To rigidify or not to rigidify? - vibration issues
-
@Nxt-1 you would be surprised how much the proper mounts help. Belts don’t really transmit vibration as they are already vibration dampers. If you don’t have good linear bearings then that will cause vibrations but i suspect you do have good linear motion components.
If you are using aluminum extrusion filling it with something viscous (hydraulic oil or honey) would help in 2 ways. One it would dampen vibrations and it would also fill the resonance chamber with a fluid that would change the speed of the sound waves.
-
@baird1fa said in To rigidify or not to rigidify? - vibration issues:
@Nxt-1 you would be surprised how much the proper mounts help. Belts don’t really transmit vibration as they are already vibration dampers. If you don’t have good linear bearings then that will cause vibrations but i suspect you do have good linear motion components.
I don't disagree that proper stepper damper can do wonders, I have actually used them with success in the first iteration of my printer (nema 17 that time). The issue here is that the motor rigidly connects to a 2nd shaft which holds the pulley. That shaft is supported by two bearings, which sit inside the whole bracket. So there is a rigid connection from the stepper to the axle through the bearings into the frame. What I said is that a damper will not solve that part of the problem.
I indeed use high quality THK SRS_WM rails, they will not be a source of significant vibrations.
If you are using aluminum extrusion filling it with something viscous (hydraulic oil or honey) would help in 2 ways. One it would dampen vibrations and it would also fill the resonance chamber with a fluid that would change the speed of the sound waves.
If I had not previously filled the towers with just pain sand, this would be a way easier solution. Right now, I don't have means of getting the sand out of the 1,5m long extrusions without turning the printer upside down, which I am definitely not doing unless it is my last resort.
-
I think what you do is nice, but I honestly don't think it will be worth the investment. Nevertheless, do you have some close up pictures of prints? Please show, what is pushing you forward. I wonder, whether it wouldn't it be smarter to reconsider the effector. Your effector looks rather heavy because of the watercooling. It might be wiser to enhance the rod mechanics and widen the arm distance.
-
@dgrat said in To rigidify or not to rigidify? - vibration issues:
I think what you do is nice, but I honestly don't think it will be worth the investment. Nevertheless, do you have some close up pictures of prints? Please show, what is pushing you forward. I wonder, whether it wouldn't it be smarter to reconsider the effector. Your effector looks rather heavy because of the watercooling. It might be wiser to enhance the rod mechanics and widen the arm distance.
Why do you think adding concrete to increase mass, in effort to reduce vibrations will not be worth the investment? (I am not sure either until I try it, just curious)
I do not have pictures of recent prints, heck this version of the printer, with the new rails, steppers etc just printed 1 little cube to confirm all was working. Unless I am missing something, I do not really see why a print closeup is relevant here? I am not trying to improve print quality, only battling audible vibrations.
The pictures/video of the effector in it current state are kind of misleading. Normally, all the water cooling stuff, as well as the extruder is fully supported by the 4th axis. Currently it is not as since the upgrade, there were some unanticipated changes how the 4th axis fits in, and well, it doesn't fit at its designed location. Rest assured once I redesign the 4th axis to fit like it once did, the weight on the effector is about identical to a regular bowden fed smart effector.
-
Why do you think adding concrete to increase mass, in effort to reduce vibrations will not be worth the investment? (I am not sure either until I try it, just curious)
Well, the effector of my Deltas is ~0.13kg. With 200m/s travels and accel values of ~3000-5000 mm/s² i get usually 0.2g of vibrations. This is nothing in comparion to my Ender-5. Movable x-axis probably 2 kg. Additionally, the case of the ender is already way less than the weight of my Deltas. I don't underestimate if I say, that the whole Ender is shaking if I set acceleration to 3000 mm/s². Haven't taken a measurement, but but acceleration is probably in the order of 1-2 g, maybe more. Print quality on both machines is decent, beside stringing issues (Deltas with 200 mm/s travel and 60 mm/s print mainly because the BMGs become less reliable to deliver filament at 0.25 mm layer heights). Also I think it is just more economical to reduce weight of the effector than increasing weight of the frame.
please post some print pics
-
@dgrat I am sorry that I am going to be blunt, but I really think you don't get what the issue is I am trying to solve. I am not looking to improve print quality, there is no issue there. Even if I remove the effector, the arms and all else you desire except for the motors and belts the frame still resonates at about 800Hz. Once again the effector and its mass, large or not, is not relevant in this case.
Again, I have no prints done with this machine yet, so stop asking for print pics. If I was trying to solve wavy prints or stuff like that, sure. But not in this case.
-
Have you enabled interpolation on your machine? I have the feel, that dampening 800 Hz vibrations is also not that easy. There are too many Alu parts which can resonate, even if you add endless weight.
-
@dgrat said in To rigidify or not to rigidify? - vibration issues:
Have you enabled interpolation on your machine? I have the feel, that dampening 800 Hz vibrations is also not that easy. There are too many Alu parts which can resonate, even if you add endless weight.
Its not about adding endless weight. Its about adding the right kind of weight.
Metals like aluminum are essentially great big crystals, that propagate vibration really well.CNC machines traditionally use cast iron, not steel, because its actually much more of a granular structure. Because the vibrations are having to cross lots of boundaries between the grains, the materials soak up a lot of the energy. Epoxy granite is used a lot because it has this property, its a mix of aggregate, with epoxy as a binder. Its almost all boundaries between rocks and epoxy, so its really good at absorbing vibration (plus you don't have to wait weeks for it to fully harden like cast iron, and you don't need a foundry).
Regular concrete isn't quite as good as cast iron or epoxy granite, but its still much better than aluminum, and its so cheap and easy compared to those others...
There are lots of people out there building cnc machines using steel or aluminum bonded into concrete or epoxy granite, its a very effective.
Will there still be bits that vibrate or resonate? Probably, but this is a test that will cost tens of dollars and is fun. And based on what the machining world has been doing for over a century, its probably going to work. -
A quick update on slab 1. After a little over 48h of curing, I demolded the slab and it looks great. The surface you see on the picture is actually the back side that was inside the mold and it is quite smooth, I'm happy. I will wait until tomorrow before I start drilling holes, that would bring the final cure to about 36 hours.
On thing I noticed it that especially the mold side surfaces are quite dusty. A quick google leads be to believe that this might be a mixing/moisture error, but I'll be the first to admit that I am not a concrete expert, not at all
. I'm thinking that if it keeps being dusty, I might seal them with the same glossy sealer coat I used on the epoxy granite base of the printer. If the concrete is effective in what we want it to do that is.More pigment for the other slabs and M8-tslot nuts should arrive tomorrow. So hopefully I can test effectiveness with one slab already by tomorrow and if successful, cast the other slabs by the end of the week.
-
Just finished drilling the holes and well, the front looks nice. Concerning the back, I see two options. Option one, I blame the first guy who comes up in my @ list for not warning me about concrete cracking. Option two, you guys never mention this again, I never mention it again and no one will ever know
.I guess I am a reformed man now. I fully believe in molding in the bolt holes. This is what I plan to do for the two remaining slabs (and also add a single rebar rob in the middle while I am at it, as suggest by others.
I will post anther update later today how effective this single slab is, as I return to the hw store and return my too short bolt for the proper length ones
-
@Nxt-1 said in To rigidify or not to rigidify? - vibration issues:
Option two, you guys never mention this again
deal!
-
As promised, the update. For starters, attaching the slab to the frame is no joke. Aligning the tslot nuts, keeping the slab at the correct height and inserting the bolts is no easy feat. That aside, I attached the slab and did some noise test moves (straight op and down). I could feel quite a bit less vibrations,, if any at all actually, in the one tower. On the other hand, I must admit that I did not really hear something spectacularly different.
As I got the materials to make the other slabs already, I will still make them, attach all three and see how that turns out.
One note that could be important. I did not source the rubber that sits between the slab and the tower yet. As the slab is not perfectly flat against the surface of the tower, I did not tighten it down very hard. I did this to prevent cracking the slab, but this means pressure is on the low side, and not all surfaces of the slab actually make contact with the frame.
For the rubber sheet, I was thinking of getting something in the order of ~5mm thick, to allow for quite a bit of compensation in the flatness of the slab. Thoughts on this?
-
@Nxt-1 said in To rigidify or not to rigidify? - vibration issues:
For the rubber sheet, I was thinking of getting something in the order of ~5mm thick, to allow for quite a bit of compensation in the flatness of the slab. Thoughts on this?
On the other hand, as I write this I am thinking whether I would not be a better idea to go with something thinner to better preserve the coupling between the frame and the slab. That is the whole point of this exercise after all. I could source 2mm EPDM or 2mm NRB quite easily from work I believe (either one, I doubt it really matters here)
-
@Nxt-1 looks nice, maybe I'll use something similar for the CNC machine.
-
@Nxt-1
I'd try 2mm first, especially if its easier to get. You can always double it up. -
@Nxt-1 Not sure how much help it is to you, but I had a go at calculating the first mode of vibration for your towers. Assuming they are individual simply supported beams (ie only looking at one tower, held at both ends) 1.5m long, and the same cross section as the ones I found on google. That gives you a first mode in bending of 1600Hz. Given how yours are mounted at either end, I'd say yours would have a shorter effective length (so a slightly higher natural frequency).
Still, it's within the same ballpark which I thought was interesting. I'm guessing that because your belts also run the length of the towers, the excitation from the motors can also travel down those so excite it from both ends?
Have you tried just bolting a length of metal to the outer side of the extrusions? (so they effectively become 60-150 at one point). You could even do this with the rubber sandwiching to get extra damping?
Adding mass tends to drop natural frequencies (as well as dampen them because exytra joints), but you might find you just shift it to a lower frequency that you still excite. You might be better adding stiffness to increase the frequency to a less audible range.Either way, the blue concrete is ingenious and looks brilliant so fingers crossed it works for you!
-
@engikeneer said in To rigidify or not to rigidify? - vibration issues:
@Nxt-1 Not sure how much help it is to you, but I had a go at calculating the first mode of vibration for your towers. Assuming they are individual simply supported beams (ie only looking at one tower, held at both ends) 1.5m long, and the same cross section as the ones I found on google. That gives you a first mode in bending of 1600Hz. Given how yours are mounted at either end, I'd say yours would have a shorter effective length (so a slightly higher natural frequency).
Still, it's within the same ballpark which I thought was interesting.Not to downplay your calculations, especially with limited information available, but 1600Hz is quite a bit from the measured 800Hz no?
I'm guessing that because your belts also run the length of the towers, the excitation from the motors can also travel down those so excite it from both ends?
I have have wondered about this as well. The belt certainly do vibrate, so I figure they are a source noise themselves. However I doubt they can really carry the vibrations from the top pulley shaft to the bottom pulley shaft. I would think that they are far from rigid enough do do that?
Have you tried just bolting a length of metal to the outer side of the extrusions? (so they effectively become 60-150 at one point). You could even do this with the rubber sandwiching to get extra damping?
Adding mass tends to drop natural frequencies (as well as dampen them because exytra joints), but you might find you just shift it to a lower frequency that you still excite. You might be better adding stiffness to increase the frequency to a less audible range.With "so they effectively become 60-150 at one point", do you mean bolting a slab of 30x150x1500mm of steel to the side? Because those would weigh 50-60kg a piece. And I am already struggling with the weight of the concrete slabs at it is
. Furthermore 30mm of thickness will not add significant stiffness to the frame, I would be better of bolting a piece of square hollow structural steel tubing to the towers. They could be much larger and thus much stiffer for the same total weight.This leads us to the start of this thread though, will increasing stiffness cause the vibrations to lessen? The consensus seems to be I wont really. Now the weight of all that steel would still be as effective as the concrete would be, with the caveat that steel is not as effective at damping vibrations as concrete. So it might become more of a resonator itself.
Either way, the blue concrete is ingenious and looks brilliant so fingers crossed it works for you!
Thanks, I use all the finger crossing I can get
-
@Nxt-1 I was meaning bolting another aluminium extrusion, either a full 30150 or even just a single 3030 to keep it nice and light. More thinking it might be a lighter weight solution!
One other thing to consider is to do all of the towers slightly differently. That way you'll get some separation of the resonant frequencies so the towers don't interact and further excite each other. I'm guessing you'll get this a bit for free with the concrete though -
I know from CNC machines that mixing materials is a good idea. They mix steel with aluminium, so the eigenfrequencies are different. One has to think of the different material expansion with different temperatures however (even between the different aluminium alloys), otherwise you bent your aluminium extrusions!
-
We have arrived at the moment of truth.
Friday I assembled the molds with, as you can see, the white pieces in there to make bolt holes and a piece of rebar in the center. Saturday I mixed up another 3 bags of concrete mix and did the pour. Then finally on Tuesday, I demolded the slabs and attached them to the printer. Before I go over the results, I'll share some more of my experiences and remarks.
Removing the wood mold pieces went very very easy just like last time. The white inserts for molding the bolt holes were not. From the 40 spread over the two slabs, just four came out with gentle percussive action. I did have a 1,5° taper angle on the pieces, but I ended up needing to use a drill to get rid of most of the plastic. The last cleanup was done using a hot air station at 200°C. The holes are not perfectly smooth around the edges, but no major issues.
Last time I noticed that the sides touching the mold wood were quite dusty and remained quite dusty no matter what I did. I did a quick google and assumed bad mixing was to blame for that. On the two new slabs I notice the same however, yet that could be due to me using the same mixing ratios. But actually I now believe it is due to the wood absorbing moisture from the concrete while it is still wet, thus effectively causing localized bad mixing. I figured this as I poured the excess concrete in a plastic bucket and after drying is was completely dust free. The wood absorbing moisture also explains why the top surface did not show the same issue.
I decided to add a piece of rebar in the center to give some added strength as the piece is quite thin compared to its length and did not want it cracking in half while working with it. The picture below shows how I suspended the rebar with some metal wire between the printed pieces. After demolding I noticed something interesting. In both slabs I can feel something moving quite distinctly when I jiggle a little with the slab upright. It really feels like the rebars are loosey goosey in there
Don't ask my why, and I guess at this point it doesn't really matter that much either.While playing with the first slab, I became clear that it did not sit perfectly straight against the tower. I got a piece of ~3mm rubber matting from a local home improvement store that sold it by the meter. I cut six pieces sized and holed correctly for the slab so I could double up on the thickness if needed. In the end I did end up just using one piece per tower, but oh well.
Now it is finally time for the moment of truth, attaching the slabs to the towers. At a little of 30kg a piece it was not trivial but I managed to get them one there. Some bolt holes are left open as one or two holes are badly located and do not line up well enough, or the t-nut fell down while playing with the bolt and I did not feel like removing the slab to put a t-nut back
. Once I attached the first slab and even the second I did some test moves and was not really impressed with what I heard. But the third was the charm I guess, definitely an improvement. I can no longer feel any vibrations with my hands on the towers.I must nuance the results a bit sadly. While the difference is certainly there and is quite noticeable, it's not perfect and not really acceptable even. I went from being able to hear the printer form all areas of the house to only audible in the next room. To combat the last noise I have some proper nema 23 dampers on order and I am most likely going to construct some sort of enclosure for the entire printer, that I line with sound dampening material.
Congrats to the people that made it this far down the wall of text, enjoy the pictures and drop an upvote if you think these longer post as worth the time
Many thanks to all that pitched in with ideas and suggestions.
Just for the curious among us, this ended up costing my about €165, of which almost half went to the M8 t-nuts and the blue pigment.
