3D Printer yet?

[xrunner]

Oh now, come on try it - nothing can go wrong. 

[Mr. Scram]

I dunno. Seems to me if it sucked that hard, why would anybody ever make a nozzle out of it? Also; a awful lot of SS cookware out there.

Much more likely a user error of some sort; either doesn't belong in this application or I didn't build the hotend right for that nozzle. 

mnem
 

Stainless nozzles are intended for filled filaments. Carbon or metal filled filament will wear out your brass nozzles in no time because the fill material is much harder than the nozzle. The different thermal properties of brass and stainless mean some engineering is required to keep filament temperature and pressure where they need to be. Hot ends and nozzles seem deceptively simple but building good ones is far from a trivial task.

Decent stainless cookware will have an aluminium or copper disc integrated into the bottom. This disc will spread the heat much more evenly to prevent hotspots, while still retaining the hygienic benefits and durability of stainless cookware. Cast iron pots and pans brute force the issue by having a large mass and therefore thermal mass, evening out the temperature and dampening thermal differences.

[mnementh]

I dunno. Seems to me if it sucked that hard, why would anybody ever make a nozzle out of it? Also; a awful lot of SS cookware out there.

Much more likely a user error of some sort; either doesn't belong in this application or I didn't build the hotend right for that nozzle. 

mnem
 

Stainless nozzles are intended for filled filaments. Carbon or metal filled filament will wear out your brass nozzles in no time because the fill material is much harder than the nozzle. The different thermal properties of brass and stainless mean some engineering is required to keep filament temperature and pressure where they need to be. Hot ends and nozzles seem deceptively simple but building good ones is far from a trivial task.

Decent stainless cookware will have an aluminium or copper disc integrated into the bottom. This disc will spread the heat much more evenly to prevent hotspots, while still retaining the hygienic benefits and durability of stainless cookware. Cast iron pots and pans brute force the issue by having a large mass and therefore thermal mass, evening out the temperature and dampening thermal differences.

Yeah; I knew the latter; and for the same reason, SS can be one of the more frustrating materials to weld. I discovered the former after doing some research once I figured out the problem. 

Still doesn't seem it should be able to make that much trouble; there's a lot of heat applied and a lot of covalent area with the threads compared to the total mass of SS. Bottom line is I used 'em cuz they came with one of my parts kits; now I know better. 

mnem
That's okay... they'll make good mini-flamethrower nozzles.

[mnementh]

Oh now, come on try it - nothing can go wrong. 

Iterative Design called for you... s/he sez "lick me."



mnem

[Ed.Kloonk]

You may have seen this, if so, sorry.

Build a house to put your 3D printer in!

[mnementh]

Now that my Diggro Alpha is actually a 3DP again and not just a machine, it's been running pretty close to nonstop for days now. *knock on wood*

The nosepieces were the ball-buster; once I got to actually printing, I was still fighting layer adhesion issues. But those first three partial prints still allowed me to revise text location and resize the LED opening to allow for squish; I actually managed 3 revisions in one afternoon before the successful (but too short by 10mm) print in the back. I'm going to take the piss on that one and blame the preceding week of testicular torment for me not noticing that it was too small in the Z axis until I had the actual part in my hand. 

That said... the nosepieces take ~7.5 hours each; so those 3 failed prints and one bonehead fail all in one afternoon actually saved me more than a day and ~$6 wasted filament.  After that, the two on the right came out exactly how I wanted them. 

   

Here you can see my progression in designing the tailpiece; the first version in back suffered some cosmetic bunge due to orientation, and it didn't actually muffle anything at all. But having it in hand DID give me the idea for what would become the final product; that channel on the bottom was originally added as an afterthought, but after chopping the ass off the piece in Fusion, what remained became an actual muffler that works.

The part on the left was the first version of the successful part; I did some cosmetic refinement (mostly reworking the fillets so they match the nosepieces and adding a tube for the LED mount) in the part on the right but the biggest improvement came from just reorienting the piece. I originally printed it foot down, figuring that would work better for the filament bridging necessary to make the slots in the muffler, and I was concerned that the support material might pull the slots apart. The bridging in the slots turned out a bit stringy, but the supports did behave. Most noticeably, the ceiling area inside the part had lots of nasty stringing, plus the finish pattern in the top is really obtrusive.

The part on the right is mostly just reoriented to print on the butt of the piece; this greatly improved the quality of the slots, making straight bridging and clean slats which are only a few lines wide with excellent layer adhesion. More importantly, the ceiling became a sidewall in the print, so no stringing... and the ugly pattern that was on the top gets nicely massaged away by the sandstone finish of the Build-Tak. Best of all, the new design and this orientation makes what was a 11.5 hour print now a 5.5 hour print. 

So at this point... it's just a matter of making 6 copies of the two pieces, and doing the modding of the actual PSUs. Nice, boring fiddly-bits work. 

mnem
 

[mnementh]

Low Tech Hi-Tek solution to printer monitoring: as in "I have this baby-cam lying around to watch the kids in the back yard; might as well sic it on my 3DP when I go to bed with a print job running."

Yeah, I know... wooden table, cardboard enclosure...  I totes need to get it up on a paving stone like my old one for safety's sake. And build a proper enclosure. At least I did put a smoke defector in there with it.

mnem
*toddles off to ded*

[mnementh]

From the SAVE MY ASS!!! Dept...

      

Had the opportunity to test the "Power Failure Print Recovery" feature in Marlin 1.9.x this morning; power went out at ~5AM for a couple hours, and this print was at approx 90%. That little booger and a single-layer-height line of overextrusion from printing to a fully-cooled part was the only evidence of any misdoing. Cleaned up, a very usable part; though some repeating blebs around the corners show that I do need to clean and tighten the X-Y tracks/rollers after running it pretty much nonstop for 4 days.

Well done, Diggro. 

mnem
*off to engage in that illusion known as lunchtime*

[Zucca]

"Power Failure Print Recovery"

Can't use it, I have happly deactivated the SD support on my Ender 5. Now I have finally plenty of nvm memory to play with other Marlin features.
Unfortunately the Linear Advance is not supported on my Melzi boad.

[Zucca]

So for an exciting start to the day, I downloaded and compiled Marlin 2.0.7.2

I am running also the 2.0.7.2, and I have activated the PID Bed jazz, works no problem.

Now the pain in the ass is the raspi 3B+ which loose the serial connection over USB about once every 3 days.
No suprises here, I had no hope on the rasp devil berry to run stable. This last octoprint experience proves my point.

What to do now? Oh I know: in FreeBSD we trust.

[Nominal Animal]

Now the pain in the ass is the raspi 3B+ which loose the serial connection over USB about once every 3 days.

That is a hardware problem; the Qualcomm Broadcom USB hardware implementation can occasionally forget USB packets in certain situations, which should never happen if it actually followed the standard.  Basically, if one is using a Raspberry Pi, all USBy communications are based on "perhaps I work, perhaps I don't".  They've done a lot of software fiddling to make it as rare as possible, but it's just lipstick on a pig IMO.

What to do now? Oh I know: in FreeBSD we trust.

Nah, just switch to a better SBC that is supported in mainline Linux.  (I mean, ARM is Tier-2 on FreeBSD, but primary tier on Linux.)
For example, a $20 La Frite (1GB RAM, AML-S805X-AC) with a $15 (8MB) or $20 (16MB) EMMC 5.x module (for OS) runs upstream Debian ARM port and a vanilla upstream kernel without vendor-specific patches.

Want moar power, ~ twice as much as R'Pi 4, for twice the price? Odroid N2+, also running mainline kernels.  (Hardkernel does have their own kernel dev tree at github they promise to maintain till Jan 23, 2023.)

[mnementh]

I'm gonna step back here and say that aside from the power outage... which my Diggro did handle pretty much in stride... it has been running pretty much nonstop for 5 days now. I do believe it feels it has something to prove.

mnem
Or maybe it saw me looking for ewaste recyclers online... 

[Nominal Animal]

Diggro no aggro, eh?

Pat, pat.  Good Diggro.

[mnementh]

Pretty much. My balls still ache from the full week of prior tho. 

mnem
Happy Halloween.

[Zucca]

That is a hardware problem; the Qualcomm USB can occasionally forget USB packets in certain situations, which should never happen if it actually followed the standard.  Basically, if one is using a Raspberry Pi, all USBy communications are based on "perhaps I work, perhaps I don't".  They've done a lot of software fiddling to make it as rare as possible, but it's just lipstick on a pig IMO.

I would like to squeeze all the rasp devil developers and let them fall free out the window.
Monkeys.

Nah, just switch to a better SBC that is supported in mainline Linux.

I have a  BeagleBone under my bench doing  nothing. I still think that a FreeBSD Lenovo would be rock solid, BUT I am curious cat: I will give it the Beaglebone a chance.

[Nominal Animal]

I have a  BeagleBone under my bench doing  nothing. I still think that a FreeBSD Lenovo would be rock solid, BUT I am curious cat: I will give it the Beaglebone a chance.

I might, too, if I had a BB.  I know the Linux kernel is rock solid (and utterly ubiquitous in everywhere you wouldn't suspect, like TVs), but the userspace is increasingly giving me the heebie-jeebies nowadays – fragile as frick, wrought with single-point failure possibilities.  FreeBSD folks are more conservative, value stability over the latest fad, which makes for good basis for appliances.

Me, I do my appliance software sides from scratch (or from minimal Debian/Devuan up), but that's way more effort than most people have time or interest for.  For me, it's comfy old slippers stuff, not "work".

[mnementh]

So... not a big fan of Java-coded Blu-Ray players, ehhhh?    Only Sony would put that on the box like it was a feature... 

My experience with the B-bone is nearly a decade old and at best mired in frustration. When it was first released it was hailed as the saviour of ArduPilot; the only way for the project to move forward. It (the ArduPilot project thereafter) was, and remains to this day, an unmitigated clusterfuck of competing visions and infighting. 

mnem
Old mother Hubbard
went to the cupboard
to get her poor doggie a bone.

But when she bent over
ol' Snoopy took over
and gave her a bone of his own...

[Monkeh]

Now the pain in the ass is the raspi 3B+ which loose the serial connection over USB about once every 3 days.

That is a hardware problem; the Qualcomm USB can occasionally forget USB packets in certain situations, which should never happen if it actually followed the standard.  Basically, if one is using a Raspberry Pi, all USBy communications are based on "perhaps I work, perhaps I don't".  They've done a lot of software fiddling to make it as rare as possible, but it's just lipstick on a pig IMO.

I like to crap on Qualcomm as much as the next guy, but the Pi is all Broadcom. Which is even worse.

[Nominal Animal]

Now the pain in the ass is the raspi 3B+ which loose the serial connection over USB about once every 3 days.

That is a hardware problem; the Qualcomm USB can occasionally forget USB packets in certain situations, which should never happen if it actually followed the standard.  Basically, if one is using a Raspberry Pi, all USBy communications are based on "perhaps I work, perhaps I don't".  They've done a lot of software fiddling to make it as rare as possible, but it's just lipstick on a pig IMO.

I like to crap on Qualcomm as much as the next guy, but the Pi is all Broadcom. Which is even worse.

Crap, I mixed the two.  I was trying to say Broadcom, not Qualcomm.  Must edit.   Thanks for the correction, Monkeh.

Edited to add:  In case anyone thinks this is just bashing a company or two for no reason, look up "GPL violation" and the company name.  GPL is not at all difficult to comply with.  It is not even a question of opening up your own source code; just stop using modified GPL code is enough.  In many cases, in appliances, there is often just a few small but important changes to GPL'd code that contains no trade secrets, just basic functionality stuff (usually horribly implemented) that would have been cheaper and faster to just do right and push upstream to the Linux kernel, and keep their Secret Sauce in the non-GPL userspace they write themselves.  Companies like Broadcom hate the idea of free/open source software for some reason, though, and cannot understand business models that leverage GPL software without violating the license. 

I'm still surprised Eben Upton even allows three or four Raspberry Pi Foundation members (James Hughes, Phil Elwell, Dave Stevenson) to submit an occasional patch to the Linux Kernel – although only Phil Elwell is listed as part of their 'team'.  Must irk Upton to no end, I'm guessing; he's a dedicated Broadcom employee.

[Zucca]

Soo I took out the Beagle Bone: one USB, no WIFI on board. Mmmm 5 minutes later is was back in the box and parked under my bench again (with a lot of Raspi that I have to sell now).

Just installed FreeBSD on my old Lenovo, the more I learn FreeBSD the more I love it. It just works if you know what you are doing, as always anyway eh  .

PS: Next I will show off my upgraded printer. It starts to make a lot of fun.

[mnementh]

I'm having fun with mine too.  Tiny Toothless is 28mm tall.

I decided to tinker with the Longer A30 profile in Cura... it has a 0.06mm ultrafine profile.  Here I've done no fine-tuning for this profile; same settings I use on the Ender 3 Pro Profile for most of my work.

mnem
};=)~~~~<

[Monkeh]

To derail a bit further as I'm looking at SBCs this evening..

For example, a $20 La Frite (1GB RAM, AML-S805X-AC) with a $15 (8MB) or $20 (16MB) EMMC 5.x module (for OS) runs upstream Debian ARM port and a vanilla upstream kernel without vendor-specific patches.

I'm struggling very hard to find a reason to buy this board other than the price (which I admit, is nice).

There's almost no documentation. No good photos, no pinouts, the schematic doesn't match half the photos I've seen.. I'm looking at two boards which both claim to be the same revision but are substantially different.

It's also rather hard to buy. Amazon US (which doesn't carry that board anyway), and their own store: Which is hard to navigate and still carries a shipping delay banner from 9 months ago. Do they not want people to buy them?

[Zucca]

I'm having fun with mine too.  Tiny Toothless is 28mm tall.

Nice beast, one day I will go below 0.2mm Z layer too, but first I have to design custom plates for the dual Z upgrade on my E5.
Oh the 0.9 steppers are coming today, ohhhh another upgrade.

[Nominal Animal]

For example, a $20 La Frite (1GB RAM, AML-S805X-AC) with a $15 (8MB) or $20 (16MB) EMMC 5.x module (for OS) runs upstream Debian ARM port and a vanilla upstream kernel without vendor-specific patches.

I'm struggling very hard to find a reason to buy this board other than the price (which I admit, is nice).

The hardware is much better (reliable) than 'Pi, at the same price range.  That's why I used it as an example.  Libre Computer completely fucked up the Kickstarter, though; I was/am a backer.  Drived off a LOT of developers/users.  (Didn't get the EMMC, which was a mistake.)  You can find the pinouts in the schematics and at the LoverPi forum.  Or you can ask Libre Computer directly; they're absolutely not hostile, quite friendly, but .. neglectful.

Now that I think of it, they fucked up the Kickstarter much the way Mnementh described Creality fucking up theirs.  (Except Libre Computer didn't get it in retail any sooner, just completely botched the time estimates.)

At the core, it is a very straightforward Amlogic AML-S805X-AC board, so the support matrix at Linux-Meson fully applies.  GPIO Pinctrl, PWM, interrupts etc. are built in to the kernel.  For example, GPIO pins are accessible via the standard kernel /dev/gpiochipN character device(s).  So, technically, as long as you can find the Meson docs, you don't actually need any manufacturer specifics except for the schematic.  This was also a core point: you are not dependent on the manufacturer support or documentation of the various subsystems; you only need to know what they are, and the community has the rest, because the hardware itself is supported at the kernel level.

If you were to ask me, I'd recommend an Odroid instead.  (I have a HC1 and a C1+.)  They have much better documentation, and the user forums are quite active (as in, manufacturer answers questions and takes advice very actively; it's not just us users talking amongst each other).  The hardware hasn't failed me yet.  I am thinking of getting a HC4 (to replace my HC1 as a firewall/server), but the price (USD $65 as of now) is higher than 'Pi.  The base C4 (USD $50) matches 'Pi 4 performance, has a rock solid Amlogic S905X3 (see Linux-Meson), and 47 GPIO pins supported in the vanilla kernels, in case you need those.

[Monkeh]

For example, a $20 La Frite (1GB RAM, AML-S805X-AC) with a $15 (8MB) or $20 (16MB) EMMC 5.x module (for OS) runs upstream Debian ARM port and a vanilla upstream kernel without vendor-specific patches.

I'm struggling very hard to find a reason to buy this board other than the price (which I admit, is nice).

The hardware is much better (reliable) than 'Pi, at the same price range.  That's why I used it as an example.  Libre Computer completely fucked up the Kickstarter, though; I was/am a backer.  Drived off a LOT of developers/users.  (Didn't get the EMMC, which was a mistake.)  You can find the pinouts in the schematics and at the LoverPi forum.  Or you can ask Libre Computer directly; they're absolutely not hostile, quite friendly, but .. neglectful.

Can't trust the schematic because it doesn't align with one of the two identically marked but blatantly different boards I can find pictures of..

Swing and a miss with those guys, sadly. Which is why the Pi is still going strong (despite being objectively a turd, it's a half way well documented turd one can easily buy). Odroid definitely much closer to the mark.