After messing with this add on for the last few weeks, somewhere north of probably 30-40 hours of attempts with it sometimes working and sometimes randomly failing even when nothing has been changed from one print to the next. It's simply too frustrating.
When the device works, it's amazing and I want to tell the world about it. But the problem is it won't work most of the time. Every time I fix an issue another pops up.
I've tested and adjusted so many things that while I'm sitting here trying to list the things I've had problems with I can't grab one thought out of the flurry, but I'll try, in no particular order.
The Y adapter is quite frustrating and causes so many issues of getting jammed in one direction or the other. I have mostly fixed this by loosening the screws on it by half a turn, as I found suggested in one thread here on the forum. But this fix only works sometimes and it will still get jammed in it. I have a Capricorn tube installed from the primary extruder to the hotend.
I can form beautiful tips when extracting the filament, sometimes. I haven't changed my extraction gcode in probably a week now and none of the tips are consistent and I have no idea why. I've tried different filament and different temps. The problem isn't getting a tipped formed, it's getting the same kind of tip repeatedly.
The final straw happened just a bit ago today, I haven't touched the extraction gcode in a while, as mentioned, and yet for some random reason the printers primary extruder didn't extract the filament all the way. It literally pulled it all the way out just fine earlier today and with no changes just randomly didn't, it left about 50mm or so in the Capricorn tube. So another print failed, that's three failed prints today, unknown number of failed prints over three weeks or so.
The control box for the unit has a really weird bug too that has driven me insane a few times. If for any reason the switch is pressed for more than 6 seconds or so, the entire device freezes up until it's unplugged, which obviously fails the print. This is problematic when it needs to be pressed for 5.5 seconds and the button is capable of being shifted ever so slightly when pressed sometimes. Again the key word being sometimes here.
My suggestion to improve the product is this; remove the printers primary extruder, have one extruder for each of the four colors you want, these all feed into a four way splitter that is not Y's stacked, but a true 4 into 1 meeting. Have the mother boards ribbon cable that controls the printers normal extruder (that is now removed) to a four way splitter. Each of these ends goes to a 4 pole relay that is controlled by the products unit. The switch that the unit uses to select which filament to feed is instead switching which relay is active, and therefore which extruder the motherboard controls and the true four way reduces the need to shape the tip so accurately. These would exponentially increase the consistency of the unit and therefore the ability to make adjustments that stick while also simplifying the overall control process. I've been an electrical engineer for 7 years, I'm probably just going to try and implement this myself, but if you'd like to make a mk3 with this, by all means please do.
I keep having the same issues as well, I had it working very briefly for the first two or three changes a few months ago and then quit fighting with it since I ran out of time and was waiting till I got a revo installed on my MKS3+. I had even adjusted gcode to get the temperatures exactly right for a certain brand of filament. I got the tips pretty decent (matching what it looks like we should have), but in general my issue seems to be the same problem that could be resolved with just a slightly better grip on the filament. If I just push a tiny bit on my spool when it pulls, it can get things right. I'm willing to pay for an upgrade cost if it makes this thing work better since I know its kind of a labor of love backed kit.
Hi Bill, will the upgrade parts (Mk3) be available to everyone. I too suffer in silence with the feeding and retraction inconsistencies.
If you’d like to send the parts I certainly won’t stop you. I may give changing the tubing out a try if I mess with it again.
As for the break circuit detection I have not experienced anything like that with my Ender 3 Max personally, but I very highly suspect that bridging with the wires at the relay with a 1k or perhaps 550 ohm resistor will keep it from tripping.
Ok... there are a few things of note in your comments... ones that are quite easy to remedy, and I have some ideas for the others.
1) Timing of the button presses. If you press it for longer than 5.5 seconds and it's going 'silent', in reality, it's being pressed for 7+ seconds. The 7 second press causes it to "forward" all commands to a second Mk2 unit. It will only cancel that mode with an 8 second press. The solution is to shorten your button presses by 1/2 of a second to ensure you don't go over that limit. The time that the gcode has is the actual time "on the button", but if your unit is triggering the button a little earlier, then the time it takes to press the button and then back away from it can also add to that total time. Shortening that press time will take care of that. Also, make sure you're trigger point is exactly .1 or .2 mm prior to the actual distance you enter into the gcode generator. If it's more than that, remember, it takes 1 second to go over 25mm... which means it can take an additional 125ms for .1 mm of travel... time two = .25 seconds just for .1mm of travel both ways... if your button press is actually .5mm late, than you're adding an additional 5 x .25 = 1.25 seconds to long. (If you simply shorten the time on the button, then you don't have to worry about these calculations... it's the total time that is important... where you subtract it doesn't matter, if that makes sense.) 2) You can't use relays to switch the motors with modern electronics. The Mk1 version of the 3D Chameleon used exactly that method, which fails with modern TMC devices due to their "break detection" circuit which causes the stepper motor drivers to shut down. They won't restart until you recycle the power, which is not an option with these electronics. Also, Creality machines (in particular) and their derivatives, tie all the enable pins on the X, Y and E stepper motor drivers together, so it makes even software resets impossible. 3) Our Mk3 already is design complete and includes the extruder built in. (Similar to the current model where it's actually installed in the extruder replacement scenario.) If you'd like, we can send you the replacement plastic parts to make your Mk2 an Mk3 (or, technically, what we're calling the Mk2+) which should give you a much stronger drive. (It's has about 3x the grip on the filament of the old design... and is tunable.) 4) The inconsistent drive... with the Mk3, we improved the grip strength by about 3x over the Mk2, which should easily over power any friction points in the PTFE or the Y adapter. Mind you, the Y adapter is 1.85mm ID on the inside... so if your filament is larger than that, it will get stuck there. Relaxing the compression on the Y adapter by 1/2 turn is the correct path there... however, if you're using Capricorn tube in the hot end, be aware that Cap tubes are 1.95 ID +- .05... so you mileage may vary on if it'll fit or not. Our 1.8mm ID tube in the hot end should constrict it to smaller than the Y adapter. Cap tubes above the Y adapter should be just fine as well as between the Y adapter and the extruder. 5) Tip shaping... with the Mk3 parts, the drive system should be significantly stronger so the tip shaping is much less of a concern. Stringing is the only real concern. With a smaller 1.8mm ID tube in the hot end, and the stronger grip on the filament, you should see far fewer issues like these.
Bill