Am I getting tower lean?

[NormL1]

New delta and home printer user here. I am trying to print an object that is out near the print limits. When I first tried, I got material floating in the air close to the edges and areas of no print. I found a discussion on belt tightness effecting lifting at the edges. Sure enough, a belt tightening where the issue was occurring did fix it. I still have areas where is does not print until the second layer though and my forum reading has me thinking it is tower lean, but, with that said I have no idea how to remedy that.

This is what a first layer looks like:
[img]http://www.imgtree.net/files/qw1gqmj4a8koohmrvg50.jpg[/img]

Oh, and while I am asking stupid questions, I am going to eventually try to print an HO scale Bilby tower. Not far enough along with the machine yet, but, I am thinking of the issues. A Bilby tower is what surveyors in the mid west used to get their instruments up very high, so, they could turn angles to distant objects. It is an instrument tower surrounded by a man tower. My issue is there isn't a lot of substance to them. Article http://oceanservice.noaa.gov/news/featu ... tower.html How would one even think about print such a spindly tower? At HO scale it is slightly over 400mm, so, I am going to have to remove a section or play scale games. This is the outer tower model I made a couple of years ago
[img]http://www.imgtree.net/files/8yx3ebfg06eg7nlsqbmd.jpg[/img]

[mhackney]

That looks more like a bed leveling issue in the Y direction (towards the Z tower in the back and opposite it in front) to me - which would be corrected with auto calibration. Have you run the autocal script? These machines do change slightly throughout the day or week due to temperature, humidity and phase of moon. I calibrate at least once a day.

On your model, I would think it would be pretty straightforward to print each of the 4 sides - probably in 2 sections due too their length - and then assemble the parts. The top section might be handled differently, probably the two platforms and 4 sides. I wouldn't even think about printing that upright with all of those spans. I suppose you could design in "zipper" support and get away with it but that would be a lot of design work.

[mhackney]

I should also have mentioned, you can check tower lean with a good machinists square. But the first place to start is with the auto calibration. Unfortunately, the Repeater autocal does not correct for tower lean or rotation. The Duet dc42 firmware's autocal does but that won't help us!

[NormL1]

I did run the script just before and should have mentioned that. I have had this issue since assembly and I started printing bigger. I do have a square that should work as a check

As far as the model, you are correct, that would be a much better way to print around the issue. I just I am still spoiled by not having the support worry with laser sinter

[mhackney]

Make sure to check your endstop switch on that Z tower. Any slight movement can result in odd calibrations. Check all of them, but the Z tower is in the primary suspect based on the print.

[NormL1]

Will do Obi-Wan, sounds like a likely suspect, thanks

[mhackney]

Here's a little old school trick...

If you have an obvious bed leveling issue in the direction of a tower you can tweak the endstop screw on the carriage to adjust it. In this case, since the first layer is thick towards the tower, you want to make the tower length "longer" so you would screw the endstop screw out a little. Start with 1/8 turn and test print and repeat until you nail it.

[NormL1]

OK, I was under a bad assumption with the screws, I thought they were software adjustments for height. When I set up the touch screws on the skates for the switches, I didn't find anything in the instructions about height. I placed mine at 11.5mm from the plastic surface behind the screws and made them uniform. I will adjust the "Z" by 0.5mm and gauge results

[mhackney]

One other thing (this must be why I have such a high post count!), how tight are your bed clamps? As I've mentioned several times lately (today in fact) 6 clamps is over constraining holding down the bed and could potentially result in a warp of the bed surface. I only use 3 of mine, that guarantees a perfect plane. Simply rotate every other clamp until it disengages the bed. You also want to make sure the support on the bottom side it is screwed to is loose so it is not interfering with level.

[NormL1]

Interesting point as I was wondering the same when I built it. Did the shoulder shrug and moved on. I have all six on, but, they are fairly loose as I just was very uncomfortable placing torque on glass

[mhackney]

Yes, the "new" way to adjust heights is compensating in software. But long before there was delta auto calibration and long before the extinction of the Carrier Pigeon, we calibrated our printers manually like this. It is quick to tweak a screw and works well.

I'm just anal retentive enough that after I run an autocal I check my tower heights and if they are off more than .1mm I tweak the offenders to get them all really close and re-run autocal. The Rostock Max (and Max Metal) are unique in this respect, the other common deltas out there have fixed end stop switches and triggers so the correction is always done in firmware/parameters.

[mhackney]

Twist 3 out of the way and tighten the other 3 to hold the plate securely. Guaranteed no warping and no plate movement. Over contained systems is one of my (many) soap boxes!

[NormL1]

Can I assume that you mean you check the results given to the Eprom to see the height difference as opposed to actual measurement? My height micrometer isn't that high I am afraid

[mhackney]

Yes, the offset parameters in the EEPROM are actually saved as "steps". You can do the math to determine the actual distance since the X, Y, Z move at 80 steps/mm. So 8 steps (or 8 in the EEPROM) would be .1mm.

The parameters are Tower X endstop offset (and Y and Z).

[NormL1]

I'm on it. Tomorrow, I will post the results. Thank you for your extensive incite

[NormL1]

OK, the most telling was the square with none of the extrusions being what I would call good. X and Y were not that far off, but, Z gapped a lot going up. I took all six covers off and all six extrusions were in fact seated on the stops. Is the only physical way to get to 90° to gap the extrusions to the stops until all are 90°, then of course loosen the tension on the belts a bit? It seems to me to be the place to start and then start adjusting the touch screw heights

[mhackney]

Hmm, I would have thought they would be pretty good with the injection molded parts. Can you loosen things up, wiggle and jiggle into alignment and then tighten? It could be that things were misaligned when tightened down and are under a little tension and might relax to a natural alignment. Worth a shot!

[NormL1]

I was able to achieve 90° by loosening the lower Z and forcing the extrusion up a few millimeters and now I want to run a test, but, Octopi.local can't resolve today

Edit, as soon as I posted, it could resolve, ... more

[mhackney]

Welcome to the wonderful world of networking!

[geneb]

Welcome to the wonderful world of notworking!

Fixed that for ya.

g.

[mhackney]

LOL, it works both ways!

[NormL1]

No like button, aargh!!

[NormL1]

Well I gave up for a while as I just couldn't get a print that was level. New firmware, a couple of ferrite beads and new delta calibration and level isn't an issue anymore and now I can just get to the actual printing issues. If I try to print a vertical wall it will be shifted layer by layer to the "Y" tower. The head assembly isn't hitting anything that would make it re-center and I am too new at this to know what else to look for. I have multiple issues though, layer separation and stringing that are going to keep me busy chasing down, but, I don't have any idea where to look for layer creep.

[mhackney]

a consistent and repeated shift along the line to a tower is a sure sign of a mechanical issue - Y tower in your case. Check the stepper pulley is secure, they can loosen. Also check belt tightness and make sure the belt is routed correctly. More than one user routed the belts wrong. Also check that the carriage is not moving and that all hardware is secure.

In general you can divide problems into several camps:

mechanical issues

extrusion issues

slicing issues

It is pretty straight forward to separate these so you can focus on the problem. If you have more than one of these, start with mechanical. Slicing issues should be eliminated by using a known working profile or even gcode prepared for your printer.

Layer separation could be due to mechanical and/or extrusion issues so start with mechanical first.

Stringing is an extrusion issue fixed by optimizing slicing parameters. Do this last.