Possible part cooling design? #20
Comments
|
That's mine :) I do plan on doing a Bear version once I've assembled the 0.5alpha to see what's what. I'm quite slow thanks to my 7 mouth old daughter though... Mostly done in Fusion if anyone want the file. |
|
Oh nice to have you here! Happy you are interested in my Bear extruder :-) . I would appreciate to have the F360 file to adapt he place for the PINDA, would you prefer to send me this in private? Be sure that if I use your shroud (or a remix of it) I will add appropriate credits and license! |
|
I've had your Full Bear on my MK3 since RatRigs first sold the kit its a great design 👍 I'll message you on Thingiverse later on today with the file. |
|
Wow impressive Rob. Hope you like it :-) . Got your message thanks! |
|
I used three washers that are ~0.5..0.55mm thick per fan screw to get everything on nice and go figure it looks just like the above image. Been too lazy to draw up a fan shaped shim to print out as I figure even that would hopefully be a temporary solution. |
|
I've been trying this nozzle out as well. I tried it both with no spacer washers, and with 3 spacer washers each on the two print fan mounting screws and the nozzle mounting screws. So far my results have been a bit better than with the stock Prusa R2 nozzle. It actually does seem to just barely fit w/o the spacer washers, but it's very close and possibly touching the PINDA, so I decided to insert them. |
|
I designed a spacer. I'll post a link once I have a chance to print it and test it out. |
|
If you are interesting trying out a Bear Extruder (v0.5a) specific version, then you can - https://www.dropbox.com/s/poudc3owi7r05y9/Initial%20Bear%20duct.zip?dl=0 It's not finished yet, but it's working as intended. |
|
@robrps I'm currently using your duct on my 3 Bear extruders with great success. Super excited to try out the Bear specific version. I'll see if I can get them printing tomorrow. I would really like the F360 files when you have it complete. Thanks! |
|
I finally made a shim yesterday. it is a quick and dirty copy face from the front of the extruder body. but works. honestly I was considering doing something similar in tpu just so I could see if it quites the fan down. printed face down so the nice smooth side is facing out. |
|
Interesting. This is the spacer that I made. @robrps - I tried this latest version and it is not even remotely close to fitting on my extruder. @ufoDziner are you using his version from the dropbox share, or the one that was posted to Thingiverse? |
|
I'll just forward this little known remix of the R3 shroud, which, if you look closely, you can see the air has less of a fight to get through, and also, less air is allowed out the first exit holes, and is redistributed deeper. |
|
I don't like the R3 fan position. With those cheap fans, it's going to cause premature bearing wear and eventual fan failure. What I like about @robrps design is that it works great with the original fan position (assuming you use the spacer, or some washers). I'd really like to see an update of it that doesn't require the spacer. |
|
Here's a version with the Bear logo cutout to save a little plastic (good idea, @trevjonez ) |
|
Yeah, my bear duct is a little too tight at the moment - it doesn’t require screws to hold it in place! It also blows a little too much on the block, so without a sock it can cause issues. I’m starting that design again but parametrically so adjustments can be made using parameters. Just getting to grips with Fusion360’s way of doing things. I’m more used to NX but I want others to be able to make changes. |
|
@robrps, It looked like the issue was the mount hole was lowered too much. I'm using your previous design with my spacer now and it's working well. BTW, @gregsaun has been making some changes to the extruder in the dev branch, mostly to improve hotend cooling, so you may want to check to make sure there aren't any fitment issues with the new design. |
|
@matthew-humphrey I'm using the dropbox version. It is very close tolerance as @robrps stated, but seems to be working okay. I mount the shroud to the extruder first, then add the fan after. Makes fitting it much easier. First pic is a normal Bear extruder, second is with my Volcano remix.
|
|
Thanks, @ufoDziner - much appreciated! Ok, I see what it is. There's a lip that needs to go behind the extruder cover. Following your advice, I put the fan shroud on first, saw that, and now it's all together. Looking forward to trying it out! |
|
@matthew-humphrey Just in case you didn't notice this; part of this shroud goes on the inside of the extruder cover. It's not like other shrouds. The inner flange goes inside, then the screw mounting flange goes on the other side of the wall. |
|
@ufoDziner - Yes, that was exactly the issue. Thanks again. |
|
I've done an update to my bear cooling duct. If you are feeling brave, you can download from the links below. I've finished it during my lunch break at work (Shhhhhhh!!) so I've not been able to print or test as yet. There are 2 version, one similar to the last, the other has a staggered exits at the side.... might be worse! Print with 3 perimeters and a print width of 0.45. https://www.dropbox.com/s/u23mmxfw3rlxaw5/Bear%20Part%20Cooling%20-%20Non-Staggered.stl?dl=0 If I can get the Fusion exporting working I'll link the file. But it's a distributed design so its not letting me. but the angle, exit width, stagger and wall thickness are all adjustable parameters. |
|
I just got the first bear one on and fit and function has been great. I'll give these a spin. only thing I wonder is if we can do anything more to have the air flow angled more downward towards the part/bed? edit: actually I am probably wrong in thinking the downward angle is not enough. I think physical perception was just off. looking at these latest STL's in slicer layer by later makes it easy to see how much downward angle is being put in already. |
|
@robrps - @gregsaun shared with me a technique to get the Fusion 360 export working when it gives you that error about the distributed design. From Gregoire:
|
|
Thanks @matthew-humphrey although I've now realised that by bringing in the whole of Greg's Bear Extruder and X-Axis means that it now is going to export it all as well!...Oh well, next iteration I'll just use the Extruder Cover file or create non associative reference/skeleton geometry. Link to the file below. Having just printed these, they are not quite ready for use - The above 2 STLs are a little too close to the PINDA, plus the little tab cut out is needed, but currently missing (along with the Bear text and strengthening ribs.) https://www.dropbox.com/s/p5dttqnfxyzx4j9/Alternative%20Part%20Cooling%20Duct2.f3z?dl=0 |
|
@robrps, Gregoire has been making some changes to the Pinda position recently. I think in the last iteration he moved it back, but you might want to check before you make any more updates. BTW, you guys have any tips on printing this in ABS? I bet some bad overhang drooping in the openings where the air comes out. |
|
@matthew-humphrey The probe position is not an issue, because it's a parametric design it's trivial to adjust. Normally I print these parts with the custom supports, but they are usual the last thing I create. The exits are now straight and not curved, so you'll need to adjust your bridge settings or use supports from the build plate, or custom supports with the new Slic3rPE. |
|
@robrps So, would you consider this ready to test? I can print a few more of them for testing. Thanks! |
|
I have the non staggered version on mine for 24 hours or more now. so far so good. I actually like it better without the BEAR text on the front. I don't have the patience to drop nozzle size and speed to get the text as nice and crisp as I would want it. |
|
So the staggered version made no difference that I could see from the non staggered version. So it's probably only useful to clear any dramatic Pinda position design changes in future versions of the Bear Extruder. I made some tweaks to the model. @ufoDziner This should be the last design change until the next version of the extruder gets released. Unless something is broken/not working of course! @gregsaun please feel free to use any or none of this model, the dependences on your parts means it's probably easier if it were put into your Fusion360 design if you wanted to use it for your Bear Extruder. |
|
@robrps setting up the latest for print now 👍 such good community here. If ever I get an mmu2 I intend to print one where the lettering is a different color. the contrast would be so so nice |
|
@RH-Dreambox wow awesome, thanks, very interesting. Very curious to see your new shroud, sounds promising! For the thermal runaway -> E3D sock! @robrps ok that sounds very nice, great news! Do you have printed another one to compare? |
|
Photo of the one in Slic3r, it's not quite as nice. Used 3 perimeters as per the above. |
|
What incredibly good prints robrps and strahlfabrik performs! How do they wear???
Thanks gregsaun, it helped! |
|
Prusa Firmware has trouble reading the fast ldo fan. Seems it is too fast or so for the reading. You might consider disabling fan check. The better cooling performance is imho worth it:-) Or you use a different firmware like marlin (which has no fan check yet:-)) |
|
Thank you stahlfabrik for that information. I will test the fan again and I use firmware 3.7.0 with the new "7x7 mesh bed leveling". Another important thing I notised with the new firmware was that my bed level got completely new and very uneven values (0.3 mm) so I had to redo the adjustment (nylock nuts). |
|
@RH-Dreambox the magnets have a very tiny effect. 0.05mm as a worst case. If you noticed the offset you saw was generally from left to right or vice-versa, it may have been caused by the way Prusa "calibrates" the Z axis. On both my printers, re-"calibrating" the Z-axis introduces a random error that I attribute to the left and right Z axis not being exactly even. You might have better results by hand calibrating it my moving axis near to and then manually adjust both sides to top (gently), then use menu to lower axis at least 10mm and then re- home. |
|
@RH-Dreambox and @stahlfabrik I don't believe the issue with the fan error is because it is "too fast" because I pointed out this issue a while back. prusa3d/Prusa-Firmware#1546. While I don't use the LDO branded fan, but rather a Runda branded fan Zaribo was selling. But from reading, both seem to suffer from the same issue and both are similar specd, ie 5000rpm and 0.3A. What I've observed is the error happens even when the fan is at low speed (I've tried 25% -35% fan speed). I also noticed an issue with fan stability, ie, the fan would run at a constant low speed for about 5-10 seconds and then briefly increase in speed for about a second and settle back down. Listening closely to the fan, I hear the increase in speed. prusa3d/Prusa-Firmware#1516 I could live with disabling fan checking to prevent the fan error stopping the print, but the change in fan speed is more of a concern to me. @stahlfabrik I know you are a firmware guru. The old PWM method seemed to have worked much better with these fans (no error and no fan speed change). Think it would be worth it to implement the old method back into the later firmwares for those that want to use the more powerful fan? I don't know if the same issue is present in the LDO blower fan available at Printed Solids that's rated for 4500RPM. |
|
@Hotpocketdeath the issue may be due to the fact that when PWM-ing the fan, it interferes with the tach signal from the fan. Newer Prusa firmware changed the PWM frequency to lower fan whine, and at the same time introduced a feature that disables PWM every 5 seconds for a 100ms interval. They also have code that rejects tach pulses occurring "too frequently". Within that 100ms interval, the tach interference goes away, and in theory the "too frequent" filter would not be triggered and sampling of the tach would happen. I suspect the faster speed of the LDO fan is causing some interaction with this scheme and falsely detecting a lack of tach pulses. I'm sure it's fixable, but personally I made the decision a long time ago to get off the Prusa firmware train. I'm running Marlin now, and can't imagine any reason why I would switch back to the Prusa firmware. I mean they are shipping bugs that burn out peoples' Einsy boards! |
|
100% agree. Since Prusa decided to lower pwm noise on the fan, it runs buggy and pulses every few moments when not at 100% duty cycle. I tested this with Prusa fan. It is not the fan. It is the bad firmware. Marlin stock - works like a charm Also: those three wire fan are regulated by voltage and not by a separate pwm signal. So it is not a fan bug. But a control board or software bug |
|
The print fan on Einsy boards is regulated by PWM-ing the ground signal, effectively PWM the current through the fan. You are right that they don't have a 4th input to allow an external PWM input signal which then causes regulation by a circuit in the fan. |
|
@stahlfabrik and @matthew-humphrey, what version of Marlin do you use? Is there anything special you have to consider when driving Marlin? |
|
Sorry for taking this offtopic. Best would be to talk about this on our discord. I run marlin 2 since a few days. Earlier was running 1.1.9. But motocoder and I have forked marlin in the PrusaOwners GitHub. We started to implement missing features. So for example we have a gcode to wait for PINDA to cool down or heat up. Which basically solves most needs. Also motocoder has brought in ambient temperature reading and is working on fan check. So we make marlin more cozy for Prusa users:-) |
|
Thanks for the info, I will in the future relate to the right topic :-) |
|
FYI I have split the overhang test in two parts:
|
|
Good for me ;) |
|
Today I finally got a breakthrough with a very good print of Gregsaun's overhang test "bear_overhang_50-70.stl". But when I tested with an old PLA filament that lay on the shelf, the printout was very good. I was a little surprised that this cheap "garbage filament" performed better than my expensive "quality filament". So my conclusion is that the filament AddNorth Basic PLA does not work for these tests. I will therefore see if I can find another "glossier" filament to use. AddNoth's filament gives a matte surface which may not work. |
|
@RH-Dreambox yeah I am at around 40 versions, this is painful :-D . I am improving the small shell test at the moment, should be pushed during the day. It is an interesting approach, not complete but I think it helps to decide which one is better. |
|
Funny that I also archived a perfect print with my cheapest PLA. But I did not do a comparison to my better plas yet |
|
In case anyone hadn't seen this already, this is the list from LDO of where you can get the various parts they sell through distributors. I picked up the 5k RPM fan through All3DMakers a while back: https://www.facebook.com/ldomotors/photos/a.1868172056829988/2216862281960962/?type=3&theater |
|
I swapped out my noctua for the sunon and the stock 5v blower for the LDO blower. WOW! What a difference. The sunon isn't much louder, but the LDO blower is quite a bit louder. There is also a 10C drop in temp when the blower kicks on. I think I'll be using a sock from now on. I haven't done the overhang test (I'm printing a bunch of other things I need), but these parts are printing better than before. |
|
@davidpesce you have to retune the PID values with the higher output part cooling fan so that it can correctly compensate for the temp drop. another great option to prevent the sharp drop is to use a silicone sock on your block. and finally you can upgrade to the copper e3d block which has a much higher thermal mass making the extra airflow easier to power through. personally I run a copper block, no sock, and a custom pid tune. (pid adjustment is super easy with klipper, not sure what it looks like on marlin/prusa fw) |
|
@trevjonez yeah, I redid the PID and it didn't help much at all. The sock worked perfectly! |
|
Hi all! Lovely thread - great discussion and actually entertaining reading... One thing I haven't seen mentioned that could explain some different results between different people's setups is ambient air temperature in the room near the printer. Do any of you have thermometers near your printers? and if so, would you mind sharing what your typical ambient air temp is during a print? Someone printing at 17°C, perhaps in the northern reaches of Europe or Canada is going to have an immediate 10°C cooling advantage over myself this time of year (27°C near my enclosure already!). |
|
I am going to close this one, massive thank you for everyone that participate to this very interesting chat! |
|
I know this isn't quite the right place, but I just wanted to mention that the LDO fan that Zaribo sells only runs at 4500rpm aswell. I've asked them. They also changed the product description to reflect that.
It seems some one batch of Prusa/LDO fans outputs bogus values over the sensor wire, which can lead to Fan Errors. I seem to have two fans from the same batch. They will show randomly-changing RPM from 1000 to over 5000. You can check the reported RPM in Support->Extruder Info. Does yours do the same? |
|
@haarp Read 19 posts above (from Hotpocketdeath) |
The comment following that one is interesting, so that's where fan pulsing at below 100% speed is coming from. However, that's a separate issue. The RPM readout problem I described happens even at 100% fan speed (where pulsing should not be a problem), and it only happens with two fans I bought from Prusa in January 2019. Older fans show the correct RPM. I do think that Prusa has had a bad batch at around that time. |
and others
Looks like you are slated to do a part cooling shroud.
Might I suggest this design as an option? https://www.thingiverse.com/thing:3063554
it will keep the fan vertical thus not having an issue with the wear and tear that the r3 angle might induce. And the design seems to be decently tested for performance already.
I loaded it up and compared with the mk3 parts and it looks like shifting the fan and shroud forward about 1.6-2mm gives it adequate clearance around the pinda.(2mm shown in image)
The text was updated successfully, but these errors were encountered: