Overhangs Sag Too Much

D
dwalton 3D printing enthusiast Level: 15 Points: 130
Posts: 20

Posted by
D
dwalton

Posted Oct 22, 2013 - 5:16 AM  

I have been having trouble printing the Ultimaker robot recently. The Delrin wheel that pushes filament against the hobbed bolt was off-center and was causing under extrusion. I also had a problem with my fan duct as it unfolded, partially melted and was blowing on the nozzle tip. I cut a new delrin wheel on my laser cutter and printed Daid's fan duct from Youmagine:

https://www.youmagine.com/...-fanduct-replacement

This fixed my problem under extruding, but now overhangs sag badly as the following pictures show:

Top
Back
Left Arm
Front

Here are the pertinent settings from Cura:

layer_height = 0.1

wall_thickness = 1.3

retraction_enable = True

solid_layer_thickness = 1.3

fill_density = 0

nozzle_size = 0.4

print_speed = 50

print_temperature = 197

print_bed_temperature = 70

support = None

filament_diameter = 2.82

filament_flow = 100.0

retraction_speed = 40.0

retraction_amount = 4.5

bottom_layer_speed = 20

cool_min_layer_time = 7

fan_enabled = True

fan_layer = 1

fan_speed = 100

fan_speed_max = 100

cool_min_feedrate = 0

cool_head_lift = False

I am using silver PLA from ultimachine. I measured the filament size with a digital micrometer.

I built this heated bed:

http://www.thingiverse.com/thing:30655

I printed some hollow calibration pyramids and was surprised that I could keep lowering the nozzle temp. now that I fixed my extrusion problems.

As you can see in the pictures, the front look better than the back. There is a little warping on the bottom which is not typical of my prints with the heated bed. The rings on the robot's left arm are sort of warped up so that they do not stick out as much as the rings on the right arm. The back looks pretty bad.

I checked my nozzle temp with a thermocouple attached to a multimeter and the temperatures seem accurate. I can lower the temperature more but 197C seems pretty low . . .

Do you guys have any suggestions for me? I would love to know what I am doing wrong.

Thanks!

dwalton

41 replies
 6649
Sort 
illuminarti You can trust this guy Level: 63 Points: 3496
Posts: 2795
Location: Memphis, United States Printers: Ultimaker 2, Ultimaker Original, Ultimaker Original+, Ultimaker 2 Extended, Ultimaker 2 Go, Ultimaker 2+, Ultimaker 2 Extended+, Ultimaker 3, Ultimaker 3 Extended
Posted by
illuminarti

Posted Oct 22, 2013 - 5:43 AM

Is your fan actually working during the print - and are you sure that it's blowing onto the print, and not sucking air off it?

Also, does it really seem to be slowing down so as to take 7s per layer?

gr5 Moderator Points: 15875
Posts: 9639
Location: Boston, United States Printers: Ultimaker 2, Ultimaker Original, Ultimaker 2 Extended, Ultimaker 2 Go, Ultimaker 3
Posted by
gr5

Posted Oct 22, 2013 - 6:20 AM

Possibly the 1.3 wall (aka shell) thickness could be the problem. You should set this to 1.2. This should always be an integral multiple of your nozzle diameter. So what Cura is probably doing is doing the 2 inner wall passes at .4mm each and then a .5mm outer pass which means it isn't touching the inner walls very well and is overextruding by 25% (.5/.4) to make up the full .5mm with a .4mm nozzle. This causes drooping. The "good" side probably got a little more cooling - or perhaps the fan air didn't push down on the filament strings as hard?

I could be wrong about this but this might explain it all.

Also are you sure this is gray PLA from UM? It looks very blue in the pictures :smile:

gr5 Moderator Points: 15875
Posts: 9639
Location: Boston, United States Printers: Ultimaker 2, Ultimaker Original, Ultimaker 2 Extended, Ultimaker 2 Go, Ultimaker 3
Posted by
gr5

Posted Oct 22, 2013 - 6:24 AM

There is a little warping on the bottom which is not typical of my prints with the heated bed.

Try adding brim and also clean the kapton tape before printing.

D
dwalton 3D printing enthusiast Level: 15 Points: 130
Posts: 20
Posted by
D
dwalton

Posted Oct 22, 2013 - 8:35 AM

Thanks for the fast response! I was hoping from help from either gr5 or illuminarti and am grateful for both of your answers.

illuminarti,

The fan is definitely blowing air out of the duct onto the extruded plastic. I have not timed how much it slows down, but it gets pretty slow when printing the horns. I'll time it tomorrow and reply.

gr5,

Thanks for catching the odd wall thickness. I am still trying to figure out how you got the amazing prints in this thread:

http://umforum.ultimaker.com/...-print-such-quality/

I'll try the print with 1.2 mm wall thickness and cleaned the kapton tape with alcohol. I forgot to add brim, but I expect the warping will be minimal with clean kapton tape.

The filament is definitely the silver PLA from Ultimachine. I didn't correct the white balance for the lights I used for the pictures. I have white LED light bars surrounding the top of the printer. My wife calls it "The alien landing zone" since it seems to be lit like the ones you see in movies. It is great for seeing what the printer is doing, but not so great for capturing colors accurately.

In any case, I have started another robot print and will let you know how it turned out when I get up tomorrow.

Thanks again!

gr5 Moderator Points: 15875
Posts: 9639
Location: Boston, United States Printers: Ultimaker 2, Ultimaker Original, Ultimaker 2 Extended, Ultimaker 2 Go, Ultimaker 3
Posted by
gr5

Posted Oct 22, 2013 - 4:57 PM

I looked at your settings again and they appear to be what I would recommend for a high quality print (other than that 1.3).

Another possibility is loose belts. My belts are looser than the ones I saw on all the UMs at MakerFaire, but obviously not "too loose" as my printer prints quite well. Have you watched and listened to this video?:

I tightened my belts the suggested way - by loosening those nuts shown in the video. It only tightens slightly but it was enough. The short belts may be more likely to be a problem and also easier to fix.

Please only change one thing at a time so we know if it was the 1.3 issue or something else (for future people who have similar issues).

D
dwalton 3D printing enthusiast Level: 15 Points: 130
Posts: 20
Posted by
D
dwalton

Posted Oct 22, 2013 - 7:47 PM

I changed the wall thickness to 1.2 and this improved the print quite a bit:

Front

Left Side

Top View

Back

My Lighting System

I also included a picture of the "Alien Landing Zone" lighting system. It really helps me see what the printer is doing, especially when making sure the first layer height is correct.

The prints are better, but still have room for improvement. The front looks much better, with very little sagging. The left arm rings are still warped up a bit. The back is better, but there is still pronounced sagging.

I watched the print and noticed that the plastic moved a little while the printer moved over it. It was as if the whole model was still soft. When I was having under extrusion problems, the fan was blowing on the nozzle and was cooling it. Now I think the fan is blowing down too much and is not getting enough air on the recently printed layers.

Thanks for suggesting the video on the belts. I think mine are reasonably tight. It is funny that you mentioned the Maker Faire. I was at the Bay Area Maker Faire this year and spoke with Erik. He was great to talk to and I was excited to see the new things coming from Ultimaker. He gave me one of the Ultimaker robots he had printed at the Maker Faire and I have been using that as a reference to which I compare my prints.

Do you have any other suggestions?

gr5 Moderator Points: 15875
Posts: 9639
Location: Boston, United States Printers: Ultimaker 2, Ultimaker Original, Ultimaker 2 Extended, Ultimaker 2 Go, Ultimaker 3
Posted by
gr5

Posted Oct 22, 2013 - 8:57 PM

Huge improvement!!

Do you have any other suggestions?

Don't use the heated bed? lol. I have one but didn't use it for that nice quality UM robot you mentioned. I just printed on blue tape (on top of kapton just like yours). Didn't even clean it with alcohol because I don't have much of a problem with this particular shape and size (small and contact with bed is larger (or at least no smaller) than the rest of the print). And the one from Erik also was not printed on a heated bed.

Well you could up the layer cool time from 7 seconds to 10 seconds although 5 is plenty for me.

I think you are at least partly right about the softness. Could it be moving it due to warping? How is your fan duct different from the standard one anyway? The one you linked to should be identical to the stock one in shape. I still use my stock fan shroud.

Well the UM2 has a fan on each side. Maybe you should go with that if you plan to use a HB.

D
dwalton 3D printing enthusiast Level: 15 Points: 130
Posts: 20
Posted by
D
dwalton

Posted Oct 23, 2013 - 1:21 AM

I'm open to putting blue tape on the heated bed for a small print. I only really need the heated bed when printing large things. I'll also try a print with a layer cool time of 10 seconds.

I am going to try the robot again with a modified fan duct as well. I'll post the results when the prints are complete.

My fan duct is Daid's replacement for the stock duct and it is pretty close to the original. Unfortunately, the original unfolded and touched the heater. It melted where it touched the heater and was blowing on the head, causing it to cool and under extrude. If I can find some plastic of the right thickness, I might cut a new folding duct on my laser cutter, but ordering a new duct from Ultimaker is a bit expensive.

I really appreciate your help. Thanks again!

gr5 Moderator Points: 15875
Posts: 9639
Location: Boston, United States Printers: Ultimaker 2, Ultimaker Original, Ultimaker 2 Extended, Ultimaker 2 Go, Ultimaker 3
Posted by
gr5

Posted Oct 23, 2013 - 1:58 AM

unfolded and touched the heater. It melted where it touched the heater

Same thing happened to me. I fixed it with kapton tape. I see you already have the yellow kapton tape so it should be easy to cut a piece and place it over the melted part and fold it back on itself. Now it can touch the aluminum block if I want and no trouble although I also rotated the aluminum block so it doesn't touch the duct anymore.

fyi kapton tape can handle something ridiculous - I think maybe 600C?

D
dwalton 3D printing enthusiast Level: 15 Points: 130
Posts: 20
Posted by
D
dwalton

Posted Oct 23, 2013 - 7:59 AM

That sounds like good advice. I'll have to see if I can resurrect the duct with the tape. I tried another duct from thingiverse and the results were disappointing:

http://www.thingiverse.com/thing:21182

I didn't even bother finishing the print because the overhang was so bad. I put Daid's duct back on and am trying another print with blue tape over the Kapton and the heated bed turned off and the temperature reduced even further. I'll post pictures when it is finished. It looks a little better, but the left arm rings still curve up and the back overhang looks a bit rough. The print was soft, even near the head (well away from the heated bed) so I am thinking I may need that second fan.

Is your bad aluminum or glass? I ended up having to wipe the blue tape with alcohol to get the PLA to stick to it. I was wondering if my aluminum bed is conducting heat away from the first layer faster and if that affected it's ability to stick to the tape.

gr5 Moderator Points: 15875
Posts: 9639
Location: Boston, United States Printers: Ultimaker 2, Ultimaker Original, Ultimaker 2 Extended, Ultimaker 2 Go, Ultimaker 3
Posted by
gr5

Posted Oct 23, 2013 - 7:00 PM

My bed is very thick aluminum (too thick - takes 20 minutes to heat to 70C). I have one layer of kapton on top and then a layer of blue tape. I never thought about that - that the aluminum can suck down the heat a bit. Thanks for mentioning it as I will keep it in mind in the future.

And again - that white robot was printed at room temp so the aluminum was around 70F (21C).

Something seems a little strange. I'm wondering if your temp is a little high. Or maybe it's just the filament - I haven't used UM gray in a long time and I'm not sure if it prints better cooler. Is there a way you can calibrate the temperature of your aluminum print head? Maybe test at what temperature it boils water? If it boils water at 80C or 90C that would explain everything.

Remember that Marlin will shut down the extruder at 170C so if you go lower than 180C be very careful it doesn't over shoot and go below 170C.

That white robot was printed with "white unspooled" PLA from printbl.com at 190C, 20mm/sec, .1mm layers. But I've printed "almost as good" at 70mm/sec 210C .2mm layers. Or at least no issues with the arm rings.

Also I have a stock UM fan and fan shroud. Fan at 100%. Hmm. 20mm/sec fan at 100% is going to be a hell of a lot of more cooling than normal.

I rarely print at 20mm/sec. I'm usually not that patient and not picky about quality (usually).

D
dwalton 3D printing enthusiast Level: 15 Points: 130
Posts: 20
Posted by
D
dwalton

Posted Oct 23, 2013 - 8:42 PM

I had a breakthrough! Repairing the original fan duct with Kapton tape dramatically improved the print quality. It is not as good as your print, but it was nearly as good as the print from Erik. My first print with the repaired duct got knocked loose towards the end of the print, but the results are very promising. I'll print again and upload a comparison of the last few things I have tried.

I attached a thermocouple to the print head with Kapton tape and attached it to my multimeter. It looks like the aluminum block is a few degrees cooler on the outside, but I think my temperature is pretty accurate. I think the problem was that the printed fan duct is not nearly as good as the original.

My heated bed is 1/4 inch thick. When I print on a cold bed, I can see the temperature of the bed go up by a degree or two as the bed is heated a little by the hot filament and the bed's temperature sensor is in the middle of the bed (where I am printing). I originally used a reprap PC board heater, but that took too long to heat up. I replaced it with this one:

http://store.qu-bd.com/product.php?id_product=61

It takes about 5 minutes to heat up, even with the thick aluminum plate. I used my laser cutter to create a plywood board that holds the pad against the plate. I tapped holes in the plate and sandwich the heater between the underside of the plate and plywood using screws to draw it together. I have been thinking about insulating parts of the plate to see if I can get it to heat up faster.

gr5 Moderator Points: 15875
Posts: 9639
Location: Boston, United States Printers: Ultimaker 2, Ultimaker Original, Ultimaker 2 Extended, Ultimaker 2 Go, Ultimaker 3
Posted by
gr5

Posted Oct 23, 2013 - 9:47 PM

It would be relatively easy for me to double the wattage of my heater. But I don't use it very often so it's fine.

I have been thinking about insulating parts of the plate to see if I can get it to heat up faster.

I thought of that but then it will cool down slower also! And it won't heat up "the chamber" as much. I occasionally tape clear plastic on the 3 sides with a box on top to get a warmer chamber when using the bed.

gr5 Moderator Points: 15875
Posts: 9639
Location: Boston, United States Printers: Ultimaker 2, Ultimaker Original, Ultimaker 2 Extended, Ultimaker 2 Go, Ultimaker 3
Posted by
gr5

Posted Oct 23, 2013 - 9:53 PM

I'll print again and upload a comparison of the last few things I have tried.

I'm really surprised that Daid's duct isn't as good as the original. Are you sure that was it? Or did you change some other things at the same time?

gr5 Moderator Points: 15875
Posts: 9639
Location: Boston, United States Printers: Ultimaker 2, Ultimaker Original, Ultimaker 2 Extended, Ultimaker 2 Go, Ultimaker 3
Posted by
gr5

Posted Oct 23, 2013 - 9:57 PM

and cleaned the kapton tape with alcohol.

FYI - I rarely do this. I just thought maybe some oils from your fingers got on there. Keep in mind that kapton tape at 25C is not very sticky for PLA. There is a threshold where heated bed at 60C is no better than 25C but by the time you hit 70C it is very sticky and works great. Once the first layer is done you can even lower the temp back to 60C and it continues to stick very very well. Lower it too much however and the part might pop off due to the aluminum shrinking.

D
dwalton 3D printing enthusiast Level: 15 Points: 130
Posts: 20
Posted by
D
dwalton

Posted Oct 24, 2013 - 5:18 AM

Quote by gr5

I occasionally tape clear plastic on the 3 sides with a box on top to get a warmer chamber when using the bed.

That is a good idea. I cut three pieces from 1/4' foam insulation used on houses, taped them together and cut out slots on the left for the power and USB cable. I can set that on the printer and quickly insulate the chamber on large prints where I am concerned about warping.

D
dwalton 3D printing enthusiast Level: 15 Points: 130
Posts: 20
Posted by
D
dwalton

Posted Oct 24, 2013 - 5:36 AM

Quote by gr5

I'm really surprised that Daid's duct isn't as good as the original. Are you sure that was it? Or did you change some other things at the same time?

Daid's duct blows down a little bit more than the stock duct. I noticed more air blowing in parallel with the bed with the stock duct. I have an anemometer and will see if I can measure the difference.

The duct was all I changed since going to 1.2 mm walls. I even used the same g-code files for the prints. I would turn the HB on and off using the Ulticontroller.

I am using the silver filament from Ultimachine:

https://ultimachine.com/...3mm-silver-1kg-spool

Here are some earlier prints:

Earlier Front Prints
Earlier Side Prints
Earlier Top Prints
Earlier Back Prints

Here are some images of the improved prints:

Better Front Prints
Better Side Prints
Better Back Prints
Better Top Prints

The left arm is still a little warped but the back and the keychain hole on the head are dramatically improved. Thanks for all of the help!

D
dwalton 3D printing enthusiast Level: 15 Points: 130
Posts: 20
Posted by
D
dwalton

Posted Oct 24, 2013 - 5:45 AM

Here are some pictures of the duct repair:

Melted Duct

This shows what happens when your duct unfolds and hits the nozzle. This causes severe under extrusion and I decided to replace the duct.

Comparison

This is a comparison of the replacement duct with the original. The replacement duct ended up not working as well as the original.

Safety Tape

This is the repaired duct, reinstalled and taped so that it won't unfold.

gr5 Moderator Points: 15875
Posts: 9639
Location: Boston, United States Printers: Ultimaker 2, Ultimaker Original, Ultimaker 2 Extended, Ultimaker 2 Go, Ultimaker 3
Posted by
gr5

Posted Oct 24, 2013 - 6:13 AM

It's a robot army!

Something is still a bit off with your cooling but this is much better. Next step: eliminate stringing:

http://umforum.ultimaker.com/...bration-photographs/

When I tried to print the perfect robot I used the "cut off object bottom" feature and printed only the robot head to save time. I printed 5 heads until I got it perfect. To get rid of stringing, lower the temp - maybe more than you have so far (try 180C) - and increase minimum layer time a little - maybe to 10 seconds. And keep it under 50mm/sec like you have been.

If you can eliminate that stringing I will tell you the final secret to get the antennas perfect. It should help the arms too.

By the way: How is your play/backlash? If you grab the nozzle (while cold!) and push it around, is it loose at all versus the rest of the head? Are the bearings loose? Are the belts loose?

D
dwalton 3D printing enthusiast Level: 15 Points: 130
Posts: 20
Posted by
D
dwalton

Posted Oct 25, 2013 - 4:09 AM

That is a great tip. I cut the bottom off like you suggested and tried prints at 185 and 181. When the temperature dipped below 180 briefly at the beginning of one print, the filament stopped coming out.

I still got stringing, even with minimum layer time of 12 seconds and at 181C. I've been swamped today (and have a broken toilet that needs urgent attention), but will try to post pictures later tonight.

I also tried a print with the cool head lift feature enabled, but that did not help.

I suspect you are right that I have a cooling issue since the horns seem to stay pretty mushy as the nozzles runs in circles around them.

I had a problem a while ago where the fan always ran and found the Darlington BJT that controls the fan had failed. The Darlington did not have any protection from back EMF (AKA flyback current) and I think that PWMing the fan (running at less than 100%) killed the transistor. I replaced the transistor and put a clamping diode in parallel with the fan to protect the new transistor. I'll check it with a scope to make sure it is working correctly. I am 95% certain that my repair will not alter the performance of the fan, but I can check the PWM signal with an oscilloscope easily enough.

I moved the head around by hand. It is a bit stiff from the belt tension, but I didn't feel any backlash. I noticed that I could slightly tighten one of the pulleys that turned the y-axis. In my experience, the pulley set screws need to be really tight or they will slip. Otherwise, the belts sound tight when I pluck them like a guitar string. I might be able to get the left y axis belt slightly tighter. I have belt tightening devices on all of the belts. I bought a set of belts with 1 less teeth to make them tighter, but haven't seen the need to put them on yet (I was concerned the belts would get loose over time).

Do you have any other ideas?

Thanks again!

gr5 Moderator Points: 15875
Posts: 9639
Location: Boston, United States Printers: Ultimaker 2, Ultimaker Original, Ultimaker 2 Extended, Ultimaker 2 Go, Ultimaker 3
Posted by
gr5

Posted Oct 25, 2013 - 5:49 AM

I think you just need a different filament if you are going to avoid stringing. UM Gray probably isn't capable of zero strings. Well - one more thing to try - try printing the head real slow. Try 20mm/sec and 185C and keep the minimum layer time down to 5sec. 12 seconds is silly.

gr5 Moderator Points: 15875
Posts: 9639
Location: Boston, United States Printers: Ultimaker 2, Ultimaker Original, Ultimaker 2 Extended, Ultimaker 2 Go, Ultimaker 3
Posted by
gr5

Posted Oct 25, 2013 - 5:52 AM

Oh!!! One more thing! Maybe 4.5mm isn't enough retraction without this clip:

Print this up and add a wire tie. It makes it so you only need 4.5mm for retraction. Without this clip you need a little more than 4.5mm - not sure - maybe 5.5mm. But best to just print it because if you do too much retraction, you get air in the nozzle and then bubbling and imperfect parts.

http://www.thingiverse.com/thing:46157

D
dwalton 3D printing enthusiast Level: 15 Points: 130
Posts: 20
Posted by
D
dwalton

Posted Oct 26, 2013 - 3:04 AM

I had another breakthrough. You suggested I try another filament so I tried UM red and had the same problem, even when printing at 25 mm/s. The melted filament for the horns kept sticking to the nozzle. I even tried putting a little machine oil on the nozzle to see if that would make the filament stick less with no success.

I also tried to make retraction occur a bit more often on the horns to see if that would help, but it did not seem to help much. I set retraction minimum travel to 1 and minimal extrusion before retracting to 0.25. I wasn't worried about grinding the filament since the robot does seem to cause much retraction.

I then decided to try another filament:

http://colorfabb.com/...h-orange-300-20.html

It is pretty expensive and I don't use it often, but it made quite a difference. The stringing is gone and the horns look very good. I am printing a full robot and so far it looks almost as good as yours. I have been distracted by a plumbing emergency, but will try to get pictures of my experiments posted tonight along with my settings.

I also measured the speed of air over the bed with the stock fan. It was between 1.9 and 2 MPH which I am certain is higher than the printed duct. I'll try to measure the wind speed with the printed duct as soon as I get a chance.

D
dwalton 3D printing enthusiast Level: 15 Points: 130
Posts: 20
Posted by
D
dwalton

Posted Oct 26, 2013 - 3:08 AM

Quote by gr5

Oh!!! One more thing! Maybe 4.5mm isn't enough retraction without this clip:

Print this up and add a wire tie. It makes it so you only need 4.5mm for retraction. Without this clip you need a little more than 4.5mm - not sure - maybe 5.5mm. But best to just print it because if you do too much retraction, you get air in the nozzle and then bubbling and imperfect parts.

http://www.thingiverse.com/thing:46157

I have this clip installed:

http://www.thingiverse.com/thing:24363

I have it pushed to the second-highest setting and it seems to work OK. Is the one you mentioned better?

gr5 Moderator Points: 15875
Posts: 9639
Location: Boston, United States Printers: Ultimaker 2, Ultimaker Original, Ultimaker 2 Extended, Ultimaker 2 Go, Ultimaker 3
Posted by
gr5

Posted Oct 26, 2013 - 3:22 AM

Is the one you mentioned better?

Not really. The only thing better about the one I linked to is that all the other ones I used broke within a month. Whereas the one with the wire tie will last forever.

41 replies
 6649
Sort 
Support