adarcher's build/troubleshooting

[geneb]

The chances that the hot end is to blame is vanishingly small.

I printed four of the small two-post parts last night at 190C and there's no strings at all.

You need a thermocouple to compare what the controller is reporting to what is actually going on. The most likely culprit here would be a thermistor that's out of tolerance and reading much lower than it actually is. It could also be that the beta value for the hot end thermistor in config.g has been changed from what was supplied, causing the firmware to read it as cooler than it is.

Stringing is caused by either retraction, temperature, or both. PERIOD. The hot end is simply the conduit for those things and cannot impact the stringing phenomenon directly. It would be like blaming the engine in your car for overheating constantly when you're feeding it 110LL Avgas.

You can get a meter with a K-type thermocouple probe from Amazon or Harbor Freight - no 3D printer owner should be without one. If you get one, set the temp on the machine to 190C and then touch the thermocouple to the spot where the nozzle enters the heater block. It's not the most accurate measurement point, but it'll get you within a few degrees of the internal temperature of the heater block.

g.

[wepollock]

Another tool for my toolbox... yeah! @ $6 on ebay that is a no brainer.

I was wondering how much play existed between reported and actual temperature. The infrared heat meter i have is good enough for the bed but sort of useless for the nozzle.

I just bought a $99 RCA windows tablet as my printer control panel.. I am going to plug my logitech webcam into it and then link the webcam into the duet browser application.

[adarcher]

BTW, I'm asking for a new hotend kit because I wouldn't be able to unscrew the thermistor without destroying the heat block. The set screw for that, the screw for the heater, the terminal blocks are all so tight they are beyond me to loosen.

I already wanted to put in a new thermistor to test that... had to give up. Stripped set screws are no fun.

*edit: just got back from harbor freight. Got a K-type thermocouple and it measured 190-195C for my 190C setting...

*edit2: Alright, good thing I complained in public about not being able to do something... While at harbor freight I got a new small screw driver set and was able to loosen the screws--with a LOT of finesse and patience. I'll need to find a new JST connector if I want to try a different thermistor though, couldn't find my extra.

[adarcher]

Recap for today's work:
Got a thermocouple sensor and confirmed my hotend temps weren't too far off from what they should be... according to the sensor.

Tested down to 165C on the same gcode I posted. I couldn't get it to stick to the bed that low, so I let it be 200 till it started the circles, then set it to 180 then 170 then 160. It finally jammed at 160 but was consistently stringing the same till then.

I got the hotend screws to loosen, so I tried my e3Dv6 just for kicks. Same thermistor since I don't have the JST connectors in my extras box. Same stringing.

Tried retraction settings varying between 3.2-8mm, of course it jammed at 8mm. Tried 45-80mm/s speeds too.

I fooled around trying to see if there was slop in the bowden tube anywhere. Found some! It was in the exit connection of the extruder... shimmed that tight. Still strings. Made new bowden clips just in case as well.

I'm wondering about the stretching needed for the bowden. I checked another spool of pla. Same thing. Does pla become more stretchy if left in the open air? I know it get brittle.

I tested the stretch of the filament by turning the extruder nob by hand back and forth. With the hot end cold and gripping it's end, I could turn ~5-7mm either direction before feeling a significant stopping force. Captubes seem to live up to their reputation... didn't feel any friction going in/out of the tube throughout the day. Even when I tried pulling a bit that was wider from being in the hotend, where the old tube would have jammed, this just took a little more force and came out with a slight pop. I think seemecnc should start using it standard

[geneb]

I told you it wasn't the hot end's fault

If you've got material stringing at low temperatures, I'd be concerned about the material or your slicer settings. Did you try printing the gcode that I uploaded?

g.

[adarcher]

Not yet, tried to modify it so it didn't do the bed cal, but after a few tries of it going nuts, I used my own. I really can't do the G32, the FRS's don't work well at all at bed temp and I don't like seeing the effector tilt to 30 degrees when probing.

I do wonder about the slicing and material, but I don't have the problems with my V2 on the same material, and the slicing was from a fresh install of cura and the profile.

I'm working from home today, so I'll look at your gcode a little more carefully at lunch.

Did it real quick just now. I had to slow it down to 50% for the first layer. Didn't want to stick. You can see the almost completely failed first set there, second from the camera.

Added a pic of my shim. That aluminum piece there moved quite a bit and even dislodged the makeshift shim during print so I doubled it up and restuck it.

[wepollock]

G posted a G-code profile for you to test. Did you try that?
Are you using a fresh roll of filament?

M. Hackeny who posts here has a blog called Sublime layers that has some worthwhile tips.

http://www.sublimelayers.com/2017/10/mu ... art-1.html
http://www.sublimelayers.com/2017/10/mu ... art-2.html


I would say the Seemecnc hotend is ahead of E3D, the effector certainly is. From where the Artemis is in terms of hardware components; any change will just be a marginal change. I am not saying a change will be a marginal improvement; some changes will actually downgrade your printer or it will require software re-tuning.

Coincidentally, to Hackney's blog, I am using blue capricorn PTFE on my Rostock Max V2. The cost of 2m of tubing and a cutter is $25. I cannot provide a metric to say whether or not I have "upgraded" or "downgraded" my printer with this change.



Some large marginal improvements;
== enable the effector to carry a large payload without backlash with the extruder on the effector
== enclosure - low hanging fruit
== software tuning (we are near the point where there is not much more can be done with hardware; so any little change for marginal improvement will add cost w/o necessarily adding benefit.)

Some small marginal changes;
There are tons of things which could be changed to provide small marginal improvements; you might be adding cost in excess of the potential benefit.

[geneb]

I bet the root cause is the bowden tube not being fully seated in the extruder. If it's moving as much as you say, it's robbing part or all of the travel of the retract.

Also check to make sure the bowden isn't moving at the hot end-end.

Once again, it's NOT the hot end.

g.

[adarcher]

Never said it had to be the hotend btw, I just like to test everything when things go wrong. I had an extra v6 and it quelled that question in my head. I also figured if I made the setup more similar to a tried and tru setup, I'd have a better idea of what to look at. Also, so far I like the v6 more than the se300 and don't see how it could be a down grade--not having a good print from the se300 doesn't help it. They are very similar, but the v6 is more mature.

Bowden is firmly inside the extruder and hotend as much as possible. I spent a good deal of time trying to verify the bowden was seated all the way in the hot end--to the best of my ability, it is/was. It was kinked a bit at the start, but not any more.

I asked for a new thermistor but haven't heard back yet.

This shouldn't be a big deal but it is. Nothing looks wrong at the moment so after all the obvious fixes have been exhausted, the non-obvious need to be looked at.

[adarcher]

Just an FYI:
When I switched back to the SE300, I looked at the captube section I put inside since the original was kinked. It was compressed, so it was a little tough to get filament through it. Looks like the captubes are slightly wider than the ptfe tubes SeeMeCNC uses. I even tried the other tubes I had from other machines. The only one that fit nicely in the heat break of the SE300 was the kinked one. I straightened it out as much as possible and made sure not to force the top end on it too much--the obvious reason it was kinked.

This is a minus for the SE300, that little insert is pretty easily bent in there. Maybe I'll ream out my captube section (now that it's mostly waste since it's so short) and try it again next weekend.

This was midday last weekend--just forgot to mention it--so it doesn't change my current state sadly.

[geneb]

Yeah, don't go messing with the PTFE tubing in the hot end. The original may have been over compressed (but still didn't affect the performance of the hot end), but the tube needs to be held firmly in place in the hot end. Don't go "reaming" it out, you'll just guarantee a poorly performing hot end, if it works at all. You really need to stop abusing that poor hot end.

[edit]
Also, what on earth is a "captube"? No 3D printer I've ever seen has anything called that.

g.

[wepollock]

Is your PFTE tube blue:

If it is a capricorn tube this might be your problem

February 2018
NOTICE: In an effort to continually improve our product, we recently began manufacturing our XS tubing with tighter tolerances. Due to this, more users are reporting filament jamming, higher friction, and things of that nature in the last couple of months. We believe many of these issues are being caused by out-of-spec filaments, but at the same time, we need to support the filaments people are using.

The bottom line is, if you have an issue with your tubing being too tight, please contact [email protected] for a prompt free replacement.

We apologize for any inconvenience this has caused.

Also did you rebuild your hotend? or change stock parts.

[adarcher]

I mentioned a few times that I got capricorn tubes to try to get rid of the friction in the long bowden. https://www.captubes.com/

It's not cheap, but not expensive either. I was talking about reaming the captube, not the original ptfe--the short piece is worthless otherwise, won't hurt to try as everything else I've done seems to have no affect. The captube is much more stiff and may handle the pressure of the top part of the hot end being screwed in. Seems like the slightest extra pressure will kink the original ptfe inside.

Thanks for the tip about tolerances. I'll look into the replacement, my filament slides pretty darn easily through the stuff I got, so I'm not sure if that would be my problem. The only compression is from the hotend heat break and top part being a smaller diameter and squishing the small piece I cut for it.

[adarcher]

I'm going to table retraction/stringing bug fixing. I might just have to live with the Artemis being a string monster for small parts.

I found my old retraction test and tried that instead of the circle/cylinder one.
https://www.thingiverse.com/thing:2240682

After swapping out the eztruder end piece and clamping it, the thermistor, and swapping out the short ptfe tube in the hot end that was kinked, I didn't have any strings on the cubes. Still strings on the cylinders though... I'll look into getting a low heat heat gun (cheap travel hair drier maybe?) for when I do get strings.

If you look at my previous picture of the eztruder, you'll notice there are no screws in the top part. Does anyone else have those? I believe those need to be filled and tight to avoid that metal end piece room to move. Gene, thanks for sending a new one, it is tighter than the old one, so not a waste.

On to the FRSs not working when the bed it up to temp--60C.



I put the big guy on it's side on the table so I could see the bottom and took some pictures of what I could:

The others had their cables neatly flat on the 8020, and out of heat's way. This guy (on the problem tower, pretty sure it's the Y, have to keep looking it up) was curled up and flat against the underside of the bed heater.

I thought, that ain't good, so I cut the ziptie and re-sat the connector along the side bar and re-ziptied it. Very hopeful, I righted the printer and heated it up... still no go on that tower at 60. Fine at room temp, just not 60. Z is really the only tower that triggers nicely. X is iffy, but Y is consistently unresponsive.

Looks like the FRS kit is a common one, so I'll look it up again and try to debug what to look for or how to protect the sensor pads.

[adarcher]

OK, my search for troubleshooting came up with people saying "make sure it triggers at temp" but not how to fix it if it doesn't trigger.

In order to test the pads, and also because of Gene's post about bed clips and triggering, I took out the bed climps and raise the bottom's top cover as high as I could and clamped it there. I needed better access to the innards down there.

I noticed that the pad that works flawlessly has the connector go out away from the bed while the other two go under the bed. I spent the afternoon dissecting the lower area of the Artemis and lengthened one of the sensor wire sets so I can have all three pads extend out and away from the bed so their wires don't touch the bed. The troublesome pads worked "much better" but still not as reliably Z tower pad. They trigger now at 60C but Y stays on or doesn't come on as often as Y and X is only barely not as reliable as Y.

I then swapped the pads (probably should have done this first since it's easier) between Y and Z. Even better, but still not great. Y, with Z's original pad, still works just as good.

This makes me think that the position of the pads in their little nook is the key.

Looking online, I noticed the pads people use are usually slightly bigger than mine. Maybe I'll look into cutting some larger cork circles and try those. I know I've got some cork rolls somewhere in the attic...

Here are some pictures of the pads and the direction of their connectors so you know what I'm trying to talk about:
The start pad, connector go up and away from the bed:

The sad pad, connector is barely visible because it goes under the bed:

Same pad, but with the bed moved aside to show the connector's direction and the small cork spacer:

[adarcher]

Phew, ok. I'm in a "workable" but not "good" state now I think.
Deviation is now .1 -- .09 at 60C and .03 -- .02 at room temp. Is that reasonable? I hope not. I'd like it to be better at temp.

I'm not sure how to tighten up the FSRs... will need some help.

I didn't redo the cork spacers yet. They are stuck on there pretty good and I'm unsure if I can/would damage the FSR trying to pull it off.

As for the cork pads... any reason they don't go to the edge of the FSR inner circle? Why just a small bit on the inside?

It's hell to take the bottom apart because of all the parts interfering with each other. I ended up moving the FSR junction to the inside because it would otherwise bend the wires when the top cover goes over it. I'll move it back once I settle all the tuning issues here.

[wepollock]

.1 -- .09 at 60C and .03 -- .02 at room temp. Is that reasonable? I hope not. I'd like it to be better at temp.

I read an article in which Johann Rochell (the google engineer who developed the Rostock), and some blog posts, which speak to the issue of heated beds with FSR. The bottom line is that temperature affects-sensitivity; from what I can glean the best results you will get is with a cold bed. Your metrics prove that out.


I also read the datasheet on several FSR's
Cold: -40ºC after 1 hour 5% average resistance change
Hot: +85ºC after 1 hour -15% average resistance change
Hot Humid: +85ºC 95RH after 1 hour+10% average resistance change
This also proves out your observation.

The JohnSL board is reading analog resistance measurements within an expected range; and while heated up the resistance is changing based on the spec sheet. If I am reading the spec sheet properly heat reduces resistance.

Using an Offset Macro to hone in on a good bed level.


(Every few prints I run my offset macro twice or three times over - the macro dials in the gap this way, it saves the last each time it runs)
(then I check with a feeler gauge)
http://www.ebay.com/itm/100B-17-Blades- ... SwpuFaE88R

G28
G31 P500 X0 Y0 Z-.xx; negative up from bed positive is towards bed make sure you use your offset number in xx I use -.43
G32
M500
G0 Z0
M117 "Measure Gap From Extruder to Bed - waiting for 30 seconds"
G4 S30
G28

The real key is repeatability; get that XX number correct in the above macro and every print will be golden.

Before running this macro I make sure the nozzle is clean; I heat up the nozzle; carefully pull the filament back a bit, and clean with a brass brush.
Sometimes I run the macro with the nozzle hot, sometimes cold; there is negligible difference.

Rather than look for a number; I am looking for a result, the best first layer I can get.. it's more meaningful than a number IMHO

I am looking for my "skirt outline" to "feel nice." I don't want it smushed on the bed; I am looking to feel a nice smooth and slightly curved top.
If I was printing something flat across the bed I am looking for smoothness as well. If I see a rough spot I am too close. When removed, the bottom layer of the print feels like glass and is shiny, there is no lip between the first and subsequent layers.

I have the thickness gauge pictured; you can also see the skirt and that there is NO stringing of any kind. I am printing right on clean glass. (sometime I will use glue after the first on glass print because I am too lazy to change glass plates - I have three plates, that I like to clean at the same time.)

.88 extrusion muliplier
1st layer at 200 degrees; thereafter at 198 degrees
Vertical lift .8 (but I would like lower)
Retraction speed up to 6000
Print speed 3000 (sometimes) up to 4200
First layer height 90% Width 100% Speed 50%
Random Start points
Cooling layer 1 0% layer 2 50% layer 3 75%
Ooze control - force retraction between layers; Only retract while crossing open spaces, only wipe for for outer perimiters
Avoid crossing outline for travel movements (this is a great setting - if you do get string it will be in the part)

The bead of the skirt is nice and smooth; it is also totally uniform (the camera perspective distorts the uniformity)
[img]http://i.imgur.com/VOwzLvc.jpg[/img]

Some good tools

http://www.ebay.com/itm/Razor-Blade-Scr ... 2749.l2649

http://www.ebay.com/itm/Pen-Style-Scrat ... 2749.l2649

[img]http://i.imgur.com/AAF2A7y.jpg[/img]

Notice I have aligned my bed clips in the correct direction DUH! the prints were not shifting even without the bed clips engaged.
I usually clean the bed with tap water, but sometimes I will will hit the bed with a water cleaning followed by some 99% isopropyl alcohol.

[adarcher]

Currently have the machine offline again. Was able to squeeze out some prints for the the FB group Eiffel Tower project though.

The parts had some pretty bad ringing in them, so I plucked the belts and they were so loose I could barely hear a low rubbery sound when I should have heard a twang of some sort.

Thought I'd adjust them afterwards. Good idea because I basically put the printer out of commission because I just don't have time to clear space to turn it apart again to find out how I should have done it instead of how I tried.

The bottom pulleys looked like the ideal place to adjust tension to me--just loosen and push down a little till the belts twang the same, then retighten. Nope, there is some hidden peg that keeps them loosely in place below the rails. No ability to move it down. I ended up unbolting it to see if I could pull the sucker off and lost the t-nut inside the tower. Patience with this machine really is trying sometimes.

Beautiful but a little hard to maintain.

[wepollock]

plucked the belts and they were so loose

It's easy to tension the belt, simply lower the belt pulley just below the linear rail; loosen both bolts and press down, then tighten. The bolts are held into place with T-Nuts that can sometimes get wedged in place and thus have to be pressed free. Assure that the little post in the middle of the belt is centered and not causing friction. I would not super-super-tighten the belt. The only objective is to assure that the belt is adequately keyed into the gear.

Even if the free side of the belt osselates a bit (during a fast move like a home) I find that the belt remains firmly seated in the timing pulley. It would take a goodly amount of force to make the belt skip out of the pulley. I actually tested the holding power of the belts in the gears...

As to the Artemis, I am not seeing much to maintain, change, or tune.

The only changes I have made so far; capricorn tubing and a geared extruder (titan). Both those changes were made, not by necessity but, only because I had the items in stock. There is other hardware things I would like to try; I am starting to maintain my Rostock Max V2 first. I put in a new heater cartridge and rewired the hot end, the belts are getting a bit long in the tooth. These machines can run for 100's-1000's hours..

The only hardware-maintenance I have done is to clean the rails with a qtip, clean the ball joints, remove dirt which accumulated on the grease, and then lightly re-grease. I am nearing the endpoint of slicer tuning. I am getting good PLA prints, and as of today good PETG prints. PETG is beyond my expectation.

[geneb]

There's nothing keeping the belt turn-around from being pushed down. The only thing keeping it in place are the 1/4-20 screws and the backing t-nuts.

g.

[adarcher]

My eyes were playing with me then. There is a hole under the pulley and when I tried pushing the fully loosened injection molded piece down, it would just pivot over the hole--I guess I'm just not as strong as I thought I was. I'll need a tensioning bolt or something.

[geneb]

Chances are pretty high you'll just break something if you force it. You should be able to tension the belt well enough by hand strength alone.

g.

[adarcher]

Hah. The universe doesn't want me to be happy with this printer...

Got the pulley back on--was a lot easier than I initially feared now that I looked at the printer in proper light. I concede I was much too hasty before on dissing the belt tensioning. Though, I am a bit frustrated with not getting much out of this machine--the most expensive machine I've every bought for myself--so please excuse me some impatience as it's been >4 months since I got the machine.

I did a small print and it was "OK", bed still doesn't calibrate great so the first layer sometimes screws up bad if there is any retraction/zhop in it. The nozzle just picks up all the adjacent filament that has already been laid down and carries it around til I cancel it. Cleaning the outside of the nozzle didn't seem to help, its just so darn attractive to filament that has already been stuck to the bed.

I figured, there was a new firmware update for dc42 with the multi-tap from SeeMeCnC, that I'd install that. Seemed to go without a hitch yet when I did my first print (multi-tap set to 4 got a really nice consistent calibration value), the machine let out a crazy whine and printed way up in the air.

I threw my hands up and got the latest Artemis FSR SD folder from git to downgrade and bah, I can't downgrade from 1.21 as easily as I hoped. I'll figure it out, there are FAQs and such on this problem, but I'm outa time for fiddling with this guy for the weekend. I hope someday soon I can "waste electricity" and have prints going while at work like I use to with my V2.

[wepollock]

My perception is that the machine was designed to be turnkey (even with some state of the work leading-edge features)

I would suggest that the machine be restored to "factory defaults." (do it yourself/ask to have it done/pay to have it done/if you are NY?)

If bed leveling is the issue, there are other ways to skin a cat via Duet software.
Manual calibration If you do not have a Z probe, you will have to calibrate manually. Using the Escher3D wizard
The Escher 3d wizard will help you calibrate your delta printer. It uses the same least-squares algorithm as RepRapFirmware, which converges faster than traditional methods of delta calibration.

From a modifiers point of view; there is some little stuff that would make the process easier
--I can see folks wanting to put an E3d hot end having to think about it a bit; then they are liable to lose the ability to go back to factory defaults.

--12v is pressing the limits of a geared extruder

[adarcher]

@wepollock, I really feel like you don't read my posts sometimes. Your comments don't make a lot of sense if you had.

There isn't much that isn't stock on my printer. I have mentioned that I like to keep it stock while I fix the issues. I've tried different extruders and hotends and always return back to the originals to make sure I can fix the issue. The hardware changes are not to fix the issue, it's to pin point the issue. If I change a part and the issue continues, I've taken that part of the printer out of the picture in regards to my issue.

The only things I've changed right now:
- fan in the top because it was loud (has no affect to the printer as a printer)
- the position of the FSR pad connectors under the bed (improved the FSR triggering since the cables were not over heating and pinched by the heat spreader)
- loosened the bed clips so it can accept my build plate. (has no affect on the leveling because it's a grip either way that is anchored to the heat spreader)
- firmware (latest firmware was an improvement, and is the current task at hand)

Things that have been changed to see if they make any difference, but returned back to stock:
- tried cap tubes (were too tight anyway, got the replacement ones, thanks)
- tried my e3dv6 (I've got a modified heat sink that fits the SE300 mount, was easy to swap in and out)
- tried my titan (along with the new inner ptfe tube, this reduced stringing a considerable amount, will most likely go back on soon)

Now, yes, this printer is suppose to be close to turn-key. But, considering that I was in the first batch built, having a few issues is expected. Having the ptfe tube in the SE300 kinked is an example of that. Not having holes tapped or screws stripped also show this. This printer came as a kit and while parts were probably tested on their own, it was not tested together or if it was, it lost it's credibility as tested once it was taken apart and shipped to me.

I'm a huge fan of this printer, me bothering to spend so much time on it is proof that I've got love for it.

Now, I'm off to look for people who have tried using the 1.21 firmware or have an idea on how to downgrade properly--looks like the firmware filename was changed and the install script wants the old name, so this might be an easy fix, or just an order of operations issue.