Maybe Belt/Motor Skipping?

[dpmacri]

According to your signature, you're still running Marlin firmware. Is that right? I'm pretty sure the majority of people here are using the Repetier firmware. Perhaps that's the major difference?

[analog_banana]

I hear you!
Yes I do want to switch to repetier - It's a question of changing one thing at a time.
I'm pretty sure it's NOT firmware - how do you think they got these things running well back in 2013 before Repetier firmware?

[bubbasnow]

on that video, @8 minutes there was some grinding noise?

those G28 homes were not skipping, have you fixed that issue?

Motion between looked pretty smooth, i was hoping it was some skipping and such....not the case.

If you were to disconnect the belt for the cheapskates how loose is it?

[Flateric]

Please also check your pulleys and see if they are cracked around the plastic where it meets the aluminium hub. I have had 2 pulleys fail in such a way and it was a major PITA to track down the cause. I switched over to all aluminium pulleys and have obviously had zero problems with this since. Although I suspect this is a relatively rare problem.

[Eaglezsoar]

Please also check your pulleys and see if they are cracked around the plastic where it meets the aluminium hub. I have had 2 pulleys fail in such a way and it was a major PITA to track down the cause. I switched over to all aluminium pulleys and have obviously had zero problems with this since. Although I suspect this is a relatively rare problem.

Not so rare, I have had two of the plastic coated pulleys fail in the same way. I suspect it is a very common problem.

[analog_banana]

Bubbasnow:
The skipping is at 15 seconds on my shorter youtube video: https://www.youtube.com/watch?v=IjgRf8SZw-s
The skipping occurs occasionally - maybe once or twice a day if I am using th eprinter for several hours a day. (Which I usually am)
I plan to check the cheapskate looseness in the next day or so as part of a complete recalibrate / ground up retest.

Flateric: I have solid metal 15 tooth Pulleys - the old versions, so I don't think they will decompose.
Out of curiosity - why were they changed in later machines. What is the problem with the metal gears?

I have been putting in an average 8 hours a day for most of January on this machine.
I think I will give this machine 2 more weeks before I remove what parts are useful and give up.

Here is the list of things I am going to try,If anyone has any other ideas for testing, do let me know:

Try different ABS Filamet
Check Wiring going out of Ramps 1.4
Re-test Hot end / Heated Bed temperature reporting accuracy
ReTune PID for hotend and bed
Check Cheapskate movement - using the video I saw somwhere to show how free they should be
Check belt tension.
ReCheck Aluminium pillar right angles and straightness
ReCheck Delta arm bearings for smoothness and wobble
Recalibrate EZstruder feedrate
ReCheck filament output length at hot end
Recheck Radius
Relevel bed
Try using a different control computer

Should I be looking at any of these:
Current going to motors??
Acceleration / Retract Acceleration??
Min Feedrate settings?
Jerk settings??
Tuning Trim pots on Pololulu??

[analog_banana]

This is my Ramps.
THe only wierd thing is that the extruder motor has to be wired in the reverse to the arm motors to work correctly.
My trimpots are all look the same - I have not touched them.

Generally all looks ok...

[analog_banana]

Retesting done so far:

Re-test Hot end / Heated Bed temperature reporting accuracy = Now very accurate
ReTune PID for hotend and bed = Very consistent over multiple tests
Check Cheapskate movement - See video below:

I think they are running ok - no signs of any real wear.
ll bearings rotate when skates are moved.
No play or looseness.
Now I must retension the belts...

[youtube]http://www.youtube.com/watch?v=AtTH4tCtDEQ[/youtube]

[analog_banana]

Bubbasnow:
The grinding at 8 minutes - I think it's just resonance causing some vibration - but I will do antoher test once the belts are retensioned.
I wont bore you with another 10 minute video - I will only uplad any anomalies if they occur.
Big thanks for watching my test! I appreciate it

I have bought some ultra-high quality abs from one of Londons many (2) 3D printing boutiques.

[analog_banana]

Has anyone come up with a legitimate way to tension the belts yet?
The manual says "I wish I had a better way of doing this"

I can't find anything thru search.
While fun, I'm not convinced the advice of using a bag of canned beans as leverage is accurate.

[geneb]

Here's a way to get the right belt tension.

[img]http://www.geneb.org/images/ok.jpg[/img]

Press a finger to the spot pointed to by the arrow with your right hand in the shape indicated. That's about as tight as the belt should be.

g.

[analog_banana]

Thanks! Very Helpful!
Apologies for that hand signal to board members in Venezuela, Turkey and Brazil.

I have just found that pushing down on each cheapskate while the machine is powered and homed helped.
There was a tangible difference in "feel" and travel on one belt, so I tightened it a bit.

[analog_banana]

Testing for pillar alignment.
I put a high quality engineer's square onto the glass bed to test squareness.
Each pillar displayed about .5mm to 1mm gap either at the top or bottom of the square.
This could just be the bed being off - which will be compensated for in calibration.
I think this is a tough call to make?
Are the pillars slightly out?
Is the bed slightly angled?

I would have to take the bed off to do more testing.
Maybe I'll save that for another rebuild attempt...

On to extruder calibration..

Look at the bottom of this photo for the small gap...

[analog_banana]

I have performed the Partial Tear- Down, Total Recalibrate as described previously,
Plus thr addition Of PLA fan, A temporary Switch to PLA (Faberdashery PLA, arguably best quality PLA available)
Temperatures written on cubes.

Settings:
; layer_height = .2
; perimeters = 3
; top_solid_layers = 3
; bottom_solid_layers = 3
; fill_density = .2
; perimeter_speed = 30
; infill_speed = 30
; travel_speed = 130
; nozzle_diameter = 0.5
; filament_diameter = 1.75
; extrusion_multiplier = 1.25
; perimeters extrusion width = 0.50mm
; infill extrusion width = 1.02mm
; solid infill extrusion width = 0.85mm
; top infill extrusion width = 0.85mm
M190 S55 ; wait for bed temperature to be reached
M104 S200 ; set temperature


It looks like maybe I have Fliament starving, as I had to up the extrusion_multiplier to 1.25 and the top is still looking a bit sparse.

The main BIG issue is the squishy, blobby expanding corners.
I made a movie to to show the effect : 0:55 in movie is a good example:
The head pushes the corners all out of whack as they rise up and get squished.

My goals are to:
learn more about measuring extrusion length and width at the hotend.
Figure out this squishiness issue.

[youtube]http://www.youtube.com/watch?v=72UtRMj_UAE[/youtube]

[DavidF]

try putting three of them on the build plate and printing again. This will give you more time for cooling between layers...

[analog_banana]

Heated bed is Borked

I AM SO CLOSE TO TAKING THIS THING AND JMPING ON IT TILL IT IS DUST.
THEN JUMPING ON THAT DUST TILL IT IS MOLECULES
THEN BUYING THE LARGE HADRON COLLIDER AT CERN IN SWITZERLAND.
AND LOADING EVERY ONE OF THOSE MOLECULES INTO THE BIG MAGNETIC DONUT.
AND REDUCING EVERY ONE OF THOSE MOLECULES INTO ATOMS.
AND TAKING THOSE ATOMS..
AND...
GASP....
HUFF......

I'M SO NEARLY OVER THIS

[analog_banana]

PS I had specially bought the Rostock Max the best quality ABS available from the Netherlands.
Why U not stay alive for special ABS I bought you??

[int2str]

Banana, why do you have your extrusion width all over the place?

Code: Select all

; nozzle_diameter = 0.5
; filament_diameter = 1.75
; extrusion_multiplier = 1.25
; perimeters extrusion width = 0.50mm
; infill extrusion width = 1.02mm
; solid infill extrusion width = 0.85mm
; top infill extrusion width = 0.85mm

With a 0.5mm nozzle, set all of them to 0.6mm for now, period.
Set the extrusion_multiplier back to 1 and print a simple single walled cube with no top. Post a picture of the result of that and measure the wall thickness.

You can see in your print video that you're way over-extruding (IMHO).
It looks more like a cube now (despite the rounded corners), so you are making progress.

Keep
Calm
and
Calibrate
On

[geneb]

Ok, so you've eliminated as problems:
The Power Supply
The Hot end (good temps)
The motors (they all work, no skips)
I'm also pretty sure you've made sure the cheapskates move properly and there's no binding or looseness in the delta arm connections (u-joints, etc).

The only things I can think of that it could be are these:
1. The RAMPS
2. The Polulu stepper drivers
3. The firmware.

The biggest issue we've got with being able to help you is the electronics you're using. There's only a tiny handful of people here that have any experience with the RAMPS board and the Polulu drivers.

Have you done any kind of adjustment to the Polulu drivers? It's my understanding that since they don't have electronically adjustable pots (the RAMBo has "digipots"), they have to be adjusted by hand. It could be possible that the drive current is incorrect, but I don't know how to check or adjust that.

You list Marlin 1.0.0 as your firmware - would you be willing to give Repetier .91 a shot?

g.

[Eaglezsoar]

Ok, so you've eliminated as problems:
The Power Supply
The Hot end (good temps)
The motors (they all work, no skips)
I'm also pretty sure you've made sure the cheapskates move properly and there's no binding or looseness in the delta arm connections (u-joints, etc).

The only things I can think of that it could be are these:
1. The RAMPS
2. The Polulu stepper drivers
3. The firmware.

The biggest issue we've got with being able to help you is the electronics you're using. There's only a tiny handful of people here that have any experience with the RAMPS board and the Polulu drivers.

Have you done any kind of adjustment to the Polulu drivers? It's my understanding that since they don't have electronically adjustable pots (the RAMBo has "digipots"), they have to be adjusted by hand. It could be possible that the drive current is incorrect, but I don't know how to check or adjust that.

You list Marlin 1.0.0 as your firmware - would you be willing to give Repetier .91 a shot?

g.

Gene, you are 100% correct about the Ramps board. Functionality is exactly the same as the Rambo except it does not have locking connectors and the potentiometers have to be adjusted manually.
With Ramps, to get the driver current set accurately you need to measure a reference voltage on the driver while turning the potentiometer. There are charts on the web that show the measured voltage to current output.
Most users "guesstimate" the current based on potentiometer position because of the difficulty in measuring the reference voltage, turning the potentiometer, and not short anything out. You need deft hands and three of them to do it right
so I usually recommend using the position of the potentiometer as a guide so that the user doesn't burn up the driver by shorting those tiny and tightly packed traces. Nothing that I have seen from the pictures would indicate that he has a stepper current
problem so my advise is not to adjust those potentiometers at this time. They are delicate little buggers and are easily broken and unless there is a specific reason they are best left alone.

[analog_banana]

Pretty much lost my temper there.
A long dark teatime of the soul led me to thoughts of stripping the reusable iron form the machine,
taking the mdf components out to the garden and lighting a big bonfire.

Anyway....

Looks like the ramps overheated by being shut inside the body of the rostock max.
Yes it has the approved fan running on it costantly.
Looks like it needs more airflow than the Rostock max casing allows.
All working again now I have the door with the electronics open.

So.. now I'm using the best ABS money can buy.
I have calibrated each component till it hurts.

Re: Extrusoin rate was "all over the place"...
because I had reverted to the settings described in the Rostock Max Manual. - which says leave on auto - which leads to them being all over the place.
I tend to go back to those pdf manual settings because - what other baseline do I have??
But thanks for the advice, I'm listening.
I have been given alot of different advice, re this, often conflicting.
I listen to it all, very gratefully, and attempt to synthesize a path from the divergent data.

Printed the Mhackney single wall cube with these settings:

layer_height = .2
; perimeters = 1
; top_solid_layers = 0
; bottom_solid_layers = 2
; fill_density = .2
; perimeter_speed = 30
; infill_speed = 30
; travel_speed = 130
; nozzle_diameter = 0.5
; filament_diameter = 1.68
; extrusion_multiplier = 1.0
; perimeters extrusion width = 0.60mm
; infill extrusion width = 0.60mm
; solid infill extrusion width = 0.60mm
; top infill extrusion width = 0.60mm
; first layer extrusion width = 0.60mm
M190 S80 ; wait for bed temperature to be reached
M104 S245 ; set temperature

Maybe my bed temp could be higher.. to which I say, bite my shiny metal ass,
I have bed adhesion, and 100 degrees plus takes forever to wait for..

So here it is:

Walls averaging .25mm!! way to thin??

So - underfeeding filament I guess...

I'm gonna give the Pololu thing some thought.
Maybe hang out on the reprap boards as well to get some info..

[analog_banana]

Might try and contact the guy who sold it to me re the Ramps issue
But he's given up selling Rostock Max. Genuine quote from vendor:

"We actually stopped selling them as while they are nice machines - they seem too fussy to get working well and I was not getting the support from Seemecnc at the time."

It's hard to stay positive when that's what the dealer tells you...

I wouldn't normally spill those beans, but I'm getting tired of this and I honestly think that UK rostock maxes were pitifully undersupported and it was a stupid idea to put different electronics in them.

There were a couple of guys on this board who bought the UK issue. They never completed their build logs. Not even first print cube photos..
They're probably more sensible than me and got on with their lives...
I have IM'd - no reply...

I might have to face the facts that I am using a frankenstein machine and all the advice you guys give, while awesome and generous.. may be for a different ecosystem.

But.. on the other hand, maybe the only difference id the pololus... I just need a pololu trimpot expert maybe...
Any other differences other than the stepper drivers??

Time to upgrade the firmware next I guess..

[JolietDelta]

Hey Banana!

I hope you didn't atomize your RMax already. :-) If you did not, here are some thoughts in the early hours. Maybe they are helpful.

Besides my RMax I have some (cartesian-style) printers running. All of them with homebrew Arduino-Shields and with RAMPS-alike A4988 (pololu) based stepper drivers. Firmware is repetier V0.8x. I know that adjusting the trimpots is a little bit tricky and I received not even two of them with nearly idendical trimpot settings.

My suggestion is: before you tweak the trimpot settings (maybe from "bad" to "Poof!"), try printing a round calibration cylinder (like this one here: http://www.thingiverse.com/thing:215101). Speed set to 15mm/s, and with no further modifications of your other settings. The idea behind, is to print an object that causes no hard positive or negative accelerations.

My humble experince is, that if you are skipping steps, then every subsequent layer will be shifted in one or two dimensions too. But this seemed not to be the case with your cube. You got round corners (and it looks like the corners in the front of your picture are more "round" than the ones facing to your Z-Tower). I had the same effect when one of my timing belts had too much play and the effector behaved like guided by a rubber band.

Skipping steps but "remembering" them is often caused when the A4988 driver is triggering its built-in thermal protection. But in those cases the printed results are looking more "ripped" and not as smooth like your cube.

I got those round corners with a) too low extrusion temp (no good layer adhesion); b) first layer printed out too high (bad adhesion on heated bed); c) underextrusion - and d) loose timing belt(s).

The key for fixing your problems is to modify only one parameter at a time. Then printing the same two calibrations objects and observe waht has changed. Everything else will make things even worse.

Don't give up. The RMax is a great machine and it can also produce great results.

(strange punctuation and grammar is intentional :-) )

Good night,
Jens

[geneb]

All things being equal, there's got to be a solution to your problems. Printing a cylinder is an excellent suggestion - it should remove the jerk settings from the motion equation. We WILL get this figured out.

g.

[analog_banana]

Thanks for the support!

Here's a cylinder test.
I don't think it's a too low temp issue - maybe It was too high, based on the experiments here:
http://forum.seemecnc.com/viewtopic.php?f=37&t=4192

This cylinder seems to be fairly straight - a liitle bit of warping just near the base due to ambient temperature / bed temperature maybe?
You can see some artefacts where the head moves from the outer wall to the inner wall of the two shells.