Speed vs Tempreture and a tale of frustration.
Posted by Mongrel_Shark
|
Re: Speed vs Tempreture and a tale of frustration. September 21, 2016 01:07AM |
Yep. Your on the money again James. The cooling effect of the fan is over powering the heater. It would still be nice to have other options than PID though.
A simple 3-5 choice if-then loop could be a lot more effective than PID in situations like this. Or use logic gates and have tuning pots. Even simpler and frees up firmware for other stuff...
Choices are nice.
For now I have it working a lot better. Blockages are reduced a lot and I can do small parts that have a few patchy layers but easily repairable with ABS paste and acetone vapour. Its intermittent and I can't work out what the variable is... I can't see corresponding temp fluctuations at the nozzle. But sometimes after retracts/ layer changes etc it under extrudes a bit. Its not a problem on bigger parts, gets lost on internal perim. But its making for some holey small parts and probably still blocking if I try and do multiple small parts. Maybe I should data log barrel temp and see if that has corresponding changes?
I played around with the fan a lot, made an airflow guide out of cardboard because I was out of ally cans. Its a lot lighter and more adjustable than a printed one and only took 2 min to tape to the fan as opposed to approx 2-4 hours for my limited CADing skills and another hour or two to print.
After that I ran auto tune 3 times and it got a bit better each time. was still oscillating a lot, but more even with the over/under, Turned up i value a bit and its staying in a +/- 2c range on thermistor, which works out to about +/- 2-4c on K probe. Its about as stable as I have had it to date. May play with small changes to the values and try to improve further.
Also got the air to the nozzle fine tuned, it gets venturi effect on one side and coanda effect on the other side. So only secondary air driven by the flow from the fan. Its really light flow and warm. I can only feel it on the back of my hand if its wet, and then it still feels warm. Its also observable with smoke from an incense but not with tissue paper. Fine point/overhang curl is gone and I can now bridge 70mm+ solid but with a few dags and 40-50mm clean.
Would still love to be able to get my Maxis k-probe amp working, on i2c or otherwise, preferably otherwise. Or any K-probe, I have op-amps and know how to use them... Thinking about running the maxis right off the ATX 5v and just feeding the output to a thermistor pin. Then swap out an unused thermistor table for the K equation. Not good with code, worried I'll get stuck with syntax errors and code that won't compile like when I defined the maxis according to the instructions.... I get a headache as soon as I start looking at what needs to be changed...
I have some no-name silver bearing thermal paste and some Arctic MX4. Both are for electronics that will let out the smoke at 150c so I doubt they are rated for the temp. Thinking about making some aluminium powder and finding a good paste to mix it with. Bit stuck of a sutable paste that will maintain high viscosity at 350c (overkill for safety margin) .
Also looking at some chunks of aluminium bits I have and thinking about making a better heat block. I got nothing thick enough to do a billet style like the cheap ones and E3D etc, but I have some 3mm cast (plasma TV parts) and a box of heatsinks I could make a club sandwich arrangement from.
Also thinking about drilling out some more M6 rod and packing bits of can with nuts for spacers on there. While I'm sure most people can get the common style heatbreak working ok. I see a lot of room for improvement there...
Also looking at my thermal insulators. This one has always been on the to do.... I have a variety of fibreglass cloth from surfboard repairs. A fibreglass fire blanket. and I think somewhere some proper glass fibre insulation from a stove/oven. I put it somewhere safe so the fibres wouldn't get everywhere. Just need to remember where that is lol. Also looking at the silicon covers some heatblocks ship with as an option. I probably have RTV silicon I use for engine gaskets thats good for 360c... Decisions decisions, and a lack of 50 hour days lol. Also sort of holding off till I get around to changing to the duel drive extruder.
I am very very tempted to try another firmware, but I am printing pretty good now. Just going to make a few printer kits, spares etc and enjoy for a bit before I go down that road. Any thoughts on tea cup? Are there any other options? I'd be more keen to try something radically different than something similar. Ideally something small and simple.
Here are my latest prints. Finally found the gears that are on my extruder. Not many 13 tooth herringbones out there...
You can see the under extrusion holes I am talking about on the small bar part with 2 holes. And around the hex on the lower large gear.
Well you can sort of see. Not very clear in the resized photos. Might have to put in microscope... So yea, very minor and easily fixed with ABS juice followed by vapour.
The small gear on the right had a light vapour treatment, the other 3 gears are raw straight off the printer.
Edited 1 time(s). Last edit at 09/21/2016 02:42AM by Mongrel_Shark.
If you can't fix it with tape and tie wire. Its broken. You need a new one.
[www.youtube.com]
[plus.google.com]
Moderators and RepRap Wiki editors have my full and unconditional permission to use any original photo's I upload for the wiki or any other way they may be of benefit to others.
A simple 3-5 choice if-then loop could be a lot more effective than PID in situations like this. Or use logic gates and have tuning pots. Even simpler and frees up firmware for other stuff...
Choices are nice.For now I have it working a lot better. Blockages are reduced a lot and I can do small parts that have a few patchy layers but easily repairable with ABS paste and acetone vapour. Its intermittent and I can't work out what the variable is... I can't see corresponding temp fluctuations at the nozzle. But sometimes after retracts/ layer changes etc it under extrudes a bit. Its not a problem on bigger parts, gets lost on internal perim. But its making for some holey small parts and probably still blocking if I try and do multiple small parts. Maybe I should data log barrel temp and see if that has corresponding changes?
I played around with the fan a lot, made an airflow guide out of cardboard because I was out of ally cans. Its a lot lighter and more adjustable than a printed one and only took 2 min to tape to the fan as opposed to approx 2-4 hours for my limited CADing skills and another hour or two to print.
After that I ran auto tune 3 times and it got a bit better each time. was still oscillating a lot, but more even with the over/under, Turned up i value a bit and its staying in a +/- 2c range on thermistor, which works out to about +/- 2-4c on K probe. Its about as stable as I have had it to date. May play with small changes to the values and try to improve further.
Also got the air to the nozzle fine tuned, it gets venturi effect on one side and coanda effect on the other side. So only secondary air driven by the flow from the fan. Its really light flow and warm. I can only feel it on the back of my hand if its wet, and then it still feels warm. Its also observable with smoke from an incense but not with tissue paper. Fine point/overhang curl is gone and I can now bridge 70mm+ solid but with a few dags and 40-50mm clean.
Would still love to be able to get my Maxis k-probe amp working, on i2c or otherwise, preferably otherwise. Or any K-probe, I have op-amps and know how to use them... Thinking about running the maxis right off the ATX 5v and just feeding the output to a thermistor pin. Then swap out an unused thermistor table for the K equation. Not good with code, worried I'll get stuck with syntax errors and code that won't compile like when I defined the maxis according to the instructions.... I get a headache as soon as I start looking at what needs to be changed...
I have some no-name silver bearing thermal paste and some Arctic MX4. Both are for electronics that will let out the smoke at 150c so I doubt they are rated for the temp. Thinking about making some aluminium powder and finding a good paste to mix it with. Bit stuck of a sutable paste that will maintain high viscosity at 350c (overkill for safety margin) .
Also looking at some chunks of aluminium bits I have and thinking about making a better heat block. I got nothing thick enough to do a billet style like the cheap ones and E3D etc, but I have some 3mm cast (plasma TV parts) and a box of heatsinks I could make a club sandwich arrangement from.
Also thinking about drilling out some more M6 rod and packing bits of can with nuts for spacers on there. While I'm sure most people can get the common style heatbreak working ok. I see a lot of room for improvement there...
Also looking at my thermal insulators. This one has always been on the to do.... I have a variety of fibreglass cloth from surfboard repairs. A fibreglass fire blanket. and I think somewhere some proper glass fibre insulation from a stove/oven. I put it somewhere safe so the fibres wouldn't get everywhere. Just need to remember where that is lol. Also looking at the silicon covers some heatblocks ship with as an option. I probably have RTV silicon I use for engine gaskets thats good for 360c... Decisions decisions, and a lack of 50 hour days lol. Also sort of holding off till I get around to changing to the duel drive extruder.
I am very very tempted to try another firmware, but I am printing pretty good now. Just going to make a few printer kits, spares etc and enjoy for a bit before I go down that road. Any thoughts on tea cup? Are there any other options? I'd be more keen to try something radically different than something similar. Ideally something small and simple.
Here are my latest prints. Finally found the gears that are on my extruder. Not many 13 tooth herringbones out there...
You can see the under extrusion holes I am talking about on the small bar part with 2 holes. And around the hex on the lower large gear.
Well you can sort of see. Not very clear in the resized photos. Might have to put in microscope... So yea, very minor and easily fixed with ABS juice followed by vapour.
The small gear on the right had a light vapour treatment, the other 3 gears are raw straight off the printer.
Edited 1 time(s). Last edit at 09/21/2016 02:42AM by Mongrel_Shark.
If you can't fix it with tape and tie wire. Its broken. You need a new one.
[www.youtube.com]
[plus.google.com]
Moderators and RepRap Wiki editors have my full and unconditional permission to use any original photo's I upload for the wiki or any other way they may be of benefit to others.
|
Re: Speed vs Tempreture and a tale of frustration. September 21, 2016 11:57PM |
Here are some pics showing the second layer problem
On this print the first layer was also a ittle patchy. I think I just didn't extrude enough before I started and there was some air in there. First layer normally goes down great so ignore that.
These steps are scaled up 150% so 7.5mm per step.
See how it took most of the second layer to start extruding at full flow? This has been happening on every second layer for a few days now.
After the second layer its all good unless there is a retract and with a speed change after retract.
This pic is 3-4 layers later. Its all going fine. Retract on every layer change. No under extrusion.
I have checked for lead screw play. Its not the problem. Still happens if I start from Z0. so always going up. At first I thought it could be because it usually goes down for the first layer then up for the second. In which case screw play would make sense. But checked it a few ways. 99% sure its not screw play. The height is right. Its not extruding at the correct flow rate... If I pause and manual extrude to get the flow back up it goes away immediately after the pause. So maybe its getting air in the nozzle???
IDK???
If you can't fix it with tape and tie wire. Its broken. You need a new one.
[www.youtube.com]
[plus.google.com]
Moderators and RepRap Wiki editors have my full and unconditional permission to use any original photo's I upload for the wiki or any other way they may be of benefit to others.
On this print the first layer was also a ittle patchy. I think I just didn't extrude enough before I started and there was some air in there. First layer normally goes down great so ignore that.
These steps are scaled up 150% so 7.5mm per step.
See how it took most of the second layer to start extruding at full flow? This has been happening on every second layer for a few days now.
After the second layer its all good unless there is a retract and with a speed change after retract.
This pic is 3-4 layers later. Its all going fine. Retract on every layer change. No under extrusion.
I have checked for lead screw play. Its not the problem. Still happens if I start from Z0. so always going up. At first I thought it could be because it usually goes down for the first layer then up for the second. In which case screw play would make sense. But checked it a few ways. 99% sure its not screw play. The height is right. Its not extruding at the correct flow rate... If I pause and manual extrude to get the flow back up it goes away immediately after the pause. So maybe its getting air in the nozzle???
IDK???
If you can't fix it with tape and tie wire. Its broken. You need a new one.
[www.youtube.com]
[plus.google.com]
Moderators and RepRap Wiki editors have my full and unconditional permission to use any original photo's I upload for the wiki or any other way they may be of benefit to others.
|
Re: Speed vs Tempreture and a tale of frustration. September 25, 2016 11:13PM |
Ok after rebuilding half the printer I have answered a lot of questions.
I re did most of the wiring looms. Fixed up the drive unit and added a quick release clamping bearing thingy.
I can also confirm that paper towel is not rated for the temp of the extruder and should not be used as a heat-block insulator.
The gear socket for the bolt head hadn’t worn. The bolt was always a bit loose. The new unused versions of the same gear had the same problem.
This led to me making a new hobbed bolt that fits better.
I did video on that because its something not many people feel confident to try and its really easy.
[www.youtube.com]
I played with various settings one at a time to see if I could fix the second layer low extrusion problem.
Printing the first layer at 100% speed took a few tries to get working. Needed a really small bed-nozzle gap. Made no difference to second layer extrusion.
Turning off retraction made no change. Extra length after retraction does improve the problem, but its a kludge because it leads to over extrusion in many places.
Higher print temp helped a lot. turned slic3r settings up from 243 to 246. This made a significant change but didn't completely fix it.
Printing fatter than nozzle width also made a huge change. Printing 0.5 wide from the 0.4 nozzle nearly completely fixed it.
Printing at 246 and 0.5 wide has fixed that.
I think I am all sorted
Aside from some serious shrinkage issues on an overnight print its been behaving itself really well. The shrinkage was a lot to do with cheap masking tape that didn't stick to the bed, and printing at 243. Have re-printed the same parts one at a time (It was a bed full of printer parts that warped and split) at 246c, with a 5mm brim, glue stick and ABS juice instead of tape, and slightly higher speeds and its seems to be sorted. Its ABS and I curently have an enclosure made from scrap timber and old towels.
here are some new prints
The smaller bolt with wing nut is M8.
I'm printing another printer to play with
If you can't fix it with tape and tie wire. Its broken. You need a new one.
[www.youtube.com]
[plus.google.com]
Moderators and RepRap Wiki editors have my full and unconditional permission to use any original photo's I upload for the wiki or any other way they may be of benefit to others.
I re did most of the wiring looms. Fixed up the drive unit and added a quick release clamping bearing thingy.
I can also confirm that paper towel is not rated for the temp of the extruder and should not be used as a heat-block insulator.
The gear socket for the bolt head hadn’t worn. The bolt was always a bit loose. The new unused versions of the same gear had the same problem.
This led to me making a new hobbed bolt that fits better.
I did video on that because its something not many people feel confident to try and its really easy.
[www.youtube.com]
I played with various settings one at a time to see if I could fix the second layer low extrusion problem.
Printing the first layer at 100% speed took a few tries to get working. Needed a really small bed-nozzle gap. Made no difference to second layer extrusion.
Turning off retraction made no change. Extra length after retraction does improve the problem, but its a kludge because it leads to over extrusion in many places.
Higher print temp helped a lot. turned slic3r settings up from 243 to 246. This made a significant change but didn't completely fix it.
Printing fatter than nozzle width also made a huge change. Printing 0.5 wide from the 0.4 nozzle nearly completely fixed it.
Printing at 246 and 0.5 wide has fixed that.
I think I am all sorted
Aside from some serious shrinkage issues on an overnight print its been behaving itself really well. The shrinkage was a lot to do with cheap masking tape that didn't stick to the bed, and printing at 243. Have re-printed the same parts one at a time (It was a bed full of printer parts that warped and split) at 246c, with a 5mm brim, glue stick and ABS juice instead of tape, and slightly higher speeds and its seems to be sorted. Its ABS and I curently have an enclosure made from scrap timber and old towels.here are some new prints
The smaller bolt with wing nut is M8.
I'm printing another printer to play with
If you can't fix it with tape and tie wire. Its broken. You need a new one.
[www.youtube.com]
[plus.google.com]
Moderators and RepRap Wiki editors have my full and unconditional permission to use any original photo's I upload for the wiki or any other way they may be of benefit to others.
|
Re: Speed vs Tempreture and a tale of frustration. October 06, 2016 07:58PM |
That is a really interesting idea Tinchus.
I have actually been wondering why I have not had moisture problems, as its a fairly humid climate here. Not as much as you, for our winter, which just ended, and its been a very dry winter but it will get very humid over the next 6 months.
I will try dying filament, I'm not expecting much change, but always good to do experiments to prove/disprove a theory.
I think you may still be onto something with the pressure though. I have recently started playing with Pressure advance settings in Slic3r in an effort to solve a bridging problem. While I still need to dial in the setting a bit, or probably more to the point, fix something mechanical that's been on my to do list for ever.
Since I tried adding some pressure advance my bridges start better, where before the origin side was not anchoring well. My second layer and other speed change under extrusion issues seem a little improved. and my overhangs are curling up less. A whole bunch of problems that where not affected by any other setting are al affected strongly by pressure advance setting.
I think I know why too. As I have a wood block between extruder and heatbreak, as heat protection for the extruder. There is a 16mm wide cavity in the bottom of the extruder where the J-head would normally mount. I have been meaning to make a wood or printed spacer to go in there for ages. When I get blocked sometimes there is a huge spaghetti jam in there... I think this cavity allows the filament to flex, like with a bowden tube. Causing some odd pressure issues. Where the drive is less direct than it could be. As well as potentially kinking the filament on the way in to the barrel.
Having experimented with cooling on the heatbreak a lot since my last post. I am very confident that was a very large part of my problem too. Since I turned the fan up and better isolated the heatblock from the cooling air, I have not really had any serious blockage issues. Unless I forget to start the fan or have it too low. I can now print much slower no problems. Although still some under extrusion issues where speed changes happen. Now I am confident its the pressure lag/advance where energy is stored in the flex of the filament.
I actually had a series of bad blockages last night testing the advance feature. Not sure it was caused by this, I think I was just overdue to clean my barrel out. Every time I drilled out the blockage and re-assembled I would extrude a few mm then it would block immediately. 7 times in a row I had to drill it out without even getting close to running some g-code with the advance setting. Lots of black stuff in there. So Its all dissembled in a jar of acetone now. I'll clean it up and re-assemble in another few hours. In the mean time I will see if I have a 16mm (or a little bigger) hole saw that can use to make a spacer to prevent the "Bowden Flex" problem. If I can't make a wood one I'll print one first print.
Then I will prepare a moisture test experiment
P.s. I do see smoke/vapour from my hot extrusion. Although this could be many things aside from water vapour. Could be the plastic I am printing, or more likely plasticiser additives. One thing that makes me think its additives is that it sets off my smoke alarm in the next room sometimes. That tiny bit of steam would not do this. I can boil a kettle near the printer and not set off that smoke alarm. So wile I like your coloured light vapour examination, further testing would be required to confirm what the vapour is. I have been looking into plasticisers. Glycerol is a common one, its boiling point is 290c and its melt point is 17-18c (https://en.wikipedia.org/wiki/Glycerol). Consider water for a moment, its boiling point is 100c, yet it does evaporate much cooler at a slow rate and steams visibly from about 60-70c, so you could well be looking at glycerol vapour evaporating at print temps under 290c! What you need to do is condense some. I'd be inclined to try some cooled glass, maybe a jar with ice in it and a good sealed lid. Hold the cold jar in the vapour and try to condense some on the outside. As water will not turn to a solid on the outside of the cooled glass, but glycerol(and likely most other plasticisers) should. You should get either a wet liquid condensation which would confirm water vapour, or a gel like gooey or soapy condensation confirming plasticiser. Now that you have a sample you could get help from a chemist to further confirm what it is.
Obviously as you have a humid environment this test must be conducted in a way that prevents the environmental humidity from condensing on your jar. As your printer is enclosed this should not be too hard. Just run a control test with no plastic extruding to confirm no moisture from the air is condensing on your collection jar. Or if there is some, and you cant eliminate it completely. Weigh it to get the background interference level, and subtract that from your result.
If you can't fix it with tape and tie wire. Its broken. You need a new one.
[www.youtube.com]
[plus.google.com]
Moderators and RepRap Wiki editors have my full and unconditional permission to use any original photo's I upload for the wiki or any other way they may be of benefit to others.
I have actually been wondering why I have not had moisture problems, as its a fairly humid climate here. Not as much as you, for our winter, which just ended, and its been a very dry winter but it will get very humid over the next 6 months.
I will try dying filament, I'm not expecting much change, but always good to do experiments to prove/disprove a theory.
I think you may still be onto something with the pressure though. I have recently started playing with Pressure advance settings in Slic3r in an effort to solve a bridging problem. While I still need to dial in the setting a bit, or probably more to the point, fix something mechanical that's been on my to do list for ever.
Since I tried adding some pressure advance my bridges start better, where before the origin side was not anchoring well. My second layer and other speed change under extrusion issues seem a little improved. and my overhangs are curling up less. A whole bunch of problems that where not affected by any other setting are al affected strongly by pressure advance setting.
I think I know why too. As I have a wood block between extruder and heatbreak, as heat protection for the extruder. There is a 16mm wide cavity in the bottom of the extruder where the J-head would normally mount. I have been meaning to make a wood or printed spacer to go in there for ages. When I get blocked sometimes there is a huge spaghetti jam in there... I think this cavity allows the filament to flex, like with a bowden tube. Causing some odd pressure issues. Where the drive is less direct than it could be. As well as potentially kinking the filament on the way in to the barrel.
Having experimented with cooling on the heatbreak a lot since my last post. I am very confident that was a very large part of my problem too. Since I turned the fan up and better isolated the heatblock from the cooling air, I have not really had any serious blockage issues. Unless I forget to start the fan or have it too low. I can now print much slower no problems. Although still some under extrusion issues where speed changes happen. Now I am confident its the pressure lag/advance where energy is stored in the flex of the filament.
I actually had a series of bad blockages last night testing the advance feature. Not sure it was caused by this, I think I was just overdue to clean my barrel out. Every time I drilled out the blockage and re-assembled I would extrude a few mm then it would block immediately. 7 times in a row I had to drill it out without even getting close to running some g-code with the advance setting. Lots of black stuff in there. So Its all dissembled in a jar of acetone now. I'll clean it up and re-assemble in another few hours. In the mean time I will see if I have a 16mm (or a little bigger) hole saw that can use to make a spacer to prevent the "Bowden Flex" problem. If I can't make a wood one I'll print one first print.
Then I will prepare a moisture test experiment
P.s. I do see smoke/vapour from my hot extrusion. Although this could be many things aside from water vapour. Could be the plastic I am printing, or more likely plasticiser additives. One thing that makes me think its additives is that it sets off my smoke alarm in the next room sometimes. That tiny bit of steam would not do this. I can boil a kettle near the printer and not set off that smoke alarm. So wile I like your coloured light vapour examination, further testing would be required to confirm what the vapour is. I have been looking into plasticisers. Glycerol is a common one, its boiling point is 290c and its melt point is 17-18c (https://en.wikipedia.org/wiki/Glycerol). Consider water for a moment, its boiling point is 100c, yet it does evaporate much cooler at a slow rate and steams visibly from about 60-70c, so you could well be looking at glycerol vapour evaporating at print temps under 290c! What you need to do is condense some. I'd be inclined to try some cooled glass, maybe a jar with ice in it and a good sealed lid. Hold the cold jar in the vapour and try to condense some on the outside. As water will not turn to a solid on the outside of the cooled glass, but glycerol(and likely most other plasticisers) should. You should get either a wet liquid condensation which would confirm water vapour, or a gel like gooey or soapy condensation confirming plasticiser. Now that you have a sample you could get help from a chemist to further confirm what it is.
Obviously as you have a humid environment this test must be conducted in a way that prevents the environmental humidity from condensing on your jar. As your printer is enclosed this should not be too hard. Just run a control test with no plastic extruding to confirm no moisture from the air is condensing on your collection jar. Or if there is some, and you cant eliminate it completely. Weigh it to get the background interference level, and subtract that from your result.
If you can't fix it with tape and tie wire. Its broken. You need a new one.
[www.youtube.com]
[plus.google.com]
Moderators and RepRap Wiki editors have my full and unconditional permission to use any original photo's I upload for the wiki or any other way they may be of benefit to others.
Sorry, only registered users may post in this forum.