Further to my last post (part 4) this is just a quick update on my findings with regard to the leakage, my first attempt at a fix, and some other thoughts and observations.
2. Removing filament
Having discovered a filament leakage problem, I had to remove the hot end assembly from the printer. To do that, I had to unload the filament but I was nervous about pulling hot, molten filament up through the heat breaks where it could cool and cause a blockage. These Slice Engineering heat break tubes are extremely thin, so I don’t think that they will take kindly to rough handling if they do get blocked. I have been given advice on how to fix a worse case scenario but I’m not at liberty to publish that here.
In the end, I decided to do a sort of “cool pull” for want of a better description. I heated the hot end to 185 deg C, retracted just 2mm of filament from all 6 inputs, let the hot end cool to 140 deg C then did a big 100mm retract from all 6 inputs concurrently. That worked well. The filament itself had a distinct break with just a fine “whisker” after the break. When I eventually got the heat breaks out and held them up to the light, all I could see was shiny metal with no sign of filament inside the tubes.
3. Leakage inspection
The leak was as I suspected, around the edges where the plate holding the heat breaks fixes to the combining block. Here are some pictures.
In the picture above, there is red filament on the side of the mixing chamber plates, but this is just where the molten filament ran down the outside, and not leakage from between the plates. I simply scraped most of that filament off with a finger nail. The next picture shows that there are no lines of filament between the mixing chamber plates.
This is the other side of the hot end.
This is the heat break plate with the heat breaks removed.
You can see filament sticking up through the centre of the holes showing a nice clean break. You can also see the 4 screws holding the central part and how the edges where the filament has leaked out, are unsupported.
To get the rest of the hot end apart, I clamped it in a vice close to the printer so that I could use heater to soften the plastic. Here is the top of the combining block with the heat break plate removed.
From that picture, I would say that the plate distortion isn’t confined to just the edges but in the centre too, because there are signs of leakage across much of the face.
4. New parts and testing
I made a new heat break plate, slightly wider so that I could fit additional screws at the ends as well as the four in the middle. Then I made a new combining block the same size. I forgot to take pictures of the individual parts but here is one showing the new partial assembly.
There are now eight screws holding the heat break plate to the combining block, instead of just four.
I installed the hot end using my “test rig” approach rather than mounting it on the printer, so that I could actually see it. Then I made up some longer temporary Bowden tubes, fitted the 1.5mm diameter “flushing nozzle”, loaded and flushed some filament through to clear out and debris. Then I fitted a 0.5mm nozzle and extruded a couple of metres of filament through all 6 inputs at 5mm/sec.
Unfortunately, there was still signs of leakage as you can see.
But it seemed to be nothing like as bad as before so I decided to fit the hot end properly and try another print. Unfortunately, this went no better than my first attempt. The leak must have got worse, because I had the same problem with leaked filament running down the side of the hot end, cooling and then getting stuck between the printed part and the carriage. So I had to abort again. Here is a picture of it anyway.
For comparison, here is a picture of the original part using the Diamond hot end. The print was aborted at a layer where the infill direction was the opposite to the finished part.
I’m still not sure if it’s valid to make any comments with a leaking hot end. But it seems like I might be making progress. Overall, my hot end is more “pink” and the contrast between red and white is nothing like as stark. So it looks like some sort of mixing is happening – just not enough.
5. Next Steps
Obviously I still have to fix the leak and to that end, I have ordered some brass to make another heat break plate out of. This should have much less of a tendency to distort than aluminium, especially at elevated temperatures. I’m reluctant to try a sealant such as Silicon RTV, although I may have to. High temperature “O” rings (eg FFKM) in grooves would work but I doubt that I could machine the grooves with the equipment I have in my garage. I’ll see what a brass part is like or maybe I could even try steel for this particular plate( but it would need to be stainless).
Could it be that my mixing chamber is too restrictive so there is simply too much pressure? If there is a restriction in the mixing chamber, the pressure would build up before it which might explain why the leak is at the junction between the heat break plate and the combining block but not between the mixing chamber plates. The Bondetch BMG extruders are very powerful when used singly, and I use them in multiples so I have the potential to build up a lot of pressure before I see any sign of filament slippage.
I also have to address the fact that more mixing is required. Can I achieve that without making the chamber even more restrictive or having an impractically large volume?
But what is bothering me most of all is that I simply cannot see anything of the hot end when it is installed on the printer. So if it does develop a leak, the first I know about it is when a great lump of plastic drops down and gets jammed between the partially printed object and the carriage plate. Also, accessing the extruders to load filament or fit Bowden tubes is really difficult, even with the expansion boards hinged up or down.
So I have a plan and most of the design work is done. Firstly, to address the extruder access, I’m planning to rotate them by 90 degrees. To minimise disruption as much as possible, I’m retaining the existing X rails and Y carriages as well as the expansion board mounts. But I need to make a completely new gantry assembly (from aluminium parts). Here is an OpenScad image of the design as far as I’ve got with it.
Obviously, I’ll also have to rotate the hot end. Again, to minimise disruption, I want to retain the existing X rails and Y carriages, but I’m planning on mounting the hot end much lower down. I’ll lose 40mm or so of Z axis travel but I have 760mm to play with and I’ve never yet printed anything that tall. Here is the new OpenScad hot end gantry design as far as I’ve got. I need to design a part cooling solution. I’m planning on retaining my method of using the nozzle as a Z probe but I have new approach to mounting the hot end such that it is constrained in X and Y but still able to move slightly in Z.
So I have finish the design work, then make a new heat break plate from brass and lots of new parts to make up two new gantry assemblies. Then I have to re-think the mixing chamber which will no doubt involve more machining.
Whatever the final outcome, I’ve decided to stick with this hot end over the Diamond. Even if I have to give up on the mixing chamber, I’ll still have a 6 input hot end, that is smaller, lighter and quieter than the Diamond. It is also “all metal” whereas the Diamond uses PTFE lined heat breaks and I can easily change nozzles. So I will be able to print at higher temperatures and/or try printing with abrasive or “exotic” materials even if I end up with another “combining” hot end rather than a mixing one.
But because of all the new parts I have to make, as well as some serious thinking, the next update could be a few weeks hence. At least it’s giving me something to do during the current Covid-19 lockdown.
I hope everyone reading this is fit and well in these difficult times.