After cutting my hand on the exposed threads of the Prusa Mendel i2 for many years, it was time for a belated upgrade. I don't want to just buy a new printer because it is not fun. So, I decided to buy a new frame for my printer. I chose P3Steel, which is a laser-cut steel version of Prusa i3. It does not have the potential squareness problem of vanilla i3, and the steel makes it less shaken than similarly designed acrylic or wooden frame printers.

I hope that my new framework can greatly improve reliability and print quality. I want less time to fiddle with it and more time to print. My best hope is that switching to M5 threaded screws instead of M8 used by i2 will improve my z-layer accuracy. I let my old printer work long enough to print out the parts of my new printer, and then happily assembled my new printer.

I did not wait for all the electronic equipment to be installed. No, it's time. I lit it. I look forward to the most square, quietest, and most accurate printing with amazing z-layer alignment. I didn't get any. There are always visible ripples in my prints. My first tendency is that I over-squeeze. Of course, there is nothing wrong with my shiny new mechanism. In addition, I made this printer, and I am an excellent printer manufacturer, and he knows what he is doing. Excessive squeezing looks a lot like a Z-axis problem. So, I adjusted my squeeze until it was perfect.

It is useless. My next goal is a threaded rod. I doubt them because the supplier from whom I bought the kit sent me a very bad smooth rod. They are unhardened, unground, non-straight, ordinary old steel bars. Useless as a bearing. I eventually had to cut the sticks from my old printer and cancelled the project I had prepared for them. So, logically, the threaded rod they sent me was also rubbish.

I ran to the hardware store, bought a brand new M5 rod, some new M5 nuts (this time the real grade), and went home. It took me an hour to straighten the bars until they rolled on the surface of my table with no obvious high points. I installed a new lever instead of the one that came with the printer. no result. In fact, the new rod is more wobbly than the old rod.

I am a little depressed at this time. I started a printout and looked at my printer as closely as possible, noting any visible deviations. I used some rubber tubes as couplers for my z-threaded rod and stepper motor. When I watched the printer run, I noticed that occasionally the threaded rod would slide half a turn in the coupler and then return to the previous direction the next time the motor reversed. Aha! I dismantled another project and snatched some spiral-cut elastic couplings. I changed the rubber tube.

I was taken aback. The print quality has hardly improved. how? I have fixed an obvious visible defect in the printer. Obviously, this sliding is repeatable, so that it is out of the equation itself. OK. Well, I noticed that since the inner diameter of my coupler does not exactly match the threaded rod, the rod is not perfectly aligned with the axis of the stepper. This may cause some swings. So I spent some time using kapton tape to make the threaded rod run as centered as possible without going to the hacker space to use the lathe.

It must be so. I replaced almost all components on Z at least once. Alas, it's still wobbly. I am about to give up any requirements for technicians, engineers, hackers, etc., to learn yodeling or consider wandering. I decided to take a closer look at the printer to see if I could detect what was wrong with Z and what was wrong.

I noticed at that time that when looking closely, the Z rod seemed to enter the guide nut at an angle of 1 degree. If the lead nut enters at an angle, it means that it must bend every time it turns. This will be the reason for the 100% swing, but there is a problem with the theory. How can the rigidly fixed rod and the precisely placed threaded rod not even align?

My first thought was to blame my old printer because this part was printed on it. i2 There was a problem printing squares. Things are usually parallelograms, especially in Z (completely my fault). So maybe the part that fixes the nut and the bearing is not square. If it tilts in one way or another, it will definitely cause this problem. So I carefully printed out the new parts and reassembled my printer exactly as before. I know that my new printer is square. I tested it. So these new parts must solve this problem. They must.

Wandering, right? I mean, fresh air, no need to wash clothes, occasional free food, just walk on the ground. Day after day life... It is still shaking despite the new parts. What else can I fix? So I sat down and thought about it about fifty times. To my surprise, I figured it out. It's time to take out my precision mechanical regulator. 

My logic is like this. One side of the mechanism holds the smooth rod of X in place. There is a hole through the part on the other side. When you assemble the printer, you should slide the rods over one part and fix them in the other part. Well, I'm very smart. I reamed the holes of both parts to make them fit tightly. (For plastic, you can put a smooth rod in the drill bit and pass it through the hole, it will actually cut a very good press fit.)

This results in both parties being able to hold the rod very firmly. The design of toolson P3 is very clever. Keep the rod stable on one side. You can slide on the other side. If the rod can slide the machine seems to be inaccurate, but it is not. The lever only keeps the nozzle in its Z position. The direction they can slide is X. However, the X position of the nozzle is determined by the belt. The rod is not involved at all. The belt is tensioned by the rod to the part where the stepper is fixed. Therefore, the X position of the right bracket does not affect the X position of the nozzle at all.

This is a lot, but in theory, this means that if one side can slide, it should automatically align with the Z bar. If the Z rod is correctly fixed in place, it should theoretically be impossible to misalign the Z rod. That is, unless you deliberately press-fit the through-hole-this is where I failed. So instead of sliding the part along the axis until it is comfortably in equilibrium. I can choose a position on the axis and let it sit down. I have obsessive-compulsive disorder, so I made sure that my shaft aligns perfectly with the end of the plastic part and bend the two rods inwards. This leads to the angle I see.

I tapped it with my precision mechanical adjuster, which gave it enough force to overcome static friction and near-balanced springs. I turned on the printer. Look, my Z has improved significantly. The header image has a before and after comparison.

I still have some problems (maybe because I broke a lot of things in order to solve this problem). My Z is not perfect. My next step is to order a certain length of threaded rod from McMaster, because their threaded rods have tolerances on the threads and are expected to be straighter. I also want to replace stainless steel nuts with brass nuts for better fit.

I also want to reprint the through-hole part and ream it to make it a sliding fit. In theory, this will solve my Z problem, but hey, this might just be another problem. Does anyone else have a similar debugging story?

Fail of the Week is a Hackaday column that celebrates failure as a learning tool. Help keep the fun by writing down your own failures and sending us a link to a story, or sending a link to a failed article you found during your Internet trip.

Great, I like the way you put photos (with annotations) where they are needed in the story.

Why stick to threaded rods, if you want decent precision, get some apex screws.

This problem ultimately has nothing to do with threaded rods, so ACME screws are of no avail. This is a planned upgrade :)

I think he was responding to your plan at the end of the story, where you wrote: "My next step is to order a certain length of threaded rod from McMaster, because their rod has a tolerance for the thread. I hope it can be more Straight. I also want to replace the stainless steel nuts with brass nuts for a better fit."

I am personally not familiar with ACME screws, are we talking about replacing threaded rods with some kind of long machine screws?

what. ACME describes a trapezoidal thread profile that is more suitable for transmitting linear motion. Mainly because its thread profile is much thicker than the triangular thread on the M5 thread. However, there is no guarantee that it is a better thread overall.

For example, the reason cheap threaded rods work in 3D printers is that lead nuts (ordinary nuts) are always preloaded by the weight of the extruder components. The fine thread and thin cross-section of M5, if used with a properly matched nut, can provide the required resolution.

The ACME screws he was referring to did not help, because ACME threads are inherently better or more precise. It helps because it is a lead screw. This means that, ideally, it must have tolerances that make it useful for linear motion. The tolerances that the M5 screw provided by it does not have are: 1. A better fit between the lead nut and the lead screw. 2. Straightness tolerance. Ordinary threaded rods are mainly used for static tension, that is, they become straight under load. The screw bears dynamic loads and defines the position. It must be straight.

The reason I use threaded rods now is price and time. It is very cheap and can provide good results with care. I also don’t have time to machine/design the necessary parts to properly fit the lead screw into the frame in a way that makes me happy.

Price should not be an issue anymore. They sell a pair of products with nuts and motor couplers for less than $10 on AliExpress, etc.

closure. 60-degree threads (English, metric) or 55-degree threads (Wyeth) are self-locking, you can tighten them and lock from the taper of the thread. For sports, you don't want it to lock. Ideally, you need square threads. The problem is that square threads are actually difficult to machine, so the compromise is a 29 degree thread angle in ACME (English) threads or a 30 degree thread angle in trapezoidal (metric) threads.

There are some lead screws that use 60-degree threads, and those that use general-purpose lead screws. I guess this is because of the backlash mechanism they use.

I bought a few of these screws (from a store I trust). From the feel (I haven't used them to print yet.) They have a gap of about 0.1 mm. That is, the nut can move so much without turning.

No, not ACME. ACME is imperial trash and will give you rounding errors when the firmware calculates the number of moves. You need trapezoidal metric threads.

One thing I don't like about the vanilla design is the way the belt tension is transferred to the Z smooth rod and tend to bend it. I have discovered this design from other people who don't like it: http://www.thingiverse.com/thing:160636 and transfer the tension to the X-bar in the direction where the X-bar is strongest. When the X end slides freely on the X rod, it will still align with the Z rod/drive screw, but now you can pull the belt tighter.

This is how it is configured. You are right, this is a better design :)

The prusa design has a serious flaw, the X belt tension tries to pull the Z rod closer. I ran into a similar problem on the Printrbot Kickstarter version. I solved this problem by adding a laminated CNC cutting aluminum "roof" to the steel. Apart from the new foundation, it goes without saying that I upgraded to metal plus the first one I got Chance. Sometimes I wonder how things like http://www.colido.com/product/colido-diy/ work!

If you put the roof on two sets of rods (threads and shafts), you will introduce vibration, z-wobble, and other noise. There is an entire section about it in the RepRap wiki.

You usually want the threaded rod to *not* lock to the top. — If you lock the threaded rod at the top and lock it together with the rest of the rods, any X/Y defects in the rotation of the Z-axis rod will be transferred to the rest of the frame and cause a lot of "noise" to the other axis of the frame.

This is basically a trade-off that allows you to use imperfect rods at the cost of reducing the vertical accuracy of the Z axis (which is by far the most accurate axis anyway), and it prevents these defects from being transferred to the rest of the printer .

As long as your X tensioner only squeezes around the belt top rod, it's fine. It will never pull the threaded rod closer while deforming the shaft or roof.

Oops, it was a typo. The last sentence should read: "If you don't deform the shaft or roof first, it will never pull the threaded rod closer."

No, no... I didn't roof the threaded rod. Threaded rods are shorter than steel rods, so this is impossible. (Like this http://www.matthewtmead.com/blog/wp-content/uploads/2014/08/printrbot-crop.png here). The steel rod foundation is installed on the laser-cut layer frame and the layer frame is not strong enough to prevent the tight X belt from bringing the Z rod close to each other, which causes the same problem as the author here. (That is, the 2-3 degree draft angle on the rod will cause the X coordinate extruder to move every few layers)

Don't buy "threaded rods". Buy some 8mm acme rods with floating brass nuts. It is far superior to standard threaded rods.

I know you think the pole has nothing to do with it, but you are wrong. Not only can the rod bend, but changes in the pitch also produce a perceptible swing. On rolled threads, the pitch changes slightly within a given cycle length. If the two screws are not perfectly matched in rotation, and considering that they are standard threads on a standard nut, this will cause a wobble, you will find a mismatch, and you will feel the wobble. Other factors also have an impact, but I assure you that the main problem lies in your screw/nut combination, coupler, and alignment. Oh, this reminds me not to use those spiral couplers, they are horrible and can cause wobble. Use rigid couplers in this application to get better results.

Over the years, all this has been mentioned many times on the Internet.

I did not disagree with your point of view, nor do I think the rod has nothing to do with it. It just has nothing to do with the specific problem I am solving. You are 100% correct, the random discrepancies I still see in the printed matter can definitely be improved with a properly installed lead screw :)

I didn't see improvement: (The last photo seems to be the same as the first one, except that it has failed.

sorry. I did not mark them clearly. There are two photos in that photo. Before the left. After the right.

I have gone through all these problems and more. It will never be perfect, but it has a balance that I can bear. The basic improvement I made was to fix the x-bar to the x-end of the box, decouple the x/y movement of the z-bar and tighten the belt on the x-bar. The belt needs to be very tight to prevent rebound.

There is a part on thingyverse that uses two nuts to fix the gap of the z-bar. There is a spring in the middle. The cylindrical shape is located below the x end and is not connected to them. Two forks loosely surround the smooth rod to prevent it from rotating. The z-bar must float freely at one end, so it is connected to two points (motor and x-end, but only on the z-axis, x/y decoupled). Any z-swing will not be converted into a nozzle, there is no z-axis backlash (the pitch is much more consistent than you think, I have checked), and the x belt is very tight but does not force the z smoothing rod inward, resulting in In the binding, the haphazzerd movement on the z axis and the jerky movement on the x.

I did not observe high frequency vibration by fixing the x end. When I light up the layers with LED lights, there are still some bands, but it is on the order of 0.1 mm or less, most likely caused by inconsistent squeezing motion, stepping rod or filament diameter.

Other possible reasons for the belt shape are the use of a bang/bang powered heating bed, the cooling/heating cycle causes shrinkage/expansion reduction/elevation of the bed, the gap in the x/y belt causes inconsistent circumference and filling, problems or errors in the setting of the slicer, The filament spool is stuck or produces inconsistent resistance. If you use Bowden settings, it is important to set the retraction speed correctly, and it is also important to set the acceleration and maximum retraction speed. You must not miss the step or binding. Keep the extrusion speed constant, and the circumference will not deviate too much from the filler, etc.

Recently, the nozzle will not start from the glass plate, but from 5mm above it. The complete cold reset of arduino and repetier temporarily fixed it. One of the wires in the z probe is broken, but in general it prints well and occasionally encounters challenges.

But I think all these 3D printers are of high quality and are significantly better than professional printers?

Obviously, my spelling quality is not high: p

They just got a bad rap.

Wow, what a story. Thanks for sharing. Impact maintenance comes to the rescue!

Well written, a thorough Q&A session, thanks to Gerrit Coetzee and all other participating users who asked him. Before setting up the basic printer, I learned a few things, so far, except for the temperature of my heater, it is ok, I fixed it about an hour ago and everything is fine. Now I can make my Fast Traxx rebuild case. {Bottom spring etc... additional rollers... Now what I need is to be able to purchase or make new treads from some kind of stable rubber (such as base material).

Like you, I upgraded my prusa2 to p3steel a few months ago.

But I decided not to use any printed parts.

I bought an aluminum X-axis assembly from aliexpress (I did have to reposition the larger box for the smooth rod because it was designed for different smooth/threaded rod spacing) and compared it to the frequently mentioned 8 Combine with mm trapezoidal rods.

This is what it looks like: http://bernieke.com/posted/p3steel_x-axis.jpg

These are the parts I bought: http://www.aliexpress.com/item/Reprap-Prusa-i3-3D-printer-parts-no-motor-aluminum-alloy-all-metal-X-carriage-X-end /32393120642.html http://www.aliexpress.com/item/3-D-printer-accessory-stainless-steel-T8-2-D8-Trapezoidal-screw-Makerbot-lead-screw-threaded-rod/32446423338. html?spm=2114.01010208.3.1.rKMCGp

Very satisfied with this setting!

Thanks Gerrit! I don't have a 3D printer, but it is on the list I will eventually use. I learned a lot about possible troubleshooting routes. . Thank you for the hammer joke. Sometimes clicking on it can solve the problem, sometimes it requires a real hard hit. If it fails, you go to the garage to get a sledgehammer!

no problem! Glad to help :) 3D printer is a useful but very frustrating project, haha. Yes, there are different sizes of mechanical adjusters, it is important not to perform accidental expansion. (Https://en.wikipedia.org/wiki/Bob_Widlar)

A serious question raised by a person who has no experience in this technology. Has anyone encountered problems due to EMF or radio noise? Normal and boring ordinary laser printers are always in a horizontal state when they are working, until the CPU moves a few inches. Just want to.

Yes. It happened. When I entered it, something strange happened to me. Now I carefully routed better.

The original Mendel and Prusa i2 had a way to lock the X smooth rod in place, usually using a screw in the adjuster to resist the force of the X belt pulling the Z smooth rod, so Z can track up and down correctly (assuming the Z rods are parallel). I haven't built i3 yet (my i2 variant works well and does not require additional height...) but I am really surprised that this key feature is not retained. However, it is best to apply a small amount of preload on the Z bearings, otherwise they will "float" and will move laterally with the movement of the X carriage. It is best to put some load on the bearings to make them work properly. Regarding lead screws and threaded rods, an important consideration is that the layer height you use is equal to the total number of steps of the stepper motor, rather than a fraction of the number of microsteps. Failure to do so may result in inaccurate movement. See http://prusaprinters.org/calculator/#optimallayer. The M5 rod is usually good because it has a pitch of 0.8 mm. Use a motor with 200 full steps per revolution, which means you can get a good layer height at 0.1 mm intervals. However, at an interval of 0.1mm +0.05, you cannot get a good layer height! If you use inches instead of millimeters, it will be difficult to get the correct layer height throughout the entire step. Use the Prusa calculator above to get the best floor height. The straightness of the lead screw/screw is very important. It is in line with the motor shaft, but "slight" swing should not be a big problem. As long as the lead screw is not restricted by the top, it is only connected to the motor and the X axis. Otherwise, the movement will be transferred to the X axis. Personally, I like a stepper motor with a built-in 8mm lead screw, as shown below: http://www.alibaba.com/product-detail/lead-screw-nema-17-8mm-lead_1762893359.html. No motor coupling can mess up, and an accurate lead screw. They provide the best movement for the Z axis.

Please be kind and respectful to help make the comment section great. (Comment Policy)

This website uses Akismet to reduce spam. Learn how to handle your comment data.

By using our website and services, you explicitly agree to the placement of our performance, functionality and advertising cookies. learn more