v3 hotend!

[alaris2 1]

so UM took a long time getting the v2 hotend to us, and worse than that, we had no engagement with the community, no prior warning and there were lots of grumbles and complaints about popped bowden tubes etc. for months and months.

then destroyer set up a very good thread as a wishlist for what we really wanted the hotend to do and I, for one, wanted to print with something a bit more exciting than PLA anyway.

I have been watching an interesting thread develop in the reprap forum regarding an all metal hotend, and decided to talk to the inventors and persuade them to do a 3mm version that would be UM compatible (the original was 1.75mm).

They did this and at exactly the same time UM promptly released the v2 hotend.

several people have still been grumbling about the high cost of the v2 hotend, the fact it doesn't address all of the problems and the fact it doesn't come with heater or thermocouple (ie. can't be used as a 2nd extruder since you have to disassemble the first to reclaim some of the parts). I was disappointed to see the continued use of the PEEK part and no active cold side (I do like snowygrouch's solution - see viewtopic.php?f=6&t=1368)

anyway, to cut a long story short, this is what's sitting on my workbench right now...

top left is the v1 hotend, for comparison only. bottom left is the v2 hotend as supplied, cost approx. £100 from UM, right hand side is the UM compatible 'v3' hotend from E3D, cost £50.

note they use a resistor for heating (which I'll replace with the superior UM heater cartridge to maintain the print speed).

height is the same as the v2 hotend, heat break is stainless steel, top heat block is actually an active cold side and there's a 3mm polished hole down to the nozzle which is a 0.4mm in this case.

Phut and I had an interesting discussion on printing with atypical materials - with an all metal hotend and active cooling, there's no reason why you can't print nylon, PC, perhaps even ptfe? at any rate, you're no longer limited by the hotend, but by the physical properties of the materials you wish to extrude.

I'm about to put all this together and will then do some side by side comparisons of the 3 hotends if anyone's interested in photos or more details?

[IRobertI 520]

You're wondering if anyone is interested in more pictures and info? What a silly question! Of course we want more info and pics

[destroyer2012 0]

Truly I'm amazed ultimaker is charging so much for the V2, when others with even more machining done are charging less (see: arcol.hu hot end; four machined parts in all metal and theirs comes with a heater resistor and thermistor).

I think the metal hot end is the way forward! Although again nobody has done any research or calculations on what is the best hot zone length and orifice length to optimize retraction at high speeds. There's been lots done on the hobbed bolt side, but nothing on the hot end. Why?

It's also possible that these numbers are different for different materials. Maybe (now this may be really pie in the sky) a hot end with an adjustable hot zone would be good? Maybe you could move the location of the fan along the heater barrel to determine where the hot zone ends? This might not be needed or practical but it's what comes to mind . Could be a good setup for research purposes too...

Keep up the good work, alaris2! Let us know how it goes! Meanwhile I am upgrading my bowden tube setup once again..

[phut 0]

The base V2 hot end kit at 56 EUR isn't too bad but it's very much the shipping that kills the value for money in my case. I'm frankly more bitter that I bought the spare hot end set when I bought my printer (And that was not long ago) and then the V2 hot end comes out right round the corner with no reprise for those who might have bought into spare V1 parts :/

From the looks of it where the materials are pushing up the costs will be the TFE and PEEK parts - the PEEK part must have cost a pretty penny to setup being moulded (?) and TFE in general is a bit more expensive to obtain and machine than most metals.

[snowygrouch 1]

I will keep an eye on this ! Looks interesting.

I would still say that a non-metallic insulator is nice, otherwise the hotend cooler will be extracting a great

deal of heat, which will end up in the hotend drawing more Amps to keep hot.

http://www.technicalglass.de/pdf/macor_ ... ramics.pdf

Perhaps this might be more suitable than PEEK.

Thermal conductivities for reference.

PEEK = 0.25 W/mK

Macor (machinable ceramic) 1.5

Stainless Steel 16

Brass 110

Aluminium 237

[alaris2 1]

I hadn't spotted the arcol hotend - nice work but probably not suited for extruding nylon (which is where I'm taking this hotend) at 320C. the use of heater resistor is very common in the reprap community.

I haven't worked out the wattage to see how it compares with the 40W heater cartridge but I note the heater block is smaller in all cases (to match the size of the resistor). that seems to suggest a limit to the maximum speed which would figure since repraps tend to run at slower speeds.

having said that, the E3D boys mention 100mm/s which is nicely in UM territory, so maybe not? but their filament is usually 1.75mm, so maybe so..

at any rate, they've done calculations and tests with this hotend, so it's not total guesswork and their approach to design and test was not unlike Bertho's which was why I liked what i saw.

I did look at ceramic and liked the idea, but the availability and machinability were issues - not all of us have the access to, or tools to work with exotic materials.

the good thing about the stainless steel heat break is it doesn't conduct much heat (as snowygrouchs numbers show) so the heat from the heater block mostly goes to the nozzle. what little goes up is removed by the fan (which is a puny 0.09A 30mm type, it's not going to shift much air), so I *think* the increase in current draw is going to be minimal, certainly compared with the benefits - I'll find out for sure very soon!

I should add, that in my costing above I include shipping and tax (ie. total price) to get the hotend to me in the UK. so it's 56 euro + about 18 tax and 29 shipping + paypal charges = £100

the E3D hotend is UK based and they include tax and shipping in the £50 price. they say they will ship anywhere at no more than the minimum shipping charge for that country (ie. no surcharge).

today I constructed the v2 hotend and to UM's credit it went together very easily apart from the clips and bowden tube which were a bit fiddly.

i did heat up the v1 hotend before removing the cartridge tho, after hearing someone had problems with the wires breaking off. no problems to report on my side tho.

first print with the new hotend and....

yes it prints. OK but so did my v1. nothing dramatic to report I confess but I'll need to do some more scientific tests to show if it's 'better' in any way. if anyone has any requests for thingiverse items to print (not 50hr prints or yoda heads please then I'm happy to consider them and I'll post the results from the v2 and v3 heads here.

I also constructed the E3D v3 head. I made one small mod to the unit as supplied by E3D - the reprap boys use a direct drive extruder, so I tapped the collar of the hotend with an M6 to fit my M6 threaded bowden tube (made that mod while it was being used with the v1 hotend and for those who haven't threaded their bowdens - I can highly recommend it).

I'll take some photos in the next couple of days - I've just ordered some heatsink compound (might as well do this properly) to make sure heating and temperature readings are more accurate.

[Daid 306]

 

having said that, the E3D boys mention 100mm/s which is nicely in UM territory, so maybe not? but their filament is usually 1.75mm, so maybe so..

Friend of mine thought he had his printrbot running at 80mm/s, but the firmware was limiting it to 60mm/s. Also, acceleration values of most printers are lower, so the same "top" speed doesn't always mean the same printing time.

[destroyer2012 0]

The two qualities of a hot end that are at odds are retraction (oozing) and speed (how fast you can push filament out).

There was a post on the reprap forums (can't seem to find it now :( ) where people were using a j-shaped solid with one outline and no fill to test high speeds (on the straight section), retraction (on the dot), and slowing down (on the curly part).

here is one example of a good retraction test piece

http://www.thingiverse.com/thing:15087

Something also that is important for multiple heads is the ability to retract the filament to eliminate oozing while the second head is idle. So try retracting and waiting to see how much oozes out (or how long before ooze starts). Maybe you can try retracting to the "cool zone" and then starting back up?

be careful though when testing speed that the machine actually is going the speed you set. For example you can set 1000mm/s but because of acceleration and minimum layer time your printer will never reach this speed.

[m-wakileh 0]

Hi there... Just posting to say that I'm working on full metal hotend myself... Had the new UM Nozzle or this thread come out sooner I would have waited... but I'm in too deep now...that I think I will push ahead (time permitting)... The design is based on the UM assembly.. its watercooled with the intent of being able to print nylon and other high temperature plastics..The bowden holder and the extruder will be redesigned as well... Here some early pictures of the incomplete prototype....

regards,

Michael

[m-wakileh 0]

...and here the top plate with a threaded bowden holder...

[msurunner 2]

Way cool Michael! I'm curious what the weight difference will be, or do you not care?

[m-wakileh 0]

Actually I think the weight will be acceptable... I do care about the weight- that is the reason why I'm first attempting to make the watercooler out of aluminium and not copper, even though the thermal conductivity isn't quite as good. The welding should definitely be more difficult with the aluminium. I will know how that goes in about a week. This first iteration has 1 design flaw and was based on some false measurements so unfortunately I cannot reuse the wooden parts that hold the linear bearing. Perhaps this is for the best as wood = firehazard. In any case..that is the part I will (re)design next so that I can use the parts I've already cut (I typically only remove 0.2-0.25mm per layer so the cnc does take some time). I will measure the weight when I have all the parts, but I am pretty confident that tweaking the motor currents will alleviate any weight issues. I will probably follow up with a redesigned watercooler for easier welding and I plan to show off a new extruder as well =)... Anyway, I really hope this will work... The most critical part in the design is an M6 threaded titanium tube which connects the hotend with the cooler...the titanium is the thermal barrier... Unfortunately I was only able to source Grade 2 Titanium locally ..so only testing will show if it can stand up to the temperature, pressure and mechanical stress...

regards,

Michael Wakileh

[chopmeister 0]

Where could I find some info on that E3D "v3" hotend? Tried googling around and found nothing. Thanks.

EDIT: Nevermind, found it. Guess their SEO sucks.

[alaris2 1]

for the benefit of others - it's

http://e3d-online.com/#/featured-products/4567186833

yes, they need to increase their presence on the web, but they are very new.

you'll want the 3mm version of course (unless you want to experiment with 1.75mm filament)

I'm hoping they'll start stocking other useful parts, nozzles, maybe even filament soon - it would be good to have a UK source of 3D printer parts.

I'm about to test the full metal v3 hotend next. I waited all week for heat sink compound to arrive and now realize I need an M3 tap - but I can't wait any longer.

I tested the UM v2 hotend most of the week, results were OK but nothing spectacular. it didn't pop the bowden or plug, but then neither did my v1 ever do that. so whilst it's not worse, it's not demonstrably better either.

One thing I did notice - the v2 hotend seems to require more pressure to force the filament through? I've had to up the steps per E slightly to get decent prints and unfortunately I still get occasional variable print quality (as I did on the v1) which I was optimistically hoping would just fix itself. oh well.

some photos in a few hours when I've done some prints with v3 for comparison.

[futurejames 0]

Looking forward to the results. I'm not especially keen on buying the new hot end from Ultimaker either. Its cost is high, I could handle it if it weren't for ridiculous shipping costs.

[alaris2 1]

it worked, but it could work much better.

short story is that I asked for a 3mm version, but forgot the ID of the UM nozzle is actually 3.2mm. for thick filament there's a fair amount of friction therefore so I'll try and buy a reamer or something tomorrow and open up the holes to 3.2mm for a smoother print.

I did get filament to come out, but it was a struggle by hand (didn't help my filament was 2.95mm tho)

[m-wakileh 0]

I've decided to post some progress on the full-metal water-cooled hotend prototype. Perhaps others could chime in as to whether my posts should be moved to a seperate thread or not. Initial design flaws were circumvented by using 3mm aluminium vs 4mm for the last pieces. If the titanium heat barrier concept and alu welding concept works out, I would likely see if a refined design can be produced locally. What would Ultimaker's stance be on that (seeing what the initial design was based on... I would release cad files under some open source license as well)?

[m-wakileh 0]

...

[destroyer2012 0]

Whoa you are turning the whole head assembly into aluminum? That's crazy! Isn't it going to become really heavy now? I thought you were changing only the hot end.

Why is it necessary to change the entire head to metal if you are water cooling the hot end anyway?

[alaris2 1]

I confess I'm not sure either, but I think this thread will get really confused if we have so many hotend designs on it.

no offence intended wakileh but do you think you could start a new thread on this one?

[alaris2 1]

Success!!!

I've been trying different configurations - here's an early version I tried, it's a hybrid between v2 hotend and the E3D metal hotend.

It worked, but I broke my bowden by overthreading and the pressure required to extrude was astronomical. this was due to the ID of the stainless steel being 3mm instead of 3.2mm (as used by UM).

So I switched to a slightly different configuration, with a floating bowden (since it's not long enough now to reach) similar to destroyer's setup. I removed the collar and switched to the brass UM threaded section for now. It means having to run the fan flat out to stop it overheating, but it works very nicely. complete success at max speed on first attempt (altho I did have to manually adjust the bed because I forgot to re-level it!).

time for a bit more calibrating and tidying up.. I've already taken it up to 320C without trouble but I need to sort the fans out before trying nylon.

[m-wakileh 0]

Nice work alaris! Will be watching this thread closely...(I may choose to abandon watercooling in favor of air cooling )

@destroyer: my current prototype weighs 245-250grams vs 95-100 before (both including bowden clamp etc).. That weight can definitly be brought down.. I built it with what I had at hand (3mm alu) - the sidewalls holding the linear bearings needed to be extended approx. 2cm. The metal bowden clamp and top plate are practical and indestructible.. I will be keeping those...

Further updates on my prototype(s) will go into a seperate thread..

regards,

Michael

[destroyer2012 0]

Success!!!

So I switched to a slightly different configuration, with a floating bowden (since it's not long enough now to reach) similar to destroyer's setup. I removed the collar and switched to the brass UM threaded section for now. It means having to run the fan flat out to stop it overheating, but it works very nicely. complete success at max speed on first attempt (altho I did have to manually adjust the bed because I forgot to re-level it!).

Awesome!

I have two questions though:

1) What exactly is your setup? Do you have the bowden tube float above the metal nozzle's cold end? Are you saying it is cool enough now to just insert the filament up top and not have it form a plug?

2) How fast is "max speed"? What mm/min are you feeding filament to the extruder?

So it looks like all we need is an aluminum block to screw on above the heater block, and a heatsink+fan combo to cool it down? That's way better than water cooling, seems like I could make it with only a drill press too. Sweet.

[alaris2 1]

you're right to ask questions - and it's time for a hastily drawn cross-sectional sketch to explain!

yes the bowden is floating above the nozzle in this config (config 2).

in config 1, I used that metal collar piece and threaded the bowden into it, but when I broken the bowden it was too short so I borrowed an idea off a sketch you made - thanks for that!

in config 1 I used the stainless steel threaded section. I believe this is crucial, but it had a 3mm ID which gave a lot of friction. manual extrusion was *very* hard.

I switched to the UM v2 brass section in config 2 and it conducts heat up to the cold end faster than my little v3 fan can remove it - it wasn't designed for that. I could put a larger heat sink and fan on I suppose, but that increases the weight. I'd prefer to revert to the v3 SS piece - anyone have facilities to ream out stainless steel to 3.2mm for me?

the reason for preferring this config is that after about an hour of printing, the top of the cold end block was getting close to 50C (ambient was 19C) which would have spelled disaster, except that I noticed PLA leaking between the v2 nozzle and threaded section (which I thought wasn't supposed to happen!!!) so I caused my own disaster by taking it to pieces to fix this and then promptly snapped the brass nozzle. ho hum.

kisslicer tells me my current 'max speed' settings are :

perimeter = 50mm/s

infill = 70mm/s

sparse infill = 100mm/s

I haven't changed these settings in ages now so they're the same I used for v1 and v2 hotend testing.

in summary, yes you're correct, you need an al. block for a cold end, with a heat sink and fan - no water cooling or lathe or other exotic tools. tho you will probably want a drill and tap/die set.

BUT - I'm pretty sure you want that SS piece too. otherwise your printing time is limited by how fast you can remove the heat and heat prefers to travel up the brass tube not down to the nozzle..

the SS part is what attracted me to the E3D solution in the first place (since I have no hope of manufacturing that part myself)

[destroyer2012 0]

You could try to drill out the stainless steel part in steps. First, use a #31 drill (3.041 mm), then 1/8" drill (3.175 mm), which should be plenty big for extrusion. Sure it won't be as smooth as a reamer but close enough? It's already got a hole in it so alignment shouldn't be too much of a problem. You might need some carbide drills though.

Plus, you should only have to drill out the part of the barrel where the filament is still cold. Once it gets hot and liquid it will flow into a smaller opening. So maybe drilling only one cm deep with a 1/8" drill into the stainless steel part is more doable?

Another alternative is to drill out a brass bolt to replace the broken V2 heater barrel, and then try to make a thermal break by filing down the threads in the right place.

I might try to drill out some stainless steel bolts just to see how insane it is... MAY be doable at home? I hope so because this hot end seems like a really good idea