to fan or not to fan

hi all,
i'm getting a bit confused here with all this.
i think i am suppose to put down the first later nice and hot to help it stick to the bed (no fan). - that seems ok.
and then i have been turning the fan on as was having trouble with the prev layers still being tacky. (so fan on ) - thats ok
but now having warping problems which happens i think (read stuff from nophead) if i put hot plastic layers onto cooled plastic layers
(i assume hence other people having heating beds).
so am confused - want the fan on to cool the plastic so next layer sticks BUT don't want the fan on because it will cause warping - arrghhhh.
please set me on the path to printing heaven.
roger

Edited 1 time(s). Last edit at 11/14/2012 08:49PM by rogerw.

---

Prusa 'Explorer' (3dStuffMaker), GEN6, J-head Mk III-B, Bowden Extruder, Marlin 1.0.0 RC2, Repitier-Host V0.84 and Slic3r 0.9.8, PLA. Live at Victoria, Australia.

this is sort of what i mean (from [blog.thingiverse.com] discussing COOL in skeinforge):

" .... checking 'enable cool' .... will cause the print head to “orbit” around the printed area rather than immediately printing the next layer ... which is good because constantly running a fan pointed at the extruder head can freeze the plastic, leading to failed builds and stripped-filament-related swearing.

Without a fan, this setting can cause trouble if the printed area of a given layer is very small. The purpose of orbiting is to get that hot extruder nozzle away from a new layer for a while ... "

---

Prusa 'Explorer' (3dStuffMaker), GEN6, J-head Mk III-B, Bowden Extruder, Marlin 1.0.0 RC2, Repitier-Host V0.84 and Slic3r 0.9.8, PLA. Live at Victoria, Australia.

apparently there is cool option in slic3r too - will try to find that

---

Prusa 'Explorer' (3dStuffMaker), GEN6, J-head Mk III-B, Bowden Extruder, Marlin 1.0.0 RC2, Repitier-Host V0.84 and Slic3r 0.9.8, PLA. Live at Victoria, Australia.

I use no fan for the first two layers then fan on for the rest if the layer time is less than 60 seconds. It's needed for overhangs and small cross-sections. I saw a significant increase in print quality with it, and it should be standard in kits people sell.

I use 1:4 watered down PVA glue on the heatbed = zero warp even for large parts.

Turning the 40mm fan on drops the temperature of the hotend down about 2 degrees C but the firmware increases the current to compensate and it recovers in about 10-15 seconds. So I don't see any problems you mentioned about freezing.

Andy

Edited 2 time(s). Last edit at 11/15/2012 04:06AM by ajayre.

great Ajayre, thank you - the 60 second rule is a good one to start with

what does I like ^^" mean?

---

Prusa 'Explorer' (3dStuffMaker), GEN6, J-head Mk III-B, Bowden Extruder, Marlin 1.0.0 RC2, Repitier-Host V0.84 and Slic3r 0.9.8, PLA. Live at Victoria, Australia.

some more discussion i found [forums.reprap.org]

---

Prusa 'Explorer' (3dStuffMaker), GEN6, J-head Mk III-B, Bowden Extruder, Marlin 1.0.0 RC2, Repitier-Host V0.84 and Slic3r 0.9.8, PLA. Live at Victoria, Australia.

+1 to constant fan on PLA after a few layers are down (as to not disturb bed adhesion)

I only use a fan when there are overhangs. When I don't, they tend to curl upwards. I just use a small USB fan next to the printer.
I the past when I had a larger hotend I had rigged two small computer fans onto the x carriage. Worked pretty well.

I use a 40mm fan for PLA, I run it continually even the first 2 layers, I find that if I start the fan after the start of the print, it takes time for the print head to recover it's temperature (~5 degrees) and I can see that in the print afterwards. I use a heated bed at ~60 degrees for PLA along with dilute PVA on the glass, and I don't have warping problems.

For ABS I don't run a fan at all.

polygonhell,
"i can see that in the print afterwards" - what do you see? does it seems to do morse code for a while and then get going ok.

also found the COOL function in slic3r which someone indicated to me prev.
it has the ability to turn the fan on/off and also slow down the print speed.
i don't have a fan (connect to electronics that is) and so that won;t work but need to see if the "slow down" bit will still work (there is only one checkbox to turn the COOL feature on - not sure if the "slow down" bit will work if i don't have fan for the fan bit)

BUT the more interesting thing for me is that it the dialog/parameters are saying (haven't got it here so going from memory) that it will turn the fan
on if say it takes less than say 60 sec for a layer. and will slow down the speed to min of 10mm/s if it takes say less than 20secs to print a layer.

This is important for me - that is all the small bloody calibrartion blocks which take about 3 secs for a leyer should be done at 10mm/s.
Why doesn't it say this on the wiki calibration page. ie. for a newbie the out of the box setting are going to 30-50mm/s which is going to cause a big blobby mess.

roger

---

Prusa 'Explorer' (3dStuffMaker), GEN6, J-head Mk III-B, Bowden Extruder, Marlin 1.0.0 RC2, Repitier-Host V0.84 and Slic3r 0.9.8, PLA. Live at Victoria, Australia.

Visually the layer printed as the hotend recovers from the lost heat looks slightly misaligned because I assume it shrinks less, or at a different rate.
I also don't change the bed temperature for the first layer because IME it causes variation in the layers.
The Morse code thing is a series if partial extruder jams, it can be caused by printing a layer or two of PLA without a fan cooling the thermal barrier.

Having said all of that my issues could be exacerbated by printing the first layer a lot slower than many here do.

polygonhell,
can you explain the morse code thing a bit more. do you mean the extruder jamming or the hot nozzle clogging/jamming - by the way my extruder is not on the x carriage is on the side of the printer. if you mean the (mechanical) extrude jamming - how does the fan (which is on the bed ) affect that?
roger

---

Prusa 'Explorer' (3dStuffMaker), GEN6, J-head Mk III-B, Bowden Extruder, Marlin 1.0.0 RC2, Repitier-Host V0.84 and Slic3r 0.9.8, PLA. Live at Victoria, Australia.

Polygonhell Wrote:
-------------------------------------------------------
> Visually the layer printed as the hotend recovers
> from the lost heat looks slightly misaligned
> because I assume it shrinks less, or at a
> different rate.

I don't see that in my prints. Perhaps try increasing the temp by 5C?

Andy

Quote

polygonhell,
can you explain the morse code thing a bit more.

I'll admit I'm guessing at the actual cause since there is no way to sede what's happening inside the hotend when you see the morse code effect.
I've only seen it with PLA, and to me it looks like a series of PLA jams, that clear themselves.
PLA goes soft before it melts and it can mushroom on the inside of the PTFE tubing inside the hotend, if the mushrooming is far enough from the melt zone of the hotend it will block the PTFE tubing causing a jam, a fan blowing over or near the thermal break on the hotend, stops the temperature further from the meltzone creeping up to the point where the PLA will mushroom. I think the morse code thing you mention is a series of small jams close enough to the meltzone that they clear themselves.

Quote

I don't see that in my prints. Perhaps try increasing the temp by 5C?

Yes I've tried a lot of different temperature variations, as I mentioned earlier it may just be the slow printing of the first layer that's causing the issue with the temperature variation.
I also don't need to print the first layer without a fan to get it to stick, so I'm not overly motivated to experiment :/
Next time I end up with a "calibration day" I'll do some experiments and post pictures.

I've seen morse code plastic on my first layer also when the nozzle is too far away from the bed.

But checking good extrusion at temperature into free air is a good pre-flight check to make sure there are no current blockages.

it is also possible to control speed of fan. set max speed to say 60%. this will blow less air. there is also bridge speed, set this to 60%. just something else to try if your nozzle is cooling down. in slicer it is in the cool profile, in skien just set fan to M106 S153 ;//this is 60% on

I've been having a bit of a think about this - please just ignore me if I'm stating the obvious

Basically what we need to do is to cool the extruded filament, if we don't, there will be heat build up and the part will melt causing all sorts of problems. On the other hand if we cool the part too much, it will quickly go to a solid state and shrinkage will cause the part to deform or at least set up stresses that try to peel the part off the bed.

So, ideally, we need to cool the part to a temperature where the material will creep easily but will not sag. This is difficult to do by blowing cold air onto the part (or by waiting for ambient air to cool it) because we have poor control. If we were to blow air of the correct temperature we could do so in a fairly basic way because blowing too much on wouldn't be a problem.

The easiest way to do this is with a heated chamber so the hot air can be recycled. The big problem with this approach for repraps is that the parts used to make the reprap would soften and fail.

So, we have two choices for indiscriminate cooling

1. Make the reprap from heat tolerant materials so we can have a heated chamber which means it's not really a reprap since it can't replicate
2. Generate hot air to blow onto the part - not an ideal solution since it will cost money and there may be issues with dealing with the waste air.

If neither of the above is practical, we need to work on how we can cool the part more precisely using room temperature air.

I'm thinking of a solution where we blow air at the immediate area around the (insulated) nozzle and also extract it from the same area so we get some preheating of the air and don't cool more of the part than necessary. The fan speed should also be linked to the time taken to print the layer - does any of the slicing software support that?

Comments?

I regularly experiment with a hair dryer that has a hot and cool setting. Because I am holding it it is pretty easy to modulate where and how much air is hitting the part. I can also compensate for nozzle temps diving by just moving it away for a few seconds when I see the temps going down. Even with quite a bit of control fans still suck as a magic bullet.

The thing that seems to work the best is putting a spatula right on the a curling area of the part. Just the slightest force is enough to flatten and the spatula acts as a heatsink. Touching a part while it is printing also gives a much better idea of how much heat build up is taking place which is often quite ridiculous and is not helping print quality any. I've thought about how to automate the spatula thing. Maybe a roller or something. I dunno it seems hopeless and is just asking for the part to be ripped off.

I think the proper solution is to just have soluble support interfaces below all overhangs and a proper modeling of the heat within a part and just be patient enough to obey the min layer time the modeling spits out.

billyzelsnack Wrote:
-------------------------------------------------------
> is hitting the part. I can also compensate for
> nozzle temps diving by just moving it away for a
> few seconds when I see the temps going down. Even
> with quite a bit of control fans still suck as a
> magic bullet.

Try using Marlin - it has PID control of the hotend temp. Mine recovers quickly and is stable once the fan is on.

Also try attaching a nozzle to the front of the fan to direct the air towards the top layer rather than the hotend. I currently don't use a nozzle but it might help if you are having issues with the temp dropping.

Andy

I've seen morse code plastic on my first layer also when the nozzle is too far away from the bed.

yes winwalk
i suppose i'm having trouble determining the problem of my morse code
ie.
1. nozzle height not quite right; or
2. extruder playing up again and not delivering properly; or
3. nozzle temp drop for a few seconds

i think the most likely culpret of these is getting my bed level.
i think it is pretty good but i have set up a stl to print large circle (single line width) on the
bed and will look for highs and lows as well, pause after each tree ring, adjust and then run again with it
print another tree ring a bit further in.

anyway thats tonights plan ....

roger

---

Prusa 'Explorer' (3dStuffMaker), GEN6, J-head Mk III-B, Bowden Extruder, Marlin 1.0.0 RC2, Repitier-Host V0.84 and Slic3r 0.9.8, PLA. Live at Victoria, Australia.