[Solved] Surface Quality Issue with E3D Cyclops

[Glacian22]

What I wanna know is this - when you're done testing, what are you going to do with 300 fan shrouds?

g.

He'll fashion them all into a very stylish hat? *hasn't slept in 24hrs*

[Tonkabot]

I am following this thread with great interest. I have been holding off buying a cyclops to see if they are working out for people. Thanks for effort gestalt73!

Along those lines I bought a really cool stepper motor with a 5.2:1 planetary gearbox, that I think would make a decent extruder motor.

@ Tonkabot. I am looking for a gearbox stepper. Do you have a link to the one you bought?

This is the one I bought, and have a second in my cart at amazon to get: http://www.amazon.com/gp/product/B00JBP ... YS4P930GVM
Hopefully that link doesn't put you into my shopping cart or anything stupid. Cause it does if I follow it.


[img]http://ecx.images-amazon.com/images/I/5 ... SX425_.jpg[/img]

[Mac The Knife]

Added some ninjaflex to your basket for you.
















Not really, it doesn't link to your account.

[gestalt73]

Hey geneb and bvandiepenbos,

My entire V7 effector weighs 173.7 grams
My entire Cyclops effector weighs 201.6 grams.

So the Cyclops assembly weighs 27.9 grams more than the V6.

And when I bolt on my overkill "triple fan cyclone of parts cooling" rig, it adds 89.9 grams to the weight of the effector, and I haven't really noticed a difference in print quality with vs without the fan.
(Although, I only put on the fans when I want to print quickly and the part isn't large enough to give me 20-30 seconds between layers.

And the punchline to the fan shrouds, is that they don't even fit on my effector, so now I have 2 well printed, and 12 horribly printed but still functional fan shrouds sitting in a zip loc bag.
But it is a nice small benchmark part. Unfortunately KISSlicer likes to reverse the direction of the exterior shell every couple of layers, so your retract and retract acceleration settings need to be really well tuned to get a good print out of it.

I have all these 0.9 degree steppers I bought for my second printer, but I think 1.8 degree steppers may be a better choice for the geared extruder.
The bowden setups seem to respond really well to crisp retracts, and I'm having a hard time spinning it fast enough with the 0.9.


I've been designing and printing out the pieces required to permanently mount the wade extruder. I'll post some pics shortly.
It didn't make much sense to tune the extruder settings when everything was jerry-rigged. I was running it from the top platform, with a stupidly long bowden tube.

[gestalt73]

Ok, so the permanent extruder mount turned out well, and it's all installed.

I'm so thankful the SeeMeCNC guys posted the dxfs to Github. It was very straightforward to modify the extruder mount design. Thanks!

I wanted it to work out so the extruder assembly could be easily removed for various reasons, but that doesn't look like it's going to happen for V1.

The e-steps are calibrated, and the new mount is a bit lower, so the bowden tube will be a bit shorter than before (yay!).

It will take me a while to dial in all the other settings for the new extruder (accel, retract speed, retract mm etc)

[img]http://i.imgur.com/c5d1QQX.jpg[/img]

[img]http://i.imgur.com/ilC59aK.jpg[/img]

[img]http://i.imgur.com/k8MQ52Y.jpg[/img]

[gestalt73]

Ok, so far I think this issue is pretty much solved. The only thing stopping me from marking the topic as [Solved] with a TLDR; is that I'm not getting quite the same consistency from the greg's wade extruder as I was from the EZstruder.
I'll play with the stepper current, and maybe reprint the gears tomorrow.

It's loud, but I have the extruder currently set up with 1500mm/s/s acceleration, with retract/prime set to 3mm @60mm/s.
The seams aren't quite as nice as with the EZStruder, but no stringing.

The new V6 benchmark print and the Cyclops print now match in terms of quality though, so I'm pretty sure that was the root cause.

I see Glacian and Nylocke have published designs that mate the EZStruder to a geared stepper. If I can't figure this out with the wade, I'll set up one of those next.

Thanks everyone for your help!

[gestalt73]

I'll be updating this topic to solved. So Solved. So, So Solved.
Like, good enough for a daily driver good.
Thank you everyone for your help.

I've been busy updating my setup over the past week. Thanks to the posts by Glacian, Nylocke, and MHackney's plate adapter stl, I set up the ezstruder with the Kysan 5.18:1 geared stepper. It took me a bit to fine tune everything, but I'm getting equivalent results now between the V6 and the Cyclops.

I am printing at 260c, which I think helps because of the two right angles the plastic needs to flow through on this hot end.

The Geared stepper is set up with the following changes from stock (These are smoothieboard settings):
- 1.1 Amps
- Acceleration of 250mm/s/s

After that, I took off the Cyclops nozzle, blowtorched it, and took a closer look.
Turns out there was a burr in the nozzle, that I carefully and gently took care of with a loose drill bit.

Here's what I'm getting from the Cyclops now. Please keep in mind that this is an "overlit" pic.
http://i.imgur.com/UXD7dvd.jpg
[img]http://i.imgur.com/UXD7dvd.jpg[/img]


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

[gestalt73]

Here's another sample part printed with the Cyclops. Green ABS and natural HIPS.

Sliced using MatterControl. The surface finish is really good, and the support material breaks away easily leaving a nice smooth finish.

The prints that I'm getting now are practically indistinguishable in quality from my V6.

What's interesting is that the natural hips will take on the hue of the abs plastic.

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]

[Nylocke]

Those are looking NICE. Glad we got it all figured out. You're doibg some pretty cool things with that hotend.

[Glacian22]

Dang Gestalt, that's looking really nice! I'm tempted to get a cyclops now too...

[626Pilot]

This hot end is an enormous pain in the ass. If you retract too far (which is how far exactly? E3D doesn't give you any docs with the kit) you have to tear it down and burn filament out of the heat breaks. If you run too much filament through one hole without plugging the other, you have to tear it down and burn the filament out of the heat breaks. Basically, every possible mistake you can make will involve tearing it down and burning filament out of the heat breaks, and hoping you haven't run out of thermal paste yet. (Which I probably have at this point, and I have to tear it down and burn filament out of the heat breaks yet again.)

I would describe my experience with this hot end so far as "punitive." Hoping for some better luck later on.

What retraction settings are you using? That would be incredibly helpful.

[626Pilot]

Add to the above list: Don't ever take the filament out when it's hot, and don't let it sit too long without printing either. (Ooooh... I can't just let it sit at 230C for a few minutes? That is not good enough.) If it jams, forget about clearing it using any sane, reasonable method. Yep - time to tear it down, burn filament out of the heat breaks and go through more heat sink compound!!! I would have to say this is about as much fun as the sensation of the needle when I'm getting my teeth numbed. It's just the same thing, over and over, and as soon as I think I've got it figured out it jams again.

I seriously have no idea how anyone can tolerate this irritating hot end long enough to get any good at using it. It punishes EVERY last mistake by making you tear it down. I have now gone through the entire packet of heat sink compound - there isn't a drop left. I've chased it all out to where I cut it open.

Oh well, I've had enough punishment for one week. Time to order more heat sink compound and try the Chimera. That can't possibly be as difficult to use as this (insert lots of four-letter words here).

I've given E3D a link to these posts.

E3D, your heat break design is no good on this hot end. It's WAAAAAAAAAAAY too easy for melt creep to creep its way far enough up the heat break, that there is no possibility to clear it without dissassembly. YOU NEED A BETTER HEAT BREAK DESIGN FOR THIS PARTICULAR HOT END. Reusing the design from the Kraken etc. is not good enough. I'm sure there are people who can get this thing running great, but the learning curve is murder. The guy who started this thread was willing to deal with far more teething issues than your customers have come to expect from you. Nobody wants to spend eight hours tearing this thing down over and over again, always thinking they've got it tweaked to where it will actually work, only to discover that their filament is jammed so hard that it breaks when they try to retract it by force - and then to come back the next day, and spend ANOTHER eight hours, and have nothing to show for it. This really needs work before you should feel comfortable handing this out with your name on it. It's just not good enough for prime time yet.

If I was to design a switching hot end, I wouldn't do it like this. I'd do it like a waterjacket. Two halves, machined separately and sandwiched together with some sort of gasket. Smooth path, no 90-degree bends, shortest possible melt zone - I get the impression that the melt zone on this thing is EXTREMELY long. Filament will flow easily for a moment, but then thicken and jam. It's just way too sensitive, probably because the heater block is so long and the heater cartridge can't keep up with heat loss when filament flows quickly. It would be better to design the heat sink and heater block with the filament paths MUCH, MUCH, MUCH closer together, like shotgun barrels. The big heat sink is great because you can also use it with the Chimera, but the cost of that is that the Cyclops is extremely finnicky. Perhaps two filament paths right next to each other, and on the heater block (which wouldn't be much larger than an E3D v6 block), they'd go into a 45 degree angle towards each other and meet at a central junction. (You could just drill that, you wouldn't have to sandwich it.) There would be much less friction, and a much smaller melt zone.

Does anyone know a good brand of heat sink compound? Arctic Silver is only rated to "150+ C". I don't know if it would function at all in a hot end at 230C. I'm done with this hot end, but I think I need it to put together the Chimera.

I'm also backing the Diamond on Kickstarter. You can get the nozzle for about 40 bucks. I already have enough E3D heat sinks and thermistors and the like, so I don't need the full kit. That hot end is designed similar to what I suggest above: drilled (not sandwiched), perhaps 45 degrees of turn in the filament path (rather than 180), and a relatively compact size that a heater cartridge should be able to keep up with more easily.

[Renha]

I already have enough E3D heat sinks and thermistors and the like, so I don't need the full kit.

That's from their FAQ:

In the beginning we had some smaller issues with clogging but after we
changed form E3D v6 to E3D lite6 the clogging issues were gone and the
print quality stayed the same.

[elmoret]

Arctic Silver is fine, that part of the heatbreak isn't anywhere near 230C.

[gestalt73]

Hey 626Pilot,

Yeah, I totally agree that this hotend is a bit more persnickety than the V6, and I'm not sure I've figured everything but here's what I have so far.


E3D recommendations are 1.5mm max retract and 2.0mm max retract on toolchange
I've run a few parts with 3mm retraction successfully. I am going to experiment to see if I can reduce it further without stringing.

On a normal hotend, when you retract it just leaves a void in the nozzle.
But on Cyclops, any retract from one filament will result in the void filling with melted filament when the second extruder actuates.
And as 626Pilot said, the penalty for not mastering this is hours of dissasembly and cleaning.

Just in case you didn't already stumble across it, they've posted some basic directions on their wiki:
http://wiki.e3d-online.com/wiki/Cyclops ... umentation

I've run into almost all the same issues that you're talking about, although I haven't had to disassemble it yet, probably the biggest pain is to keep both filaments seated.

I've only recently gotten consistently results from the Cyclops, and I don't yet know which parts of my process are sound, and which are voodoo, but here's what I'm currently doing.

For Loading [ABS, HIPS] Filament:
- Load in both filaments cold, make sure you're able to press them into the heat break
- Heat up hot end to 260c (I'm also experimenting with 270c)
- Carefully manually push your primary filament a bit (should only be a few millimeters)until you feel "the give", do the same for the secondary filament.
- This is the trickiest part. If you start jamming in filament on one side before the other is fully seated, melted filament is going to backflow up the other heatbreak, and you're going to have a bad day.
- If I can't manually advance the filament at all, I'll remove that bowden, and carefully use a jewelers screwdriver 1.5mm diameter or less to manually push down the filament in the heat break. This seems to happen most often with my HIPS filament.

For Unloading or Changing [ABS, HIPS] Filament:
- Lower the hotend temp to 180c
- Depress the bowden tube retention clip, and remove the bowden tube with the filament

KISSlicer settings:
- I've toyed with MatterControl and KISSlicer, and I'm settling on KISSlicer, using the Short Wall to prime filament between switches
- MatterControl has retract, and retract on toolchange settings, E3D's official recommendation is 1.5mm retract / 2mm max toolchange
- I think MatterControl also does an obscene retract at the end of a print for some reason, but I haven't verified that yet.
- I'm currently using 3mm 60mm/s retracts and primes for both the ABS and HIPS filament
- I'm currently printing at 260c-270c, I read somewhere that Stratasys prints at 270c with 70c heated build chamber, so I've been printing hotter than usual when needed.

[gestalt73]

The part I have on the Cyclops now is running with a 2.5mm 60 mm/s retract for both filaments, and I'm not getting any stringing. I'll try 2mm on the next part when it completes in about 5 hours.

Update, just restarted it with 2.0mm retract for both filaments, and I'm not getting stringing. Looking good so far.

When I started using cyclops, I was using mattercontrol, which has this silly setting "Minimum Extrusion Requiring Extraction" which means that if your extrusion volume is below a certain threshhold, don't bother retracting before movement. It resulted in string city, an the hotend moved from feature to feature.

I think that's why I started bumping up the retraction before I figured it out. 2.0 looks good so far, when I'm bored I'll try knocking it back down to 1.5mm for kicks.

[626Pilot]

I already have enough E3D heat sinks and thermistors and the like, so I don't need the full kit.

That's from their FAQ:

In the beginning we had some smaller issues with clogging but after we
changed form E3D v6 to E3D lite6 the clogging issues were gone and the
print quality stayed the same.

Ah, crap. Well, if I have trouble with the heat sinks I have now I can just order some lite6 ones.

[TFMike]

Gestalt, would you mind sharing the files for the nifty herringbone greg's extruder? I need something like that for an experiment tomorrow

[gestalt73]

Hey TFMike,

The Greg's Wade I tried is this one: http://www.thingiverse.com/thing:18379/#files

fwiw, I couldn't get a really consistent surface finish with it, so I went with the geared ezstruder.

[duvdev]

Hi and thanks for all the info. I orders 2 geared nema 17.

Can some one please share the configuration changes that I need to do after installing the Cyclops.
I use the stock Rambo board.
Thanks.

[626Pilot]

Hi and thanks for all the info. I orders 2 geared nema 17.

Can some one please share the configuration changes that I need to do after installing the Cyclops.
I use the stock Rambo board.
Thanks.

This guy has done it. Maybe try replying to this post: http://forum.seemecnc.com/viewtopic.php?p=62274#p62274" onclick="window.open(this.href);return false;

[626Pilot]

I tried getting the Chimera to work, but my effector is not tram with the bed. If I align it at bed center, it's okay - but the further out I go, the more one of the nozzles is going to be lower than the other. So, I decided to give the Cyclops another shot.

You must use geared extruders with this hot end. A regular EZStruder doesn't have enough torque, nor does it place enough pressure on the roller (although there's a mod you can get that compresses the spring more). When I was using EZStruders with this hot end, I endured the pain I posted about above. With geared extruders, and the hobbed bolt being a few millimeters wider than stock, you get both ample torque and "pinching force." I have been running PLA and HIPS through the hot end. When I was using a stock EZStruder, the HIPS just wound up getting stripped as soon as the going got tough, which was practically all the time. It never ran for more than a few minutes without screwing up. However, with the geared motor and the fatter hobbed bolt, the HIPS doesn't get stripped. It just goes right through.

Another benefit to using geared motors, with lots of pinching force on the rollers: I was able to clear a nasty jam just by extruding 1mm of filament at a time. I had also tried ramming an Allen key down the jammed side, and I did manage to get the filament to go down a little, but it immediately jammed again. Feeding 1mm at a time, I could see the Bowden tube being pushed more and more - but after I had extruded 4 or 5 mm, the jam cleared!

I noticed that on a test print that with 2.7mm retracts, there was almost no stringing whatsoever. This is noticeably better performance than my E3D v6. It ran string-free for most of the print, and there were one or two very fine strands towards the end. I would have to say that as prints with retraction go, it came out cleaner than almost everything I've ever printed.

I'm gonna try that dragon with the egg thing next. Is there a way to find two-color stuff on Thingiverse? I thought they used to have a category for it, but I can't find it.

[bot]

I also recently installed a Kysan geared stepper, though a slightly different model than the one posted here (I got mine from tridpirinting). I notice a tiny bit of backlash in the 8mm shaft, at certain points of the roataion. Is anybody else experiencing that? It doesn't seem to hurt the quality of the prints so much, but I'm worried it will make retractions sloppier than need be. Should I look for a new one, or do all geartrains have some amount of backlash?

[626Pilot]

I got two of those from the same site, with their bigger hobbed bolts. One of them has a little transmission noise when it gets to a certain angle of rotation. Other than that, I haven't noticed anything. Retracts have been very, very clean so far.

[bot]

Can you get your finger in there and wiggle the shaft a bit and see if it moves? I had the same "rattling" kinda noise at parts of the rotation, and it turned out to be a tiny amount of backlash. Everything does seem to work nice, so it's likely a non-issue. I might have to increase retracts by .2mm to compensate, though I have only done three prints on the new nozzle, so it's hard to tell where my retracts stand in regards to the new nozzle and stepper.