G_T's build log - trials of an insulated hotter Trident

[G_T]

I've got the bed oriented less tilted than in this pic, but you can see the wave I've mentioned.

Creeping the temperature higher. It went a touch higher than shown in the pic for a bit.

The last pic is just printing a CF-PC part with these settings. It is a bit overextruded but still usable. It isn't being printed for its beauty. And this filament is never going to yield very flat smooth surfaces IMHO. The filament itself is lumpy. Extrusion rate won't be constant in practice, it will have rapid little variances. There's also a lot of carbon in this one which doesn't help in that regard.

It's overextruded because I had changed from a 0.4 to a 0.6 nozzle and not recalibrated. In the end I had to reduce the extrusion multiplier by 5%.

The interior picture isn't quite current -  the rear bento box has been removed. There's now one more layer of insulation on the top, and some fans have been swapped for quieter fans as mentioned previously. That made the printer much more pleasant to be around!

There's still ABS in there but less of it, and at some point those echains have got to go!

Gerald

[G_T]

I've printed the parts for an internal spool holder. Some more bowden tube should arrive tomorrow and then hopefully spools can be inside when printing. That should eliminate moisture uptake during longer prints. I had been printing from a filament dryer but that was causing a lot of failed prints. The CF and GF filaments are moderately stiff and quite capable of expanding out off the reel and forming a birdsnest. That would result in a jam and an air print.

The default voron hanger off the back of the printer doesn't have this issue because there is additional drag on the spool - no bearings. The internal holder is a support frame for the hang-off-the-back spool support arm. So it should provide the same little bit of drag which helps so much at keeping stiffer filament under control.

The minus is going to be extra awkwardness putting a spool in the printer. Oh well. We'll see how it goes.

The hanger part has already been installed in the printer. I just had to print some tube guides since the ones I had on hand didn't orient correctly. Hopefully the ones cooling in the printer now are gong to be appropriately sized and shaped.

I want to crank out all the replacement Z parts soon (the ones that aren't already metal due to the kinematic bed mounts). The new Z1 preload HIWIN rails arrived today so I'm ready to improve the Z's. The X and A, B have already been done during the gantry and X, Y motion rebuild.

The new bed is supposed to arrive tomorrow so that will possibly be part of the lower rebuild.

Bootstrapping a printer is a bit of work. Push the printer to its limits so you can print the hotter parts needed to push the printer further. And hope it lasts long enough at each stage to make it to the next stage.

Gerald

[G_T]

I had to do a lot of fiddling to get this to work acceptably, but the roll of filament can now be inside the printer. I had issues with interference of the reverse bowden tube and the loop of wire coming out the back of the print head. I had interference with the upper loop of tube and the light strip on the left.

I came up with a compromise configuration that has the tube pushing against the extrusion just above the left light fixture when the head is in the back left corner, and a bit of extra tube between the vertical guides and the top pivoting guide so the tubing moves to the back and out of the way as the head passes across the middle of the bed when all the way back.

It works; we'll see how well and for how long.

The parts are a mixture from different projects hosted on this forum.

I HOPE there is enough drag on the spool...  This filament wants to birdsnest if you look at it wrong and it isn't a full spool. I have no way to see the spool until the print bed drops. So I can't tell. I may have to add a small low window on the front door just so I can check on  the spool during a print.

Last picture - still working the chamber temp higher, very slowly! I want to know the problems before they become big problems! This was from a previous print.

Hope I didn't screw up the print in progress. I was wondering why the chamber temp was only in the mid 60's when it should have been at least mid 70's. Turns out I had the exhaust filter on low, not the chamber filter. At least I caught it early. It took 2 minutes to get back to 70C and is slowly climbing up to normal.

Gerald

[G_T]

I had been forgetting about that filter grill in the back. Since it is high up it gets pretty hot. Reprinted last night and installed this morning. I print with a skirt and brim for PC variants, and left the rows of smaller holes top and bottom filled. That should reduce circulation heat losses to the backpack filter unit, while still providing plenty of opening for when the exhaust fan is used.

Layer adhesion with this 3DXTech CF-PC at these temperatures is very very good. One can't just peel the brim off. It won't peel along layer lines and it doesn't want to separate. It has to be cut off. I've been impressed with the mechanical properties of this filament, though not at all with its cosmetics. In the printer I care a lot more about how it functions so I'm fine with that.

In contrast, filament like Polymaker PC-ABS prints beautifully once dialed in and if one takes care of adhesion since it loves to warp. The wire cover on the Stealthburner was reprinted in that filament, but I've since removed the wire cover. It traps heat and I can't afford that.

Gerald

 

PS - Had another print failure, the first one in a little while. This was after an hour and a half of preheating and three hours of printing at 77.5-78C and it just quit extruding. At least the two larger parts completed fine and this was a small part. I pushed filament through the hot end by hand; no clog. I wonder if the extruder motor overheated? Whatever the reason, it isn't the same as previous reasons for failed extrusion which I've already solved.

Edited May 5, 2024 by G_T

[G_T]

I completed printing the parts for the lower internal rebuild. Time to replace a bunch of ABS with PC-CF. That 750g spool is now pretty nearly done for.

FYI - Bolts on original ABS parts will not stay snug. I have to check frequently and re-snug, or print quality goes downhill. With the replacement PC-CF parts, the bolts are set a little more snug than I dared with the ABS parts, and they haven't loosened at all at these temperatures. That's a big improvement. I also have zero concern with these breaking... I'm not sure how people are breaking PC parts, unless it is due to using lower temp PC which is blended with other stuff to lower the temp, or printing it at too low a chamber, bed, and possibly extruder temp to get well bonded low stress parts. Even small parts don't want to break by hand, even with considerable effort. BTW drying this filament is important. I'm heating to 70C for a while then sticking in a vacuum chamber overnight.

If a printer is not well aligned, these PC-CF parts won't flex to accomodate like ABS will do.

Not shown, there are also HIWIN Z1 preload rails to replace the Z's. They need cleaning and regreasing.

It seems like the printer is now having occasional dropouts of the extruder motor. That's not really a surprise, since it isn't a high temp motor. I don't intend to be using this print head all that much longer anyway! I'm thinking to go with a sherpa mini or micro, or hextrudort, using the newer RIDGA V2 gears and a high temp LDO motor. Hopefully this sort of extruder will natively run a fair bit cooler, and therefore be able to be run in a hotter environment.

Gerald

[G_T]

I'm doing something slightly different than usual perhaps. The Z parts which attach to the linear rail slides, and their bottom cover pieces, are designed to bolt together and then never come apart. In my case they are a very good fit. However since they are not supposed to come apart, I decided to wick in some rubber reinforced superglue. Very little would go in since the gap is microscopic and due to print texture.

This is not something I recommend doing! Superglue has quite a range of melting temperature. I did this in one other spot in the printer and it's had a lot of hours at heat so far. That location hasn't melted, migrated, or softened, so I figured to use a little here. Again, I do not recommend doing this! Some superglue melts at too low a temperature. Most of them probably for my target temperature range.

Anyway I'm letting you all know in the spirit of disclosure. A better approach would be to wick in a drop of acetone. I'll probably do that anyway, since it is a solvent for both PC and superglue. It should improve the bond. If you do this, be very careful. A little goes a long ways and you want to control where it goes which is rather difficult with acetone!

The wire hole covers and the Z echain anchor have been replaced. While I was at it I added one more link to that echain. It always seemed a bit tight when the bed was all the way up.

I've done the first round of cleaning the new Z rails. I use WD-40 spraying into the rail slides to pressure wash the bearings etc. Then I move it around some, wipe it down, move, wipe, paying particular attention to the rail ball grooves. If it is really clean, then the rag doing the wiping should end up pretty clean as well. If not, or if you see any evidence of metallic particulate (different from oxidation being cleaned out) then cleaning is not done!

Usually I've had to alcohol soak and perform repeated cleanings to get a carriage to slide smoothly, tested at slow speed. Not with these rails. They actually were pretty clean from the factory. I'll inspect again tomorrow to see if any more cleaning is warranted, then the grease goes on.

I'm not sure what grease I'll use this time - either a high temp grease (have some Krytox on hand) or a heavy-duty weapons grease since those are made for harsh environments and high temp exposure. I've learned to stay well away from lithium grease at elevated temperature... The roughly 130C continuous operating temperature is not sufficient for a hot printer. Bearing surfaces generate higher temperature than ambient. The grease essntially solidified over time and flaked off when I used it on the lead screws. It also acted as a minor plastic solvent at the elevated temperatures.

The Y rails I replaced a bit ago with these same genuine HIWIN Z1 rails, and I used a weapons-grade grease. The Y motion is super smooth now without being loose the way it was with the previous rails. It's had enough ours for an initial test of the grease so I'll probably just continue to use Slip 2000.

One of the new rails feels closer to a Z2 than a Z1. I guess I'll put that one in the back. The other two feel just like the two currently on the Y axis. Real HIWIN rails have better manufacturing and QC than most other affordable (a relative term) rails we can get our hands on, but they still are not high end linear rails.

Gerald

[G_T]

That batch of parts has been swapped. I left on the original rear Z rail since it has the highest preload of the original set and is pretty smooth, and one of the new rails was a little too stiff for the job IMHO. So that's a good batch of ABS removed from the hot box. I forgot to mention when I replaced the wire hole covers I added some bits of Kapton tape to reduce air flow.

The following remain to be done, in no particular order:

1) Bento Box in higher temp material, probably 3DXTech CF-PC since that's what I've been printing lately and though the first spool is essentially dead now, I do have a second one unopened.

2) Tensioners. I'd have already done them if I knew what I wanted to do! I need to print them in something structural that is not brittle, and doesn't creep under pressure and heat. It also needs to have a smooth non-abrasive surface. I'm surprised the ones I'm using now have held up to the heat.

3) Extruder. I have the parts for it just need to print. I'd like it to be in CF-PEKK but I can't print that yet, so first round will probably be 3DXTech CF-PC.

4) Generically, the hot end. I don't know what I want to do yet. Actually, I sort of do, but don't know if it will work. I'd like a mini-stealthburner probably v2, to reduce size and weight. But I want to keep the TAP.

5) Base heater. I haven't decided how I want to control it yet, but I can assemble the heater part. Hopefully this will reduce the 1.5hrs warmup time to something more reasonable.

6) Replace the bed. I have the parts. I do need to print a higher temp replacement for the PC-ABS wiring connector, or just get rid of it entirely.

Actually, it sounds like I have a long way to go with this build.

Gerald

[G_T]

Finally a usable part out of PA12+GF15. I'm having real bed adhesion issues with this Fiberlogy filament. I can print small parts, just not large or thick ones.

This is the new under bed wiring block. The print quality isn't pretty but the part is quite strong and definitely usable so I'm using it. Print settings could use a little tweaking. Currently 265C nozzle 110C bed 72C chamber 0.6 nozzle. Magigoo PA on textured PEI. Skirt and brim.

Time to replace the bed with a hotter flatter bed.

Gerald

[G_T]

I've had little time, but am working on swapping to the new bed.

The bed is a Mandala Rose Works ultraflat, no magnets. I did a little informal testing of one with magnets and concluded the magnets are spaced too far apart for my liking. Some materials I've printed warp very hard. It didn't take much force to cause even a thicker build plate to bow up some between the magnets. That's not a behavior I wanted so I went for magnet sheet instead. This particular sample is marked as 0.0015 out of flat. I'd hoped for a bit better, but it is better than what I've been using.

Speaking of the bed, the drilled and tapped holes could have used a bit better deburring in a few cases. Also when one drills holes, the less fresh the drill bit (ir milling cutter) and the more off the feeds and speeds, the more the metal just outside the hole is displaced. This results in a slight rise in the surface at entrance and exit of holes.

I used a small fine flat file to remove this rim around the holes, and then used some reasonably coarse sandpaper with water to scratch a swirl pattern into both sides of the plate.

The magnet sheet is from Fabreeko and is marketed as having higher temperature adhesive, rated 149C long term. We'll see how it does. It may turn out not to hold magnetism when I ramp it up over 120C. I wouldn't be overly surprised. If that turns out to be the case I'll figure a way to remachine the bed to put in lots of little samarium cobalt magnets. That's the type you want for high temp. Harder to find and cost more, but obtainable. I'm being lazy and skipping that. Hopefully the magnet sheet works well enough.

Cleaning the bed for the magnet sheet was with dye- and perfume-free clear dishwashing detergent followed by acetone followed by denatured alcohol.

Then I used a strong flashlight at a grazing angle to spot all the dust, and swiped that away with an acid brush before applying the magnet sheet.

The magnet sheet goes on first, then the heater. The reason is the heater isn't very flat but the magnet sheet is pretty flat. One can place the new aluminum bed on a flat surface - I used a granite surface plate - and then there is a flat backing to prevent bowing of the aluminum plate when pressing the magnet sheet down.

It can be flipped over and the magnet sheet protected - perhaps by a build plate - and the heater then pressed into place. So the bed never gets bent while adhering stuff to either side.

I've been using a kinematic bed mount system - something I consider quite useful for a printer that can cover a fair temperature range. I didn't want to put holes in the magnet sheet so I mounted the balls and springs to the plate before attaching the magnet sheet. The spacing between the balls is just wide enough to go over my surface plate. I assembled them with blue loctite and don't plan to ever remove them.

I trimmed the oversized magnet sheet using single edge razor blade and a number of strokes to cut much of the way through. Then the blade would glide along the plate to remove the excess. It is definitely nicer working with an oversize magnet sheet!

I've left this sitting in my printer with a build plate on top. I'll get to the heater when I get the chance.

The heater is from Fabreeko. It is fitted for the Voron printer, trident in my case. Edge-to-edge, and glue rated 200C continuous usage. 600w. Built-in 150C thermostat. It's not supposed to need RTV Silicone sealing the edges. I may do it anyway. I'm figuring on max of perhaps 145C anyway so this looks about perfect. Again, we'll see how well it does in practice. I do expect to add a 150C thermal fuse. Anyway with this printer and only 600w I don't think a thermal runaway of the bed would matter. It would just get a little toastier inside.

I already had a thermal runaway with the print head. It was pushing a puddle of polycarbonate. My fault really - I was printing above the thermistor's rating. So I switched to a PT1000 afterwards. Not necessarily the smartest move putting a 300C thermistor in a 500C heat block and printing over 300C routinely!

First pic is bottom of the bed I removed. Second is prepped plate with kinematics mounted. Third is plate on the granite surface plate, with the magnet sheet adhered. Fourth is trimmed and placed back on the mounts.

Gerald

[G_T]

Heater attached. Same process.

Gerald

[G_T]

Bed prep...

Working on the wiring. I've wired in a 150C thermal fuse. It may not really be required since there is a 150C thermostat built into the bed which will also cut the power at that temp, but a bit of redundancy is ok. I used crimp connectors and two layers of high temp teflon heat shrink tubing.

The crimps are barrel connectors, and I ran the wires all the way through so they have a double crimp.

I used some Astro 9477 crimping pliers that I picked up several years ago off Amazon. They work wonders for all sorts of crimp connections. I don't know if they are still made the same, but mine are very nice. I used Type B jaws (jaws are easily swappable) for the barrels.

I used a torch, gently, to shrink the teflon tubing - outdoors. I have no desire to inhale the fumes of overheated teflon! I think it is a bit crazy to use it for heat shrink but it did shrink and it definitely took some heat to get it to do so. Anyway it is rated for higher temperature than it will see, whereas normal heat shrink is a bit borderline with this proximity to the heater.

I preserved the wire shield figuring it was a good idea to keep it.

The plan is to run the bundle of wires through the bed's echain and wire the bed directly to the appropriate power connections in the electronics bay. The connections are all right next to where the wires will enter the bay so it is convenient enough. The thermistor wire though is not likely long enough and would be more annoying to run. So I'll probably use a connector between the bay wire and the chamber wire. If I do I'll locate the junction around the echain exit, not sure which side.

So the net result is no more plug connections under the bed. I feel better about that, since I'm raising the temps. I figure to set the Klipper max at about 148C and limit myself to 145C. Of course most of the time it'll be well below that. Everything in the bed is now rated for these temps; we'll see how that plays out.

Gerald

[G_T]

I completed the bed side of the wiring through connecting the echain. Except for the ground wire none of the wires are connected to the electronics bay yet. I cross-wired the controller to the old bed's thermistor so I could run a test bed mesh. I was curious...

Two things became apparent. First is that somehow my Z chain is about an inch too short. I have no idea if this has been an ongoing thing or not. In any event I'm out of echain so I just ordered some more IGUS. That means it will be sometime next week before the bed is functional.

As for the bed mesh, the canyon-like dip running across near the backside of the bed is gone. The new bed is smoother. However the back right shows a notable dip, increasing as the probing approaches the corner. This is a Trident with a kinematic mount, so this most likely means I need to tighten the front left bed retention spring to keep the bed from tilting under probing load. Tridents are not the friendliest printers for kinematic mounts.

I'll run another mesh after the bed is properly wired. And tighten that one spring now while I'm thinking about it!

Gerald

[G_T]

Work has kept me busy for the last several weeks so only a little progress to report.

The new bed checks out as flatter. The two back corners test as turned down, but I think that is mostly deflection from the probing force. The Trident has support at front corners, and back center. So don't necessarily believe the turndown shown at those back corners! I'll do some more checking later.

One thing I learned is the X linear rail had some bow to it. I suspected it when I made a bed mesh after the bed swap and still saw the bed rolled off on the left and right side. I was thinking to flip the rail upside-down and check the result, but when I had the carriage free, it was moving very freely, and seemed like it might be time to regrease.

Since I was taking it off anyway, I decided to swap it instead. And since the Tap had to be disassembled including pulling the little Z carriage off the little rail, I want ahead and regreased that one. Warning - careful reassembly of that little tap z is required! Grease does help keep the balls in place, but vertical is a bad orientation to be working. I used a toothpick to reposition balls as necessary and reassembled very slowly. Then I wiped off the excess grease.

The new X rail is stiffer and feels more solid. It also got rid of nearly all that bow. I could pull it off and flip it as a test but the bed is essentially flat enough. This mesh was taken with a rather textured plate on. I'll take a better mesh sometime with a smooth plate and probably also with garolite and glass beds.

BTW, with the old bed remember how I had to deliberately misalign the A and B Y rails vertically, to remove an apparent twist in the bed? So the first mesh I took with the new bed (not shown) showed opposite direction twist. So I got to undo the twist I put in to compensate. That's good news.

Since I have the corner braces and side braces on the frame, it's actually a pretty easy operation. One could literally look at the bed mesh differences, left edge and right edge, to compute the twist. Multiply by somewhere around 1.5 or so, and use calipers to check while resetting one side of one of the rails by that amount.

I was lazy though and didn't bother with the calipers this time. I just loosened the three vertical bolts on the brace piece and tapped it with a rawhide mallet. Took a mesh, and repeated. It took me three tries to get it essentially perfect. Again, those brace pieces which greatly stiffen the frame and make this sort of adjustment easier, are NOT part of a stock Voron build.

The first mesh was the new bed after detwisting, and the second after replacing the X rail.

By the way, I really like the CNC tap unit. That's the one part in the print head I do not really want to replace. It should be able to take the intended heat just fine.

FWIW, after regreasing that little Z rail carriage, the occasional overly inconsistent data point stopped happening. I probably should have greased it better before assembling it last time around!

Gerald

[G_T]

I made one other change but didn't document it. When I reprinted the A and B motor mounts in CF-PC (modified mounts, with top bearing support for long shaft motors), the right mount had a location set for a switch to act as a Y limit. That pocket sticks out a few mm from the face.

When using sensorless homing, this would result in hitting the right motor mount at that pocket, and then racking the gantry as the motors tried to pull the left side all the way back. This is far from ideal.

Since I have the printer torn down I wasn't inclined to reprint at the moment... I used a sharp chisel to progressively shave off the protrusion. I used masking tape carefully to trap and collect all the little bits I shaved off. So now the Y homing is cleaner.

BTW that CF-PC is tough. It took some notable mallet and chisel work to shave it down, and the chisel had to be sharp. It wasn't sharp by the time I was done. I have ZERO concerns about this material structurally.

Funny when I tore the printer partially down for this recent round of bed and gantry work, I found a little screw riding under the gantry's echain. I have no idea how it got there.

Eventually those echains will be removed. But not until I make the new print head.

I'm not expecting that to be the next job though. I have all the parts for a chamber preheater that will go on the bottom. It is made from a bed heater, a stainless steel plate, a 7" copper short finned heat sink, and some bottom insulation. I've had the parts for a while, so I think it will be the next upgrade. It should knock down the 1.5 hour preheat time needed for hot prints. With the chosen materials and the power output, there should be no fire hazard even in a short. It could be left full on and not be a problem. It won't get hot enough to be a problem. I'll be testing of course.

I'm sort of hoping I can get the printer to where I can print the new print head from CF-PEKK. I have a 500g roll for the attempt. But we're not quite ready yet! And I'll probably have to print an intermediate head in PC or Nylon first. Seems a waste to print a head just to print a head, but that's the way it goes with bootstrapping.

I probably should have just saved my money and bought a commercial high temp printer, but oh well. If I get everything worked out, then the Voron community has an exemple of how to get it done DIY style. I really do intend to print at least smaller and perhaps medium parts in materials such as PEI (Ultem etc) and PEKK. PEEK will be out of reach. I'll not achieve enough chamber temp to prevent it from warping itself to smithereens. And I've no interest in extruding some plastic just to say I've extruded some plastic.

Gerald

Edited June 7, 2024 by G_T

[claudermilk]

8 hours ago, G_T said:

Since I was taking it off anyway, I decided to swap it instead. And since the Tap had to be disassembled including pulling the little Z carriage off the little rail, I want ahead and regreased that one. Warning - careful reassembly of that little tap z is required! Grease does help keep the balls in place, but vertical is a bad orientation to be working. I used a toothpick to reposition balls as necessary and reassembled very slowly. Then I wiped off the excess grease.

Didn't you print and use the little tool in the Tap repository for this? They have a part just for sliding the carriage on to so you don't lose your balls. https://github.com/VoronDesign/Voron-Tap/blob/main/STLs/MGN9_Assembly_Tool.stl

[G_T]

I've got one around somewhere. IIRC a short plastic piece of rail came with the LDO kit. It won't work with every brand of rail. There are dimensional variations. I forget offhand which brand of carriage won't go over the plastic rail, and one is overly loose. The MGN specs just seem to be treated as ballpark numbers by the various manufacturers when it comes to rail dimensions.

Right now the parts and boxes make an impressive disorganized pile. The tool is somewhere in there. I just wanted to get the job done! I should have mentioned the tool, or just using another section of rail, but generally I don't bother with it. Z0 and Z1 preload units are pretty easy if you have steady hands. It's good that you linked the STL. I remember the first time I took a carriage off a rail I was pretty nervous.

[G_T]

The printer is reassembled. Need to PID calibrate the bed, and check belt tension at elevated temps, but it should be able to print again. I've set the bed max at 145C since everything in the bed (now) is rated for at least that. There are also now no connectors in the wires in the chamber running from the bed.

I've started heat soaking it for breakin, initially sitting at 95C for a bit to remove any moisture, then I'll warm it up. I consider the practical printing max at the moment to be 125C due to a few remaining parts in the chamber that will not like it that hot or hotter. I have a little more printing and replacing to do! Bento Box, print head, and belt tensioners. And I still need to get rid of the X/Y echains.

Unrelated to all the work to make this a hot printer, it turns out I have either a bad or a loose upper bearing on my rear Z motor. There have been some noises from somewhere in the box over much of the Z travel but I couldn't locate it. Now I have. I can grab the bottom of the rear Z shaft and move it laterally. The other two are rock solid. Also leveling probing near that shaft seems to have a higher failed point rate, another sign of a problem.

Too bad these 350mm integrated teflon shaft LDO motors seem to be unobtainium at the moment.

Gerald

 

PS - I've taken the bed to 141C and no strange smells or smoke. Actually I think it smells less than the old bed did at 120C. I dropped the bed to about 80C and had chamber at 60C for initial conditions; running a bed pid calibration for 125C now.

No, this isn't the way it is normally done. But conditions inside this printer are too far off normal to run the calibration like normal. If I started with a cold chamber for instance, but then later tried to print at 140C, the temperature would stay above 140C. It takes a lot less power with say a 90C chamber to run a 140C bed then it takes with a 50C chamber. This is probably not as much of an issue with a normal printer.

The same thing applies for calibrating the hot end. IMHO it should be done with the chamber in the neighborhood of printing conditions, not cold. For my printer, that means I'd calibrate the hot end with the chamber at no less than 75C which is below where I'm normally printing.

Edited June 9, 2024 by G_T

[G_T]

The new bed is working beautifully. The existing print heat, not so much so. Erratic thermistor behavior at times, rather evident in the temperature traces. A machine reboot often but not always solves it. Sometimes it just doesn't want to heat up to the temperature I'm printing at (305C for PC-CF) but other times it does perfectly. The problems only seem to occur once the printer has been run for a number of hours. Before that it does fine. I'm not sure what is really going on.

Rant On!

I do know I plan to replace every single part in the print head except the nozzle and the little Z rail for tap. With extreme prejudice. I'll probably keep the same optotap board since it has been working perfectly. Actually the CNC Tap has been working perfectly. I have no complaints with it at all, except the magnets might be a little stiff.

But the hot end, the heater, the thermistor, the extruder, the boards and wiring - all is going. I'm tired of dealing with this.

So I warm up the printer, and set print head to 150C. I make sure it stays nice and stable at that temp. Then 180C, and again make sure it is stable. Then I go to 305C and print. This is a 500C rated hot end so it isn't hot for it. But it is air cooled and has a small heat break (Mosquito) which is not ideal for these chamber temps. I go a little hotter than should be done with the air cooled variant of the Mosquito IMHO.

I'm trying to print the replacement hot end parts. The last print failed right at the end of the print, right as that spool was running out of filament. Can't win. It started skipping extruding - probably a touch of overheating of the extruder motor. Historically it might start doing that somewhere around 4 hours of being hot. That has made it so I basically print a part or two in a session since that's all I can get away with.

Re-running PID_CALIBRATE with the chamber at 68C and starting at 150C going to 280C. Hopefully this smooths out the temp behavior. It's like all the messing around I did with the bed made the head not track temperature as well as it used to. Not sure what's up with that.

Not ok, failing to calibrate for 280. Trying 260 from a colder extruder. I've had plenty of issues getting PID happy with the hot end in a warm chamber. The behavior isn't quite the same as for typical printers. Heatup is faster, cooldown is slower, due to ambient temp. Since Klipper HAS the ambient temp through a chamber thermistor, wouldn't it be nice if that were taken into account? PLA temps and fan running, expect rapid cooldown. PC temps and no fan, and expect rapid heating and slow cooling.

Heck, even just having some additional parameters to pass it clues would be useful:

PID_CALIBRATE HEATER=extruder TARGET=300 CHAMBER=75 FAN=OFF

That would provide enough info to tailor the expected heating and cooling behavior so PID_CALIBRATE wouldn't find behavior out of expected range just because its expectations are faulty.

It is REALLY ANNOYING that Klipper does a fail when PID_CALIBRATE doesn't go according to its expectations. It should just turn off the heater and check to see if temp is dropping after perhaps a second. If so, then there is no emergency, probably just a bad run. If not, THEN kill the power.

Anyway it looks like I have another thermistor that is misbehaving. This one only lasted a few months. PT-1000 from Slice this time. It's temp reading shot up from about 250C to over 400C in about 2 seconds so Klipper quit again. As a test I had PC-CF in there, just retracted a bit. It didn't ooze any. The temp was nowhere near that high. It wasn't high enough to perticularly melt it. And it didn't read all that high when Klipper came back up.

I've had it sitting with bed and hot end powered down for a while now. The hot end still reads 90C with the bed cooler than that, and the chamber notably cooler. It's fried.

... Looks like it's reading about 30C high when the printer is fairly cool.

Another one bites the dust. I guess I'll be rebuilding the Mosquito again. But I'll be out of printing for a few days. Hopefully I have another PT1000 around here somewhere. I'm not sure that I do.

I wish we had a 2nd thermistor in the print heads, so we could detect when one starts to fail. I wish we had extruder motor current as an output on moonraker. Actually, current from all the motors, the bed, and the hot end. Those things would let us see problems as they develop, perhaps before they start costing prints.

If an accelerator were present, then as part of print prep, give the gantry a good shake and look for changes in resonance indicating stuff is coming loose or out of adjustment.

I wish the hot end calibration included filament flow rate and heat requirements so it could be predictive rather than reactive. Then perhaps heater wouldn't cycle around so much in output even when it is otherwise doing a good job of holding the temp fairly stable. A sustained steady extrusion rate of filament fundamentally requires a constant power input. I've yet to see anything resembling that, at least not on this printer. Power supplied to the hot end looks more like a pinball game under even fairly steady-state conditions.

End Rant!

Well, at least this time I didn't find it stirring a polycarbonate pudding puddle on the bed, like last time a thermistor failed.

Anyway, on the plus side, it seems a good part of the stinky from the printer was from the old bed. The printer doesn't stink the same way any more. It's a nice improvement.

Gerald

 

[G_T]

PS - The extreme prejudice part... I should explain.

Looking at the print statistics, the failure rate with the current hot end for longer prints is making my effective filament cost about $250/Kg. It should be around half that for the filament I've been using, 3DXTech CarbonX PC-CF.

Everything except the print head seems to be working perfectly and reliably in the temperature range I expect to be printing most of the time (nylons and polycarbonates, the higher temp variants). Until the print head can take the heat, I can't test at higher temperatures.

70C or hotter, 4 hours or longer, seems to be the failure numbers for my Stealthburner. Around 78.5C chamber is the highest I've printed so far. About 330C hot end is the highest I've printed so far. I do have the shroud removed which did seem to help the numbers a little. Before that the failure rate was higher since the shroud traps heat and is otherwise useless. IMHO the Stealthburner is too much about cool factor and not cool performance. Sorry; that's just my opinion.

Note, my hot end is not mounted in ABS or ASA. Never try the temps I'm running without at least replacing those critical parts with something that can take a bit more heat.

So, excepting the print head, the printer now seems very reliable at below 80C chamber temp. I've not been having to make any adjustments. The printer, it just works. And it's nice and quiet, and seals the fumes in pretty well.

The print head, it just fails.

Gerald

Edited June 11, 2024 by G_T

[G_T]

I put in a new Pt1000, but it made no difference.

Initially it read room temp about 20C high, just like the one I removed was doing in the end. But otherwise the temperature readings were pretty stable. So I stairstepped temperature while warming the printer up. Then PID calibrated. The temperature trace didn't look quite right but it did generate some calibration values.

So I loaded some filament. By now the printer had been running for a couple hours. Crash after crash. The second pic is a trace of the cooldown, in the end with the door open. It is clear something is going seriously wrong with the temperature readings from the hot end.

This was with the print head stationary. So it is rather unlikely to be a connection or wire. I jiggled the wires anyway and that didn't make a difference that I could tell.

So my guess is something flaky with the octopus on that temperature port. Why that one and not any other temperature sensor, I have no idea. I'll have to change it around and see what happens. But not tonight.

If people have had this sort of behavior before I'd be curious to hear about it.

Gerald

[TitusADuxass]

On 5/4/2024 at 10:28 PM, G_T said:

I had to do a lot of fiddling to get this to work acceptably, but the roll of filament can now be inside the printer. I had issues with interference of the reverse bowden tube and the loop of wire coming out the back of the print head. I had interference with the upper loop of tube and the light strip on the left.

I came up with a compromise configuration that has the tube pushing against the extrusion just above the left light fixture when the head is in the back left corner, and a bit of extra tube between the vertical guides and the top pivoting guide so the tubing moves to the back and out of the way as the head passes across the middle of the bed when all the way back.

I want a similar spool holder for my Vz 330. 

Have you had issues with the filament tension?

[G_T]

No it's worked pretty well so far. The very end of a roll of very stiff filament could be an issue since it wants to expand off the reel so much and could snag itself. But since this arrangement doesn't use any bearings, just sort of slides the roll over a piece of teflon bowden tubing, it keeps some tension. It seems to be enough.

I like the arrangement and so far haven't seen any reason to modify it. It does require parts from a few different projects though!

The horizontal arm reel holding part is the typical Voron rear spool holder without mods. I just took it off the back of the printer and reused it.

The top pivoting support for the bowden tube is the Voron Trident Support by Exerqtor: https://www.printables.com/model/460608-voron-trident-support

I'm using the Super Slim LED Bar with Simple Bracket for light and retention. I did widen the two at the end quite a bit. I rather like this lighting solution and these brackets protect the lights from the bowden tube when it would otherwise rub them when the print head is in one region of the bed:

I use the Trident Simple Internal Spool Holder - but just the spool holder not the Bowden Guide:

And I'm using the two bowden tube guides from this internal mount project:

So it's a hodge-podge of parts but it works very well. I didn't want the reel on bearings because I'm printing fiber reinforced filament which is pretty stiff. It would birds-nest, almost guaranteed. But the sliding action of the system I'm using keeps it nicely under control. The reel may not "roll" smoothly like it would on bearings, but I see no artifacts in the prints. I like it and am not inclined to change it.

You do have to have the bed up near the nozzle to get the roll in or out of the printer. You can probably also put in or remove a spool with the bed all the way down but honestly I haven't tried. I prefer to work under the hot bed than over it, since I run the bed hot enough to give a person burns.

I do not store spools in the printer. When I'm done for the printing session I pull out the toasty spool of filament and pit it into a vacuum chamber. It stays there until the next print session. BTW, this vacuum chamber is working extremely well for me, though it only holds 3 rolls. It is a version with a glass lid. My vacuum pump is an old Gast pump I've had for years. https://www.amazon.com/gp/product/B08PSFNZ1V?psc=1

Some of the parts from that vacuum chamber are not the highest quality but not the lowest either. Surprisingly, it will hold a vacuum for weeks. I'm not kidding. The needle hardly moves over a week. I got this thing when I started to deal with nylon but I no longer worry about damp filament. I do heat the spool of filament the first time before it goes into the vacuum.

I hope this helps you with your project!

Gerald

[G_T]

Things took a turn for the seriously weird.

Gerald

[G_T]

So from the above thread, the neopixels and their wiring is now gone.

Gerald

[G_T]

Since I'm trying to track down sensor problems, I went ahead and added a couple more temperature sensors for Klipper to monitor:

In printer.cfg:

 

 

[mcu]
##  Obtain definition by "ls -l /dev/serial/by-id/" then unplug to verify
##--------------------------------------------------------------------
#serial: /dev/serial/by-id/{REPLACE WITH YOUR SERIAL}
serial: /dev/serial/by-id/usb-Klipper_stm32f446xx_080038001850535556323420-if00
restart_method: command
##--------------------------------------------------------------------

[temperature_sensor Octopus]
sensor_type: temperature_mcu
min_temp: 0
max_temp: 100

[temperature_sensor raspberry_pi_4B]
sensor_type: temperature_host
min_temp: 10
max_temp: 100

 

Since I'm running Noctura fans for electronics cooling, this will let me see whether that is sufficient on long hot print jobs. The electronics bay has plenty of room for air flow so I expect it is fine. But this way I can be sure. Once I know what the temperature ranges look like during long duration load, it might be cool to rig up a macro to perhaps alter the chamber lighting to indicate potential electronics overheat.

Gerald