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

[G_T]

 

I need to provide a real build log here when I get the chance (have plenty of pics), but continuing from

I printed a part in PC-CF with 120C bed 300C extrusion and chamber hovering a hair under 70C. It was a nozzle scrubber to take a little more heat. The part came out perfect. But, the part wouldn't quite fit due to kinematic mount, so I modified the part to print it again.

This is getting into the lower end of the range of materials I want to be printing. So I was quite pleased. I also guessed the numbers rather than making any calibration prints. Sometimes it's better to be lucky than to be good! I gave the part a good stress test. It is quite stiff and solid.

The second print showed intermittent layers where underextrusion appeared to be happening in short stretches every few layers, and then the printer went into air printing. I cleared the clog and went at it again, this time increasing to 305C and slightly decreasing the already slow max print speed. It went into air printing without warning after the first few layers.

What I think is happening is the hot end cooling fan is not handling these temperatures very well and starts to slow down, and then heat creep progresses past the heat break terminating the print. The slow printing speed is helping the filament transfer heat. Although it is also possible it is the extruder motor which is having the issue. Or both. I don't think it is filament slipping. It doesn't take much force to push the filament at these speeds and temps (hand push test) and the extruder seems to be gripping it just fine.

Of course I tried to clear the filament path by running the temp higher and manually pushing filament with more force... That has worked before, but not this time. Now it is seriously clogged! I might have to replace that heat break and I don't have a spare.

It was funny - I could manage to slowly ooze mostly clear PC out the nozzle, but the carbon fiber was left in the heat break and nozzle! I'm guessing it is at least 40%CF up there now, probably quite a bit higher. I'm resisting the temptation to reassemble without the nozzle or cooling fan, (actually just dangle the mosquito) and run the temp up higher to see if I can drip the massive clog out.

Gerald

 

[G_T]

I began this project in November. It's hard to believe it has taken this long and I still have a ways to go. It started life as an LDO 300 cube, but with corner braces added (a bit of remachining of braces available off Amazon), a 6mm polycarbonate front door, and https://www.amazon.com/dp/B0CFW3NZ84?psc=1 for most all the other panels.

This is what can happen to the standard POM Z nuts when running at somewhat elevated temperature when grease is present. The plastic was essentially slowly dissolving, and Z moves were getting very squealy. Z artifacts were increasing. This didn't start until the chamber temp was over 60C roughly, perhaps 65C. Above 65C the parts deteriorated very rapidly. Bed leveling would sometimes fail. So I machined drop-in replacements out of PEEK. These work better.

Delrin (POM) is softer so it absorbs vibration better. So I did have to reduce Z rapids or the rods would go into resonance. I should just clog them up with POM dust and grease to dampen the resonance! It won't hurt the PEEK. Anyway having slower Z rapids is a non-issue from my perspective. So little of the print time is spent doing Z rapids...

The printer picture is not current but it will give an idea of the beast. Mostly I've just been upgrading parts to higher temperature parts as I've found or anticipated the need. There are lots of updates to come, but I need to print some of the parts first. I'm intending to replace ABS anywhere it might get any load or be close to the bed, with PC-CF or aluminum. Some parts are already replaced with PC-ABS as an intermediate material. I may have to go back to that approach rather than jumping straight to PC-CF.

I attached a pic of one of the sides before I drilled the mounting holes. You can see I had to contour the thickness of the insulation to provide room for the belts and the Z motion. The panels are working well, but I should have probably used a little more insulation. But if I had, then I'd be having more thermal issues at this point. I can always improve the insulation later once the thermal issues are all worked out. That's if I even need to.

The front door is funny. Polycarbonate has a decent coefficient of thermal expansion. Just having the view port in the front door means that part tries to warp when hot, causing the door to hold itself somewhat open. I added a miniature bungee-like strap to hold the door closed. I'll probably end up adding another layer of insulation to the door and reducing the viewport size.

Insulation is pretty simple in this one. I tiled the panels on the inside with 4" x 1mm self-adhesive cork squares, and then added thin metalized fiberglass on top. Amazon again - https://www.amazon.com/gp/product/B08M6BQ1HT and https://www.amazon.com/gp/product/B08M6BQ1HT I should perhaps have used slightly thicker insulation. TBD. I can fix that later if I have to.

Bootstrapping a printer...

While I'm getting the hot end working again, I'm also going to be replacing the lights. Those stick-on LED strips don't stick when you get up around 70C. I'm replacing them with light bars. We'll see how those handle the heat.

Lots of other changes are in store for this one. Unless I give up!

Gerald

[claudermilk]

Hmm, now you got me thinking. My Trident is getting squealing Z on large Z movements (like when parking). I'm on my original POM nuts, which are now at (<checks history>) a bit over 3000 hours. I wonder if they are wearing out. 

[G_T]

It is possible. The squealing can be accompanied by resonance in the threaded rods. You can see them vibrate rather than just rotate on axis. So it can also be wear in the stepper bearings, or lead to premature wear of the bearings.

What you can do is grab each Z mount and see if you think the slop is excessive in any of them. Replacing the nuts is a minor pain. If you are not having trouble getting consistent Z motion you may still be ok. You can test that directly with TAP, or set a paper's width off a nozzle, move down some distance then move back. Check fit of paper... Try different distances... If it isn't consistent, it's time for some maintenance IMHO! If it is consistent, you could just slow down your max Z speed until the sound disappears.

Having spares on hand though might be a good thing.

Looks like I might have trashed my Mosquito. I cannot get the heat break out of the heater block. It is fused. And to add to the fun, my order of EPDM 6mm Gates belts seems to have been stolen before I got home from work. Just not my day.

Gerald

[claudermilk]

Well, for one I have WobbleX so the lead screws are free to wobble around as they need. I'm also playing with Orca 2.0 and running the 0.25 nozzle & 0.1 layers; I'm currently getting absolutely gorgeous first layers on print jobs. The squealing is occasional during big, fast Z moves, never during a print. I did find I have a full set of POM (and brass) nuts in my parts bins, so I'm all set if an when they are actually a problem. I'll leave it for now--I have two working printers again after a couple of weeks of struggle so I just want to print for a bit.

[G_T]

FWIW, I have a small collection of brass and pom nuts as well. The quality varies. One set of the brass nuts has the thread tilted off axis. Those would be unusable. So check quality before use!

I hit another interesting issue. I've been working towards a rebuild step, replacing more of the ABS parts with higher temp parts - aluminum, polycarbonate, high temp nylon... Along with the update I intended to replace the Gates belts with Gates EPDM higher temp belts.

I've bought these belts now three times, from different suppliers. I have reason to think they are genuine. They are also IMHO not suitable for our use. They appear to be good quality, but here's a test to perform on any belts you want to use:

Take two lengths of belt, and intermesh the teeth. Keep hold of one place or clamp them together at a spot. Then run your fingers down the pair of belts keeping them meshed. Once you get to moderate length I'm finding large error.

This is going to result in oddities of the motion, since a given rotation of a stepper will result in inconsistent length being moved, courtesy of the pinion being splined. We rely on belt consistency and the Gates EPDM belts don't seem to have it. That's why I have the regular belts in the printer at present. It may be luck of the draw. The first batch of regular belts was quite a bit more consistent so that's what's in there. None of the EPDM Gates have been usable.

Now if the glass filled high temp nylon will arrive, and the replacement heat break for my mosquito, I can get back to printing parts.

I did repace the lights, but haven't really tested it much yet.

I also splurged and picked up a UHT Revo (the 400C High Precision model not the 500C version). I doubt I'll have real use for going above 400C. I'll never get the chamber hot enough to warrent using such filament!

I can't use this hot end yet since I only have an ABS mount for the Revo Voron. That's something I plan to print out of the GF high temp nylon. I don't want PC there since PC burns too well once lit.

But I'm not really sure I want to stay with Stealthburner anyway. For me it's too much form over function. I don't need a decorative faceplate. I don't need voron pattern lights in the front. I don't need parts lights - my chamber lighting is plenty bright if I want it to be. I'd prefer smaller printed parts, and less weight.

I'd like something like a WristWatch or G2SA (printed in CF PEKK or other) or LDO Orbiter 2 bolted on top an aluminum combo fan mount and heatsink for the heat break. Put a fan on the heatsink. Add a second fan with duct for parts cooling, but perhaps make that assembly easily removable. Generally I won't be needing as much parts cooling, as I don't plan to be running all that much PLA and the higher temp filaments don't seem to need much if any cooling for engineering parts (avoiding bridges and overhangs). Then I want this assembly bolted onto a Vitali Tap. I'm running that tap version now and it is rock solid and reliable. There's no play when it is down.

If I still had my Bridgeport vertical mill I'd have already made it.

Gerald

[G_T]

New record for chamber temperature, peaked at 76.5C and stayed close for a few hours printing. Unfortunately I had another nozzle clog at about 4 1/2 hours in. It seems the 0.4 nozzles clog about every 16 hours so far when printing 3DXTech PC+CF. Hot end 305C, bed 120C.

It's being a race at the moment. I'm printing hot enough that ABS fumes are coming out of the printer, in spite of it being pretty well sealed, and in spite of not printing ABS. The ABS parts in there are outgassing. I'm pushing it really near the limit, printing some parts that will allow me to take it hotter.

The batch I'm working on is A/B motor mounts with pin mod and upper bearing mod. I have some long shaft steppers that I'll be switching to. The bearing mod puts a support bearing in the upper parts so the shaft is fully supported.

These PC+CF parts aren't going to win any beauty contest, but they are strong. I couldn't break a flow test tab in my hands. Layer adhesion is quite good, probably due to printing slowly at good temps.

When I swap these in I'll also be swapping in an all metal gantry. That leaves the tensioners. As much as I'd like to print them in PC-CF that would be no good for sliding parts. I'll see what I can do with PA12-GF but I expect the same result. I'll probably end up reprinting the RamaLama-II in 3DXMax PC.

Sorry the short video is upside-down. I've done very little video editing before.

There is a temperature gradient in the printer, the top being hotter than the bottom. So I'm replacing parts top-down at this point.

I think I mentioned it before, but the first thing I had to replace was the POM nuts with ones I machined out of PEEK. The POM nuts were sort of dissolving away when I went above 70C. The holder for the Mosquito is high temp nylon. I have the hatch removed from the Stealthburner for a bit better cooling. The bed mounts are kinematic which helps given the wide temperature range involved.

The magnet sheet is down to at best half its original strength. When I get all the ABS parts replaced then ramp up the temps, that magnet is dead. I'm planning on seeing if things will be stable at 90C chamber temp. I'm not really planning on going above that.

The last pic was during construction.

Gerald

Edited April 15 by G_T

[G_T]

Tonight I was trying to print some replacement parts in PC-CF instead of the original ABS. After a couple instances of air prints and a couple firmware shutdowns, I switched from a 0.4 nozzle to a 0.6 and got a small part printing. Once it built up about 1/8" thickness I figured out what is going on.

I'm thinking the thermistor is failing. I have no idea what temperature I was really printing at, and neither does the firmware. 40% infill high carbon fiber content polycarbonate, and it made a textured pond. I'm thinking the intended 305C was probably more like 400C.

I had struggles getting the A and B motor mounts reprinted and the quality wasn't what I would have expected. But, the parts are very strong and everything bolts together ok. I finished printing those parts a couple days ago. They are mock-assembled at the moment. These are pin mod, with the top bearing mod (a bearing is added at the top of the output shaft). The new motors are LDO's Super Power HT motors with long output shafts. The shafts are about an inch longer than I want so I'll cut them down a bit before installing in the printer.

The PC-CF parts appear to be roughly 2x the stiffness of the ABS parts and I couldn't break them in my hands. I have zero concerns about these failing from either heat or stress.

Gerald

[G_T]

The printer is now torn down for rebuilding. I've replaced the A and B drives with the new ones mentioned above. You can see the upper stepper bearing shaft support, and the pins for bearing shafts. They went in fine. I'm not 100% sure of the required pinout with the motor change though. TBD which way they turn for instance! I tried to test with just this update, but without a thermistor in there, Klipper shuts down quite quickly.

Since I was messing with the wires I took the opportunity to get rid of a lot of extra wire length. The new A and B wires are made to length.

The teardown picture shows the corner braces pretty well. They add a great deal of stiffness to the structure, and of course the panels add more since they are bolted on rather than clipped on. It is a very rigid printer.

After these pictures were taken I started in on the gantry. When I removed the belts I found the gantry had been reduced to a wet noodle. It could be racked half an inch with hardly any force. It is amazing it would still print usable parts, though at only 50mm/s max speed. That's probably why I had to run it so slowly to get usable prints.

I expect all the ABS holes had gotten ovalized, and at the chamber temperatures I've been running none of the bolts would stay all that tight.

The new mounts in PC-CF were printed with 5 walls and 5 top and bottom instead of the usual 4. On the Trident they are not moving parts, so overbuilding just costs a bit more time and plastic. I put a pic in of one of the pieces being printed.

The last pic is what a part looks like when a thermistor fails. My own fault really. It was only rated for 300 and I've used it up to 330C. It's days were probably numbered. There were more parts I was planning on replacing this round, but since I could no longer print... Teardown time!

The gantry is being replaced with an all metal gantry - the one by Common Anomaly. More about that later. For my purposes it looks like a huge improvement.

Gerald

[G_T]

The gantry has progressed...

With the old ABS and extrusion gantry, it could be racked a half inch with very little force. About half of that was due to accumulated deformation of the ABS parts and the inability to keep bolts tight.

The new all metal gantry racks about a mm for the same force. It is otherwise as smooth, so I consider it a rather large improvement. It should be impervious to the temperature issues I was having with ABS.

I had decided to use this gantry when I was sourcing for this build. However initially I went with the stock gantry design. I figured to get it going, then change later.

A big reason is I wanted to go umbilical rather than e-chain. This new gantry is not designed to work with e-chain. I jury-rigged it using some 30mm M5 bolts, some washers, and some nuts. The underside of the e-chain on the gantry is stabilized with some double sided tape. Hopefully this works for a little while at least. I still plan to change to umbilical but I need to print some parts first.

I also wanted to show that e-chain could be done with this gantry, though not necessarily that it should be done!

Belts are the next step. More about that later.

Gerald

[G_T]

Well, I debated and ended up replacing the Y rails which were pretty good for consumer rails, with some genuine Z1 preload HIWIN rails (still consumer rails, but better QC and a touch smoother machining). I replaced them one at a time, using one side to align the other. I have done this before with larger machine tools and much beefier rails. However I should have flipped the printer upside-down to make it less awkward!

The motion is now even smoother and the rack resistance is even slightly higher than before - courtesy of Z1 preload. The previous rails I would have called Z0.5 if such a designation existed. They were pretty decent rails. These are just slightly better. It's also a touch quieter when being pushed around by hand, probably due to fresh grease.

The - now - old Y rails were the best of my MGN9 rails when I did the initial assembly. I'm considering using them to replace the Z rails in the front. I'm not sure it's worth it.

But if I were to replace them... though it is really not likely to be worth it, I might just get 3 more of the HIWIN Z1 preload rails for the Z. I will be rebuilding the Z once I've reprinted the parts.

BTW, I'm trying Slip 2000 EWG for lube this time around. It's good for high temp and harsh environments, well beyond what a printer should experience.

Gerald

[Buurman]

4 hours ago, G_T said:

The gantry has progressed...

With the old ABS and extrusion gantry, it could be racked a half inch with very little force. About half of that was due to accumulated deformation of the ABS parts and the inability to keep bolts tight.

The new all metal gantry racks about a mm for the same force. It is otherwise as smooth, so I consider it a rather large improvement. It should be impervious to the temperature issues I was having with ABS.

I had decided to use this gantry when I was sourcing for this build. However initially I went with the stock gantry design. I figured to get it going, then change later.

A big reason is I wanted to go umbilical rather than e-chain. This new gantry is not designed to work with e-chain. I jury-rigged it using some 30mm M5 bolts, some washers, and some nuts. The underside of the e-chain on the gantry is stabilized with some double sided tape. Hopefully this works for a little while at least. I still plan to change to umbilical but I need to print some parts first.

I also wanted to show that e-chain could be done with this gantry, though not necessarily that it should be done!

Belts are the next step. More about that later.

Gerald

 

What is the Y gantry piece you have there, looks great!

[G_T]

Common Anomaly Ultralight X-Beam with the aluminum X-Y joints. Ti backer on the X beam, on the back side, to reduce thermal induced bending. The Ti backer is specific to that gantry since the dimensions are a bit different than would be the case if it were an extrusion.

I did not weigh it, but the new gantry is a little lighter than the original. It is certainly prettier!

I also used the bearing mod for longevity, and used Fushi ABEC-9 bearings that I've been pleased with, along with some precision shims. The carriage is CNC Voron Tap by Vitalli, which you can get off etsy and probably other places. All these parts are very nice.

Or do you mean something else? I'll try to provide whatever details you are looking for!

 

Different subject - belts. I have a big issue with the quality of most of the "gates" belts I've been able to obtain. I've bought "genuine" gates belts from several different reputable sources, both the regular belts and EPDM. I want to use the EPDM belts in this printer since they are a little more tolerant of higher temperatures.

I have a simple test for belts. Intermesh teeth on two lengths. Squeeze between fingers so the teeth stay intermeshed, and draw out a length of belt. You'll almost universally find that one side ends up notably longer than the other. If used in a core-xy printer one diagonal will move a different distance than the other, if given the same stepper shaft rotation. I'm not sure how to compensate one diagonal length in clipper vs the other diagonal length. Turning parts 45 degrees in some cases and scaling could work for compensating in the slicer.

To make it more fun, the absolute moves would depend on where the location is on the bed. One would need a travel scaling mesh much like we already have a bed mesh.

Of the several sets of EPDM belts I have, some have larger text, some smaller. Otherwise they look the same. However, one belt length has a slightly different feel in the hand. It feels more rubbery when sliding between fingers. That length of belt lines up very nicely with itself when I perform the intermesh test. So of course that's the set that will get installed.

Too bad I have only enough for one pair of belts and I don't know who I got it from!

I might have as much as 20m of belt here that fail this simple test. All obtained from different places, all claiming to be genuine Gates.

I wonder if only that one length of belt in my collection is genuine gates, or if one of the available knockoffs is actually of higher quality?

Gerald

[claudermilk]

Interesting. That's not been my experience with Gates belts. I did the interlock teeth trick to make sure my A & B belts were the same length for both the 250 Trident and the V0.2. Both interlocked perfectly the entire length. The V0 was whatever LDO includes in their kit--I presume genuine Gates. I believe I bought the belt for the Trident from West3D. Sounds like either bad pieces or not truly genuine in your case. 

[G_T]

The belts are in, but right now tensioned possibly a hair less than what I'll use eventually. I want to set the tension with the printer hot anyway and I'm not ready to close it back up just yet.

Since I replaced the thermistor with a new PT1000 I made an initial pid calibration but at only 200C. The results would be way off anyway calibrated to 300C without the enclosure and preheating to 70C chamber. I'll recalibrate when it is all reassembled.

It is alive! Somehow the motor wiring ended up correct and the motors even turn in the correct directions.

Sensorless homing needs some work though. The settings are going to end up nowhere near the old settings due to the motor change. I finally got it to go to the limit but it slips the belt after that. Too much torque and it's not stalling so there is no stall current to detect... In other words, the settings are still way off.

But it moves, and it is quieter than before the motor swap. Anyway the loudest thing in the printer by far is the pair of fans cooling the electronics area. If anyone has any suggestions for good quiet replacement fans I'm all ears!

Gerald

[claudermilk]

Nice.

For the electronics bay fans, I swapped to Noctuas (and switched the ports on the Octopus to 12V after briefly wondering why they were running so hard).

[G_T]

I just ordered a pair of the Noctuas... wish they were 24v but oh well.

The motion now makes less noise with the panels off that it used to make with the panels on.

I've spent a few hours now messing with basic tuning. Sensorless is being more of an issue with the new motors than the old ones. There is a weird behavior when the head bumps an axis and then pulls away. There is a couple mm diagonal in the motion. I've managed to reduce it to that level but not eliminate it so far. It's usable anyway and seems to reliably home. I'll mess with the settings for a while and see if I can get it better, but not tonight.

I've run several iterations of bed mesh now. I hadn't run one in months. It seems the bed has slightly deteriorated since it was new.

When it was new it registered around 0.12mm out of flat. Now it is more like 0.216mm out of flat. Mostly it's falloff to the back left.

The biggest problem is the shape of the bed resulting in the three probe points I'm using for leveling being less than ideal choices. I wonder if it makes sense to choose oddball points for bed leveling, rather than ones near the supported corners of the bed. If I did that I could flatten the bed mesh a bit.

The main culpret is a bit of sharp falloff right at the front right corner. That gives a false three point flattening. If I moved that probe point in a couple cm it might improve the resulting bed overall paralellism with the rails.

But really I probably should consider replacing the bed. I've never been satisfied with the shape of the mesh. Having an overall error on the order of a layer's thickness just seems to me to be a bit much. No matter what I do I can only alter the tilt. I have replaced the X and both Y rails with slightly better rails and it didn't make all that much of a change in the bed profile. Though, it is slightly smoother than it used to be so that's something.

I did one round of probing without a top sheet and it didn't make much difference - other than leaving a bunch of little dimples in the magnet sheet.

I suppose if I removed the magnet sheet I could hand flatten the bed with a surface plate and some elbow grease.

Around the beginning of the year I picked up another bed. I could easily see how out of flat it was with a straightedge. Honestly I generally got flatter plate when it wasn't machined. These things are not what I'd call well machined. If I still had my vertical mill I'd fix them.

Does anyone make an honestly good flat bed for a 300mm trident at this time?

So I tried relocating the probe points a bit to see what would result. After numerous attempts it came out with .171mm. Better, but not good. It does look like the two Y rails are slightly out of parallel now after the rebuild. But fixing it won't improve the overall mesh error by much. I'll give it a shot though.

Have to break out the big calipers to see where it is best to make the change. I tried simulating it by putting a block under one of the feet but even with the panels off this thing is too rigid. It made no difference.

Gerald

[G_T]

I flew the white flag.

I've spent about 8 hours fiddling with deliberately unsquaring the y rails to compensate for the bed twist. I can improve the apparent flatness that way a little, but there is nothing I can do for the wave shape of the bed mesh. I've used two sets of linear rails for X and Y and not made much improvement, though the curves are a little smoother with the HIWIN Z1 rails - possibly just because they are Z1 not Z0. Right now I'm leaving the y rails .08mm out of square vertically to introduce a twist in the motion as compensation. This gives 0.166mm apparent wave to the bed which is the best I've gotten it recently. I can't make it any better.

In any event all these cycles of homing and bed mesh have given a little breakin period for the new belt, motors, pulleys... and new settings for sensorless homing. The machine is working pretty well though sensorless Y pulls off the back on a bit of a diagonal. So there is room for a bit more tuning. Nevertheless it is functional, and moves nice and sooth/quiet. Did I mention quiet?

Noctura fans should be here tomorrow then I can get rid of the whiny electronics cooling fans. I'm really looking forward to making that change!

I had another bed here, that didn't pass the straightedge test though it was touted as a good bed by the seller. It certainly had a smooth surface finish - which is usually a sign of some mass finishing process applied post-machining. That makes it look good, but lose precision. In this case the bed was out of flat by a little under 0.4mm. Ugly. It is now in the scrap metal pile for whenever I need a piece of aluminum plate. Expensive scrap.

I ordered a Mandala Rose Works ultraflat magbed. Hopefully that lives up to its specs. I'd be happy with 0.1mm total error. I'd also be happy with being able to get rid of the deliberate twist I've put into the machine to compensate for the current bed.

The magbed will get a full size 750w heater and a 150C thermal fuse. My target max is 140C bed which is slightly under where the magnets should start to die.

Gerald

[claudermilk]

I have a standard MRW bed on my Trident 250 and it's plenty flat. An old mesh still saved shows 0.07 total range. Of course with tap and KAMP, I don't really care--I get perfect first layers on this printer almost all the time. Any issues are not from bed flatness (or lack thereof).

[G_T]

Mine is an LDO bed which I understand are generally pretty good. It isn't real bad but not quite as good as I'd like. However when I run the temp up to 120C and the chamber heats up, it does get a little worse. It's on a Kinematic mount... It keeps exactly the same shape just has a little more deviation.

I am having one weird thing happen though, and it is not completely consistent. Bed leveling doesn't look exactly level in the bed mesh. It is always off the same direction a little bit, but not always exactly the same amount. That didn't used to be the case before the recent printer reconstruction. It looks like two of the three probe points are level but the third it might be triggering as much as approaching 0.05mm early, resulting in a somewhat tilted bed mesh if a new run is made.

I've spent a bit of time now with the printer hot, and running it in after this round of rebuild. Everything seems to be working fine except for that slight bed tilt issue. We'll see if it persists, or if I can figure out a reason for it.

Calibration... Since I replaced the thermistor I needed to repeat PID_CALIBRATE for the heater. I don't think I've ever gotten it to calibrate at 300C or higher which is where I'm often printing. I had to calibrate for a lower temperature, and increase the separation from the bed (at 120C). Is there some way to get the algorithm to throw away its assumptions (PLA to ABS range) and just get the coefficients? It's annoying that it just shuts down the firmware.

I suppose I could make a bunch of calibration runs at a range of chamber and heater temperature that were low enough to make it happy, then extrapolate. I'm not sure I want to go through the work.

I put the Noctua fans in, but not as variable voltage. Just full voltage using 2 wires... And forgot that these are 12V fans not 24V! I'll have to open the electronics bay back up and change some jumpers. I was wondering why they were basically as loud as the original fans, though they did have a more pleasant sound. Lucky I didn't fry them.

Gerald

[claudermilk]

10 hours ago, G_T said:

I put the Noctua fans in, but not as variable voltage. Just full voltage using 2 wires... And forgot that these are 12V fans not 24V! I'll have to open the electronics bay back up and change some jumpers. I was wondering why they were basically as loud as the original fans, though they did have a more pleasant sound. Lucky I didn't fry them.

Gerald

 I did the exact same thing. I did kind of try to warn you.

I would expect any LDO parts to be high quality, so you are probably fine with that bed. Running it at 120 is pretty much maxing out the printer capabilities/

[G_T]

Yep, felt pretty silly when I realized what I'd done! And I remembered your warning... afterwards of course!

Nah, this printer will go a fair bit higher by the time I'm done! I plan to run the bed up to 140C for some filament types, and the chamber to perhaps 90C. I'm figuring on adding extra cooling for the extruder motor, and possibly even an active chiller for it if I decide to go even higher. The internals already use a lot of metal, with CF-polycarbonate and GF-nylon for some parts. The hot end I'm planning on reprinting with something that will go higher than that. CF PEKK is one that I expect to attempt for those parts, assuming the roll ever arrives.

I've switched the jumper for the electronics fans.

I decided to go ahead and add the next layer of insulation to the top at this time. That will bring it up to 3 layers of insulation plus the panel. But I'll also be trying to set up something in Klipper to use the exhaust fan as a thermal limit for chamber temperature. I don't want it to be going much over 75C until all ABS is removed from the chamber. I've printed above that but that's really pushing the limits at this time.

I've also got the parts to build a chamber preheater that is going to go on the bottom. Heavy stainless bottom plate on top of a quarter inch layer of bed thermal mat, with a 300W bed heater and a copper heat sink. It's about 7 or 8 inches square. TBD on electronics control at this point. It's not a lot of heating power but should cut down the preheat time by a fair bit. It's low enough power I can use it as a warmer and keep the printer at a modest temp inside, perhaps 50C, and then power the rest of the printer up when ready to print. It should also allow not running the print bed quite as high to achieve the desired chamber temperature. Less thermal gradient is a good thing.

Gerald

[claudermilk]

 140 for the bed? I assume you've swapped the thermal fuse on the bed to an appropriate one for that temperature. What are you planning on printing that requires that kind of heat (I don't recall that being stated)?

[G_T]

I'll be installing a 150C fuse when I swap the bed out, or possibly 135C until I get the rest of the thermally vulnerable parts out of the chamber.. I'm guessing that will be in about a week. Right now it's just a 125C fuse which is fine for the moment printing mostly 3DXTech PC-CF.

Materials - Filaments like CF-PEKK, Ultem 9085, PSU, PPSU, PPS etc. Engineering plastic. Mostly specialty parts for rockets although I may offer some other parts on occasion. Perhaps self-extinguishing high temp hotend mounts for those who want to go a fair bit hotter than ABS.

Gerald

[claudermilk]

Impressive. So (inside joke warning), just make sure the front doesn't fall off...