[Solved] Surface Quality Issue with E3D Cyclops

[gestalt73]

TLDR;
1. The Cyclops requires a geared extruder, because of the force required
2. Upgraded the EZStruder to use the Kysan 5.18:1 geared stepper
3. Also checked the nozzle, and removed a burr from nozzle interior
4. Extruder drive lowered to 1.1 amps, max acceleration of 250mm/s

http://i.imgur.com/UXD7dvd.jpg
[img]http://i.imgur.com/UXD7dvd.jpg[/img]

http://i.imgur.com/gQr2gFE.jpg
[img]http://i.imgur.com/gQr2gFE.jpg[/img]

http://i.imgur.com/pOKZc5r.jpg
[img]http://i.imgur.com/pOKZc5r.jpg[/img]

More details on my latest post here:
http://forum.seemecnc.com/viewtopic.php ... =50#p67879


Original posting left below for posterity.

Anyone else out there running a Cyclops?

I'm seeing a reallly rough surface finish for models printed with the Cyclops vs my V6.

In the attached shot, I'm running the exact same gcode file with both the V6 and the Cyclops.
Same Gcode, same filament, same temps, same bowden, same extruder, same everything except for the hot end.
I even swapped out the 25w cartridge for the 40w cartridge in the Cyclops.
Both platforms are very sturdy, no wiggle, everything is tight and secure.
Watching the temps on both prints, they both hold to within a few tenths of a degree.

The print on the left was done with the V6. It's about 5cm tall, and there are some small variations in the surface finish, but overall the print turned out really well. Very smooth.

The print on the right was done with the Cyclops. Notice the angled irregularities. It is quite rough to the touch.

I wish I could say that the lighting is exaggerating the roughness, but it's not. I couldn't release a print like this to a client.

Does anyone have any ideas?

[img]http://i.imgur.com/BOMUjX4.jpg[/img]

[Mac The Knife]

can you swap the nozzles?
Watching their presentation video from the MMRF or whattever that was, they mentioned changing the internal design of the nozzles.

[gestalt73]

Unfortunately the nozzle is proprietary to the Cyclops, it's not a standard V6 or V5, so it won't be possible to swap them around.

Although I didn't really examine the nozzle before installation, so maybe I can order a spare.

I just printed a single walled calibration cube, and comparing the outside to the inside, it's not an alignment issue, it looks more like variances in print volume.
If it was some type of alignment or other hardware issue, you'd see a bulge on one side, and a void on the other side. This isn't like that.
Almost like it's not squirting plastic at a constant volume for some reason. Very strange.

[Mac The Knife]

Turbulance.

[gestalt73]

That would be... disappointing, as it would suggest that it's a design issue. And that it's not something that can be addressed by the user.

Although it would make sense, as the melted filament travels through 2 right angles between the bowden and the nozzle.

I'm also beginning to suspect that E3D already knows about this. The only two photos of models that they've released are very organic (the dragon guarding the egg, and the two colored frog).
No straight edges, or faces, so it would be alot harder to see.

These prints were all done with 20mm/s perimeter, 40mm/s internal perimeter, 60mm/s infill, so the settings aren't crazy fast.

Another possibility is that it's the second bowden slightly changing the chamber pressure.
Both bowdens need to be filled with filament, otherwise the melted plastic would just squirt out the other bowden hole.
Maybe as the head moves, slight changes in the curve of the second bowden tube very slightly push the 2nd filament in or pull it out a bit, resulting in pressure changes.
That would be equally disappointing, as it's a problem endemic to the design.

Although, I've printed two copies with the Cyclops, and the pattern in the roughness looks totally different between the two examples.

[TFMike]

try a different filament color / type and see what happens, it's worth a shot

[gestalt73]

Yeah, I'm already seeing it across two different filaments. With the V6 it's an outstanding print, with the Cyclops it's too rough and irregular.

I decided to eliminate the second bowden as a contributing factor, redneck style.
It won't make a difference, as it would still be a design issue, but it would be nice to know for sure what's causing it.

I stuffed a bit a filament in an extra short length of bowden tube, then crimped the end. This will totally eliminate 2nd bowden tube flex as the culprit.

If this print has better quality, then we know that the second bowden tub flexing was changing the melt chamber pressure as the hotend moved around.

[img]http://i.imgur.com/C6Ssdgl.jpg[/img]

[TFMike]

Yeah, I'm already seeing it across two different filaments. With the V6 it's an outstanding print, with the Cyclops it's too rough and irregular.

I decided to eliminate the second bowden as a contributing factor, redneck style.
It won't make a difference, as it would still be a design issue, but it would be nice to know for sure what's causing it.

I stuffed a bit a filament in an extra short length of bowden tube, then crimped the end. This will totally eliminate 2nd bowden tube flex as the culprit.

If this print has better quality, then we know that the second bowden tub flexing was changing the melt chamber pressure as the hotend moved around.

[img]http://i.imgur.com/C6Ssdgl.jpg[/img]

redneck style is best style, have you tried running a flame through nozzle? Maybe there is some build up in there that is contributing to this as well

EDIT: Didn't know there were rednecks in LA though haha

[gestalt73]

Well, no change with the pinched off bowden, so it can't be 2nd bowden tube flex causing the pressure change.

Also ran the 2nd half of the print at 230c to see if maybe temp was a factor, with no change.

Temps are still holding withing a few tenths of a degree.

Well, crap. I'm kinda outta ideas.

Here's what I've got so far.
The first two were run with the V6, the last three on the Cyclops.
I'm running one more with the recommended retract setting of 1mm, which seems low. It's string-city so far, but let's see if there's an effect on the quality of the print.

[img]http://i.imgur.com/wTWa2IX.jpg[/img]

[TFMike]

Have you messed with the extrusion multiplier at all? Also that black bowden coupling piece from e3d still tends to move a bit when I print with it here on my end, maybe it is doing the same for you? Maybe try sticking the larger screw in piece in its place and sealing it off? Have you tried pla, nylon and abs or maybe something more exotic just to see what happens?

[Glacian22]

Maybe contact E3D and see if they have any suggestions?

[gestalt73]

Ok, so retraction settings had no effect, and printing it hotter or cooler made no difference. Prints at both 235c and 260c have the same surface finish.

I'm not sure that changing the extrusion multiplier would make a difference, it's the exact same gcode file for all the prints.
The multiplier was calibrated using a print off the V6,
And because the surface is so irregular, I can't use a digital caliper to calibrate the print.

When I heat up the hotend, and manually advance the filament, I'm feeling maybe only slightly more resistance than when I do the same with the V6, so it doesn't feel like a blockage or restriction in the nozzle.

But I suppose I could remove the nozzle and blowtorch it, then do the mhackney thing to check the nozzle channelling.

I'm seeing the exact same problems with both blue and my orange ABS.

Tim with Filastruder hasn't had a chance to work with the Cyclops yet, and I just sent a message to E3D with a description of the issue and the URL to one of the pics.

Over the next couple of days, hopefully there are one or two others running Cyclops that might have some tips.

[TFMike]

Another idea would be to pull it off and stick it onto another printer if you have one and see how it prints on a different unit. I'm all out of ideas. What is the Mhackney thing?

[JFettig]

Your heater block is awfully close to the end effector platform. Pull that off and see if its warped - my first revision I was having a heck of a time getting to work and that is why, it caused the whole thing to warp and give me print artifacts and various other problems.

[gestalt73]

Hey TFMike,

Unfortunately I haven't finished my second printer build, so I only have one printer to try. Everything turns out beautifully with the V6, so it's unlikely to be an issue with the rest of the printer hardware.

Hey JFettig,
I guess I can't totally rule out platform warping or flex, I've just pulled everything apart to examine.
Were you seeing print wierdness like what I'm getting right now?
Here's a shot of the clearances between the heater block and the effector platform.
I've just pulled everything apart again, and I'm not seeing any melting or warping. The effector platform looks flat to me.
The heatsink does get warm compared to the V6, but nothing hot enough to warp the bracket.
At temperature, both my V6 platform and my Cyclops platform feel about the same in terms of give and stiffness.

I see that rpress and Bubbasnow are also rockin' the Cyclops, I'ma gonna drop them a note to see if they've seen anything like this.


Here's the clearance I've set up with the Cyclops.
[img]http://i.imgur.com/jC89V8O.jpg[/img]

[geneb]

Might it be the heat radiating from that huge heater block? Try putting some of that silicone muffler tape on it and see if that cures the layer roughness.

g.

[Polygonhell]

I'd be tempted to start by printing a largish hollow cylinder and look at the surface finish on that.
The reason I'd do that is because i agree with genes assessment, there is a good chance the filament is staying molten too long, and on a larger cylinder, it should have less effect, since little of the print stays under the Hotend.
The idea being to try and understand what the cause is then find a solution.

[gestalt73]

Hey geneb,

Ya know what? It can't hurt to try. I'll wrap the block and try a print or two to see if there's any effect.
Even without the wrap, I'm within a few tenths of my target temperature, but it could help if the platform is warping due to heat.

Hey PolygonHell,

I was wondering the same thing, but in both my samples I'm running with the exact same part cooling setup for both hotends.
They're running the same gcode and everything, and the V6 print turns out great.

I'll set up a wide single wall cylindrical model to print to try it out though, one that won't require any cooling.
(EDIT) oooh, you're saying that the heat is radiating *downward*. hmm, didn't think of that. I'll let my hug-ish calibration square complete, and we'll see how it turns out.

I'll also redesign the effector and bracket today to be extra, extra rigid and bulky to try to rule out flex. I suppose I could also widen it up a bit to give more clearance.

[rpress]

Yes, I do indeed see this artifact on my Cyclops as well. I can't tell if mine is as bad as yours; not to be rude, but your photos are pretty blurry. Do you have a better camera, maybe something that can do macro?

My theory is that the filament extrusion is not consistent due to the stick-slip friction. The Cyclops has a much more convoluted path than the v6 and I have noticed that the required extrusion force is higher.

I'd say try these:
1) Canola oil. This will reduce friction in the bowden and also reduce the sticking at the heat break. Not exactly a long term solution but worth a shot.
2) Better heat sink cooling. Do you have even more problem with PLA? The heat sink on the Cyclops is kinda wimpy. I also noticed the fan has turbulence because it's too close to the fins. I spaced mine off by 1/4" and put a piece of Kapton to close the gap on the bottom. I've also overvolted the fan.
3) Double-check PID tuning. The Cyclops will need higher gains because of the increased mass, also you're running a 25W vs a 40W element?

[gestalt73]

So, that's interesting. the large-ish cube turned out ok.

Here's a comparison shot with the large cube, the original V6 benchmark print on the left, and the Cyclops print on the right.
During reassembly, I wrapped the block in silicon tape, and rerouted the thermister wires and fan wires completely away from the block and in different directions.
I'm running another test print now to see if that was the trick.

Sorry about the Mr. BlurryCam shots, the originals were not blurry at all, but maybe between resizing, imgur, then the forum things got a bit off.
I'll include a link to the original shot on this one below.

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

[gestalt73]

Ok, I'm out of ideas.

rpress has mentioned that he's seeing similar quality issues with the Cyclops hot end.

Has anyone gotten good quality prints out of their Cyclops?
My hope was that I could use it as a daily driver, and then just use the second extrusion for HIPS support when needed.
I don't see how that's possible with the quality of prints I'm seeing.

I've tried everything, and nothing has had an impact on the print quality.

Here's my final print of the night, multiple parts at the same time.

What's interesting is that if you light it just right, it looks fine. I have intentionally lit the photo so you can see the surface defects.
The full sized photo is available through the URL below.

The dual extrusion aspect does work very well with an appropriately sized prime tower. No complaints there.
The few test prints with ABS and HIPS as breakaway support worked really well.

I did send a support message to E3D, I'll let you know if we figure anything out.


http://i.imgur.com/hKfrs8w.jpg
[img]http://i.imgur.com/hKfrs8w.jpg[/img]

[rpress]

I'd say try these:
1) Canola oil. This will reduce friction in the bowden and also reduce the sticking at the heat break. Not exactly a long term solution but worth a shot.
2) Better heat sink cooling. Do you have even more problem with PLA? The heat sink on the Cyclops is kinda wimpy. I also noticed the fan has turbulence because it's too close to the fins. I spaced mine off by 1/4" and put a piece of Kapton to close the gap on the bottom. I've also overvolted the fan.
3) Double-check PID tuning. The Cyclops will need higher gains because of the increased mass, also you're running a 25W vs a 40W element?

I've tried everything, and nothing has had an impact on the print quality.

So you have tried the Canola oil, spaced the fan back, and re-did your PID tuning? None of these had any effect?

[gestalt73]

Hey rpress,

(3) PID Values / 25w heater block One of the first things I did to narrow down the issue was to swap out the 25w heater block with one of my spare 40w heater blocks. For each print, the temperature variance never goes more than +/-0.5c, and normally holds within a few tenths of a degree using the existing PID values. My current PID values do result in a bit of initial overshoot, but settles quickly and stays put for the duration of the print. This was the case for both the 25w block as well as the 40w block.

(1) Canola oil I had assumed that was suggested only if printing in PLA, I've never heard of it used for ABS prints. I don't have Canola oil handy, but I dipped the end of the ABS in Olive oil and I'll give that a shot right now.

(2) Fan changes I did note that the cold end was always warm. I only have the materials on hand to space the fan out by 1/8", so I've set that up, and placed a piece of kapton tape underneath the fan to help direct airflow. I don't have any > 12vdc sources handy, so I can't overvolt it. But I'll give the spacing a shot right now.

Thanks for your help! I'll let you know how it works out in a bit.

I'd say try these:
1) Canola oil. This will reduce friction in the bowden and also reduce the sticking at the heat break. Not exactly a long term solution but worth a shot.
2) Better heat sink cooling. Do you have even more problem with PLA? The heat sink on the Cyclops is kinda wimpy. I also noticed the fan has turbulence because it's too close to the fins. I spaced mine off by 1/4" and put a piece of Kapton to close the gap on the bottom. I've also overvolted the fan.
3) Double-check PID tuning. The Cyclops will need higher gains because of the increased mass, also you're running a 25W vs a 40W element?

[gestalt73]

Hey rpress,

They were all great suggestions, unfortunately I didn't notice any differences on this print.

I can pick up some actual Canola oil and longer screws tomorrow to try again, but the olive oil and spacing out the fan didn't seem to make a difference tonight.

Next time you run your Cyclops, can you grab some overlit shots?

I'm starting to wonder if it's just me, or if it's an issue endemic to the design.

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

[KAS]

You can see some of the same artifacts in the left part only exaggerated in the right part. The layers appear to start inset then gradually move towards the exterior. On the next layer they start on the exterior and move towards the interior. The mounting tab towards the bottom looks decent on both parts. Although you can still see the line created in the base of both parts just below the tab. I'm thinking mechanical calibration, mainly because you can see some of the artifacts on the original item.

Can you print this STL and post a pic?

75mmDisk-pt4mmtall.stl

(24.89 KiB) Downloaded 132 times