Are there any dis-advantages to extruding the aluminum heatsinks rather than completely CNC’ing them on a lathe? This way it saves on costs that can be passed down.
Are there any dis-advantages to extruding the aluminum heatsinks rather than completely CNC'ing them
Just the cost of a custom die. And to extrude it with fins in the proper orientation (perpendicular to the filament path) would be very difficult.
Extrusion is all about volumes. Not sure the market can handle the volume required to make it viable.
Die cost is not a problem (for us at least). In the long run this process saves money. Regarding the fins, if the extrusion design is balanced as it is in this case, I don’t think there should be issues with the fins extruding. The only potential issue would be the die becoming un-balanced due to uneven forces and reducing its life span therefore, but again, I don’t think this applies to this case since the heatsinks are uniform.
@Whosa_whatsis @ThantiK Feedback? Ideas?
Assuming you mean for a hot end?
A lathe allows you to make circular fins around the filament path, with lots of surface area and an axial thermal break between each fin and the next. Extruded heatsinks would be easy to make with fins running end-to-end, but that makes the airflow more problematic and conducts heat between the top and bottom of the heatsink.
You could also use a section with horizontal fins, and drill a hole through a central core, but you’d still get more top-bottom heat conduction, and because it would mean that you couldn’t produce them by bar-feeding your stock, it would likely negate any cost savings.
@Whosa_whatsis Yeah the top and bottom I had concerns about because it would be flat up against the fan and block airflow. The advantage of extruding, is you can get thinner fins potentially and therefore stack more fins together. Here’s a part that was extruded: http://shop.dyzedesign.com/product/full-kit-dyzend-dyzextruder-accessories-107385
Yeah, like I said, you can build one that way, but you lose the thermal isolation top-to-bottom, and the machining that you still have to do becomes more difficult.
There are also diminishing returns on thin fins. If they can’t conduct heat to the ends, making them thinner so that you can add more doesn’t help.
Yes, that is very true.
Is it possible to twist an extrusion as it comes out of the die? Spiral fins…
Crazy thought, making one big, helical fin with the same thickness and spacing as separate circular fins should perform almost identically, but it may be stronger (and thus able to be made thinner), and it may be possible to improve the efficiency of the process by using the techniques used for cutting screw threads…
edit: Ha, @Eric_Cha had almost the same idea as I was typing, though even if you could twist an extrusion, I don’t think you could twist an extrusion enough to have the desired effect.
I use extruded cpu heatsinks for the heads I make (pentium3). The fans are vertical with 2 holes either side for dual heads. The heads are drilled heatsinks are drilled 4mm so the bowden tube runs through it upto the barrel of the hotend. I have no cooling problems the vertical fins mean that airflow is directed down onto the heat block I cover the opening with kapton tape so air flows up only, I have done it so it blows down and then ducted to the tip for cooling but a separate fan is much simpler. The blocks never get warm unless the fan fails.
The is a picture of one on my website http://3done.co.za in the slider on the main page.
@Whosa_whatsis I also just now typed a while big thing about revolved extrudate and the advantage it could bring. . Without reading on. Thought I was on to something big haha. Glad I’m not alone though.
Edit: can we use a fan that blows down the “screw” on to the print, directly from above? Assuming we can keep it off the heat break/block
I think what @Shai_Schechter was suggesting was to run the filament path across the extrusion not along the axis, this would keep the fins perpendicular to the filament path, the only question I would have is how the central section of the extrusion where the filament path is would affect heat transfer.
@Shai_Schechter
Just an idea : what about high pressure Aluminium castig. The tool price is almost the same. The cycle time is slightly longer than extrusion. But you’ll get very accurate shapes, no material waste and you are relatively flexible with the design. We use this process for some parts in our company where the requirements are comparable with hotend heatsinks. And we were able to replace machining with this process.
U get a lot of chinese clone e3d hotends which are diecasted. You could order some and test them .
I have a box of LED laser heat sinks I ordered to try this- all sorts of sizes and shapes. I think it has merit, but gotta build one to see. You can buy them with solid centers to drill out and tap. The fan - if it needs one, would need to go from bottom to top- a challenge. Probably needs to have Teflon liner though. Carl is working on a Ubis ceramic replacement - sort of a hybrid… May be a candidate for this, not sure.
We also have a new tip we are testing … Fun hot end announcements to come
@Whosa_whatsis Why do you think the alternating fin/break design is optimal? If you have a proper heat-break between the heatsink and hot block to constrict conduction, and if the goal is to provide maximum cooling at the bottom fin, what you’d want is a big heat pipe to carry heat from the transition zone to as many fins as possible. (I’ve never understood this part of E3D’s designs – they’re very sensitive to bottom-fin cooling, and that seems like a design flaw.)
Compare the Replicator 1/2 style cooling bar where a big piece of aluminum stock first carries heat away from the thermal barrier tube, and then the entire cooling bar is uniformly cooled by an attached heatsink. It works just as well, if not better in my experience, than the E3D circular fin approach.
@Ryan_Carlyle That design also has a relatively short zone where the aluminum is in contact with the barrel. I have anecdotal evidence that this design doesn’t perform as well if you make that aluminum block taller so that it contacts more of the barrel (which is a bit counter-intuitive). The trick is to get the heatsink as cool as possible, as abruptly as possible, and using fins that are more thermally isolated from one another while still being coupled strongly to the barrel seems to help with that.
I’m also partially basing it on the fact that the heatsinks on the chinese E3D clones, which always seem to have a thicker core around the barrel than the genuine ones, don’t seem to perform as well if you control for quality of the barrel screwed into them.
Surface area of the fins, and the ability to conduct heat away from the barrel and into them is what’s most important. I discussed the MK7 design with the guy who designed it at MBI, and he agreed that the heatsink probably wasn’t doing much, and that a strategically-cut flat aluminum plate would probably be just as effective, if not more so. The heatsinks they used were actually pre-existing parts for something else that they found and were able to source easily, which is why they are missing more of their fins than they should be for this application.
One advantage the later replicators I’ve seen seem to have (not sure if it was a stock part or third-party) is that they clamp a smooth barrel rather than using screw threads. I always thought that the screw threads were sub-optimal because only one side of each thread will really be in strong contact. In early versions of the design, the whole thing was mounted using a jam nut on the threads between the thermal break and the heatsink, clamping to a steel plate (and, at least in some builds, there would be jam nuts top and bottom, which could potentially act against one another and cause neither side of the threads to have optimal contact with the cooling block). I always thought this design would work better if the nut and plate were made of a aluminum to conduct heat away from the area just above the thermal break. Of course, a clamping a piece of hypodermic tubing should be even more optimal (especially from a manufacturing standpoint), and I never understood why the designs that clamped the barrel still kept the clamped section at a diameter of 6mm when they removed the threads. It seems like this would just serve the increase the thermal isolation between the inner and outer walls of the barrel.
@Whosa_whatsis Based on years of experimentation by Jetguy and others, the threaded thermal barrier with MORE thread engagement length performs better than smooth tubes or shorter lengths. Specifically, the optimal arrangement seems to be a single brass nut on the bottom of the cooling bar, which produces maximum contact force on the threads near the heat break, and then as much aluminum bar thickness as practical. Thinner bars demonstrably lead to PLA jams.
You’re absolutely right about using both top+bottom nuts, Rich Webb recently proved the theory about causing the threads to float and make poor contact.
The problem with smooth-wall clamped tubes is that you can never get enough contact stress to fully conform the surfaces to each other. A torqued thread is self-centered and has sufficiently high contact stress on the flanks to flatten surface asperities to produce a good continuous contact helix with low thermal resistance. Whereas a smooth clamp approach has between 1 to 4 narrow, low-stress longitudinal contact lines and maybe a set screw. Surface asperities greatly reduce heat transfer because it’s impossible to achieve very high clamping stresses. The clamped smooth barrel of the R2x is practically incapable of printing PLA, and the set-screw’d smooth barrel of the widely-used mk9 and mk10 clone hot ends required the addition of a PTFE liner to print PLA. The only really successful PLA-friendly smooth-bore designs I’ve seen (like the Polystroooder) use extremely low-conductivity heat breaks to reduce the necessary heat flux through the heat sink and internal non-stick coatings to compensate for the corresponding loss of heat-shedding from the filament across the cold zone of the heat-break.