Diamond hotends heatsink
Posted by fma
|
Re: Diamond hotends heatsink October 02, 2017 04:33PM |
Registered: 8 years ago Posts: 622 |
I'll shortly be writing a post on my blog about my recent experiences with the 5 colour Diamond hot end. There are certain issues with "mixing" hot ends (that is to say those which have multiple inputs but a single output) that are caused by occasions where one of more filaments may be heated but unused for a long period of time - an hour or more depending on the object being printed. It is under these circumstances that strange things start to happen which "single filament" users will not have experienced.
I won't go into things like PLA starting to hydrolyse and other strange happenings, but one thing that might be pertinent to this particular thread is that the 5 colour suffers from an issue that I think is caused by heat creep and which does not happen with the 3 colour version. What happens is that over time if the filament does not move forward but is heated to print temperature, it tends to swell in the area of the heat break. It doesn't melt and then harden which can be caused by too much retraction drawing molten filament up into the cold section, but rather it just seems to swell to the diameter of the PTFE liner. With PLA, this doesn't seem to cause a problem but with PETG, the filament tends to "bind" and it takes a lot of extruder tension to push this "swollen slug" of about 5mm through into the hot end. After which, it flows freely again until it sits and soaks at print temperature for an hour or so, whereby the problem will re-occur. The trouble is, setting a high extruder tension for the entire print is not a viable option as it grinds the filament away.
Apart from the fact that the 5 colour Diamond has two more heat sinks than the 3 colour, in order to fit these heat sinks into the same circumference, the lower fins have been reduced in diameter. So one theory which might explain why the 5 colour has the problem and the 3 colour does not, is that the reduced diameter of the heat sink fins has had a negative impact of the heat transfer efficiency (from the cold side of the heat break). I tried to measure the temperature in that area by putting a thermocouple between the lowest two fins. It was in contact with the lower fin but only just touching so is not by any means an accurate measurement. However, with the hot end heated to around 200 deg C, the temperature of the 5 colour small fins was around 66 degrees C but with the 3 colour and the larger fins it was around 42 degrees C.
I think that a higher air flow over the heat sinks might be the solution. However, I'm using the highest flow rate 24V 50mm fan that I can source so my next step is to modify the mount to take a 60mm fan which will enable me to use 3 to 4 times the airflow that RepRap.me recommend. I'm also re-designing the fan ducts and shrouds. I'll let you know how I get on but probably via my blog rather than these forums.
[somei3deas.wordpress.com]
[www.youtube.com]
I won't go into things like PLA starting to hydrolyse and other strange happenings, but one thing that might be pertinent to this particular thread is that the 5 colour suffers from an issue that I think is caused by heat creep and which does not happen with the 3 colour version. What happens is that over time if the filament does not move forward but is heated to print temperature, it tends to swell in the area of the heat break. It doesn't melt and then harden which can be caused by too much retraction drawing molten filament up into the cold section, but rather it just seems to swell to the diameter of the PTFE liner. With PLA, this doesn't seem to cause a problem but with PETG, the filament tends to "bind" and it takes a lot of extruder tension to push this "swollen slug" of about 5mm through into the hot end. After which, it flows freely again until it sits and soaks at print temperature for an hour or so, whereby the problem will re-occur. The trouble is, setting a high extruder tension for the entire print is not a viable option as it grinds the filament away.
Apart from the fact that the 5 colour Diamond has two more heat sinks than the 3 colour, in order to fit these heat sinks into the same circumference, the lower fins have been reduced in diameter. So one theory which might explain why the 5 colour has the problem and the 3 colour does not, is that the reduced diameter of the heat sink fins has had a negative impact of the heat transfer efficiency (from the cold side of the heat break). I tried to measure the temperature in that area by putting a thermocouple between the lowest two fins. It was in contact with the lower fin but only just touching so is not by any means an accurate measurement. However, with the hot end heated to around 200 deg C, the temperature of the 5 colour small fins was around 66 degrees C but with the 3 colour and the larger fins it was around 42 degrees C.
I think that a higher air flow over the heat sinks might be the solution. However, I'm using the highest flow rate 24V 50mm fan that I can source so my next step is to modify the mount to take a 60mm fan which will enable me to use 3 to 4 times the airflow that RepRap.me recommend. I'm also re-designing the fan ducts and shrouds. I'll let you know how I get on but probably via my blog rather than these forums.
[somei3deas.wordpress.com]
[www.youtube.com]
|
Re: Diamond hotends heatsink October 02, 2017 04:45PM |
Registered: 13 years ago Posts: 300 |
Thanks for the feedback! I was monitoring you blog for that 
What about trying water cooling? A simple solution would be to use a little silicone tube wrapped around the base of the heatsinks... There won't be a good thermal contact, but it may help. A copper maidenhair would be better, but you will need a powerfull pump...
Frédéric
What about trying water cooling? A simple solution would be to use a little silicone tube wrapped around the base of the heatsinks... There won't be a good thermal contact, but it may help. A copper maidenhair would be better, but you will need a powerfull pump...
Frédéric
|
Re: Diamond hotends heatsink October 02, 2017 08:46PM |
Registered: 9 years ago Posts: 978 |
Quote
deckingman
... one of more filaments may be heated but unused for a long period of time - an hour or more depending on the object being printed. It is under these circumstances that strange things start to happen which "single filament" users will not have experienced.
It seems to me that a workaround shouldn't be difficult -- ensure that every 30 minutes or so, all filaments are extruded by a few mm. This could be done as part of infill, or on a separate 'wipe' tower. Probably a g-code post-processor for Slic3r or Cura could be written?
|
Re: Diamond hotends heatsink October 03, 2017 03:58AM |
Registered: 8 years ago Posts: 622 |
Quote
frankvdh
Quote
deckingman
... one of more filaments may be heated but unused for a long period of time - an hour or more depending on the object being printed. It is under these circumstances that strange things start to happen which "single filament" users will not have experienced.
It seems to me that a workaround shouldn't be difficult -- ensure that every 30 minutes or so, all filaments are extruded by a few mm. This could be done as part of infill, or on a separate 'wipe' tower. Probably a g-code post-processor for Slic3r or Cura could be written?
Yes for sure, and I've already written a little script to purge filaments at tool change. This seems to be essential when using PLA as it start to hydrolyse (gets more and more runny over time). So when it comes time to use a particular filament, what has been sitting in the melt chamber for a long time, it all comes out at once as runny blob. That's a bit of a bugger because it impacts on my idea of advancing the tool change point in the gcode file so that one can print without using purge towers. The script itself works fine and I'm so close but for this issue. PET-G works much better in this respect (that is, it doesn't hydrolyse as much) and I'm looking at experiments with other filaments such as ASA. My goal (which may not be achievable) is to be able to print multi-colour without using purge towers.
The filament swelling in the heat break area could also be addressed by periodic purging but it's wasteful of filament and increases print time. So I'd rather address the fundamental cause rather than accept it and come up with some sort of work around. Hence the plan to try a higher flow rate fan with modified ducts and shrouds.
Edited 1 time(s). Last edit at 10/03/2017 04:00AM by deckingman.
|
Re: Diamond hotends heatsink October 05, 2017 09:11AM |
Registered: 13 years ago Posts: 300 |
Ian, I just read your last blog post, about the Diamond 'full' color. Very impressive feedback!
[somei3deas.wordpress.com]
About using more perimeters to allow the purge, you could also try to print infill first, so you have plenty of filament length to do that purge, so 2 perimeters should work. Except for the last 2 layers of each top surface where you need to purge outside of the print (I already made the system you described, purging in a box, with a metallic brush to clean the nozzle - the brush was in an old Stratasys). How to detect the last 2 layers of top surfaces is another story! Best way should be to do that in the slicer, but you may find a solution to do that with your scripts...
About the Diamond fan, how is oriented the air flow? From top to bottom (blowing on heatsinks), or from bottom to top (extracting air through the heatsinks)? I always use the second config, as I read a paper (not about 3D printing, but for electronic in enclosures) explaining that if you blow on obstacles, you may not have the air running on all opposite surface of the blower (dead zone). By extracting air, you in fact reduce the pressure, sucking the air, which then go around all surfaces in a better way. It's worth a try... Or build a water-cooled stuff...
Again, thanks for these valuable feedback about the Diamonds! If I go that way, it will save me (and other people) hours of headache!
Edited 1 time(s). Last edit at 10/06/2017 12:55AM by fma.
Frédéric
[somei3deas.wordpress.com]
About using more perimeters to allow the purge, you could also try to print infill first, so you have plenty of filament length to do that purge, so 2 perimeters should work. Except for the last 2 layers of each top surface where you need to purge outside of the print (I already made the system you described, purging in a box, with a metallic brush to clean the nozzle - the brush was in an old Stratasys). How to detect the last 2 layers of top surfaces is another story! Best way should be to do that in the slicer, but you may find a solution to do that with your scripts...
About the Diamond fan, how is oriented the air flow? From top to bottom (blowing on heatsinks), or from bottom to top (extracting air through the heatsinks)? I always use the second config, as I read a paper (not about 3D printing, but for electronic in enclosures) explaining that if you blow on obstacles, you may not have the air running on all opposite surface of the blower (dead zone). By extracting air, you in fact reduce the pressure, sucking the air, which then go around all surfaces in a better way. It's worth a try... Or build a water-cooled stuff...
Again, thanks for these valuable feedback about the Diamonds! If I go that way, it will save me (and other people) hours of headache!
Edited 1 time(s). Last edit at 10/06/2017 12:55AM by fma.
Frédéric
|
Re: Diamond hotends heatsink October 06, 2017 11:01AM |
Registered: 8 years ago Posts: 622 |
Frederic,
Glad you appreciate the blog. These things take a fair bit of time and effort to write up so I appreciate the feedback.
Also, yes the Diamond hot end isn't perfect by any means but in my opinion, it's not as bad as some approaches to printing multi colour. Until something better comes along, it's the best of a bad bunch IMO. At least it has taught me a lot and I think I know what's needed to make a better mixing hot end but I don't have the time or resources to make one myself.
I had considered reversing the air flow (which is normally top to bottom) but I discounted it on the grounds that the air would take the easiest path and so most of the flow would be drawn through the upper fins because they are closer to the fan. From what I've managed to glean, it's the lower fins that dissipate most of the heat, which makes sense when you think about it, so this is where a higher air flow will be of most benefit.
The fan(s) I am looking at using have a very high airflow but also quite a high static back pressure. That means that I can mask off some areas and channel the air flow to where it's needed. The thing I need to be careful of is not blowing too much air over the brass block and especially, not blowing cold air over the heater cartridge which is in the centre. That would just to wreak havoc. So like the 3 colour version, there are shrouds in the areas between the heat sinks where they don't touch, to channel the airflow through the fins, but unlike the 3 colour version, there is also a central vertical tube to mask the top of the brass block and the heater cartridge. It's difficult to explain but I'll post pictures on my blog. Hopefully, I'll be in a position to do some tests soon.
I've considered water cooling but because the heat sinks almost touch, there isn't any room to fit jackets around them. So it would have to be a custom cooling block and then how do you connect that to the 5 inputs which are angled at 22 degrees? It'd probably mean abandoning the brass part altogether so you might as well design a whole new hot end from scratch (one that actively mixes and has replaceable nozzles would be good). On the other hand, water cooling just means that you are transferring heat to water instead of to air. Both are fluids and unless you refrigerate the water somehow, both are at ambient temperature. So I think the main reason for using water cooling would be in an enclosed chamber where the air temperature is higher than ambient.
If the alternative fan doesn't work, I'll likely just give up on the 5 colour and stick with the 3 colour version.
Ian
[somei3deas.wordpress.com]
[www.youtube.com]
Glad you appreciate the blog. These things take a fair bit of time and effort to write up so I appreciate the feedback.
Also, yes the Diamond hot end isn't perfect by any means but in my opinion, it's not as bad as some approaches to printing multi colour. Until something better comes along, it's the best of a bad bunch IMO. At least it has taught me a lot and I think I know what's needed to make a better mixing hot end but I don't have the time or resources to make one myself.
I had considered reversing the air flow (which is normally top to bottom) but I discounted it on the grounds that the air would take the easiest path and so most of the flow would be drawn through the upper fins because they are closer to the fan. From what I've managed to glean, it's the lower fins that dissipate most of the heat, which makes sense when you think about it, so this is where a higher air flow will be of most benefit.
The fan(s) I am looking at using have a very high airflow but also quite a high static back pressure. That means that I can mask off some areas and channel the air flow to where it's needed. The thing I need to be careful of is not blowing too much air over the brass block and especially, not blowing cold air over the heater cartridge which is in the centre. That would just to wreak havoc. So like the 3 colour version, there are shrouds in the areas between the heat sinks where they don't touch, to channel the airflow through the fins, but unlike the 3 colour version, there is also a central vertical tube to mask the top of the brass block and the heater cartridge. It's difficult to explain but I'll post pictures on my blog. Hopefully, I'll be in a position to do some tests soon.
I've considered water cooling but because the heat sinks almost touch, there isn't any room to fit jackets around them. So it would have to be a custom cooling block and then how do you connect that to the 5 inputs which are angled at 22 degrees? It'd probably mean abandoning the brass part altogether so you might as well design a whole new hot end from scratch (one that actively mixes and has replaceable nozzles would be good). On the other hand, water cooling just means that you are transferring heat to water instead of to air. Both are fluids and unless you refrigerate the water somehow, both are at ambient temperature. So I think the main reason for using water cooling would be in an enclosed chamber where the air temperature is higher than ambient.
If the alternative fan doesn't work, I'll likely just give up on the 5 colour and stick with the 3 colour version.
Ian
[somei3deas.wordpress.com]
[www.youtube.com]
|
Re: Diamond hotends heatsink October 15, 2017 05:32AM |
Registered: 13 years ago Posts: 300 |
Ok, I found a way to insert the PTFE into the Diamond hotend with a very short stuff:
My first idea was to drill/thread a M6x6 hexagonal screw, but as I ordered a few weeks ago these fits, I just used one to make the job without Lathe.
I just cut the top part (press fit), and threaded the bottom M6 part: the nice thing is the hole was already 3.4mm, so perfect for M4! And as it is brass, it is very easy to do (stainless steel would be better for heat, and I will try my first idea using a M6x6 screw if this one does not work). Then, I just had to screw the PTFE tube in it, and screwed this part on the Diamond.
I will first make a test with a custom heater block as Lite6-like head, and see how it works. It may require a fan to blow on the fits (the insulator should prevent the diamond to cool down); but maybe not.
For the diamond, I plan to use the 3 lateral M3 holes to attach it (these holes are here for cartridge/thermistor fix; long headless screws and nuts for the fixation.
Frédéric
My first idea was to drill/thread a M6x6 hexagonal screw, but as I ordered a few weeks ago these fits, I just used one to make the job without Lathe.
I just cut the top part (press fit), and threaded the bottom M6 part: the nice thing is the hole was already 3.4mm, so perfect for M4! And as it is brass, it is very easy to do (stainless steel would be better for heat, and I will try my first idea using a M6x6 screw if this one does not work). Then, I just had to screw the PTFE tube in it, and screwed this part on the Diamond.
I will first make a test with a custom heater block as Lite6-like head, and see how it works. It may require a fan to blow on the fits (the insulator should prevent the diamond to cool down); but maybe not.
For the diamond, I plan to use the 3 lateral M3 holes to attach it (these holes are here for cartridge/thermistor fix; long headless screws and nuts for the fixation.
Frédéric
|
Re: Diamond hotends heatsink October 16, 2017 02:28PM |
Registered: 13 years ago Posts: 300 |
|
Re: Diamond hotends heatsink October 16, 2017 03:41PM |
Registered: 8 years ago Posts: 622 |
|
Re: Diamond hotends heatsink October 16, 2017 03:46PM |
Registered: 9 years ago Posts: 978 |
[www.aliexpress.com] lists
Teflon tube 030060PT 3*6mm 3mm 6mm Gas or liquid; `-190 to 280 degrees Celsius
If 3mm ID is too big, you could maybe use a 2*3mm tube inside that?
Teflon tube 020030PT 2*3mm 2mm 3mm Gas or liquid; `-190 to 280 degrees Celsius
Or probably better would be a 2*2.9mm metal inner tube. [www.aliexpress.com]
AFAICT, PTFE is only sold in metric sizes.
Teflon tube 030060PT 3*6mm 3mm 6mm Gas or liquid; `-190 to 280 degrees Celsius
If 3mm ID is too big, you could maybe use a 2*3mm tube inside that?
Teflon tube 020030PT 2*3mm 2mm 3mm Gas or liquid; `-190 to 280 degrees Celsius
Or probably better would be a 2*2.9mm metal inner tube. [www.aliexpress.com]
AFAICT, PTFE is only sold in metric sizes.
|
Re: Diamond hotends heatsink October 16, 2017 03:52PM |
Registered: 13 years ago Posts: 300 |
Quote
deckingman
I'm still not sure where you are going with this? Why don't you want to use the standard configuration and just fit 3 lite 6 heat sinks? They work fine on the 3 colour version.
Yes, but if they are only usefull for fixation, why bother with such huge piece of metal? And if it works, it could solve your problems with the 5 colours... And make a cheaper config.
Edited 1 time(s). Last edit at 10/16/2017 03:58PM by fma.
Frédéric
|
Re: Diamond hotends heatsink October 16, 2017 03:57PM |
Registered: 13 years ago Posts: 300 |
Quote
frankvdh
[www.aliexpress.com] lists
Teflon tube 030060PT 3*6mm 3mm 6mm Gas or liquid; `-190 to 280 degrees Celsius
If 3mm ID is too big, you could maybe use a 2*3mm tube inside that?
Teflon tube 020030PT 2*3mm 2mm 3mm Gas or liquid; `-190 to 280 degrees Celsius
Thanks for the links and suggestion of tube in tube!
You're suggesting to put a short brass tube inside the 6x3 PTFE tube, in the Diamond threaded space? And then go up to the extruder with the 3x2 PTFE tube?
Quote
AFAICT, PTFE is only sold in metric sizes.
I once had a small piece of 6.35x2 PTFE tube... But I don't know where it came from.
Frédéric
|
Re: Diamond hotends heatsink October 16, 2017 07:32PM |
Registered: 9 years ago Posts: 978 |
Quote
fma
You're suggesting to put a short brass tube inside the 6x3 PTFE tube, in the Diamond threaded space?
Yes. My thinking is that 6mm PTFE might not have much depth into an M6 thread, so the brass tube would stop the PTFE from collapsing and therefore stay in the fitting better. Also that pushing a 3mm OD PTFE into a 3mm ID PTFE tube would be difficult (although I've never tried it).
Quote
And then go up to the extruder with the 3x2 PTFE tube?
I hadn't even thought about how to get from there to the extruder.
|
Re: Diamond hotends heatsink October 17, 2017 01:09AM |
Registered: 13 years ago Posts: 300 |
Quote
frankvdh
My thinking is that 6mm PTFE might not have much depth into an M6 thread, so the brass tube would stop the PTFE from collapsing and therefore stay in the fitting better.
Well, it seems that some people are using threaded 4mm PTFE on the extruder side. The problem on hotend side is the temperature, which may soften the PTFE. I will test this config next week. If it does not last long enough, I will try 6mm. But looking at this page, it appears that the effort is not that important. I will statically test how much load I can handle with the 4mm threaded PTFE...
Frédéric
|
Re: Diamond hotends heatsink October 17, 2017 03:11AM |
Registered: 8 years ago Posts: 622 |
Quote
fma
Quote
deckingman
I'm still not sure where you are going with this? Why don't you want to use the standard configuration and just fit 3 lite 6 heat sinks? They work fine on the 3 colour version.
Yes, but if they are only usefull for fixation, why bother with such huge piece of metal? And if it works, it could solve your problems with the 5 colours... And make a cheaper config.
Seriously? You think E3D only fit heat sinks to their V6 as a way of fixing it? The 5 colour Diamond has a problem with heat creep and my latest tests (https://somei3deas.wordpress.com/2017/10/14/the-diamond-5-colour-part-2/) seem to confirm my theory that it's because the lower fins of the 5 colour heat sinks have a reduced surface area compared to the 3 colour version. You are proposing to reduce that surface area to nothing by doing away with them altogether. Even if you manage to find an alternative to heat sinks, you then have to devise a way to mount it all rigidly. Seems like a lot if time and effort to find an alternative to something that already works but hey, if it makes you happy who am I to criticise.
|
Re: Diamond hotends heatsink October 17, 2017 04:10AM |
Registered: 13 years ago Posts: 300 |
Not the V6, the *Lite6*... Yes, I do think that the *Lite6* heatsink is mainly here to be able to mount the hotend. It is a low-cost drop-in replacement of the V6. The heatsink as the same dimensions so you can use the same mounts, extruders... The V6 heatsink is another story, as it is not the same design.
Frédéric
Frédéric
Sorry, only registered users may post in this forum.