JohnInOttawa
-
Posts
368 -
Joined
-
Last visited
-
Days Won
3
Content Type
Forums
Events
3D Prints
Posts posted by JohnInOttawa
-
-
Thank you @Torgeir. It's too bad there is so much distance. It would be good to be able to chat about some of this, and a number of other areas of your clear expertise.
You'll clearly understand the background behind my question. My imager is uncooled, a Fluke TI32, 7.5 μm to 14 μm, so toward the long end of medium wave. While I can certainly treat the glass to enhance emissivity with a matte finish, my hope was to try to use an in-service material, like a PVA coating, to do this. Mostly because I am too lazy to have to remove and retreat the surface afterward 😉 , but also because I can get a nice big spot size that way and control reflections.
What I may do just for fun is take a glass build plate and stripe it. Clean, PVA and dry erase black. I'll try and post the result here.
Thanks for your post - Inspired experimentation!
Cheers
John
-
3 hours ago, robinmdh said:
1) the bed temperature: does it also show as way too hot on the temperature logging page?
Could I ask two (hopefully) related questions regarding troubleshooting the heated bed?
a) I'm assuming the only correct way to validate the heated bed temperature is independently, ideally via thermal imager or IR thermometer. Is there a recommended surface treatment - i.e. PVA glue stick? I know glass is opaque in the wavelength range of these uncooled IR sensors, but as a smooth, specular surface it would still be prone to reflection errors. Would a matte glue finish be better?
b) In the event one does find a delta (cold or hot), is there any way to correct this without replacing hardware?
Thanks and sorry for any inadvertent thread drift.
John
-
There seems to be a number of threads now talking about print bed temperature for the UM3 and 3X. I wonder if the worldwide fleet is aging together and hitting the same issues in batches. We're seeing a lot of worn and broken capacitive sensor wire questions for example.
Would there be any interest in trying to consolidate the print bread temp stuff together on a topic? Maybe with a guide to diagnosing issues and possible fixes?
J
-
Just an update. This is a very loooong read.
TLDR version - it is possible to install a new nozzle on a print core. With caveats.
Longer version.
I am printing with the re-nozzled AA 0.4. All printing aspects appear to be as expected, including active levelling, so I'm happy!
One aspect that I will likely have to live with is that the body of the old print core had sagged in the area under the cooling fins, which made installation and removal a challenge in the #1 position before this effort began. The new nozzle of course did not fix this.
Briefly, here are my thoughts about taking on nozzle replacement on an OEM print core.
Before starting
1) The video from @gr5 is the place to start. I suggest several plays of this resource before attempting anything.
2) Things are far more predictable if you have a vice that can hold just the heat block while allowing you to put a wrench on the heat break nut and a socket on the nozzle. The heat break nut cannot be allowed to move when removing the old nozzle or the heat break is done.
Removal and cleaning
3) The nozzle holds the heat break and heat block together. Once the nozzle starts to loosen, continue to hold the heat break nut until it disengages from the nozzle. You'll know once the nozzle has disengaged from the heat break.
4) Unlike the video, a print core with a worn out nozzle has a lot of carbon and deposits in the threads of the heat block and heat break, as well as on the surface inside the heat break connection that the nozzle seats against. In order not to risk the nozzle binding before it got fully seated, I cleaned the threads and that seating surface.
5) The heat block, heater and sensor will all be fairly caked with carbon and long melted filament. You'll have to remove the heater and sensor anyway to reinstall the new nozzle, so this is a good opportunity to clean everything up so you get a good thermal connection. In my case, I had to carefully clean the socket on the tiny cap screw that holds the heater and sensor in place before I could get an allen key in there.
6) The top of the heat block in my case almost looked like I had some pitting or metal transfer from the heat break to the block, but in the end, it was really just heavy tarnish.
7) I took the opportunity to clean the heating block, especially the top where the heat break was going to thread on later. Again, the priority was to ensure nothing caused the torque to max out before the thread was fully seated. When cleaning the brass, I decided not to use any sort of metal polish, as I did not relish that stuff off-gassing later. Instead, I went with an old school method - ketchup. Worked great. I finished the last of the surface with a quick pass on wet 2000 grit automotive standpaper and lapping compound. The goal was to remove residue with as little effect on the metal as possible.
Installing the new nozzle
8 )The nozzle thread passes through the heat block into the heat break. Both the heat block and heat break are threaded. This means that the only way to get a tight threaded connection all the way through is to first fully install the nozzle on the heat block, then install that assembly onto the heat break. The key interface to prevent leakage is between the top of the nozzle threaded portion and the internal mating 'ring' inside the heat break, so I decided to fully thread the nozzle into the heat block to get as much thread past the heat block as I could. Reference my post above, because of where the seal is, there is no value in teflon tape or otherwise sealing the threads.
9) A test fit of the new nozzle felt like it was binding a bit, so I decided to go with a two-step process of freezing the thread and heating the receiver. So for the first connection, I froze the nozzle, and heated the heat block to 200C. They went together without issue and the nozzle seated fully in the block as hoped.
10) I then froze this assembly while I heated up the heat break/cooling tower assembly, again to about 200C. While it was heating, I got the print core body, heat break wrench and nozzle socket wrench ready.
11) The heat break and tower were fitted back into the print core body, the wrench held the heat break nut stable and I hand threaded the heat block /nozzle assembly on as far as I could, in order to feel for cross threading. Torque was very low and I was able to feel when the thread reached a physical hard stop inside the heat break.
12) Reassembly from this point on is exactly per the @gr5 video. I had to pay special attention to the heater and sensor wires, they are quite stiff, and if you don't position them correctly, they'll push the nozzle forward and complicate loading and unloading the nozzle.
Testing and final thoughts
One final step, and again, shout out to @gr5, for the first print, I used white PLA and only ran it for a few minutes (less than 10). I then stopped and checked for leaks. Finding none, I lengthened the print run to 30 minutes.
The nozzle appears to work pretty well. It's plated brass which is supposed to be more wear resistant, I guess we'll see.
While I'm happy to have salvaged a print core, as you can see, this is a time consuming thing to do. I happened to have the tools to hold everything square, safely clean the brass components, chase the threads without shaving metal and isolate the fragile neck from torque, so my direct cost was just the new nozzle. I'll likely do this for the next core that wears out. In the end, though, the achilles heel of the print core may be the plastic parts and not the hotend. I don't yet have a fix for that.
Thanks for sticking with the long version!
John
-
23 hours ago, sbfdfmdk said:
The only consequence of that is that one of the print cores was so clogged that it has to be changed
I'm glad you are back in business. I had a print core blocked solid also from too close bed levelling and was able to recover it with a nozzle cleaning needle and some advice from @gr5. It prints completely normally once again. Hopefully you are able to do so as well.
All the best
John
-
So the strange question of the day. I bought a scrapped three stage STACO autotransformer. It was brand new, got damaged during installation.
The nature of the damage was that the unit was dropped on the end that had the control knob, broke the knob and I am considering that stage to be scrap. The other two stages are unaffected and are fully functional variacs. What I lack is a control knob.
Despite repeated attempts to order said knob over the past number of months, it just is not happening, so I think it is time I just print one. The issues are, of course, in service durability and electrical safety. The original knob was bakelite, a great insulator but almost a ceramic in how it broke. The replacement part does not need to be super strong, but tough and non conductive would be ideal.
I have nylon/CF and nylon/glass here that I would like to consider for this project. nylon/CF is a better color (black) and the manufacturer says it is not conductive. Are there better choices though?
Thanks!
John
-
Are you running an Ultimaker 3 or 3 Extended? Your issues seem very familiar, have a look at these threads.
https://community.ultimaker.com/topic/35978-differing-layer-thickness-firmware/
-
If this is an Ultimaker 3 extended, the 'update firmware' might be contributing to the issue. Have a look in this section, I will try to find the links and update this post.
I've just gone through an issue with my UM3X where active levelling was resulting in a value way too tight. So tight that laying down the initial layers clogged a print core so tightly that hot an cold pulls could not clear it. I had to heat up the tip and use a tip cleaning needle.
I also thought a firmware update would improve so did so. Active levelling resulted in a slightly higher initial position but an inconsistent height on the build plate.
In the end I followed the advice of other posters and reverted all the way back to 4.3.3 via recovery image. For some reason, the ultimaker 3 extended is affected by this. My other printer, an Ultimaker 3 (not extended), had no issues.
Capacitive sensor wise, if you can get a good picture with the front fan door open and the print cores removed, the capacitive cable might also have a pinch or fray at the connection to the sensor, I picked this up on my UM3X during this process as well.
This forum definitely helped me solve these issues and get the printer back into service.
John
-
I feel your pain.
Just went through this, repaired and working. link below. Good luck!
-
1
-
-
So, what's better than one questionable plan? Why, two!!
I note that among the concerns with DIY disassembly and reassembly of a print core is the potential for a lower quality seal than original and hence leakage.
Given that there is a PFTE liner between the cooling tower and the nozzle, I'm wondering what the pros and cons would be of using teflon thread tape in the nozzle to cooling tower connection. The stuff I have says it is stable to 280C and then slowly degrades as the temp approaches 400C so I think it would be safe, especially since the section I would use this on would be effectively sealed from the surrounding air.
Note this is with regard to an OEM AA 0.4 print core that is going to trash otherwise. (Not as lucky with the 0.8 experiment, but that may be an option for the sister of this one...)
John
Update per post below. Thanks to some good mentoring I learned that the critical point for seal is above the threads, so there is no need to complicate this process with teflon tape.
-
Just an update. Back to 4.3.3. Normal operation restored. Will stay there for both printers.
Looks to me like a completely different levelling process between 4.3.3.3 and 5.2.xx No heating cycle in 4.3.3 I'm thinking perhaps the capacitive sensor is perhaps a bigger factor in levelling with the newer firmware?
John
-
1
-
-
So I downgraded back to 4.3.3 via recovery image.
Same print, same filament, active levelling, same print cores.
I think it's the firmware. Sticking with 4.3.3.
John
-
I made a restore image of 4.3.3 and then went ahead with 5.2.16. No improvement and I am seeing a new flow issue that I have not fully diagnosed yet. I am going to restore back to 4.3.3 today and run the same print with the same cores (both Ultimaker OEM). I'll report back.
Meanwhile, my UM3 (not extended) remains at 4.3.3. I think I am going to leave it there. I can't afford to have both machines in trouble.
I understand UM3 is end of manufacturing life. UMO firmware is available from folks like Tinkergnome. How difficult is it do go this route with the UM3/3X family? I get the feeling that as the S line diverges from the legacy printers, the UM3 and 3X offerings are just going to become more problematic. Maybe we just need to form an aftermarket community that can focus on the UM3 line and not have its resources drawn in multiple directions.
Sorry to have to say this. I bet it is just as frustrating for the UM team that made these great machines to see this happening and be unable to address it as they would like. My guess is the situation is not their choice.
John
-
Things are progressing with the UM3X. Active levelling was an issue, I think at least partially corrected going to the current firmware. But a new issue is emerged.
Printing a simple, flat design, 0.2 layer height, 0.4 mm AA core, Ultimaker PLA. Bed at 60C and imaged at that temp, so correct (but slow to heat). Active levelling appeared to complete.
The pictures tell the story. transparent in the middle, much too thick at the sides, as if overextruded. Eventually things settle down, but the final product has an uneven top layer.
I'm aware the print head was changed before I got this and. looking at the rods, the X axis rod appears too long and the Y axis too short, so I am wondering if the axis rods are part of the issue - maybe they were swapped, or possibly warped.
Thoughts?
-
I'd apprecate an update on this - it seems to affect my UM3X and not my UM3 for some reason. Wondering if the latest firmware for the UM3X will fix this?
Thanks
John
-
Much appreciated!
John
-
Thanks for the link! I'll give it a try tomorrow with a fresh pot of coffee and a fully recharged sense of humour.
Will be interesting to see how the Bondtech DDG updater works with the new setup. Fingers crossed.
J
-
That's encouraging. I'll give it a try on one printer, get it stable there, then try the other one.
From 4.3.3 as you were, there is still an intermediate firmware step? If you can spare the time to just indicate what you did and the order, it will probably help me and a host of others who have been exactly where you were.
Thanks for the update !
John
-
This is related to my other thread on the issue. I'm at 4.3.3 following a recovery as well.
I understand the wish to have everyone on the same level but for some reason upgrading is hazardous to the stability of these machines and there does not seem to be a path that allows reversion to the previous, stable state. It's all the way back to 4.3.3 or, given the latest link on the 'bricked UM3' thread, 5.2.16.
Some middle ground, please.. How about the latest stable variant of 5.2.11?
J
-
Thanks. I note this is only for the newest models but it is a start. Hopefully something similar can be rolled out for the UM3 and UM2.
Cheers
J
-
Good afternoon from frozen Ottawa!
With apologies if this has been beaten to death.
I've been reading the extensive threads on firmware issues. While likely very rare, considering the numbers of units in the wild, firmware failures are so disruptive to any sort of commercial process that, like many I suppose, I'm running a number of versions behind, simply because it is stable and I don't want to rock the boat.
Of course running with obsolete firmware will only work so long before the point where I can't print what Cura outputs and may not be able to directly update from the firmware I have - or that update will fail.
Given that these printers are increasingly marketed as professional grade, I would like to suggest a more robust upgrade process - one that allows recovery without hours of lost time. For example, something like a reset image written to a USB key prior to a new firmware update that allowed a back-out would reduce the hit from hours to minutes.
I'm sure there is a myriad of reasons why this request is impossible, but then, not so long ago, 3D printing was impossible too. These things are all relative... 😉
J
-
2
-
-
While the UM3X' project printer' is now passing all diagnostics and active levelling, I notice a significant difference in time to target heat for its bed vs my UM3 (non extended). My initial inspection doesn't show up any connection problems or signs of arcing, but I would like to be able to observe as much as I can. I'll have an imager on the unit for the next test, but as I also have a watt meter that I can put inline between the power supply and the mains outlet, I wondered if it would make sense to compare each of my printers' wattage during a heating cycle.
Would it make sense to expect both the Ultimaker 3 and 3X to use the same power to heat? The only difference I can think of would be the length of the cable from the board to the bed and I can't imagine that radically altering consumption.
Maybe related, I have noticed that active levelling executes differently on the UM3 and 3X, with the 3X doing a protracted heating cycle first, then levelling quickly, while the UM3 does not seem to do much of a heating cycle at all. I have yet to synchronize the firmware, so that may be a factor.
Any other suggestions welcome!
Thanks!
John
-
Just an update: Back in business. Shout out to Erin at frbc8, and Ken and Felix at Shop3D.ca!
TD/DR version: Here is how:
First, the good news - the official part, 2014 is not expensive. Depending on the of the print head, I would wager it's a question of when, not if, so I will be planning on having one in my troubleshooting supplies.
The bad news, right now the part is not in stock locally and transborder mail is deadly slow.
Still, not all was lost, I was able to narrow down the wire to TR64 UL style 1007, 300V, 90C (more on this in a minute). 7 strand, either 22 or 24 gauge. I looked around my pile of salvaged parts and found a number of close fits.
I had never had to work around the capacitive sensor before and did not know which parts were critical. Erin was able to explain that it measures heat from the bed and give me an idea of what the risks were around the connection. Protecting the connectors on the capacitve board takes a good magnifying glass and fine tweezers. Felix was able to fill in the electrical characteristics once we realized there wasn't going to be stock for a while, so I realised I could go a bit long (under a cm) and relieve some strain on the connector ends.
The next choice was whether to try and re-use any part of the existing cable. One look at the connector and the crimp-on fittings and I decided the risk of a bad connection and frying a PCB was not worth it. I found everything I needed on digikey (Canada, but I think this is mutinational). The cost of everything was swamped by the $8.00 shipping fee so I ordered spares for DIY wastage.
The two pin female connector is a JST part, available here:
https://www.digikey.ca/product-detail/en/jst-sales-america-inc/PHR-2/455-1165-ND/608607
These are used in all 4 positions at the back of the print head, I picked up a few as I have two UM3s and suspect I am not done fixing them.
Crimp on sockets for 22-26 gauge:
https://www.digikey.ca/en/products/detail/jst-sales-america-inc/SPH-001T-P0-5L/1300245
Unless you are very good at crimping, I suggest getting a strip of ten. My first attempt lasted under 15 seconds, the second and third ones went on fine.
I decided to try two cable assemblies, one pre-assembled with the same female and and the other, the mate. My hope was that I would save myself the hassle of stripping, tinning etc. I was partially successful.
The assembly with the female end, which I hoped to be a direct replacement, was too light gauge on the wire, 26 gauge. That would be fine for a fan but I elected not to try it on the capacitive sensor.
The mating male assembly was perfect.
https://www.digikey.ca/en/products/detail/adafruit-industries-llc/3814/9380221
This assembly is a perfect match for the wire spec at 22 gauge. I was thrown for a bit when the wire was rated at 80C, not 90C as the original, but then I determined that 1007 wires can be rated at 80C UL and 90C CSA for the same wire.
After cutting off the length of black and red cable I needed (being careful to measure the length untwisted), tinning the ends, crimping on the sockets and mounting them into the JST connector, the male connector still had enough wire on it to serve as a continuity tester for my new cable. I could plug the new female connector into the male and I had 4 wire leads to test the circuit. The same male remnant also fits the fan of course and allowed me to troubleshoot a balky axial fan at the same time.
It's important to twist the two wires as they were in the original to avoid signal noise. With that, everything went back together and autolevel and all fans worked first try.
Now onto the next challenge, the very slow-heating print bed!
John
-
1
-
-
Thank you. The parts manual is a handy link. I'll ping my reseller(s) tomorrow. My guess is I have a suitable wire on hand, I note the twist likely to combat signal issues and wondered if there was anything else special about the setup.
I'll report back here on what I learn, this little length of wire does seem to be prone to wear out and breakage with age and use.
Cheers!
J
Substitute material for bakelite - electrical control
in Materials & profiles
Posted
Those are all good considerations. I am wondering if a mold is the right solution here as well. One aspect I had not considered when I posted originally is that the air gaps present in normal FDM printing can defeat any intrinsic insulating qualities in the filament itself and can contribute to shock hazard around higher voltages.
This VARIAC can increase output voltage above the source value and is rated over 500V so to your point, maybe I need to weigh what the original designers considered and go with their solution too.
J