Should needle valves be used at all?

[nebbian @nebbian I have a RoboxDual Margaret River, WA, Australia]

I’ve been getting a little frustrated at the inner workings of the Robox post processor, it seems to starve the nozzle after opening the valve. It just doesn’t seem to be doing what it should be doing.

Never one to be content with mere “good enough”, I’ve been working on a solution that just uses slic3r (no AutoMaker post processing at all). This gives me more control over the filament deposition process.

On the left: the best I can achieve from the default software (AutoMaker using its default Cura engine and the Robox post processor).

On the right: Pure Slic3r profile, no needle valve movement at all (just retraction). I haven’t spent much time tuning this either.

 

It’s making me wonder whether the needle valves should be opened and closed at all during a print (apart from when switching nozzles).

What is everyone’s thoughts?

Attachments:

You must be logged in to view attached files.

Compulsive tinkerer

[click @click My Robox is a Green Kickstarter Limited Edition]

Interesting finding @nebbian. I was always worried that retraction on Robox wont work the same way at on straight-through heads due to chamber and physics of it.

Also, using them (needle valves) between colours only won’t do in the same way as in example above. Just imagine each peak being different colour - you would end with exactly the same issue as in the picture on the left. Just look at 2 colour benchy 🙁

It seems that the robox would produce the best results if used with wiping towers in case of two colours - again no needle valve involved at all.

Or, alternatively, maybe, fix the workings of needle valves and slicing? Hm…

[Mike @17bt CHORLEY, Lancashire]

Why is it if you reply to a post, it doesn’t display it in the thread? I can see it elsewhere, but not here? Just a test 😕

[nebbian @nebbian I have a RoboxDual Margaret River, WA, Australia]

@17bt wrote:

First of all that’s very impressive differences, especially if I consider my attempts using the same little torture test and the default AutoMaker Cura, and the Slic3r port you’ve been part of, and the layer on layer quality looks very good, the only area that’s worse is the blob on the pyramid top, and that can be tuned out by that layer time setting in Slic3r, which gives more cooling time.
I must admit I have seen the CEL needle valve assembly, and I was surprised at the crudeness of it all, and it’s very basic, and I was looking at removing it altogether, and using purely retraction, as a control measure and seeing what it does.
The needle valves on the CEL system are not the real problem, the control system is and the feedback loop is insufficient to deal with the hysteresis aka filament compression residuals in the system between the nozzle and extruder, so that when it opens and starts a new layer, it splurges out and when it closes and moves, it strings. As its stands, the precise point these flow & close is somewhat variable, and needs more refinement.
I personally don’t think needle valves are the best solution to this problem, as they can never be compatible with the more abrasive filled elements, so I’d be looking in a different direction for “valving” mechanisms. They are out there, but it’ll take a step up in lateral thinking, and if I was designing at CEL that’s what I’d be looking at 🙂
If you can get better print results not using the needle valves, and only relying on them to stop the ooze past the nozzle or keeping the dormant nozzle primed for extrusion, then the control systems will automatically get better and the profile/settings will also better match the requirements of the printing process. It’s a win win, because the pressure on the closed nozzle will be less so it’ll splurge less aka not at all, and when it’s closed there will be less stringing as it moves from A to B, because the valve will close cleaner. You will also at the same time strip out the unnecessary variables that someone thought they needed when they first conceived the hardware & software, a bit open and closed loop engine management.
Good stuff @nebbian keep tinkering 😀

Thanks Mike, I found your post inside your profile, no idea why it didn’t end up in this thread.

There are two main problems with my print so far:
1) The blob on top of the pyramid (as you noted), and
2) There is a bulge on one face in the rightmost square tower.

I think they’re related. I’ll keep tinkering, but I feel like I’m on the right track here.

This code might find its way into the RSX. It would be great if we could plug into AutoMaker to disable/replace their post processor, it would make it easier for mere mortals to be able to run this sort of code.

I agree with you that the control system is the issue here. It’s clear to me that it hasn’t been tuned as well as it could have been. I’m pretty happy with the hardware (except for some niggling intermittent electrical connection issues).

Compulsive tinkerer

[Mike @17bt CHORLEY, Lancashire]

@nebbian “I found your post inside your profile, no idea why it didn’t end up in this thread.”

I think it was my edit to the line “a bit open and closed loop engine management.”, which should have read a bit like open vs closed engine management. It’s not the first time it’s happened, and the Forum architecture is a little frustrating at times, as it is extremely hard to see Recent Posts, and the Recent Topics doesn’t update on the fly either, but that’s an old horse chestnut, and brings to mind the tail trying to wag the dog. 😉

[Enginwiz @enginwiz My Robox is a Blue Commercial Version Germany]

For dual material prints the needle valves are a very elegant solution.

I also own a Stratasys Mojo. Cooling down, wiping clean, heating up, priming, wiping clean, priming, wiping clean on every change between material and support makes dual printing slow on the Mojo and wastes a lot of very expensive filament.

For single material prints retraction might be a better and faster alternative to opening and closing the needle valve. I did some single material prints on the Robox with Simplify3d gcode and retraction was completely deactivated in S3D. The Robox is capable of fast travel moves and can pull away fast enough with minimal stringing even if retraction is disabled and the needle valve is continuosly open during the whole print.

The Robox could use whatever is best in the specific region of the print.

[Mike @17bt CHORLEY, Lancashire]

I agree for dual material prints, but having seen a DEV1 head operate today, needle valves definitely have their uses, compromises, and restrictions, but I was surprised how good a 0.6mm nozzle worked without needle valves, and the test prints were a lot faster.

 

[Mike @17bt CHORLEY, Lancashire]

I’m at early days with the direct to print Cura 3.2 QuickFill tinkering, but it is quite impressive already, just using a generic G1 B1 to open the valve at the start, but this will change and mimic AutoMaker post processed robox.gcode in the next few days.

OK, my torture little @nebbian it isn’t as good as your Slic3r version, in couple of areas, but the speed was way too high in this PLA 0.2mm layer height print. Poor pic too, as ambient light was on. :/

[Author]

Posts