trident Chris' 250 Trident Journey

[claudermilk]

Quick update: Not too long after posting this I got an email from E3D that they are convinced it's a bad heater and will send a replacement. The very next email was from FedEx that a package was coming from the UK to me. This morning I saw another email from FedEx that customs issues were resolved and it's on its way--scheduled to arrive this afternoon?!  I understand and expected the hoops I was asked to jump through to make sure it really was their part that is the issue, but I did not expect them to move that fast once the replacement decision was made. 

[claudermilk]

Update to the update. The Trident has been running since I took advantage of the modular design of Revo; I stole the old, working heater from the Micro and stuck it on the Voron. Easy peasy.

FedEx delayed delivery by one day, probably because of the customs hangup. No big deal, it took all of 3 days from E3D approving replacement to having it in my hands. From the UK to west coast US. Just as easy to install in the Prusa Mini and now that printer is back up and running. So according to Nero I now have one printer. 

[billbr]

Chris, I enjoyed reading your journey. I have traveled the same road (Nevermore, OV-cam, Revo, Klicky) on my 2.4. The biggest difference is I gave up on Revo. Tried Voron Revo and 6-Revo with no great success. Everything was under-extruded unless I ran the printer at Prusa speeds. So I went back to my DragonHF and am waiting to try the E3D HF nozzle. I have a Rapido but am waiting to finish my 2.4-r2 retrofit before making another massive change at the same time. See ya on the chats! 

Bill

[claudermilk]

Thanks. I hope it's useful seeing how I gto from complete noob to running a reliable printer. I subscribe to Nero's outlook: I'm not running speedboat races with this. I am getting really, really nice parts at reasonable speeds.

As an example, here's a couple proof-of-capability prints I ran last week. These files are meant for resin printers and are for a tabletop wargame. A friend asked if I could print files he bought; the answer is "surprisingly well." This is with a Revo 0.25mm nozzle and 0.1mm layers, and just a quick removal of supports with the standard printer kit side cutters+string removal. Zoom the image.

 

[Demosth]

Wow those are some high quality prints! VERY nice.

I think you should add high quality photos of these to the https://www.teamfdm.com/gallery/category/2-voron-print-galleries/ Voron Print Gallery!

If you can photograph them with a coin or ruler it will give some context as well!

[claudermilk]

When I get a chance to set up a proper photo session, I'll do that.

[claudermilk]

So I had to do a minor update. A couple of days ago, out of the blue I started getting recurring undervolt warnings for the Pi. To review, I am connecting the Pi to the Octopus via GPIO and only had one of the 5V pins connected. For months and about 570-ish print hours it's worked fine. Now it starts to complain. So, depin the 4-pin connector on the Pi and replace with a 5-pin and add in the second 5V wire. No smoke, and no more warnings so far. I got one good test cube of the new material with no issues.

Oh, and I totally forgot to get better photos of those miniatures, and have given them to my friend to test proper cleanup & painting. I guess I'll have to print more (gee, darn).

[claudermilk]

Making some lemonade out of the blob of doom incident. Part of the fallout from that is the Klicky probe tool head mount go snapped across layers right at the bolt holes. Big surprise. I was able to superglue it back together, but that's just a temporary fix.

I printed up KlickyNG parts to replace the broken stuff. This new version works quite well, and no glue needed as advertised. It mounts exactly like OG Klicky, and the macros work without any changes. 

 

I have an itch to try out UnKlickyNG now.

[claudermilk]

Another (really) minor tweak. Customizing KlipperScreen menus.

Added Filament menu within Actions menu

Calls to macros in Klipper for nozzle cleaning and setting printer to different filament definitions. Mainly used for load/unload & give a warning on sliced vs loaded filament mismatch.

[mvdveer]

On 10/7/2022 at 12:26 AM, claudermilk said:

have an itch to try out UnKlickyNG now.

Scratch that itch, it won't go away till you do 

[Buurman]

On 10/12/2022 at 10:32 PM, claudermilk said:

Another (really) minor tweak. Customizing KlipperScreen menus.

Added Filament menu within Actions menu

Calls to macros in Klipper for nozzle cleaning and setting printer to different filament definitions. Mainly used for load/unload & give a warning on sliced vs loaded filament mismatch.

Can you help me how to do this?

I REALLY like to add custom pre-heat settings, since the defaults dont really come close to what I use.

And your just looks awesome!

[claudermilk]

Thanks.

This site is helpful: https://klipperscreen.readthedocs.io/en/latest/Configuration/

For the background, I dredged around Google image search until I found a PC desktop I liked and resized it. Then under ~/KlipperScreen/styles I created a directory for my custom style and an images directory within that. i copied all the default .svg files from the set they provided that I liked and added my background .png. Then I made a little .css file to set up the graphics.

 

button {background-color: rgba(0,0,0,0); border-radius:2em;}
window {background-color: rgba(0,20,80,0.5);  background-image: url("/home/pi/KlipperScreen/styles/BlueTrident/images/Hexes1.png");}

 

For the menu, I grabbed the defaults.conf file from ~/KlipperScreen/ks_includes and saved it as KlipperScreen.conf in my main printer config directory. That was as much to study the menu definition structure as anything, but it gives a complete copy to modify.

I started with tweaking and adding to the preheat settings as those are easy. For example, I've added a Nylon preset:

[preheat NYLON]
bed = 50
extruder = 250
gcode: SET_FAN_SPEED FAN=nevermore SPEED=1

As you can see, you can add a gcode line to run gcode commands. In this case I turn on the Nevermore fan to circulate the chamber air to accelerate heating and even it out.

An easy addition to the panels is I added probe Attach and Dock calls for Klicky probe. These were added to the Homing menu:

[menu __main homing attach]
name: {{ gettext('Attach Probe') }}
icon: arrow-down
method: printer.gcode.script
params: {"script":"ATTACH_PROBE"}

[menu __main homing dock]
name: {{ gettext('Dock Probe') }}
icon: arrow-up
method: printer.gcode.script
params: {"script":"DOCK_PROBE"}

The key to remember is that "menu __main homing" section (note the double underscore). Basically you have to give the fully qualified "path" to your menu item. Then add the item title which has to be one word. Then add the parameters--reference the docs page I linked for complete information.

For the filament menu I added, first start with defining the menu:

[menu __main actions filament]
name: {{ gettext('Filament') }}
icon: filament-spool

Then add items under it, for example:

[menu __main actions filament setabs]
name: {{ gettext('Set Filament ABS') }}
icon: filament-abs
method: printer.gcode.script
params: {"script":"set_filament index=0"}

(Note that the icons are custom ones I generated)

Hopefully that helps. It took me a while to figure out what It wanted. It didn't occur to me right off that this is still just Jinja2 syntax like the rest of Klipper's configs. The key is the menu item definition format.

Edit: I almost forgot (found I hadn't posted on this). I also set up the fancy-pants boot screen. Because of course I did.

For this, I simply followed samwiseg0's instruction on the Voron site. I tweaked the .psd image to my liking, and added my serial then saved as a png.

Edited October 25, 2022 by claudermilk
Add boot screen comment.

[claudermilk]

On 10/12/2022 at 3:25 PM, mvdveer said:

Scratch that itch, it won't go away till you do 

So...KlickyNG self-destructed and forced me to scratch this itch. One of the magnets pulled out of the pocket, then crushed the top of the retainer bit, so I had to try gluing it in. Which of course I borked up. So the KlickyNG probe is now non-functional. I tried rebuilding the old Klicky probe and that's also non-functional. Fortunately I had the old OG UnKlicky sitting in my toolbox. That one works. So I had to finally figure out the settings changes to account for it being taller. Printer is back running using UnKlicky.

I'm not staying with it, because of the height I now have to dock the probe after z_tilt_adjust since it won't clear the purge bucket and will get peeled off doing nozzle scrub. Then attach again to do the z_calibration. With regular Klicky, I can leave the probe attached through the z_tilt_adjust, adaptive bed mesh, scrub, z_calibration sequence.

[polygonprint]

Thanks for documenting the progress with your Trident claudermilk! This will be very helpful as I progress with mine. 

Edited December 6, 2022 by polygonprint

[claudermilk]

Quick update. I did get UnKlicky working for a bit. I managed to get it working well enough to print up replacements and spares for KlickyNG which it's running on right now.

I also discovered one of my XY joints was cracked; this explained why I was getting weird Z banding out of nowhere. Replaced that and it's printing beautifully again.

Current plans are after the holidays I'll pull out KilckyNG and install Tap; I have the rail and the other parts are on preorder with KB-3D. I also won some parts from West3D's Meet the Maker stream; since I am swapping out Klicky, I swapped the offered Klicky PCB for the hartk 2-piece tool head PCB. So I'll be re-wiring the tool head too. That is kind of due as I'm seeing wire insulation dust collecting on the X extrusion (1200+ hours print time now). I'd rather put in fresh now than wait for a wire to break and get weird errors.

[billbr]

Chris, sounds great. I finally went with Euclid on the 2.4 and Annex on the switchwire. The hartk 2 piece is awesome. It just takes a bit of study to get the right wires in the right places. Taking a wait and see on the Tap. Need to rebuild my v0 over the holidays before taking on Tap. Looking forward to hearing how your Tap works out. Also watching the discussions on shallow dovetails. Looks like an interesting alternative.

Bill

[claudermilk]

The first of the 2023 mods have begun. This weekend was Ramalama's idler pins mod and Tap. Since they both involve messing with the belts it made sense to do both together.

First was the idlers. That turned out to be really easy and now i have a bag full of 5x18mm shelf pins looking for more things to be used on. I also have a 150mm long hardened steel rod I need to figure out what to do with--I discovered I don't have tools available or accessible right now to cut that.

So, the idlers went so quick I was almost done before I remembered to take pictures. It's really easy. The way the tensioner works, feeding the belts is super easy. I used my handy Prusa tension meter to make sure both belts are evenly adjusted. I went between the belts a couple times at the end to make sure they were even. Easy-peasy.

  

Next up was a bit harder. Tap.

I pre-assembled the two main subassemblies ahead of time while waiting on last parts. The carriage went on fairly easily, and the belt routing is so nicely done.

 

No extra trimming needed, the little bit of spare insurance made the swap relatively painless and I can go back to inductive/Klicky if I really need to.

Now for the first problem. I intended to rebuild all of SB/CW2 with RC1 parts to replace my last-ish rev beta parts. But...

Houston, we have a problem. I didn't notice the change in the front cover-to-hotend mount interface. And I didn't print a new hotend mount. Fortunately I noticed this early on and bailed on that plan. Sticking with  current, working SB assemblies.

I had also planned updating CW2 with a modded main plate with a filament sensor switch embedded and M6 PTFE tube retainer. Neither of those was going to work either, so unbuilt that and rebuilt the old CW2 again. I'll have a think and revisit that idea later. I like the concept of the filament runout switch right at the feed gears instead of  the back of the printer. I also am annoyed by the PTFE coming out of the tool head every time I unload filament. But the fitting interferes with the cable cover. Need some more CAD work.

Now on to trying to get Tap installed. After some messing around (I think @smirkwould say faffing about) I got the magnets to give positive engagement that I could feel. I built a wire harness with the suggested 2.2ohm inline resistor (yay, got to use my new TS101 iron!). So with that, plugging in to my existing printer harness was uneventful. No sparky sparks, smoke, or fire. Just a baleful red LED. Which I cannot get to turn off.

I tried coloring the tab on the printed part with black sharpie, putting heat shrink over it (bad, jams up the works), and bits of black electrical tape. Nothing changed behavior. I now that the PCB hanging loose and has jammed some black cardboard in the gap, still to no avail. I've swapped the signal and ground wires a couple of times also no effect. I'm getting to it being a bad PCB.

So next steps are plan B & C. B is UnTap, which I printed up a couple of sets of parts. Just in case. Plan C is wired optical sensors, which I found Digikey has one of the suggested models in stock so a couple are on order.

[mvdveer]

4 hours ago, claudermilk said:

the suggested 2.2ohm inline resistor

I assume you have the 24V board and using this to bypass the surge that occurs in some boards.

4 hours ago, claudermilk said:

I'm getting to it being a bad PCB

As I understand, the resistor is only needed iy you do not have the PCB as this has a built in resistor, or if you have a faulty 24V board.

The inline resistor In the BOM is otherwise for connecting the optical sensor directly to the controller board, bypassing the PCB, 

Two options of wiring:

1. PCB - Optical sensor gets soldered on in the 5V version, no need for soldering an inline resistor the sensor. PCB 24V - as above OR with 2.2ohm resistor to temporarily bypass the surge issue (Issue#42)
2. Optical sensor connected to the printer controller board - need to solder on an inline resistor

The light on the PCB will always be on, except when it is triggered. Annoying I know.

[claudermilk]

Yes, I have the Linneo V2.1 24V board. In the Github issue #42 tanaes has a proposed workaround using the resistor. The idea is to lower the inrush voltage to stop blowing up the regulator. I am getting the led constant no matter what, so for some reason the optical sensor isn't triggering for me. I will eventually be using the hartk 2-piece PCB, so I am stuck with using 24V as far as I can see, and right now I'm just using the OG inductive sensor wiring in the printer--which is 24V.

I tried implementing the UnTap option and that kicked my butt until I had to walk away before doing something unfortunate to the printer, tools, and surrounding room. I eventually got it to where I could test out the function successfully on my multimeter with only the carriage installed. By the time I got all the SB bit bolted on and cover door closed it failed. Argh! Coming back after a sleep and think I did some additional testing with the door open. Dang if bolting that sucker down doesn't currently gum things up just enough to bind the wires and move the little sliding bit to be triggered. I'll have to re-do the wire harness again. This time I'll use some of the spare 30AWG silicone wire from the SB LED so it's super flexible and take special care to route them so they don't get jammed when everything is buttoned up. Once that's done, I think it will be fine.

I did get a few successful taps, so I'm close.

[claudermilk]

Today is face-palm day.

So @Penatr8tor got me to double check the resistors I ordered. Turns out they are 2.2K ohm resistors! So I guess killing too much power. I have the proper 2.2 OHM resistors on order and will try that again.

I also went at UnTap again and coming with rested, fresh eyes I saw my problem. First, I built a new contacts carriage with some extra 30AWG wire from my SB LED harness build. No way those little, floppy wires were going to jam things up. So now I'm getting good homing! Yay!

Now I go to z_tilt_adjust, and the fails start again. I took a look and realized I needed to move the front locations back a bit now that I'm probing off the nozzle and not the back of the carriage. The nozzle was right at the very edge of the bed! Shifted that and did some break-in tapping and it's now functioning!

The fly in the ointment right now is that I'm not seeing the crazy low standard deviations. Here's the last 100-tap probe accuracy results:

probe accuracy results: maximum -0.087500, minimum -0.106875, range 0.019375, average -0.092062, median -0.089375, standard deviation 0.005243

Good enough, but I want that crazy 0.000x dammit! I'll wait for parts to arrive and try optical again. But at least I have the mechanical backup available, and it is easy to swap between them.

[claudermilk]

The saga continues. The correct 2.2 ohm resistors showed up and I fixed my wire harness. I also looked at the through hole solder joints and added some just in case. No change in behavior.  I have 220 ohm resistors on the way (double checked before clicking order), so I'll try eliminating the PCB variable and add my own capability to follow instructions and solder. If that doesn't work I'll have to re-examine the connection on the controller side. After that I'll be out of ideas.

[claudermilk]

Time to reprise this:

I gave up on the PCB optical sensor, bit the bullet and wired up the standalone sensor (OPB991T55Z). Which of course Digikey is now out of stock on.

I forgot to take a shot before heat shrinking, but I am impressed with my improvement in soldering, the red-white-resistor threesome actually came out pretty clean. I also had to double check on the Octopus and swap the J38 jumper to shift the voltage to 5V--glad I re-read the Tap manual carefully there! I was wondering after reading the sensor specs at 16V.  I also briefly thought about cleaning up the connection down there--I have power & gnd on J40 and signal over to J34 (PG15). But, read on the nice color pinout image which warns that PB7 shares signal with the BL Touch port! Since I'm running my case NeoPixels on the BL Touch port, I am leaving things alone.

Anyway, without further ado: the Trident optically Tapping!

 

 

 

 

[smirk]

I wonder how long before some one writes the macro to tap out tunes...

[mvdveer]

5 hours ago, claudermilk said:

gave up on the PCB optical sensor,

Just wondering if the dsign team made a mistake by trying to design a 24V board to please the masses. I now have two TAPs going, both with the original 5V board. Mmmmhhh....

 

5 hours ago, claudermilk said:

Anyway, without further ado: the Trident optically Tapping!

Excellento!

[claudermilk]

I think they are wondering that as well. I understand why; it's to give a plug-and-play option for jumping from inductive setups or tool head boards.

Now if  I want to make use of the 2-piece tool head board set I have sitting, I'm going to have to come up with something. The probe traces are 24V and that is it. I'm thinking solder a jumper wire on to the 5V pin of the main connector or the front face connector for the LEDs and make a little 1-wire harness to split off. It's a bit of a bodge, but it might work.

Now going back to the start and doing a full Ellis calibration regimen. I had to move live Z UP quite a lot but ended up with perfect first layer results. Continuing today and this weekend as time allows.

There is something wonky with the printer though. The left edge of the bed now comes up as higher than the rest. I'll have to double check the gantry rail on that side, it almost seems like it got shifted.