Do we need extrusion backer for Voron 3D printer?
Vložit
- čas přidán 1. 06. 2024
- I usually see some kind of metal backing on Voron 3D printer gantry and I wonder why. People claim it helps to prevent bimetallic thermal expansion effect on their 3D printer axis due to different materials, aluminum extrusion and steel linear rail. In this video, I am doing little experiment to see if one really needs to purchase the titanium backing plate to curve the bimetallic thermal expansion effect. Do we really need it? There is no deny about bimetallic thermal expansion but how much does it affect when most Voron chamber temperature is about 50°C to 60°C at most.
I almost melted my linear block plastic if I didn't turn off the heater.
Mitutoyo Measuring Tools: ((affiliate links))
Caliber: amzn.to/48G9zuG
Dial Test Indicator: amzn.to/3tOl44z
Surface Plate: amzn.to/48CLuoQ
Chinese Measuring Tools ((affiliate links))
Dial Test Indicator: s.click.aliexpress.com/e/_DEO...
Micrometer: s.click.aliexpress.com/e/_DkS...
Caliper: s.click.aliexpress.com/e/_Dnd...
▬ Get Connected ▬▬▬▬▬▬▬▬▬▬▬▬▬▬▬
► CZcams: / lavithhcm
► Twitter: / lavitht
► Facebook: / lavitht
► Instagram: / di.ai.why
► Personal Website: www.thinhla.com
► Github: github.com/cucac
I think people like to remove some variables. I believe most understand that the extrusions are not perfect, the rails are not perfect, and the mounting hardware and fasteners are most certainly not perfect. But heat expansion is one variable that is difficult to test for, and backers are one way to minimise compounding tolerances. It is not necessary for sure. But it is a hobbyist machine, people will tinker and mod this to the moon. For most, the out of the box (BOM) is good enough. But some people chase perfection by any means, and for them it is a different kind of fun, to watch the improvement to their mesh graph. But as with anything, the closer to 100% you get, the more cost there is involved, and the diminishing returns.
In short, this information is useful to those to determine how much expansion deflection to expect. But chasing .05 mm is up to the individual to determine its value.
Thanks for sharing!
Yup...yup we are not making rockets to go to the moon. 🤣
I like these type of tests because I don't have the equipment to do it myself. Thanks!
No problem! Thanks for dropping by and comment ❤️
It's cool that you're bench testing this and it's neat seeing the bimetallic flexing measured.
Modeling and measuring the bowing on the y axes has typically resulted in values of 0.1mm to 0.2mm deflection with a 40° temp delta depending on the printer size. It's not a huge amount but is enough to effect large first layers or adhesion between large first and second layers.
Backers truly aren't necessary as a long heat soak and a bed mesh will accomplish the same thing. You just need thermal stability once the print starts. I just don't feel like waiting that long before each print so I choose to use backers instead.
Also when this phenomenon was discovered the x axis had a bottom mounted mgn rail which doubled the effect making backers even more valuable.
Anyone can test this in their own Voron printers to measure the effect. Take a mesh with a hot bed before the chamber has heated up. Then center your toolhead, turn on your part cooling fan and wait 30 min. Take another mesh following a heat soak. Compare them side by side and the change in the mesh should be very obvious.
If you don't see a change in the mesh then congratulations! You don't need backers. Or if you don't mind waiting for heat soaks you don't need backers. For impatient and particular folks like me it might be worth it.
There is few micron difference between hot bed (100°C) and cold bed (room temperature) with my bed mesh. I just did a scan with cartographer probe. I 'll post a picture of heighmap soon. For me, I'd use the budget for the backers towards more "premium" Mitutoyo measuring tools. 🙂 because I can reuse the tools more often than the backers
Changed my gantry to a titanium tube. Titanium has almost the identical thermal expansion to steel, leading to negligible bi-metal effect.
Right now the extrusion can freely expand. Tests with extrusion ends bolted on may show different results. Any chance you can repeat the test with ends fully constrained?
I would recommend putting a scribe mark on the extrusion and rail at each end of rail and the middle of the rail before heating. I suspect that your rail has slipped on the extrusion, that is the screws holding rail to extrusion are not fixed resulting in the rail slipping on the extrusion. Not that this is could not happen on the printer, but would account for the low deflection.
For example assuming the rail is 500mm long and made out of steel with a coefficient of expansion of 10.8mm/mm/C then a 30 degree change in temperature would result in 162um expansion of the steel rail in length. While the aluminum with ~24e-6mm/mm/C thermal coefficient of expansion would have increased in length 360um. This should result with the back of the linear rail being longer than the front by ~200um (0.2mm), which would cause larger deflection in the center than you measured. My high school math, which may be wrong, says that an 8mm thick rail that is 200um longer on back side should have ~1.5mm deflection in the center. However if the steel rail was not mechanically secured the aluminum extrusion then bolts could have slipped on extrusion and resulted in less bowing. If this was/is the case then the scribe marks made before the heating would not align when heated, and would account for lower deflection measured in this experiment.
Additionally measuring of the length before heating and after would confirm the temperature coefficients and help validate the test.
Note that engineers will argue for days over the result of an experiment that takes 15 minutes. The reason engineers do this is that engineers develop mental and mathematical models of systems. They use these models to predict behaviors. In fact a rule good engineers have is that they never do an experiment which they have not predicted the results of the experiment before hand. That is purpose of experiment is to prove or disprove an hypothesis.
In your test you had hypothesis that the bimetal thermal expansion was not significant. However you did not predict what the bowing should be, additionally you did not define what is a significant amount before the test and then justified significant amount afterwards. This is not bad and I commend you for doing the test! However predicting results before hand and then documenting why it is not what was predicted would provide greater understanding and knowledge.
For example if the answer is to have some T slot nuts so that the linear rail can slip on extrusion with thermal expansion, then that would have been good information. Maybe make T-nut out of PTFE to plan on the thermal expansion and slippage.
Thank you very much for checking this.
This convince me that Formbot kit are good.
Regards
Great video. Finaly some tests with proper tools for machining. Do you think you could test the slop in the car of those rails and maybe compare them to machinist grade rail. That would be vary interesting.. Thanks for the great content.
"Slop in the car?" You mean the "play" in the sliding block then? Honestly, I feel no play at all. If you feel any play from the sliding block you should replace it. They (manufacture) should "preload" the sliding block to the point that there is no play. Too much preload, it will experience premature wear, too little or none it will be sloppy and might have play between block and rail.
@@lavithhcm yes that’s true but I do here a lot of the voron and vzbot builders about buying Hiwin rails because they supposedly have more preload and are better then the AliExpress rails.
Bit lower than the 0.15 i guessed but that could be due to mgn9 vs 12.
But still less than the variance on most beds so its fine if you heat soak the printer before probing.
Colour me impressed. I did expect the distortion to be much greater , by a factor 10 ;) ( IE 1 mm instead of 0.1 mm )
My point in the video is I am happy with 50°C chamber temperatute and my ABS print comes out just as nice if not better. So, i care not about having to add titanium backers.
50 degrees is pretty cold. Try throwing a blanket over your printer, or add insulation on panels, plug blind joints, add bed fans for convection effect and you can easily get over 70 degrees.
Can you heat it in owen to 70 C with hot air and then measure it?
Also, that is an mgn12 rail right? that would go on the x, with mgn 9 on the y. Its the y deflection that matters most, so the mgn9 would be less stiff. Its also not that extreme of a case, my chamber temp reaches 60C. If you are into ABS this will be an issue.
Yeah, you gotta watch out when your chamber temp gets high. Mine doesnt get that high though. I guess there will be more deflection deviation for thinner material.
FYI: while 0.05mm difference is not enough to break something, it is very much noticeable.
It probably won't cause any mechanical problems, but it will result in reduced cosmetic quality.
Also FYI, 0.003" is 0.0762mm
But your point is mostly correct.
No backing should be ever used and these titanium plates are plain stupid and a marketing scam.
The actual real solution is to just preheat all the parts before tightening the bolts.
And this is actually super easy to do and it doesn't require any special tools or equipment.
Just keep all the bolts loose, turn on the heat, let it saturate the parts and then quickly tighten em.
Aluminium and steel can store more than enough heat to let you even a couple of minutes to tighten the bolts.
So right on point! !👍 we are not making rocket engine to go to the moon. If some want to have a perfect look, they should look into finishing process. Dimensional wise, parts come out the same with or without the backing plates as some have already said. Bling wise, yeah it is a must. Money wise, expensive with paper weight value. 😁
Wait, is it 100 microns difference? 0.1mm? Damn that's a lot. I have 70C in my chamber.
Well, i think cold nozzle vs hot nozzle is more than that. Anyways, important thing is that i heat soak my printer chamber 15-20min before print_start or start_print regardless of my chamber temperature. With or without backers, i got no issue.
I heat soak until my chamber reaches 70 C. I have a lot of insulation and bed fans. It can reach 80-ish.
@@sabahoudini yes i heat soak my printer as well ~20min to make sure the spring steel plate is heated evenly and actually 100°C. And this can be easily indicated by my chamber temperature sensor. I wonder how you insulate your printer to get chamber to 70°C. I plug all possible "holes" but my max is 50°C. I guess you must use foam material for panels.
@@lavithhcm Yeah, insulation panels and bed fans. The bed fans will increase temp by 10C alone without insulation but I only use them when heat soaking. The chamber temp will not reach max after 20 minutes, you can maybe get to 55 after 45 minutes or so. Also for ABS/ASA I use 110 degrees bed. The difference in print quality is huge. Even 350mm objects with tricky geometry will not warp and the parts will be much, much stronger even when you print at 300mm/s.
You want to do this test with the print head on since that adds a lot of weight, there is usually quite a bit of bowing just from that alone.
No matter what we do, there will always be thermal expansion in aterial. Even with real CNC machine, there is usually 25-50 micron difference shown in machined part between cold and warm machine. We adjust Z offset in tool settings to accomodate for that thermal expansion.
@@lavithhcm Which is why CNC machines are adjusted to be true when warm especially those which are used for mass production like Swiss lathes the first parts will always come out a bit out of spec until the machine comes up to temperature.
Just another way to think about it, 0.1 is half the layer height. If you are printing something with a large first layer that means over 30-40 min, gantry can deflect by 20-50% of layer height, and first layer tuning is generally .05 between to high and perfect.
0.1mm is extreme printing enviroment, you gotta evaluate how yours is. i know mine is somewhere 50°C and i have no need to go higher chamber temp. Besides, that deviation is negligible well at least in my use case.
@@lavithhcm Yeah, im shooting for 60C as I print mostly ABS on the Voron. many people who are primarily interested in ABS will be shooting for that temp range. Perhaps that's why people start to feel backers are necessary. You really do start to run into first layer issues if you don't figure something our. There is actually a software solution as well that 'characterizes' the thermal change. It's called virtual gantry backers. Its free and adds no weight, and works well.
You'd usually thermal soak the printer anyways if you were doing larger parts, and how slow are your first layers if they take 30-40 minutes? If the z-offset is correct 100mm/s@5k is more than doable for most materials.
@@twanheijkoop6753 You don't have to heat soak to far if you use 'virtual gantry backers', and z-thermal compensation. I use both. I get the chamber up to 45C, and the frame is usually in the high 20Cs. I have a 350 and print fairly full beds. My first layer speed is 110mm^2 and it can take over 45 to do first and second layer. That's where you will see inconsistency from thermal expansion, first and second layer. It will look like under extrusion, but its not.
Your experimental method is suspect. Have you looked at the results from the guys who actually measured deflection on beams installed in enclosed printers and allowed them to equilibrate?
Well i might try that. Poke a hole on the side "foam" panel and keep measuring with the indicator.
@@jimmer411 if i have extra bling i'd spend it on Mitutoyo dial test indicator which not many people would consider it a must. Chinese made are ok but they are not anti-magnetic.
I'd say its more for aesthetics than anything else at this point. Maybe few years ago they did something but not anymore
Why would time make any difference? Physics do not change.