Diamond hotends heatsink

Hi!

I plan to use a Diamond hotend (3 or 5 colors) in Lite6 config.

I'm wondering: are 3 (5) heatsinks really needed ? As PTFE tubing goes to the Diamond, what about making a custom heatsink (don't know yet how), in order to attach the Diamond on the carriage, and put the (all metal) pushfits directly on the Diamond?

Thanks,

---

Frédéric

very Needed...

You don't want hot plastic up your PTFE, that will really wreck your day.

If you try and extrude from one extruder and the other non used extruders PTFE tubes are empty or liquid plastic filled, then your attempted extrusion is just going to go up the PTFE and not out the hotend nozzle. (well it will try and escape threw all orifices it can)


they use Lite6 as its an easy solution.

If you look at all the angles involved down the centre of the Lite6, Some sort of single heat sink is going to be a nightmare to manufacture. (think CNC with > 4 axis)
Or 3d printed alu

This is a thing of nightmares...

Edited 4 time(s). Last edit at 09/12/2017 07:55AM by Dust.

As the heatsink arround PTFE is hotter than ambiant air, I don't see how it can cool down the filament better than ambiant air...

---

Frédéric

About the heatstink and mount, I'm thinking to something very simple, with screws and large washers...



Edited 1 time(s). Last edit at 09/12/2017 12:10PM by fma.

---

Frédéric

The point of the heatsink is to cool the part part of the heat break outside the heater block, so that filament temperature changes abruptly at that point from solid to molten.

The heatsink functions better than air because it conducts heat *much* better than air, so that the heat can be dissipated to air over a very large surface area.

The heatsink in your image won't cool filament at all, and you don't even show the heat breaks.

Probably the plastic parts in your Bowden connectors will melt.

You say that heatsink conducts much better than the air. I agree. So, it is also directly heated by the nozzle, because of the direct contact with it! Plus, PTFE is a good thermal insulator; so how can the heatsink cool down the filament? Don't forget that I'm talking about the Lite6 version, where there is no heatbreak inside the heatsink... and this Diamond version works fine.

Look at this hotend: [forums.reprap.org], there is not heatsink at all, and no need for a heatbreak, as carriage is in metal. And it works fine. And you can notice that the guy uses plastic caps pushfits: [forums.reprap.org]

BTW, I plan to put a fan on my 'heatsink' (more a heatbreak, in fact), which will also blow directly on the PTFE tubes. And, of course (and as I said), I will use full metal pushfits.

---

Frédéric

Quote
fma
So, it is also directly heated by the nozzle, because of the direct contact with it! Plus, PTFE is a good thermal insulator; so how can the heatsink cool down the filament? Don't forget that I'm talking about the Lite6 version, where there is no heatbreak inside the heatsink... and this Diamond version works fine.

The Lite6 *does* have a heat break [e3d-online.com] The heat break is made of stainless steel because that conducts heat relatively poorly. The heat break is also very thin, to reduce heat conduction. The heat sink screws onto the heat break, and does not contact the nozzle. The heat sink conducts heat away from the heat break (and therefore PTFE and filament) and dissipates it in the air via radiation. A fan blows/sucks to continually replace the warmed air with relatively cool air. The whole purpose of this is to create the steepesp possible thermal gradient across the heat break at the point where it enters the heater block.

Quote


Look at this hotend: [forums.reprap.org], there is not heatsink at all, and no need for a heatbreak, as carriage is in metal. And it works fine. And you can notice that the guy uses plastic caps pushfits: [forums.reprap.org]

BTW, I plan to put a fan on my 'heatsink' (more a heatbreak, in fact), which will also blow directly on the PTFE tubes. And, of course (and as I said), I will use full metal pushfits.

I don't know how that hotend works. I note that water cooling is mentioned. I note also that the guy was having trouble with it, I suggest that you contact him directly to get more information.

My purpose here was just to try to correct what appeared to me to be a misunderstanding about the purpose and function of the heatsink. I'm not sure why you ask for advice and then proceed to (incorrectly) refute the advice you receive.

If you believe that your idea will work, I suggest that you build it and see.

No, the Lite6 does not have heatbreak! Your are describing the E3Dv6 standard... The Lite6 heatsink is screwed directly on the heat block, and touches the nozzle. Then, the PTFE tube goes down to the nozzle too. I own a Lite6, so I'm sure of that.





That's why the Lite6 is limited to 250°C, where the E3Dv6 can go up to 300°C or more, because the PTFE never reaches that temperature, thanks to the heatbreak. That E3Dv6 heatbreak also ensures the filament does not melt above the heaterblock.

As you said, the best is to try. First, I will make a test with my Lite6 without fan on the heatsink (I will maintain it with a plier), and see if it still works. I'm pretty sure that the fan / heatsink are here to be able to maintain the hotend on plastic part; it acts as a heatbreak, but not for filament.

BTW, you're right, the hotend I pointed uses a water cooling. I missed that. So the pushfits may not reach a very high temperature. I didn't found metal pushfits with M6 thread; I will have to use a M8 or 1/8", and re-thread it.

Edited 3 time(s). Last edit at 09/13/2017 03:37AM by fma.

---

Frédéric

The heatbreak on the lite6 is integrated into the heatsink but it performs the same function of minimising the amount of plastic in the semisolid condition. The length of semisolid plastic (soft enough to expand and press on the walls of whatever's guiding it but not liquid enough to flow easily) is responsible for a significant amount of backpressure that the extruder has to overcome. Using a PTFE liner reduces this friction and insulates the filament from the heat that does pass through the heatbreak allowing the heatsink to be made of stainless steel in one piece with the heatbreak, reducing cost.

Idris

---

{Precision Piezo} Accurate, repeatable, versatile z-probe plus piezo discs, endstop cables, pt100, 50w heaters.

Quote

Using a PTFE liner reduces this friction and insulates the filament from the heat that does pass through the heatbreak

So, the heatsinkbreak tends to heat up the filament, not to cool it down! That's exactly what I said before. So, removing the heatsinkbreak, and have fresh air on the PTFE can only improve things.

Ok, I launched the test without fan on the heatsink: so far (almost an hour later), it still extrudes at 10mm/s! The heatsink is very hot at the bottom, but not yet hot at the top (as you say, stainless steel has not a very good thermal conduction; I can still hold it in my hand)... I will insulate it to speed up the process (and to avoid the air cooling it down).

---

Frédéric

I didn't find stuff to insulate the Lite6. But after 2 hours running at 210°C, the top part of the heatsink (groove mount?) was still cold! So, I will re-designed a support without all the fan duct, in order to print real things (the above test was just extrusion in the air), and see if the print remains correct.

I also ordered these connectors: [fr.aliexpress.com]. I will drill a custom heater block, with a metal support in order to make a hotend without any heatsink at all, as I plan to do with the Diamond.

---

Frédéric

Quote
fma

Quote

Using a PTFE liner reduces this friction and insulates the filament from the heat that does pass through the heatbreak

So, the heatsinkbreak tends to heat up the filament, not to cool it down! That's exactly what I said before. So, removing the heatsinkbreak, and have fresh air on the PTFE can only improve things.

Ok, I launched the test without fan on the heatsink: so far (almost an hour later), it still extrudes at 10mm/s! The heatsink is very hot at the bottom, but not yet hot at the top (as you say, stainless steel has not a very good thermal conduction; I can still hold it in my hand)... I will insulate it to speed up the process (and to avoid the air cooling it down).

I beg your pardon... I was looking at the wrong page for information on the Lite6. You're quite right; the heatbreak is integrated into the heatsink.

The bottom part of the heatsinkbreak is inside the heater-block, so will tend to heat up. That heat will conduct back up the metal, where it will be dissipated to air quite well, due to the large surface area of the heatsink fins. So there will be a stable temperature gradient, with the bottom part hot and the top part cold. If you replace that heatsink with just PTFE tube, then, because PTFE conducts heat worse than metal, there should be a good steep temperature gradient too. However, you lose all of that surface area. So it's possible (but not certain) that having fresh air on the PTFE will improve things. It will be interesting to see the results of your experiments. The conservative part of me says that all the Reprappers who have gone before believed that they needed the heatsink -- they wouldn't have put it on otherwise. The other part of me says that maybe they were all wrong. Good luck!

Bear in mind too that heat is conducted up the filament, so that it's possible/likely that the filament inside the PTFE will be warmer than the PTFE, especially closer to the heater block. When the filament is feeding fast, the speed of the filament feed is subtracted from the conduction rate, giving a steep temperature gradient. The worst case is when the filament is stopped, e.g. whilst moving the head without extruding. Heat will conduct up the filament, conduct/radiate to the PTFE, from where it goes to the air. If the last step is too slow, the dynamic balance of the heat gradient will shift, and the PTFE and filament temperatures will rise. If the filament temperature rises too far, it will soften, expand, and potentially jam.

Quote
frankvdh
The conservative part of me says that all the Reprappers who have gone before believed that they needed the heatsink -- they wouldn't have put it on otherwise. The other part of me says that maybe they were all wrong.

The E3Dv6 came first. This design, which can reach high temp, does need the heatsink, to cool down the heatbreak, and keep the PTFE cool. Then came the Lite6. My feeling is that they build it as a drop in replacement of the E3Dv6: same size and so, so all existing design of extruders can use it. There are a lot of example of things not really optimized, which never changed for years, because of compatibility. IMO, it is a non sense for makers, as printing new custom parts is really easy, and maybe one of the major interest in 3D printing! It is a little bit different for sellers, though, which need to sell as much products as possible.

Quote

Bear in mind too that heat is conducted up the filament, so that it's possible/likely that the filament inside the PTFE will be warmer than the PTFE, especially closer to the heater block.

Yes, sure, but I think it is also the case with the heatsink.

Quote

When the filament is feeding fast, the speed of the filament feed is subtracted from the conduction rate, giving a steep temperature gradient. The worst case is when the filament is stopped, e.g. whilst moving the head without extruding. Heat will conduct up the filament, conduct/radiate to the PTFE, from where it goes to the air. If the last step is too slow, the dynamic balance of the heat gradient will shift, and the PTFE and filament temperatures will rise. If the filament temperature rises too far, it will soften, expand, and potentially jam.

Right. But yesterday, I let the hotend heatup for a long time, with the filament inside, and I was still able to extrude (with a 3:1 reduced Bowden). Slower at the end of the test, as the heatsink was hotter and melted the filament closer to the top, but still fast enough for a quite fast print. And it never jammed. I printed the new mount, so I will be able to make real prints without fan (my previous fanduct design touches the heatsink all the way, so I couldn't just turn off the fan).

---

Frédéric

Quote

Then came the Lite6. My feeling is that they build it as a drop in replacement of the E3Dv6: same size and so, so all existing design of extruders can use it

the E3D team came up with the V6 lite to counteract the influx of clones on the market, and to give us budget builders/PLA users an affordable option for a genuine e3d hotend.

FMA what was the test print ? was there a lot of retraction and at what speed, am just thinking that if fila allowed to heat up through the tube then the fila becomes elastic resulting in a sluggish retract leading to hairy prints! your method could work with a primitive object with a fast print where heat creeping aint to much of an issue, but hows it fairing with slow detailed prints? .

now dont get me wrong am intrigued to how this goes and rooting for ya, but I have my doubts.. may I make a request for a test print cube time

Quote
jinx
the E3D team came up with the V6 lite to counteract the influx of clones on the market, and to give us budget builders/PLA users an affordable option for a genuine e3d hotend.

Yes, true.

Quote

FMA what was the test print ? was there a lot of retraction and at what speed, am just thinking that if fila allowed to heat up through the tube then the fila becomes elastic resulting in a sluggish retract leading to hairy prints! your method could work with a primitive object with a fast print where heat creeping aint to much of an issue, but hows it fairing with slow detailed prints?

now dont get me wrong am intrigued to how this goes and rooting for ya, but I have my doubts..

I didn't make real prints yet. As said, yesterday, I just removed the Lite6 from the fanduct (which also act as a mount), and extruded in the air, to test jamming. I printed a new mount, and I plan to print real things tonight. I will choose objets needing retractation (but I retract at each move, so every print retracts a lot).

Quote

may I make a request for a test print

cube time

Ok. I will first print it with my current design, to see differences. I will do that for all tests. If you have more suggestions (small objects, if possible), let me know.

But remember that these tests won't be done with the optimal settings: the final goal is to totally remove the heatsink, as it only act as mount, and uselessly heatup the filament.

---

Frédéric

Arrgl, the test failed. Not because of the hotend itself, but because of the PLA mount. This mount prevents the natural air flow I had when I tested the hotend in the air, so now, the top of the hotend becomes too hot for the PLA... I could print with some filament having a higher Tg, but I think it is better to build a new heater block, as I suggested, to address the 2 problems: mount temp. and heatsink heating up the PTFE all along. So I will be close to what I want to do with the Diamond.

I have to wait for the metal pushfits.

---

Frédéric

I'd try it with 3 throats with heatbreaks and try & mount the metal feed fittings on that if poss.

I just made this:



Stainless steel screws, alu heatsink... I should receive my metal pushfits on tuesday.

---

Frédéric

I made some tests yesterday: not an epic fail, but not an epic success neither. The problem is the length of the metal pushfit I'm using: it maintains the PTFE hot on a too long distance. So, retractation sucks, as it was mentioned.

I'm waiting for new connectors ('à olive', in french. Union?), but I'm not sure they will be better. Best would be something similar to E3D system, but with metal clip, directly inserted in the Diamond...

---

Frédéric

thinking even with new conns your still gonna have the same issue "isolating the heat in the heater block" the connectors will/are acting as the heatsinks with the current system. I be tempted to use an hybrid setup of e3d heat break and peek then see how small i can get it. and dont rely on ambient room temps for cooling just aint gonna happen PID gonna work that heater and heat just gonna climb the tube...
all credit to the hotend makers "current and past" they put a lot of R&D into the current generation of HE, what you doing now is pulling out the key aspect of cooling from the D...unless you immerse those connectors in a water cooled system, and that be adding complexity and baggage
think my first test would be see how far I could thread the ptfe tub then thread some peek on the tube then how small I could go.
2nd be a washer stack on a e3d heatbreak then the connector,still aint alot of active cooling.

" Best would be something similar to E3D system, but with metal clip, directly inserted in the Diamond..."
am having an hard time following that idea you got a pic yet?

I used for years a GRRF hotend using a PEEK heatsink (with PTFE, pretty close to the Lite6 design), without any active cooling. It was a problem to mount it, because the PEEK got very hot, but it was working fine. I had no issue (but I was not in a Bowden config).

About the E3D clip, I'm talking about this:



With such clip, the PTFE would not be heated up on a too long path...

BTW, have you ever see a Stratasys hotend? It is very surprising:



As you see, there is a very long path in the heatblock. And the tubes turn 90° right after the extruders, before entering in the heatblock!!! There is a huge fan blowing (through a big tube) on the rear of the heatblock. And catridge are 2x250W!!!

---

Frédéric

"have you ever see a Stratasys hotend? It is very surprising:"
no still cant find a decent image in the search for it

I guess they keep their technology away from public eyes

What is very strange is the very long path in the heater block. On the other hand, the fan blows directly on the filament, just before it enters in the tube. So the cold/hot transition is not that long.

I tried this solution (without the curve), but I think my stainless steel tube was not smooth enough inside, and I had a lot of friction... Don't know where to find such small smooth tubes. If you have an idea...

---

Frédéric

Just kind of stumbled across this thread. As a long time user of the Diamond 3 colour and more recently 5 colour, I think I know its little quirks quite well. Basically, like any hot end, you want to melt the filament in the chamber but directly above that melt zone, you need the filament to be as cold as possible. If it gets soft, it will simple buckle and bulge when you try and push it into the chamber and that in itself can cause a jam. However good the transition between hot and cold, there will always be a section of filament that is "warm" so in order to be able to push it into the hot end, it needs to be supported. Hence the need for some sort of tube. BTW, RepRap.me (the inventors) recommend using the Lite version of E3ds heat sinks rather than the V6 for "ease of extrusion" - their words, not mine. Generally, heat sinks do a better job of conducting heat away from the filament rather than simply blowing air over the tube because of the far greater surface area. Also, having a metal jacket around the PTFE tube, which could itself get warm and "soft" will support it. Another consideration is the need to retract filament for non-print moves and with a mixing hot end, it is important that all filaments are retracted together, not just the one that is in use. So all the filaments can be subjected to a lot of "pushing and pulling" so it needs to be well supported. For example, as part of my testing I wanted to know how many retract/unretract cycles there were on the "Arai the Dragon" that I was struggling to print in PET-G (see below for the reason why) and was surprised to find that it was in the order of 5,400 using notepad++ "find". That meant that the last filament was retracted and unretracted about 4,000 times before it started to move forward. Also, too much retraction can cause molten filament to be drawn up above the heat break where it will cool and solidify causing a jam. So I think it's important that there is a distinct heat break rather than a gradual transition from hot to cold and I wonder how this could be achieved by simply blowing air over the PTFE tube (without cooling the brass melt chamber and nozzle). That's why there is also a thermal blanket between the brass part and the heat sinks (to keep the brass part hot).

Having said all that, the Diamond 5 colour does have issues that the 3 colour doesn't, at least from what I've seen so far. The problem we have is that with multi coloured prints, the filaments can be left in the hot end, at print temperature, for hours before they are needed, (so without moving forward). From the testing I have done so far, it seems that there are heat creep issues at temperatures higher than 200 degC with the 5 colour. I'm not sure if the filament actually melts in the heat break zone or whether is just softens but the net result is that about 5mm or so of filament in the vicinity of the heat break "swells" enough to almost jam. My Titan extruders can just about overcome it but only by racking up the tension. This in itself can lead to problems of grinding the filament away. The problem doesn't appear to occur at temperatures below 200 degrees C which of course rules out printing 5 coloured parts in many filaments such as PET-G. PLA at around 190 deg C is fine. Compared to the 3 colour which doesn't seem to have these issues, the big differences are that the heat sinks are closer together which may affect the air flow around them, and also the lower fins of the heat sinks are of a smaller diameter in order to get 5 of them to fit around the same circumference, which may affect their ability to dissipate heat.

So all in all, I guess my concerns would be about how the filament is supported and how to keep to keep the brass nozzle hot if you simply blow air over the PTFE tube or filament. Oh and of course if you want to print at temperatures above about 240 deg C you didn't ought to be using PTFE tube. Having said all that, the Diamond does have room for improvement as regards heat creep (at least the 5 colour version) IMO.

Thanks for the precisions about Diamond 5...

That's said, I'm still not convinced what you guys are explaining about the heatsink keeping the PTFE cold. The Lite6 heatsink is heated up by the heatblock. Right? The fan tends to keep it not too much hot, so it does not itself heat up the PTFE (so the filament). Am I still right?

But the base of the heatsink is much more hot that the ambiant air. I guess it is cooler that the PTFE. So as there is a gradient, there, the PTFE gives its heat to the heatsink. That's said, I really don't know if this gradient is better than fresh air. And I don't know how you can affirm that, without any measures.

The tests I made do not lead to one conclusion or to another, as I'm still not in a good configuration: my metal pushfit is too long, so it is like having the heatsink without fan. Which works (see above, I made some extrusion). Sort of. The only problem is retract. And, as you mentioned, it can be a mess with Diamond.

So what about having a (tiny) heatsink arround the PTFE, just above the heatblock, but not touching it? With a little fan, it can only be better than the default config, as the heatsink will be colder. Right? But this can only work with a very short pushfit, to get a short cold/hot transition. This is my next move: making this pushfit. I may have some ideas to test, but I will need to have access to a lathe.

You may ask: what is the point? It is to remove those large heatsinks, badly used: they are heated up but we want to keep them cold! They only serve as PTFE guide, and mounting. And I'm convinced we can do those 2 parts in a much smaller space. Heatsinks around PTFE could only have a diam. of 10mm and 4 or 5 fins. And the mounting part could be done through 3 threaded rods, which could allow a better and more rigid solution to mount the hotend, rather than through the fan duct (and plastic necklaces!).

Having no rigid connection between heatsinks and heatblock may also allow us to straighten the PTFE direction, and have them nearly // above the hotend, which can also allow shorter Bowden tubes.

Anyway, this won't improve the retract problem you mentioned with the Diamond 5... Maybe such (pseudo) mixing hot end is not the way to go. I'm wondering if some guys are working on an ink system to change the color of the filament in a mixing chamber, but only using 1 white filament? This would be much easier... I don't know how pigmentation is done for our filament (powder?). Has it been discussed before, on these forums?

---

Frédéric

Various people have been working on ink-based colouring systems for quite a while, I think mostly focused on inkjet technology. A soon-to-be-released example is [www.xyzprinting.com]

Getting back closer to the Diamond approach, I wonder about an "asymmetric" hot-end. Have one large (e.g. 3mm?) base-colour filament, and several smaller (1.75mm? or maybe 1mm?) filaments. These smaller filaments would really just be solid dye-carriers, to change the hue of the base colour. Perhaps they might be a different material from the main filament, but mixable with it. The physical properties of the dye-carrier material don't matter much; it just has to behave well during under extrusion and retraction. Pack as much dye as possible into the carriers, so that they are "super-intense"; hopefully the dye density would be 20 times what is needed to make a rich solid colour. Mix by volume up to 5% of various dye-carriers into 95% base filament to get any hue and saturation desired.

A transparent base would produce all kinds of translucent colours, including solid-looking colours if the dye density and the mixing were good enough. With a white base colour, you could produce shades from pale pastels through to maybe intense solid colour. Kindof like the way paint shops mix paint colours to the customer's requirement.

Quote
frankvdh
Various people have been working on ink-based colouring systems for quite a while, I think mostly focused on inkjet technology. A soon-to-be-released example is [www.xyzprinting.com]

Interesting!

I also found this:

[3dprint.com]

Quote

Getting back closer to the Diamond approach, I wonder about an "asymmetric" hot-end. Have one large (e.g. 3mm?) base-colour filament, and several smaller (1.75mm? or maybe 1mm?) filaments. These smaller filaments would really just be solid dye-carriers, to change the hue of the base colour. Perhaps they might be a different material from the main filament, but mixable with it. The physical properties of the dye-carrier material don't matter much; it just has to behave well during under extrusion and retraction. Pack as much dye as possible into the carriers, so that they are "super-intense"; hopefully the dye density would be 20 times what is needed to make a rich solid colour. Mix by volume up to 5% of various dye-carriers into 95% base filament to get any hue and saturation desired.

A transparent base would produce all kinds of translucent colours, including solid-looking colours if the dye density and the mixing were good enough. With a white base colour, you could produce shades from pale pastels through to maybe intense solid colour. Kindof like the way paint shops mix paint colours to the customer's requirement.

What about using very very fine powder, as colors?

---

Frédéric

This week-end I printed a small part, and at the end of the print, I just realized that the E3D Lite6 fan was unplugged!

But the part printed without problem, without stringing or so. The PLA mount didn't suffer. Go figure...

---

Frédéric

Quote
fma
This week-end I printed a small part, and at the end of the print, I just realized that the E3D Lite6 fan was unplugged!

was hoping to find someone to do that, the e3d team mentions it the other day with a chat with Tomas that it was possible " just in case you missed it chat
what temp was you printing at?

The temperature was 200°C. What puzzle me is I made, a few days ago, a spacial support, to have the heatsink free, and I had issues with retract. Here, all was fine.

---

Frédéric