How Do I Calibrate My Titan Extruder and Test Print?
-
During a print, the extruder speed is indeed relative to head movement. What Tony was referring was the fact that you said
"putting the test print aside, simply extruding 100mm of filament shows it's being fired out at speed, and only 80/100mm at best."
So when you do that - i.e. extrude filament via the web interface, it will come out at whatever speed you selected in the web interface. Generally, anything above 5 mm/sec should only be used for unloading filament.
Edit. If you just use a search engine in put in "calibrating extruder", you'll find numerous examples and YouTube videos of how to do it. To get you somewhere in the right ball park, the Titan will need in the order of 400 steps/mm using 1.8 degree motors and 16 X micro stepping. If you use 0.9 degree motors, you'd need to double it to around 800.
HTH
-
OK, I see.
I've done an extruder calibration before. The problem I'm having is that it's not showing the results as expected, which is why I'm checking on a DuetWiFi specific process.
For example, I just adjusted the extruder steps like so:
M350 X16 Y16 Z16 E16 I1 ; Configure microstepping with interpolation 16
M92 X160 Y160 Z1600 E420 ; Set steps per mm 16This extruded 80/100mm, so I changed it to this:
M350 X16 Y16 Z16 E16 I1 ; Configure microstepping with interpolation 16
M92 X160 Y160 Z1600 E380 ; Set steps per mm 16This extruded 82/100mm, so I changed it to this:
M350 X16 Y16 Z16 E16 I1 ; Configure microstepping with interpolation 16
M92 X160 Y160 Z1600 E200 ; Set steps per mm 16And it finally extruded 48/100mm.
I can't get it closer than 80% feed.
Can I steal someone's config.g who has a Titan? I need to compare my gcode I think.
-
I have 5 Titans (Diamond 5 colour hot end). With 1.8 degree motors, the calibrated steps per mm are
M92 X80 Y80 Z3200 E393:397:403:400:401 ; Set steps per mm
So between 393 and 401 on my machine. If you are running 0.9 degree steppers, then I'd expect your value to be double, so in the order of 800.
Edit. I'm concerned about the way you are calibrating the extruder. Something is very wrong. You start with 420 steps which you say gives 80mm. Then you drop it to 380 steps. That should give you (80/420*380) 72mm yet you got 82 mm. So something isn't right. Are you doing this with the extruder connected to the hot end and the hot end heated or do you have the extruder disconnected form the hot end? If it's the former, do the latter. Also, check your extruder tension - something may be slipping. Do you have a good mesh between the gears? Are you simply trying to do it at too high a speed? Do it at a slow feed rate like around 5 on the web interface. Don't use the same piece of filament over and over again. Once it has been through the extruder, the hobbed bolt will have made grooves which will affect the diameter of the filament if you try and use that same section again.
-
OK, I just took a closer look considering some of your pointers.
I found that the extruder filament cog spins halfway through the 100mm extrusion. I think the filament is backing up in the hotend and preventing the fed filament from being pushed in by the extruder.
The nozzle is a .2mm, I think. What could be causing the congestion? It's been heated to 210c before extrusion and is brand new. Is there an indicator for the nozzle size somewhere? Or is this another issue?
-
It sounds very much like a partial blockage and/or you are simply extruding too fast.
0.2mm is very small and rather specialised - designed for very small intricate parts that are to be printed slowly. Unless you really need very fine details, I'd suggest you go for a 0.4 or 0.5mm nozzle. Are you really sure it's 0.2mm? - that's unusual for a Volcano which is really intended to take 0.8mm upwards. Also, is it a Volcano nozzle and not a "normal" V6 one. I believe that they are not interchangeable and you must use Volcano nozzles with Volcano heat blocks. Wait…...........
..............I've just done a quick check on E3D's site and I can see that the smallest Volcano nozzle is 0.4mm and they go up to 1.2 mm, so your nozzle can't be a 0.2mm Volcano.
So to calibrate your extruder steps per mm, disconnect the extruder from the hot end, extrude 100mm and alter the steps per mm until you get 100mm. As I said before, to get you in the right ball park, start with 400 steps/mm (at 16X micro stepping) if you are using 1.8 degree steppers (200 steps per rev) or 800 steps/mm if you are using 0,9 degree steppers.
Once you have the steps per mm calibrated, then you can go on to investigate the cause of the extruder grinding away the filament, some possible reasons for that are detailed above. I would suggest that you do not alter the steps per mm again - all that is doing is trying to correct for a problem that lies elsewhere, rather than address the root cause of the problem.
HTH -
I meddled with the extruder steps for a while, and it calibrated fairly well a 440 steps/mm. I just used the 5mm extrude in the console, and counted 20 extrusions for a 99/100mm total.
I'm not sure what the deal is with the speed of the extruder. I've never had this problem with the Tevo factory extruder or hotends. It's a Volcano yes, and came with the Tevo Titan Volcano kit. Changing the extruder speed on the console down to 1mm/second still didn't seem to slow the 100mm extrusion down. It just continues to power through, causing congestion.
At least at this point I know the calibration is rightfully E440. It shouldn't congest on 100mm extrusion, like the earlier versions of this printer's setup. I have to fix it before I move onto slicer troubles. Cura gcode makes the Duet squeeze out an entire layer of filament in each corner of the test cube print, before moving to the next corner and doing the same thing.
-
Well the Cura gcode thing is likely because, as has been mentioned earlier, Cura uses absolute extrusion whilst the Duet is expecting it to be relative because there is an M83 command in your config.g file. So, as has also been mentioned, you need to put M82 in either the Cura start gcode or if you are always going to use Cura, put it in you config.g instead of M83.
I really can't understand why using 1mm/sec on the DWC does not slow the extrusion speed down. It works for everyone else.
As you have managed to extrude 100mm (by doing 20 x 5mm extrusions) with the hot end disconnected then it's a fair bet that the "congestion" is caused by a blocked or partially blocked nozzle or the temperature is too low to melt the filament don't you think? It's easy enough to check. Just remove the nozzle, enable cold extrude (can't remember the gcode for that - you'll have to look it up), then try extruding 100mm of filament. If that works, then the only thing you have taken out of the equation is the nozzle (and heater) so that's where the problem lies.
-
The start code has always included M83 because someone said that was relative, not absolute.
I changed it to M82 and it started printing the test cube in the correct configuration! It started with the brim, then the cube itself. A bunch of small things to do with the print quality need to be adjusted, but another problem occurred.
After the base layers were completed, it started extruding very slowly, which ran the filament so thin it stopped building up layers. The extrusion step setting is correct, and there's no blockage, so I don't know where this issue is coming from…
-
The start code has always included M83 because someone said that was relative, not absolute.
I changed it to M82 and it started printing the test cube in the correct configuration! It started with the brim, then the cube itself. A bunch of small things to do with the print quality need to be adjusted, but another problem occurred.
After the base layers were completed, it started extruding very slowly, which ran the filament so thin it stopped building up layers. The extrusion step setting is correct, and there's no blockage, so I don't know where this issue is coming from…
Progress - that's good.
M83 IS relative. BUT, Cura slices it using Absolute so you have to "tell" the Duet that the extruder coordinates are absolute by using M82.
Ref the last issue - that's a slicer setting somewhere. You really need to absolutely sure of you nozzle size before going much further.
-
Ah, right. Geez. 3D printing is definitely not my forte. I'm an animator and 3D designer. I've been on this technical trip because it gets me creating tangible stuff. And every time I see someone printing finer quality stuff I spent more time and money upgrading.
I took the nozzle off and it's engraved 0.4mm, but I did find Cura had been told the printer has a 0.2mm nozzle as well. I'm sure if that printer management setting finds its way into the print gcode though. What I do is save out the file from Cura and upload it to the Duet, and print it from there.
After 3 hours of running more tests, and having difficulty with over-tightening the extruder to the mount, and cleaning out the hotend, it still stops extruding the filament after the base layer of the test cube. It will test extrude plenty. However, after attempting to print, the filament jams, and the extruder gripping wheel just grinds away on the stuck filament.
I then release the filment clamp and attempt to pull the filament out by hand and it won't budge. The hotend is running a 210c, and feels pretty hot, so I don't get it. There should be nothing restricting the movement. It should be a molten pipe of PLA.
EDIT: I hit unload filament one last time, and it's perpetually 'loading' the filament, which of course is still stuck. So it's unloading in the wrong direction. As of me doing today's tests, somewhere along the way the extruder started going the wrong direction.
-
When you slice something, regardless of what slicer you use to do it, the resultant output (gcode file) will be a list of instructions to move the print head so many mm in X, Y (and Z on layer change). Most of these move commands will also include movement of the extruder so that nnn mm of filament gets extruded while it is moving the print head. Now the Duet doesn't "know" what size nozzle you are using. It just has an instruction that "tells it" to move the print head a certain distance, in a certain direction, at certain speed, while at the same time moving the extruder motor a certain amount. It also doesn't "know" what layer width you've chosen.
The slicer settings are what determine how much filament is actually extruded. The slicer calculates this from the nozzle size and/or the layer width you have chosen and the layer height. With that information, it will calculate how much filament the extruder needs to move and this is what goes into the gcode that the Duet will read and act upon. So, you can see that the slicer needs to know what size nozzle you are using. If you tell it that you have a 0.2mm nozzle but you actually use a 0.4mm nozzle it won't know any different so will calculate the extrusion amount based on a 0.2mm nozzle. The Duet doesn't care either. It just reads the gcode file which tells it that it has move the print head a certain distance in X and Y while at the same time moving the extruder and certain distance.
Ref the jamming after the first layer. This is a classic example of molten filament finding its way into the cold zone of the heat sinks and solidifying. It could be for any number of reasons, like too much retraction, or incorrect assembly, or maybe yi are just asking it to extrude too much filament.
So I suggest you sort out you slicer settings then maybe "google" E3D V6 Volcano Blocking" or some such.
I have no idea why your extruder is running the wrong way. About the only thing I can think of that could do that would be if you changed the motor direction in your config.g (M569).
-
Thanks for the breakdown. That makes it perfectly clear now.
You're right about the hotend blockage as well. Apparently there's a cold spot that causes the squashed filament to cool enough to form a handle. I pulled the hotend apart and the filament is stuck, and the nozzle won't come off either, so I'm going to have to buy a replacement. The basic hotends I used never had this problem, but I doubt this is a great product volcano anyway.
For what the base layers were, they were pretty good. This is actually the first time I've had the Duet working since I bought it early in the year, so it's good to see some action.
-
I found the specific blockage. The screw between the heat sink and the hotend has a gap in it. Somehow the plastic liner wasn't installed properly, or is damaged, and created a gap in the middle of the piece. I'd have to drill it out and put a new liner in, except I don't have either of those solutions. I replaced it with a V6 piece I had, so I didn't have to order a different mount for the V6 type extruder.
It's currently printing and extruding really well, so I'm happy about that. I'll have to look at print/slicer settings, because it appears to be printing with more filament than it should, such as 0.1mm layers but 0.2mm extrusion. It's currently squashing and mashing paths into the previous layer, which is probably too much filament. It says it's at 40%, but the 25mm test cube is only 5mm high….
-
More progress which is good - you'll get there in the end.
Ref your last paragraph. It could be over extruding but it sounds to me more like the steps per mm on your Z axis are not correct. That's easy to check - just move the Z axis say 50 mm (away from the bed) and check how far it actually moves.
-
You know what, I made a mental note about looking into the Z steps, because I had become convinced it wasn't correct towards the beginning of this process. However, I forgot after the obsession with the printing had kicked in.
It was definitely out, and is now fixed. It went from Z1600 to Z3200… a minor adjustment
I have a new hotend on the way, hopefully including an additional replacement by Tevo. Only then will my successes and failures continue the saga.
-
I finally received my replacement volcano hotend. It's been screwed together thoroughly, and after another test print it's jammed again.
I manged to disassemble the hotend and found melted filament in the nozzle, filling up the liner as well.
This issue's got me beat. The test cube print file includes 0.4mm nozzle setting, and the Duet console says it's running at 210c. The brim and first layers print perfectly, so I doubt the extrusion speed is fast. I can confirm that the joining cylinder between the heat sink and heat block meets up with the nozzle, flush. The PTFE liner in the heat sink meets up with the heat block liner, which meets up with the nozzle as well.
Theoretically, I would think that's it's not hot enough, which means the console is lying to me, or the 210c isn't enough with a volcano hotend…
There's photos of the issue on the Facebook page.
-
Maybe worth posting this on the E3D forum?
What retraction settings are you using, and how long is your Bowden tube if you use one? The all metal E3D hot ends are prone to jamming if you use too much retraction, at least when printing PLA.
-
I finally received my replacement volcano hotend. It's been screwed together thoroughly, and after another test print it's jammed again.
I manged to disassemble the hotend and found melted filament in the nozzle, filling up the liner as well.
This issue's got me beat. The test cube print file includes 0.4mm nozzle setting, and the Duet console says it's running at 210c. The brim and first layers print perfectly, so I doubt the extrusion speed is fast. I can confirm that the joining cylinder between the heat sink and heat block meets up with the nozzle, flush. The PTFE liner in the heat sink meets up with the heat block liner, which meets up with the nozzle as well.
Theoretically, I would think that's it's not hot enough, which means the console is lying to me, or the 210c isn't enough with a volcano hotend…
There's photos of the issue on the Facebook page.
That sounds like heat creep. As David pointed out, it's more likely an E3D issue than a Duet issue but there are a couple of things "duet wise" to check. Firstly double check that the thermistor values are correct for the thermistor that you are using (M305). Secondly, again as David pointed out, too much retraction can lead to molten filamnet being drawn up above the heat break where it will solidify and cause a blockage.
The Volcano has a larger melt zone so unless you are printing at high volume flow rate (which you won't be using a 0.4 mm nozzle) then you'll find that you can use lower temperatures than you have been used to. So bizarrely as it might seem, using a lower temperature will help (if it is heat creep which I believe it is). Try 190 deg C or even lower if you can get away with it.
I don't have a V6 or volcano myself but if it were me and based on my tests with another hot end, I'd try a larger flow rate fan to shift more air over the heat sinks - especially the lower fins.
-
Maybe worth posting this on the E3D forum?
What retraction settings are you using, and how long is your Bowden tube if you use one? The all metal E3D hot ends are prone to jamming if you use too much retraction, at least when printing PLA.
I've posted a thread on the E3D forum, so I'll wait for some ideas there.
I bypassed the extruder and hand fed the hotend, but it still eventually blocks up. That would mean that the retraction setting shouldn't be the issue, considering I never retract the filament, even after waiting several minutes.
The molten filament never goes above the PTFE liner in the middle screw. It just packs up and becomes in penetrable, even with my hand. And this is running the hotend at intervals between 190c and 235c. I'm going to test an E3D hotend with my hand to see if it has the same results.
-
Update: the volcano nozzle inner diameter, in the top of it, not the outlet, is visibly larger than the typical E3D nozzle. So the difference between the two nozzles is that they are different lengths, and have different internal capacities. Testing the E3D out yesterday showed no blockages, but two different brands of volcano kits failed the exact same way.
My guess is that the volcanos I've purchased have an inner capacity that causes a backup/congestion of too much filament, as opposed to the E3D, which allows less molten filament to restrict the loading filament.
It maybe a brand/design/supplier fault, and may not be present in the common volcano, or OEM brand volcano. As I mentioned, the E3D works perfectly fine over the same period of time the volcano congests, and it has less nozzle capacity. FYI.
