Palette 2 anounced  

Re: Palette 2 anounced

Interesting!
Particularly the implementation of the Canvas 2 Hub... I wonder how much they manage to reduce the need for an purge tower.

Re: Palette 2 anounced

Interesting!
Particularly the implementation of the Canvas 2 Hub... I wonder how much they manage to reduce the need for an purge tower.

The Hub won't reduce the need for a purge tower, but it does mean that the Palette 2 should do a much better job of staying in sync with the print.

The big bugaboo with the Palette+ is that even the best-tuned setups do only a mediocre job keeping the colors in sync. In my experience, the best I've been able to do is one out-of-sync color change about every 1,000 splices. That may seem pretty good, but a large print can have 500+ splices, and just a single layer with the wrong color ruins the print.

It took me a lot of experimentation to get to that one-in-1,000 error rate and I had to make some modifications to the filament path to keep it more stable. I think the more typical customer can expect one wrong splice every 100 splices or so. So not good at all except for things with just a few splices (keychains, bookmarks, stuff like that).

So I'm hoping the Palette 2 will get that error rate close to zero. There's some other changes the new design makes possible which I've already suggested (automatic splice break detection would be very helpful), but at this point I still think the MMU2 is a more promising approach.

I don't want to sound too negative on the Palette--it works a lot better than it should, and the company does a great job listening to customers and helping them fix problems. But as I see it, there's just fewer things that can go wrong with the MMU2's filament swapping method as compared to the Palette's splicing.

Re: Palette 2 anounced

I saw the video last night with 3D Printing Nerd hosting. They also have a 2 pro. The problem I would worry about is the getting out of sync. My question is how does the inconsistency of filament size play into this?
As the filament varies in size "plus or minus .03mm" how can that affect the output of the Palette line? Over time and a bad batch of filament color could be put in the wrong places.

Re: Palette 2 anounced

That may seem pretty good, but a large print can have 500+ splices, and just a single layer with the wrong color ruins the print.

Printing in 4 colours at 0.2mm layer height = 15 changes per 1 mm. "500 splices" is not a very large print. The Mk3 + MMU can have 3000 tool changes per print.

Sorry to be pedantic!

Peter

Re: Palette 2 anounced

As the filament varies in size "plus or minus .03mm" how can that affect the output of the Palette line? Over time and a bad batch of filament color could be put in the wrong places.

Filament is extruded by length and not by diameter. It will not matter what the diameter is.

Peter

Re: Palette 2 anounced

Filament is extruded by length and not by diameter. It will not matter what the diameter is.

Peter

Good to know, thanks.

Re: Palette 2 anounced

That may seem pretty good, but a large print can have 500+ splices, and just a single layer with the wrong color ruins the print.

Printing in 4 colours at 0.2mm layer height = 15 changes per 1 mm. "500 splices" is not a very large print. The Mk3 + MMU can have 3000 tool changes per print.

Sorry to be pedantic!

Peter

LOL, maybe I've been trained to avoid too many tool changes as it would guarantee failure!

Re: Palette 2 anounced

One thing that concerns me greatly about the Palette 2 (non-Pro version) is that it only comes with a 90-day warranty. It costs $99 to extend that to 1 year or 50,000 splices, whichever comes first. IMHO that just seems like a trick to make the non-Pro version look $99 cheaper than it really is. And if large prints require 500+ splices, that means even the extended warranty will have expired after fewer than 100 large prints. That does not inspire confidence in the longevity of the non-Pro splice core. The Pro version has a 1 year warranty with no splice limit. You might as well spend the extra $100 to get the Pro over the non-Pro with the extended warranty. I'm sure Mosaic is counting on many people doing just that.

The Palette 2 was announced in late August and is claimed to start shipping in October. The Prusa MMU V2 was announced in March and was originally supposed to ship in May, then delayed to late July, then slipped again to late August. (I'm not holding my breath...) I'm not putting any money on this, but if you were to bet, which one do you think will ship first? 🙂

Re: Palette 2 anounced

I'm not putting any money on this, but if you were to bet, which one do you think will ship first? 🙂

Well, I'm hoping that by the time the Palette 2 starts shipping I'll have a pretty good idea of whether it might be worth buying, based on my experience with the Palette+ and (I hope) the MMU2.

Re: Palette 2 anounced

Well, I'm hoping that by the time the Palette 2 starts shipping I'll have a pretty good idea of whether it might be worth buying, based on my experience with the Palette+ and (I hope) the MMU2.

I believe the Pallette will work with 4 filaments and the MMU2 with 5 (or 4 + soluble). That extra filament is a big bonus, but the downside is that the filament has to move a long way during a tool change, and that may be a whole load of time.

Let's say that that filament can be moved at 50mm/sec over 500mm. That's 20 seconds just to move the filament for a tool change. Then the "selector" (I suppose it's called that) has to move to the correct filament position. I could see tool changes (excluding the Ramming and Purge) taking at least 30 seconds and for a large print (say 2000 tool changes) that's a lot of time.

The MMU1 takes a lot less time to load/unload, but I guess it's down to how much you value reliability and, of course, that fifth filament.

Peter

Re: Palette 2 anounced

Well, I'm hoping that by the time the Palette 2 starts shipping I'll have a pretty good idea of whether it might be worth buying, based on my experience with the Palette+ and (I hope) the MMU2.

I believe the Pallette will work with 4 filaments and the MMU2 with 5 (or 4 + soluble). That extra filament is a big bonus, but the downside is that the filament has to move a long way during a tool change, and that may be a whole load of time.

Let's say that that filament can be moved at 50mm/sec over 500mm. That's 20 seconds just to move the filament for a tool change. Then the "selector" (I suppose it's called that) has to move to the correct filament position. I could see tool changes (excluding the Ramming and Purge) taking at least 30 seconds and for a large print (say 2000 tool changes) that's a lot of time.

The MMU1 takes a lot less time to load/unload, but I guess it's down to how much you value reliability and, of course, that fifth filament.

20-40 seconds is about what the Palette+ takes for every color change. The splicing method requires rather a lot of purge because even if you're dialed in exactly right, you can't be sure that the splice will hit the extruder at exactly the right moment every time. The default Palette+ settings have 80mm of purge every color change, but I presume the Palette 2 will allow that to be reduced a lot. I can't see it going much below 20mm of purge, though. It takes at least that much to ensure the old color is completely purged from the hot end. The MMU2 should be able to do better, since it pulls the old color out and (I hope) minimizes the amount that needs to be flushed through.

It possible that the MMU2 will be able to move the filament a lot faster than 50mm/sec when loading and unloading. The MMU1 had to use the same stepper for loading and extruding, but the MMU2 uses different steppers. In theory the load/unload path could be geared up since it doesn't need the precision for extruding. There's no physical reason why the filament can't be loaded fast, at least until it engages the extruder gear.

Re: Palette 2 anounced

Slic3r for the MMU V1 was hard coded 60mm of purge, IIRC. Sometimes even that much wasn't enough, e.g. when going from dark to light colors. I don't think the MMU V2 will change anything in this regard as the action in the nozzle is the same as before. In the latest Slic3r beta you can configure the load and unload amount on a per-filament basis, so you can use less purge when going between similar colors, and more purge for dissimilar ones. That should work for either MMU V1 or V2, assuming they allow it.

I thought I'd read that the Palette+ purge was much larger, like 100-120mm, no? All else being equal, the Palette should always require more purge than the MMU, since it doesn't pull anything out of the nozzle, and since it never knows exactly when the splice will reach the nozzle.

Re: Palette 2 anounced

You may be aware, that I prefer to use KISS and I wrote a post-processor for that which minimises the purge amount on a per-filament basis.

Slic3r now also has a (rather complex) method of doing the same thing.

Ramming is a way to get as much old filament out of the hot end as possible, but you still have the "melt zone", where filament cannot be retracted from. And it really does depend on colour/pigments as to how much purging is required. If you change from a dark green or dark red to white, you really do need to purge 100mm or more. (Nice thing with KISS is that is that where possible, it changes tools optimally, reducing purge where possible.)

If the Pallette does not know what colour/pigmentation each filament has, then it simply has to work on a worst case basis. Things have moved on from that.

And I was not aware that the Pallette had to pause the print for a filament change; I was under the impression that it had a loop of filament to prevent this.

Peter

Re: Palette 2 anounced

Is there a technical reason why MMU2 is not marketed as a universal multi-material solution, but rather as a mk3 addition? I mean, simply put it is an automatic filament changer (or am I missing something)...

Re: Palette 2 anounced

Is there a technical reason why MMU2 is not marketed as a universal multi-material solution, but rather as a mk3 addition? I mean, simply put it is an automatic filament changer (or am I missing something)...

It requires support in the printer's firmware to communicate with the MMU V2 control board. I'm sure Prusa Research doesn't want to be responsible for modifying and testing other printers' firmware. They want to sell more Prusa printers. 🙂 I expect someone will eventually implement it in Marlin like they did for the MMU V1's motor multiplexer board.

Re: Palette 2 anounced

Ramming is a way to get as much old filament out of the hot end as possible, but you still have the "melt zone", where filament cannot be retracted from. And it really does depend on colour/pigments as to how much purging is required. If you change from a dark green or dark red to white, you really do need to purge 100mm or more. (Nice thing with KISS is that is that where possible, it changes tools optimally, reducing purge where possible.)

If the Pallette does not know what colour/pigmentation each filament has, then it simply has to work on a worst case basis. Things have moved on from that.

In the current version of the Palette software you can specify whether a color is "strong," "weak" or neither, and it will adjust the purge length accordingly. I definitely agree that it takes a lot of purge to completely flush out a dark color when switching to a light color, and 100mm isn't unreasonable.

With the Palette you need to add to the purge to allow for errors in the synchronization. The software targets the splice hitting the nozzle at 40% of the way through the purge block, so if you need 100mm post-splice to purge colors, you need to specify about 165mm of purge. It would probably be better for the Palette to target the splice at a fixed length inside the purge block (i.e. always 30mm or so), but that's not the way the software works today.

I assume that the Palette 2's better synchronization won't need to put the splice quite as far into the purge block, but the splice is always going to have to be at least 10mm into the purge. I just can't see how you can make this approach any more precise than that.

And I was not aware that the Pallette had to pause the print for a filament change; I was under the impression that it had a loop of filament to prevent this.

The current version of the Palette pauses about 20 seconds on most color changes for synchronization. That's because the Palette doesn't directly communicate with the printer, so the software adds pauses to the gcode, and the Palette (usually) detects the pauses to figure out exactly where in the print the printer is. Those pauses will be eliminated if you use their new "Canvas Hub", which is essentially a Pi running a modified version of Octoprint, and allows the Palette and printer to communicate directly.

Re: Palette 2 anounced

It requires support in the printer's firmware to communicate with the MMU V2 control board. I'm sure Prusa Research doesn't want to be responsible for modifying and testing other printers' firmware. They want to sell more Prusa printers. 🙂 I expect someone will eventually implement it in Marlin like they did for the MMU V1's motor multiplexer board.

Ok! I thought it operated more like the Canvas hub thing; I did not realize it needs firmware support. Makes sense to focus it on Prusa printers.

[...]With the Palette you need to add to the purge to allow for errors in the synchronization. The software targets the splice hitting the nozzle at 40% of the way through the purge block, so if you need 100mm post-splice to purge colors, you need to specify about 165mm of purge.[...]
Very nice write up (the entire post). Thanks for the information! Somehow I assumed Palette would have less purge requirements, but I can see why it needs more.