of the Z height. Ok, not so bad really. My FIL used the three individual tower scripts, but I switched over to the triple tower single script method right away.
At this point there were no apparent issues. Calibration went well, I started by printing calibration cubes. I was getting ABS settings dialed in pretty good on a 20mm cube. Printed a Benchy and a few other things that are under 3" diameter.
However, I then noticed the cubes didn't measure 20x20. I wasn't too concerned at the time, but I went on the hunt to figure out what was going on. I actually never arrived at a good answer, but I wound up discovering the Z tower leaned back away from the machine a bit.
While looking into that I noticed the X & Y towers also leaned a hair. It didn't seem like any of it was a major concern nor did it seem to affect the printed parts visibly thus far. I left it alone for the time being.

I then started printing in PLA and was mostly toying with dialing in settings using calibration cubes and other small things. I got stringing dialed in great and cubes were looking awesome. So I proceeded to larger ventures.
I started with making some lightsaber parts which turned out pretty nice! SeeMeCNC translucent green PLA looks very nice! MatterHackers glow PLA made a great handle for the saber. Life was looking pretty good. Then things starting going downhill...fast.
I noticed that the lean in the towers must be afflicting my prints more than I thought as I put a few up against a square and sure enough by a few inches tall they were angled upwards of 1/8" off. So I looked into how to square up the towers.
An interesting process that was. I was just barely able to get them close to what I'd be happy calling square. (At this time there two underlying problems I was yet unaware of helping to make the squaring problem worse)
So I went back to it, printing a few small bits and being fairly happy. On the advice of a friend I ordered the injection molded carriages and performed the upgrade. Seemed like a much better setup to me and on I went.

It was then that I tried to print in a bit in Taulman T-Glase and found that the stock hot end just wasn't up to the task despite being spec'd to hit the lower limits of T-Glase temp. Not sure why, but it just wasn't getting there with any dignity.
Unbeknownst to me at this point I had started softening the PEEK a bit and was about to blow out the heater resistors. Apparently they had not been packed with enough silicone and had air gaps...they burned out.
So I figured what the heck I want an E3D v6 anyhow. Who doesn't want to print in PETs, nylon and so forth right? Especially with this glorious build volume. So I managed to squeak out a printed mount for the E3D using T-Glase Orion Blue. Oddly that was about the best print I'd had yet!
Sadly it was then that the heaters gave up the ghost completely. I'm guessing the air gaps and prior damage were simply too much to maintain that level of heat for long enough to do the print. (Also note that I had redone the PID tuning several times by this point)

I happily installed the E3D v6 and started messing with T-Glase and a bit of Colorfabb XT. COOL STUFF!! But of course I needed a lot more calibration now with the new nozzle height and such. I also needed to print new cooling fan ducts and so forth.
I think it was around this point that I came to the realization that printing anything larger than 3" diameter was absolutely TERRIBLE!!! Horrible bed adhesion past that diameter. So ok...what gives. SO I start trying harder to get perfect calibration.
Paper was used, feeler gauges were used, each was checked and triple checked with my calipers. I then realized that the glass bed was bowed up in the middle! Ok, glass should be absolutely flat right.
So I get on the support to SeeMeCNC and they say this is normal and can be fixed by adjustments to the screws while pressing down on the Onyx. Ok, fair enough heat made the thing expand too much and the screws weren't holding it flat.
I must have tried several dozen times trying to get it to stay flat to no avail. Tighter...nope...loose...nope. To make matters worse as I started printing a disc about 6" diameter to gauge what was going on it showed a very odd pattern.
The bed was not only domed in the middle, but it was sort of a...hump that ran from the Z tower straight to the front of the bed, but it was completely uneven. So after giving up on getting the bed to stay flat, it was however now better off than before, I moved to try software leveling to see if that could help.
That in and of itself (via MatterControl was its own fresh nightmare and I don't plan to go into that as I don't feel it has impact here, suffice to say I ended up ramming the bed with the nozzle a zillion times, dragging the glass off the Onyx, printing super weird wavy patterns..etc)
So after much communication with MatterHackers about the leveling function, features etc I abandoned that. I went on the hunt for more info about flattening the bed manually. After a load of reading, which at this point the amount of Google and forum searching has me absolutely miffed as it is way beyond what I feel should be necessary,
I finally come upon some info that helped with getting it flat. My Onyx is now dead flat even at 85C and I have ZERO, ABSOLUTELY ZERO, desire to mess with it further or in any way disrupt the stasis achieved!!
I had previously even tried using spacers on certain mount points and...I digress...

Ok, so now I have a flat bed, I should have very close to square towers. I believe it was at this point the real troubles started showing their uglyness. Calibrate Z height to feeler gauge, tight on a 0.1mm. Go run 3 tower script with Z set to 0.1mm and the nozzle is more like 0.15mm off the glass.
Make adjustments to screws, rerun Z height, now each tower is off farther than before? Rinse and repeat more times than I can count. I've made adjustments to the horizontal radius so many times it hurts. I've been everywhere from 140-144 for the HR.
Prints still come out with weird humpy/wavy pattern out past 3". Then somewhere along the way it stopped connecting infill to perimeter walls. No matter what adjustments I made I could NOT get it to connect. And none of the models I used were super weird or full of holes etc.
I have seen loads of people print the very same models on the very same printer with AWESOME results. No matter what I've done, my results continue to be shit and get worse.
On 4/7/2016 - Installed Trick Laser CF rod arms (300mm) with zero lash straps. I did this because after way too many hours reading, seriously at this point I'm beyond frustrated that it requires this much research to get the thing to print ANYTHING outside 3", it seems that the stock arms in NO way can get you to the outer limits with good results and may in fact have something to do with me not being able to calibrate properly.
-Set diagonal rod length to 300mm in firmware.
-Set max Bed size/print area to 240x240 under Printer>Print Area
-Spacing at top section of TL CF arms is adjusted to ~36.93mm between brass washers

Also note that my Z height and HR seem to fluctuate a lot. One calibration run they're all perfect, try to print and something is wrong. Run the Z down and it's off. Tower script..off again, so I adjust HR...round and round.

Now prints are WORSE once again. After letting the machine sit for almost a month because I simply do not have this much time to put into something that is supposed to work out of the box, which I see so much of people assembling this thing and have beautiful prints.
I did discover that two of my motors had a few screws that were loose, not sure if that was my fault for not tightening enough after installing the Trick Laser cork dampers or if they just came out..whatever. So I cranked them all down.
Then I went on a quest to see if maybe my belts were not tight enough and to see what the CORRECT tension should be. Wow what a deep well with NO answers that I could find. So I went technical for the GT2 belts which appear to have a spec of 6lbs of tension for operation.
Ok, that's a LOT higher than I'd had before...a LOT. I had used a trigger pull gauge to make sure they were at least all even previously. I had measured up from the top plate of the base 13" and deflected the belt 1/2" with the gauge and averaged about 14-15oz.
I am using a cam over belt tensioner I found on Thingiverse which seems to work really well. So anyhow, now it seems they need to be a lot tighter so..
On 5/17/2016 -Belts tensioned: 3lbs 3.4oz @13" height, 1/2" deflection. I did 3lbs for two reasons. First is that 6lbs would be INSANE and I felt the melamine structure bending as I approached this tension. Second, there are multiple pulleys due to the placement of the steppers etc, so that has got to change things significantly. So I settled on this tension as it didn't appear to side load the bearings in the steppers and still allowed pretty free, but stiff, movement of the carriages. Anything higher and I could really feel the strain on the stepper bearings.

So here I am, I have no clue what is going on at this point. I've spent entirely too much time on this machine for what it has been sold as and for what I see others doing with it, using WAY less effort than I have put in.
I'm beyond frustration. I'm about to sell it for scrap or throw it outside. I still want to love it, but at this point I can't tell if I got a total lemon or if I'm doing something wrong.
I'm not a moron, I'm a software engineer and I tinker with EE and build my own micro controller circuits etc. I've also spent a FEW years as an Air Force aircraft mechanic...I feel like there shouldn't be ANYTHING this printer can throw at me that I can't handle, yet here I am.

Help me out experienced folks. SeeMeCNC support has been friendly and helpful up to this point, but I'm just not getting anywhere despite their suggestions I just end up right back where I started. Is it me? Is it a lemon? What do I do???

I hope you can throw some questions my way and I'll answer the best I can. I likely have pictures of a lot of this and I'm happy to post settings files etc if anyone wants them. I REALLY do not want to strip the machine back to stock after putting this much time and effort into upgrades if at all possible.

Please excuse any misspellings etc...it's really late, I'm REALLY tired, and I'm at my whits end with this printer!

[lordbinky]

It's late for me, and this is all guess work so take it with a brick of salt, and I've had my Rostock V1 for several years now (but really just the skeleton at this point...yay mods!). I see two likely issues, one is the endstops. Mine lost precision from homing too fast and knocking them around. They seemed secure, but i stopped being able to stay calibrated, which led me to replace the mechanical endstop micro switch with plain magnetic hall sensors and magnets glued to the endstop screws. Inexpensive upgrade and they were something like the ~$1 OH180U sensors I wired to the board and no other fancy electronics. Well worth the increased precision. The other possibility is you're missing steps...which is a lot easier to quantify with a z-probe. (I made mine from a Sanwa arcade controller switch hot glued into a spare mounting bracket it served me well well for a couple of years, although I have FSRs to upgrade to at some point). So.. if it's the missing steps, here's a block of text that I hope helps more than hurts.

That current setting being right makes as much of a difference as the belt tension when you're going for that consistency you want. Since I don't like having to revisit issues, I just threw some spare 120mm fans at under behind my motors, and cranked up the current. Now for the caveats to that statement. I have 2.4 amp .9° steppers and an Azteeg X5 mini, which is also actively cooled with tall little heatsinks on top of drivers (and the actually appropriate places beneath them on the underside of the board).

If your drivers or motors cut out because of overtemp because you're trying to push too much current through them, you'll miss steps. If you're current is too high or too low, you'll miss steps. So, let's go about comforting yourself that everything is fine.

For testing this, Throw a little script together or cut and paste (you'll be doing it a few times) to go home then drop to 50mm or so above the bed, move to each tower, return to center (ex: G0 X0 Y0 Z50) from which you will lower the nozzle and (I use a metal pocket ruler at this point just to make sure I'm not going to crash being off drastically) then do you're bed height check. Do this a few times and if it's changing more than a few steps (You can calculate you're resolution here at prusaprinters.org.) you're on the right track to fixing something important. If a few runs are everything's within a couple steps Congrats you're tension and motors and even speed settings are doing quite well and you can rest easy knowing you're issue is somewhere else.

If you are skipping steps, then we begin our divide an conquer strategy. For the sake of sanity, temporarily lower your speed settings, to something you feel comfortable won't be causing issues, for ex: 100 mm/s^2 for acceleration, 500 mm/s max speed, and a really low jerk setting. With those out of the way, you can repeat the same process as earlier and fine tune you're motor current settings to stop skipping any steps (and if you wanted to know you're sweet spot for that tension setting you worked hard on, this would be a good time to see the impact it has as well). Once you're good there, you can start increasing your settings back up. I suggest getting the speed where you want it and keeping a precise position, then adjusting your acceleration back up. Once your movement is consistent, then it's time to look elsewhere for an issue (off the top of my head, it looks like you covered most of those though).

[Xenocrates]

I agree with the above. It does seem like an endstop issue, either the carriage mounts, or the switches themselves are most likely. There may also be issues with the endstop wiring being too close to the wiring for the power to something or another. (I understand that EE's should know better, and I know plenty that do, But I've also run across EE's who have been spoiled by shielded twisted pairs for too long to understand the potential issue).

It is very unusual for H-radius to float unless there is some major problem, as it is inherent to the physical geometry of the machine. There's no carriage adjustment to check, although it does sound like there is mechanical slop somewhere in the system. I would go over the system, bottom to top, and check that very screw is solidly in place, pulleys aligned, and all that jazz, if you haven't already.

It may also be as he mentioned, missed steps.

[captainpotatofudge]

Ok so I'm going to respond to your posts in a bit, but for the moment ...

Here's my current MatterControl slicer settings for T-Glase I've been using most recently. I won't say they're good because I can't get a decent print at the moment. They WERE good as I did have several fairly decent calibration prints. I've done the temperature towers and found this to be the best etc. I even printed an RC airplane winglet that came out pretty decent all things considered at this point. But then things go batty at random. One good print..then ten or more that won't start or are horrid, then another one or two good ones. I did have good luck printing the Cura robot with this T-Glase...not sure what difference that makes exactly.
avoid_crossing_perimeters = 1
bed_shape = circular
bed_size = 280,280
bed_temperature = 69
bottom_clip_amount = 0
bottom_solid_layers = 4
bridge_acceleration = 0
bridge_fan_speed = 40
bridge_flow_ratio = 1
bridge_speed = 6
brim_width = 0
build_height = 390
cancel_gcode = M104 S0\nM140 S0\nG91\nG1 Z10 E-8.0 F12000\nG90\nG28\nM4
center_part_on_bed = 1
complete_objects = 0
connect_gcode = M120 S4 P5\nG4 P80\nM120 S4 P5\nG4 P1000\n
cool_extruder_lift = 0
print_leveling_solution = 13 Point Disk
print_leveling_method = Manual
print_leveling_required_to_print = 0
manual_probe_paper_width = .1
cooling = 1
create_raft = 0
default_acceleration = 0
disable_fan_first_layers = 2
end_gcode = M104 S0 ; turn off temperature\nM140 S0\nG91\nG1 Z10 E-8.0 F12000\nG90\nG28 \nM84 ; disable motors\nM117 Done!!\nM120 S6 P10\nG4 P40\nM120 S6 P10\nG4 P40\nM120 S6 P10\nG4 P40\nM120 S6 P10\nG4 P40
external_perimeter_speed = 20
external_perimeters_first = 0
extra_perimeters = 1
extruder_clearance_height = 20
extruder_clearance_radius = 20
extruder_count = 1
extruder_offset = 0x0
extruder_wipe_temperature = 0
extruders_share_temperature = 0
extrusion_axis = E
extrusion_multiplier = 0.96
extrusion_width = 0.6
fan_always_on = 0
fan_below_layer_time = 60
filament_diameter = 1.747
fill_angle = 90
fill_density = 15%
fill_pattern = rectilinear
first_layer_acceleration = 0
first_layer_bed_temperature = 90
first_layer_extrusion_width = 0.4
first_layer_height = 0.3
first_layer_speed = 10
first_layer_temperature = 230
g0 = 0
gap_fill_speed = 30
gcode_arcs = 0
gcode_comments = 0
gcode_flavor = reprap
gcode_output_type = REPRAP
has_fan = 1
has_hardware_leveling = 0
has_heated_bed = 1
has_power_control = 0
has_sd_card_reader = 1
show_reset_connection = 1
heat_extruder_before_homing = 1
include_firmware_updater = None
infill_acceleration = 0
infill_every_layers = 1
infill_extruder = 1
infill_extrusion_width = 0.6
infill_first =
infill_only_where_needed = 0
infill_overlap_perimeter = 0.07
infill_speed = 25
infill_type = TRIANGLES
layer_gcode =
layer_height = 0.2
layer_to_pause = 2
max_fan_speed = 45
min_extrusion_before_retract = 1.5
min_fan_speed = 20
min_print_speed = 10
min_skirt_length = 4
notes =
nozzle_diameter = 0.4
only_retract_when_crossing_perimeters = 1
ooze_prevention = 0
output_filename_format = [input_filename_base].gcode
overhangs = 1
pause_gcode = G91\nG1 Z10 E-10 F12000\nG90
perimeter_acceleration = 0
perimeter_extruder = 1
perimeter_extrusion_width = 0.6
external_perimeter_extrusion_width = 100%
perimeter_speed = 20
perimeters = 1
post_process =
print_center = 0,0
raft_air_gap = .2
raft_extruder = 0
raft_extra_distance_around_part = 2
raft_fan_speed_percent = 50
raft_layers = 0
raft_print_speed = 100%
bed_remove_part_temperature = 35
randomize_start =
repair_outlines_extensive_stitching = 1
repair_outlines_keep_open = 1
resolution = 0
resume_gcode = G91\nG1 Z-10 E10.8 F12000\nG90
retract_before_travel = 2
retract_layer_change = 0
retract_length = 2.6
retract_length_tool_change = 2.6
retract_lift = 0.3
retract_restart_extra = 0.09
retract_restart_extra_toolchange = 0
retract_when_changing_islands = 1
retract_speed = 20
skirt_distance = 0
skirt_height = 1
skirts = 4
slowdown_below_layer_time = 45
small_perimeter_speed = 30
solid_fill_pattern = rectilinear
solid_infill_below_area = 20
solid_infill_every_layers = 0
solid_infill_extrusion_width = 0.6
solid_infill_speed = 30
solid_shell = 0
spiral_vase = 0
standby_temperature_delta = -5
start_gcode = M104 S240\nM190 S69\nM109 S240\nM120 S1 P15\nG4 P40\nM120 S4 P5\nG4 P40\nM120 S4 P5\nG4 P40\nM120 S4 P5\nG4 P40\nM120 S4 P25\nG28 ;home all axis\nM117 Printin' stuff
start_perimeters_at_concave_points = 0
start_perimeters_at_non_overhang = 1
support_air_gap = .3
support_material = 1
support_material_angle = 0
support_material_percent = 50
support_material_create_internal_support = 0
support_material_create_perimeter = 1
support_material_enforce_layers = 0
support_material_extruder = 1
support_material_extrusion_width = 0.5
support_material_infill_angle = 90
support_material_interface_extruder = 1
support_material_interface_layers = 1
support_material_interface_spacing = 0
support_material_pattern = rectilinear
support_material_spacing = 2
support_material_speed = 20
support_material_threshold = 90
support_material_xy_distance = 0.7
support_material_z_distance = 0.15
support_material_z_gap_layers = 1
support_type = GRID
temperature = 240
thin_walls = 1
threads = 4
toolchange_gcode =
top_infill_extrusion_width = 0.6
top_solid_infill_speed = 20
top_solid_layers = 3
travel_speed = 110
use_firmware_retraction = 0
use_relative_e_distances = 0
vibration_limit = 0
wipe = 1
wipe_shield_distance = 0
wipe_tower_size = 0
z_can_be_negative = 1
z_offset = 0.08

[captainpotatofudge]

Current EEPROM settings:
Baudrate|250000
Filament printed [m]|1339.976
Printer active |1618620
Max. inactive time [ms,0=off]|1800000
Stop stepper after inactivity [ms,0=off]|0
Steps per mm|80.0000
Max. feedrate [mm/s]|300.000
Homing feedrate [mm/s]|120.000
Max. jerk [mm/s]|35.000
X home pos [mm]|0.000
Y home pos [mm]|0.000
Z home pos [mm]|0.000
X max length [mm]|250.000
Y max length [mm]|250.000
Z max length [mm]|357.281
Acceleration [mm/s2]|1850.000
Travel acceleration [mm/s2]|3000.000
Diagonal rod length [mm]|300.000
Horizontal radius [mm]|141.900
Segments/s for travel|80
Segments/s for printing|225
Tower X endstop offset [steps]|0
Tower Y endstop offset [steps]|0
Tower Z endstop offset [steps]|0
Alpha A(210):|210.000
Alpha B(330):|330.000
Alpha C(90):|90.000
Delta Radius A(0):|0.000
Delta Radius B(0):|0.000
Delta Radius C(0):|0.000
Z-probe height|39.910
Z-probe speed|2.000
Z-probe x-y-speed|150.000
Z-probe offset x|0.000
Z-probe offset y|0.000
Z-probe X1|100.000
Z-probe Y1|20.000
Z-probe X2|160.000
Z-probe Y2|170.000
Z-probe X3|20.000
Z-probe Y3|170.000
Autolevel active (1/0)|0
Bed Heat Manager [0-3]|1
Bed PID drive max|255
Bed PID drive min|80
Bed PID P-gain|86.677
Bed PID I-gain|2.207
Bed PID D-gain|852.323
Bed PID max value [0-255]|255
Extr.1 steps per mm|93.914
Extr.1 max. feedrate [mm/s]|100.000
Extr.1 start feedrate [mm/s]|45.000
Extr.1 acceleration [mm/s2]|6500.000
Extr.1 heat manager [0-3]|1
Extr.1 PID drive max|200
Extr.1 PID drive min|60
Extr.1 PID P-gain/dead-time|29.9633
Extr.1 PID I-gain|4.5733
Extr.1 PID D-gain|49.0700
Extr.1 PID max value [0-255]|255
Extr.1 X-offset [steps]|0
Extr.1 Y-offset [steps]|0
Extr.1 temp. stabilize time |3
Extr.1 temp. for retraction when heating [C]|150
Extr.1 distance to retract when heating [mm]|0
Extr.1 extruder cooler speed [0-255]|255
Extr.1 advance L [0=off]|0.000

[captainpotatofudge]

Adding some images to the fray.
Ok so here's some pics of very early ABS calibration cubes. Not bad at all I thought. Maybe need to dial in some filament specific settings, but I was happy. http://imgur.com/a/qXWjw
Some transparent PLA http://imgur.com/a/JwgKP
First T-Glase runs http://imgur.com/a/IbGNb
This shows some of the dome effect, slightly better than it was, but not as good as it is now. http://imgur.com/a/NB5V1
Here's more about software leveling and the doming http://imgur.com/a/zXfAf

[Polygonhell]

Doming like that is probably just a bad Delta radius setting, unless it's an actual ripple, i.e. if goes low/high/low, in which case it's a mechanical issue.
I would suggest using one of the online calculators with a Dial indicator to calibrate, it's a lot easier than messing around with grabbing pieces of paper or even feeler gauges (which are much better).
I like this one http://escher3d.com/pages/wizards/wizarddelta.php , if you do use it with a dial indicator you will have to negate all the readings from the Dial indicator, because it's designed for use with a Z Probe.
For point of reference it did better in one iteration than I could mange doing it manually, and I understand what all the variables so it's not trial and error.

[captainpotatofudge]

Adding pics for the measure of the tower squaring relative to the glass bed currently.
http://imgur.com/a/ISdiG

[Ail]

I feel your frustration, I banged my head against mine for the first few weeks after assembly and still spit out profanity at it from time to time . I definately was not prepared for the amount of post build learning and troubleshooting required. From reading your post it does look like you have tried about everything (and selfishly I'm glad I haven't had all your problems) but I do see some similarities with issues I have had.

Ill start of by saying after a lot of trial and error the best thing I did when I got to the "I'm about to uppercut the shit out of you printer!" Stage was to just start over at page one of post assembly instructions. It seems like such a huge step backwards but after all the learning / troubleshooting it's fairly quick and painless and helps guide your troubleshooting.

Reload fresh firmware, tighten the end stop switch screws and screws on the carriages all the way down and run through the instructions (with the addition of firmware changes for your upgrades) calibrate a little and print a mediocre fan shroud you don't even need. Once I got back to that spot physical build items that needed to be taken care of (I had a loose z belt and a humped onyx when it was at 90) were more apparent and actually stayed fixed for whatever reason. From there just having to tackle / dial in the materials went a lot smoother than before and I really started to get in sync with the beast. After shot gunning so many fixes and changes and getting pissed, hitting the reset button really helped me and probably prevented me from having to fix a hole in the wall / some other near by item that was the focus of my frustration :p. Sorry to write a book with no real ideas on the issue, but if you get to the fuckit point again, it's something to consider, if only for mental health.

[captainpotatofudge]

Just sliced with MatterControl/MatterSlice, exported to SD and printed. This is AFTER I JUST calibrated Z, adjusted all three towers until they all matched 0.15mm feeler gauge in the middle and at each tower.
I then checked the Z and it was higher than what I had JUST set it at 0.1mm. I even had to LOWER my HR from 141.9 to 141.6 to get them even.
I watched as it layed down the first skirt pass and could see that the filament wasn't even making direct contact with the glass. There was a large gap. I slipped a 0.25mm feeler in there and sure enough it was easily another .1mm above that.
I continued to watch and you can see the results in the pictures. I measured the straight portions of the print with calipers and get 0.44-0.48mm.

http://imgur.com/a/ZlcOm

; Generated with MatterSlice 1.0
; filamentDiameter = 1.7285
; extrusionWidth = 0.4
; firstLayerExtrusionWidth = 0.42
; layerThickness = 0.2
; firstLayerThickness = 0.3
; automatic settings before start_gcode
G21 ; set units to millimeters
M107 ; fan off
T0 ; set the active extruder to 0
; settings from start_gcode
M104 S220
M190 S70
M109 S220
M120 S1 P15
G4 P40
M120 S4 P5
G4 P40
M120 S4 P5
G4 P40
M120 S4 P5
G4 P40
M120 S4 P25
G28 ;home all axis
M117 Printin' stuff
; automatic settings after start_gcode
T0 ; set the active extruder to 0
G90 ; use absolute coordinates
G92 E0 ; reset the expected extruder position
M82 ; use absolute distance for extrusion
; Layer count: 75
; LAYER:0
M107
G0 F11700 X2.756 Y-87.71 Z0.379
; TYPE:SKIRT
<CUT OUT ACTUAL GCODE HERE>
M107
M104 S0 ; turn off temperature
M140 S0
G91
G1 Z10 E-8.0 F12000
G90
G28
M84 ; disable motors
M117 Done!!
M120 S6 P10
G4 P40
M120 S6 P10
G4 P40
M120 S6 P10
G4 P40
M120 S6 P10
G4 P40
; filament used = 20421.1
; filament used extruder 1 (mm) = 20421.1
; filament used extruder 2 (mm) = 0.0
; total print time (s) = 101581
; MatterControl Version 1.5.2 Build 1.5.2.6797 : GCode settings used
; Date 05/20/2016 00:00:00 Time 14:46
; avoidCrossingPerimeters = True
; outsidePerimetersFirst = False
; bottomClipAmount = 0
; centerObjectInXy = True
; continuousSpiralOuterPerimeter = False
; endCode = M104 S0 ; turn off temperature\nM140 S0\nG91\nG1 Z10 E-8.0 F12000\nG90\nG28 \nM84 ; disable motors\nM117 Done!!\nM120 S6 P10\nG4 P40\nM120 S6 P10\nG4 P40\nM120 S6 P10\nG4 P40\nM120 S6 P10\nG4 P40
; zOffset = 0.08
; extruderOffsets = [[0,0],[0,0],[0,0],[0,0],[0,0],[0,0],[0,0],[0,0],[0,0],[0,0],[0,0],[0,0],[0,0],[0,0],[0,0],[0,0]]
; extrusionWidth = 0.4
; fanSpeedMaxPercent = 100
; fanSpeedMinPercent = 80
; filamentDiameter = 1.7285
; extrusionMultiplier = 1
; firstLayerExtrusionWidth = 0.42
; firstLayerSpeed = 25
; firstLayerThickness = 0.3
; firstLayerToAllowFan = 1
; outputType = REPRAP
; generateInternalSupport = False
; generateSupportPerimeter = True
; generateSupport = True
; supportPercent = 50
; infillExtendIntoPerimeter = 0.1
; infillPercent = 15
; infillType = TRIANGLES
; infillSpeed = 25
; bridgeSpeed = 10
; bridgeFanSpeedPercent = 100
; retractWhenChangingIslands = True
; raftFanSpeedPercent = 50
; raftPrintSpeed = 27.5
; infillStartingAngle = 90
; wipeAfterRetraction = True
; insidePerimetersSpeed = 30
; layerThickness = 0.2
; minimumExtrusionBeforeRetraction = 1.5
; minimumPrintingSpeed = 10
; minimumLayerTimeSeconds = 15
; unretractExtraExtrusion = 0.1
; minimumTravelToCauseRetraction = 6
; numberOfBottomLayers = 4
; numberOfSkirtLoops = 4
; numberOfTopLayers = 3
; topInfillSpeed = 20
; outsidePerimeterSpeed = 15
; outsidePerimeterExtrusionWidth = 0.4
; numberOfPerimeters = 1
; positionToPlaceObjectCenter = [0,0]
; enableRaft = False
; raftExtraDistanceAroundPart = 2
; raftAirGap = .2
; supportAirGap = .3
; retractionOnExtruderSwitch = 3
; retractionOnTravel = 3
; retractionSpeed = 140
; retractionZHop = 0.6
; skirtDistanceFromObject = 0
; skirtMinLength = 362.016203522682
; startCode = ; automatic settings before start_gcode\nG21 ; set units to millimeters\nM107 ; fan off\nT0 ; set the active extruder to 0\n; settings from start_gcode\nM104 S220\nM190 S70\nM109 S220\nM120 S1 P15\nG4 P40\nM120 S4 P5\nG4 P40\nM120 S4 P5\nG4 P40\nM120 S4 P5\nG4 P40\nM120 S4 P25\nG28 ;home all axis\nM117 Printin' stuff\n; automatic settings after start_gcode\nT0 ; set the active extruder to 0\nG90 ; use absolute coordinates\nG92 E0 ; reset the expected extruder position\nM82 ; use absolute distance for extrusion
; toolChangeCode =
; supportExtruder = 0
; supportInterfaceExtruder = 0
; raftExtruder = -1
; supportLineSpacing = 2
; supportExtrusionPercent = 125
; supportInfillStartingAngle = 90
; supportMaterialSpeed = 20
; supportType = GRID
; supportXYDistanceFromObject = 0.7
; supportInterfaceLayers = 1
; travelSpeed = 195
; wipeShieldDistanceFromObject = 0
; wipeTowerSize = 0

[captainpotatofudge]

So After that I just checked Z and it is perfect at 0.1mm again. I run this script:
; Rostock Max Tower end-stop calibration script
G28
G1 Z0.2 F15000
G4 S5
G1 X-77.94 Y-45 Z0.2 F2000
G4 S5
G1 X0 Y0 Z0.2
G1 X77.94 Y-45 Z0.2
G4 S5
G1 X0 Y0 Z0.2
G1 Y90 Z0.2
G4 S5
G1 X0 Y0 Z0.2

And I can slip a 0.3mm feeler gauge under the nozzle with a nice bit of resistance at the center, at X, at Y, and just a hair more resistance at Z. SO to ME it seems like it is perfectly calibrated at the moment. Except...the script says Z should be stopping at 0.2mm...not 0.3mm!! This is the something I am experiencing repeatedly. Ultimately infuriated that I can get calibration to STAY put one minute, but then run a script or a print and it's off by 0.1mm.

[Jimustanguitar]

You can change the Z height on the LCD panel, no firmware, scripts, or manual EEPROM entering required. Easy peasey.

Hairspray, glue, tape, kapton, etc will vary in thickness enough from print to print that you'll occasionally want to adjust the Z height for that perfect first layer anyway.

[captainpotatofudge]

I AM changing it in the LCD panel with the "Calibrate Z Height" option....Scripts run from SD card per every instruction I can find written or chatted or forum. Prints run from exported gcode onto SD card. Doing my best to eliminate the software from the equation. I've had good luck in the past printing PLA direct on glass, my issue is not adhesion or layer height based on glue. I ONLY use glue so far and I don't cake it on to the point of being off by 0.1mm randomly. It is displaying repeatedly the same height differentials and differential in the SAME locations making me think it is either electronic (missed steps or something like that previously mentioned) or mechanical somehow.

I'm not sure I catch your drift.

[captainpotatofudge]

I see two likely issues, one is the endstops. Mine lost precision from homing too fast and knocking them around. They seemed secure, but i stopped being able to stay calibrated, which led me to replace the mechanical endstop micro switch with plain magnetic hall sensors and magnets glued to the endstop screws. Inexpensive upgrade and they were something like the ~$1 OH180U sensors I wired to the board and no other fancy electronics. Well worth the increased precision. The other possibility is you're missing steps...which is a lot easier to quantify with a z-probe. (I made mine from a Sanwa arcade controller switch hot glued into a spare mounting bracket it served me well well for a couple of years, although I have FSRs to upgrade to at some point). So.. if it's the missing steps, here's a block of text that I hope helps more than hurts.

I've thought the same for a while, I don't know if I trust those mechanical switches or not. I'd like more info on how you did this hall effect setup! I'm fairly certain I have a few laying around here somewhere!

I'm not convinced I'm missing steps from heat...they do not at all appear to get hot to the touch. Maybe after many hours of running, but definitely not right away. But I'm also not convinced they're NOT skipping steps OCCASIONALLY for some reason.

Since all my troubles started I lowered print speeds to sub 40mm/s. Mostly in the realm of 10-25mm/s hoping it would show me something...it has not. I've had movement speed@195mm/s for some time as suggested SOMEWHERE on this forum as a good high speed without going TOO fast and jerking everything around.

[captainpotatofudge]

I agree with the above. It does seem like an endstop issue, either the carriage mounts, or the switches themselves are most likely. There may also be issues with the endstop wiring being too close to the wiring for the power to something or another. (I understand that EE's should know better, and I know plenty that do, But I've also run across EE's who have been spoiled by shielded twisted pairs for too long to understand the potential issue).

It is very unusual for H-radius to float unless there is some major problem, as it is inherent to the physical geometry of the machine. There's no carriage adjustment to check, although it does sound like there is mechanical slop somewhere in the system. I would go over the system, bottom to top, and check that very screw is solidly in place, pulleys aligned, and all that jazz, if you haven't already.

It may also be as he mentioned, missed steps.

Yea, it's a rat nest in there and I've hated it from day one! If I had built it...no way would it be like that! Wires would have been properly twisted, maybe even loaded with toroids as necessary, shortened to decrease crossover, etc. I wouldn't be a bit shocked if there was capacitive build up in the wrong places etc. Right that's why I don't understand how the HR could change on me like that. Makes no sense from what I understand and have checked thus far. I keep hoping someone will say "hey check XYZ which can affect your HR like that".

I can't tell if there's slop, sure doesn't feel like it to me. Belts are SUPER tight, carriages cannot wobble at all, arms are solid with no give. Not sure what else could cause it. I've checked every screw recently. That's how I found out about the few motor screws being loose. I feel like I'm out of ideas...other than to rebuild.

[captainpotatofudge]

Doming like that is probably just a bad Delta radius setting, unless it's an actual ripple, i.e. if goes low/high/low, in which case it's a mechanical issue.
I would suggest using one of the online calculators with a Dial indicator to calibrate, it's a lot easier than messing around with grabbing pieces of paper or even feeler gauges (which are much better).
I like this one http://escher3d.com/pages/wizards/wizarddelta.php , if you do use it with a dial indicator you will have to negate all the readings from the Dial indicator, because it's designed for use with a Z Probe.
For point of reference it did better in one iteration than I could mange doing it manually, and I understand what all the variables so it's not trial and error.

Ripple! YES that's the descriptor I'm looking for! For example it may go LOW between the X/Y and then HIGH and then LOW at the Z etc. So I get super thin areas and then super thick areas despite having the triple tower calibration script showing near perfect settings. Which I attributed to mechanical stiffness as I take it from what I've read. However I cannot find anywhere in my machine that it makes sense. The ball ends of the CF rods are lubed and free, the belts and everything else move without binding by my feel....so idk.

I do not have a dial indicator at this point and I'm not sure what to do for a mount since I can't print one. Plus I simply do not understand how so many people just assembled it and it works...not using dials, not using feelers and here I am trying to use precise measurements and...I get crap. lol

I've even previously used a number of your suggestions I've found around here...some helped...some landed back to the start. This is why I feel more like there is something wrong electrically or mechanically from the mfg? or maybe from initial construction? ...looks like I know what I have to do.

[captainpotatofudge]

I feel your frustration, I banged my head against mine for the first few weeks after assembly and still spit out profanity at it from time to time . I definately was not prepared for the amount of post build learning and troubleshooting required. From reading your post it does look like you have tried about everything (and selfishly I'm glad I haven't had all your problems) but I do see some similarities with issues I have had.

Ill start of by saying after a lot of trial and error the best thing I did when I got to the "I'm about to uppercut the shit out of you printer!" Stage was to just start over at page one of post assembly instructions. It seems like such a huge step backwards but after all the learning / troubleshooting it's fairly quick and painless and helps guide your troubleshooting.

Reload fresh firmware, tighten the end stop switch screws and screws on the carriages all the way down and run through the instructions (with the addition of firmware changes for your upgrades) calibrate a little and print a mediocre fan shroud you don't even need. Once I got back to that spot physical build items that needed to be taken care of (I had a loose z belt and a humped onyx when it was at 90) were more apparent and actually stayed fixed for whatever reason. From there just having to tackle / dial in the materials went a lot smoother than before and I really started to get in sync with the beast. After shot gunning so many fixes and changes and getting pissed, hitting the reset button really helped me and probably prevented me from having to fix a hole in the wall / some other near by item that was the focus of my frustration :p. Sorry to write a book with no real ideas on the issue, but if you get to the fuckit point again, it's something to consider, if only for mental health.

It just seems like this level of frustration and problems is FAR from normal. I've read NUMEROUS build threads on here and none of them had anywhere near these problems.

I've gone back to manual so many times it hurts to look at it. Firmware reload...yup done that. Screws tightened yep don that. I've calibrated with cubes, triangles, pyramids, circles...robots...loads of stuff. I'm just at the point of following the 'tear down and rebuild' option...

[Xenocrates]

If it's such a rat's nest, I would take out the screws that hold the top of the bottom plate on, disconnect everything, then pull the wires such that they hang out of the sides while you re-do the cabling and add tie points. If the endstops are falsely triggered, then it can really mess up your printer's idea of where it is in space. If you don't mind, a picture of the cabling before, during, and after the cleanup may also help.

As far as how most people just build it and it works, I think the big issue is that there are people (Like the EE who helped your father-in-law), who think they know a whole lot, and can do better than the manual, while they lack the skills to actually improve on the manual, or the domain knowledge to understand why a particular thing is done the way it is. And because they are educated, they don't bother to look stuff up, or listen when the manual says "Take care with this"

[captainpotatofudge]

If it's such a rat's nest, I would take out the screws that hold the top of the bottom plate on, disconnect everything, then pull the wires such that they hang out of the sides while you re-do the cabling and add tie points. If the endstops are falsely triggered, then it can really mess up your printer's idea of where it is in space. If you don't mind, a picture of the cabling before, during, and after the cleanup may also help.

As far as how most people just build it and it works, I think the big issue is that there are people (Like the EE who helped your father-in-law), who think they know a whole lot, and can do better than the manual, while they lack the skills to actually improve on the manual, or the domain knowledge to understand why a particular thing is done the way it is. And because they are educated, they don't bother to look stuff up, or listen when the manual says "Take care with this"

Will do! I'm not sure when I'll actually get started, maybe I'll start with little pieces on the teardown portion until I get more time to rebuild the entire thing. Very well could be an over confidence problem!

I don't look forward so much to pulling all the wiring out of the towers and such, but it's for the best. I ran twisted wire down the X tower for LED lighting I installed on the top, but I've never been completely happy with it. Rush job. Now I'll make it better.

Thanks everyone for the input! I know what I have to do, I just don't want to squeeze in the time for it right now. I also really want to find more on how everyone has done their Z probes, dial indicators, etc.

[marcom]

Sounds frustrating. The two times I've had problems with skipped steps were the following issues:

(1) Loose grub screws - you mention tightening the motor screws but not sure if you meant the mounting screws or the pulley (grub) screws - if you haven't done those give it a try, and verify that they're locked down with blue loctite. I had this problem with another printer where the instructions didn't call for it.

(2) Motor currents - make sure your firmware settings match the motors you actually have (Kysan or ATI or Wantai or whatever) - a mismatch there might also cause lost steps.

Apologies if you've already covered both of these points in your troubleshooting but just sharing my thoughts in the hope it helps you or someone else who comes across this in the future. Best of luck!

[DeltaCon]

I am skipping over the last few pieces of advice here, but if I understand correctly the issue is that if you set z0, home and go to z0 it actually does not read z0, correct? I have had this crazy making issue before and I nailed it down to mattercontrol. Since then I turned to repetier host and never saw that again. Later I related this issue with probably auto leveling set on. It is a mattercontrol feature that actually intervenes in the height the head is going to. I think if you do some manual calibration auto leveling is not correct anymore.

[captainpotatofudge]

Sounds frustrating. The two times I've had problems with skipped steps were the following issues:

(1) Loose grub screws - you mention tightening the motor screws but not sure if you meant the mounting screws or the pulley (grub) screws - if you haven't done those give it a try, and verify that they're locked down with blue loctite. I had this problem with another printer where the instructions didn't call for it.

(2) Motor currents - make sure your firmware settings match the motors you actually have (Kysan or ATI or Wantai or whatever) - a mismatch there might also cause lost steps.

Apologies if you've already covered both of these points in your troubleshooting but just sharing my thoughts in the hope it helps you or someone else who comes across this in the future. Best of luck!

1) Yup I made sure those were tight also. No loctite added by my and idk if there was any used during initial build. I definitely plan to use it during the rebuild! Thanks for the reminder.

2) Hmm...I can say for sure I didn't check that when I downloaded firmware and reflashed...crud...Thanks for pointing that out!

[captainpotatofudge]

I am skipping over the last few pieces of advice here, but if I understand correctly the issue is that if you set z0, home and go to z0 it actually does not read z0, correct? I have had this crazy making issue before and I nailed it down to mattercontrol. Since then I turned to repetier host and never saw that again. Later I related this issue with probably auto leveling set on. It is a mattercontrol feature that actually intervenes in the height the head is going to. I think if you do some manual calibration auto leveling is not correct anymore.

That's correct, it often changes what it perceives as Z0. I've had a few strange things occur with MatterControl, but I kept updating to the newer version anyhow. Back on 1.4 that I started with I didn't have ANY problems...aside from the physical doming and leaning towers. I've also experienced on quite a few occasions, totally random movement by the steppers. I should have put that in my initial write up also! There have been times when I'm homing everything and they will just go haywire right as they hit the end stops or just after. No pattern to it, just random movements sometimes up to 1/3 of total movement range.

So it goes like this: I hit Home All in MatterControl, all carriages move upward at a normal pace nice and smooth, they hit the end stops and immediately let's say the X carriage drops super fast 1/4 of range, the Z drops 1/3 and then ZIPS right back to the top super fast and repeats, while the X slowly moves back up at the same time. Super weird stuff and since it's been totally random so far I haven't been able to capture it on camera. I REALLY wish I could have. And it seems each time it happens if I close MatterControl, the behavior stops immediately.

And now that you mention it, I had used stand alone Cura, not slicer built into MatterControl, a few times and it seemed to behave perfectly as I recall for Z repeat ability...hmm.

[DeltaCon]

Does it look anything like this?

http://forum.seemecnc.com/viewtopic.php ... 258#p91066

In this example I still have to move the skate myself, but in your case the skate could drop by itself as the belt tension is too low.If it looks anything like this it is a much too high motorcurrent on your steppers. In my case that was due to Repetier changing the syntax from x/255 to percentage values in the motorcurrent statement in the firmware.

[captainpotatofudge]

Does it look anything like this? In this example I still have to move the skate myself, but in your case the skate could drop by itself as the belt tension is too low.If it looks anything like this it is a much too high motorcurrent on your steppers. In my case that was due to Repetier changing the syntax from x/255 to percentage values in the motorcurrent statement in the firmware.

I think you missed an attachment or something....