Re: Odd circle test results
The good news is that if this works there will be a printable solution. I'll take it a step farther after that and upgrade my carriages entirely. Why stick with junk when it isn't necessary
You are not logged in. Please login or register.
SoliForum - 3D Printing Community → Solidoodle Discussion → Odd circle test results
The good news is that if this works there will be a printable solution. I'll take it a step farther after that and upgrade my carriages entirely. Why stick with junk when it isn't necessary
I'm not sure about that , but at this point who knows.
here's a theory I have which might explain the deformation to not equal amount of force between the left side of the back rod and right side which I think causes bending of the rod. Its very exaggerated in this pic to make it simpler.
------
just a thought.
Also, I've reflashed my printer. Same.
That's exactly the idea. That bending rod causes loss of linear motion at those points where the direction changes, causing circle dents and sloppy corners. The reason you only see this in direction changes is eventually the linear force overtakes the torsion, but there is a delay before this occurs.
Wait. Not quite man. I'm not talking about the back Y rod. I'm talking about the 2 X rods. Those are what is bending more so than the back Y rod. And they aren't bending so much as ending up at an angle temporarily during direction changes. If you were to plot the printer's y position on a graph from left to right where this direction indicates time, you'd have a sine wave with flat tops and bottoms. Does that make sense? The flat portions would indicate the points where torsion causes loss of energy and the derivative at that point would be 0, but rather than being an instantaneous point (as a usual wave) it would have a longer period.
This issue has nothing to do with the back Y rod. This is all about the way the carriage connects to the two parallel rods.
I'm not sure about that , but at this point who knows.
here's a theory I have which might explain the deformation to not equal amount of force between the left side of the back rod and right side which I think causes bending of the rod. Its very exaggerated in this pic to make it simpler.
------just a thought.
Also, I've reflashed my printer. Same.
I think you are missing the "bushing" on the side that the rear y rod slides into on the frame of the enclosure. One exists on each side even though my pic only shows the left side.
Both of you are very confused. Perhaps I have not explained it clearly enough. This is related to the two parallel rods running along the side of the printer. Look at the way the Y carriage attaches to the right rod. Notice there are two Teflon blocks, one on top and one on bottom. Notice also that you can move the right side of the carriage up and down. This is most likely the reason for our trouble.
I'm not good at drawing 2D images, but think about what happens during direction changes. That side of the rod has 2 unnecessary and defeating degrees of freedom: it can move up and down, and it can come out of square with the left side.
The "bushing" I mentioned may in fact be a linear bearing. It's located where the X motor is.
If you still don't get it, wait until I can get home and I'll design something to replace that bottom Teflon block. It will basically be a square with an 8mm arc cut out of it to fit on the rod and remove the extra degrees of freedom. The challenge is designing it so it's not too loose and not too tight. It will also have a limited life because I'm pretty sure lithium grease will destroy the ABS over time.
This will only be a hack. The only permanent solution is to upgrade the carriage, which is my next step after this experiment. SUCH A SHAME. I don't know how a company that's been designing printers for years and has had such an awesome community showing them better ways to do things is still screwing it up. I can't WAIT to see how they screwed up the Press!
OT: this is why the e3d rocks. I couldn't get this to print without supports with the sloppy stock hotend. E3d had no trouble
As far as the carriage rocking up and down goes, This gadget worked wonders for me on my Solidoodle 2. Now sure how similar it is with newer models though:
As far as the carriage rocking up and down goes, This gadget worked wonders for me on my Solidoodle 2. Now sure how similar it is with newer models though:
I originally was trying to think of a solution like this. The problem is that was designed for the X-carriage where you have two parallel rods and bushings on one and Teflon blocks on the other. This helped offset non perfect rods and actually worked alright. That thingiverse part helped take up the up/down motion while leaving enough flexibility to deal with the imperfect rods.
In this case it is used on the right y-carriage and even though there are 2 rods parallel to one another. It is simply too far apart from the left side with the bushings. This is actually allowing the 2 x-axis rods to twist and in doing so binding the teflon blocks on the right Y-carriage. This explains why jagowilson was continually dealing with the bolts holding the blocks coming loose. There are a whole lot of different forces at play here so anything that can be done to stop the twisting of the x-rods at the right Y-axis rod carriage should help.
jagowilson has a good plan to design and test this theory so for now I'd just stand by and wait.
I'm not sure about that , but at this point who knows.
here's a theory I have which might explain the deformation to not equal amount of force between the left side of the back rod and right side which I think causes bending of the rod. Its very exaggerated in this pic to make it simpler.
------just a thought.
Also, I've reflashed my printer. Same.
This is a great illustration and exactly why that drive rod needs to be supported well. I understand that it is exaggerated to make your point. However, if your rod is actually bending on that side then you need to look into why. The bushing on the right side is more than adequate to support that Y-axis belt. That doesn't mean that it couldn't be better it just means that if your belts are that tight than you should loosen them some.
The left side is different in that you have the drive belt to the motor pulling down and it is usually run rather tight to help reduce backlash. When combined with the belts pulling forward as well it would deflect that rod just enough to cause binding in the left bushing. This was more noticeable on newer machines as things were still tight.
To work around this problem we added bearings in pillow blocks to help hold the rod aligned. Somewhere along the line SD adopted this idea and put another bushing in there on that bracket. The thought was a good one but I feel they maybe fell short on the real world testing. With bearings this works phenomenally well. The added resistance of bushings and the likelihood those bushings are NOT perfectly aligned means in the real world it creates more problems than it solved.
I bet the right side of your carriage moves up and down. Try it.
No, it doesn't
There are two screws on the right side of the right side of the carriage. The bottom one can be adjusted. You can loosen the bottom screw (I think it was the bottom one), move the white nylon bushing (or whatever its called) with the screw going through it higher, then tighten. it won't move up and down anymore.
If I understood you correctly.
So you think my assumption that its a slight bending of the back rod causing this problem is wrong?
I can't think of anything else.
If I understood you correctly, the right side of the carriage being able to go slightly up and down is not the case either because mine doesn't and you can adjust yours too...
We know exactly where the symptoms lie, has anyone pushed them? When I get home, I'm going to try and exacerbate the error we are seeing. I plan to do this by drawing a line across the platform from right to left with a degree offset of 1 degree over the length of the platform and see if the line printed forms steps instead of a straight line. If the printer is capable of performing this test, it could just be a software problem.
The STL files I attached will be my test. One test is with 1 degree, the other is half a degree.
So you think my assumption that its a slight bending of the back rod causing this problem is wrong?
I can't think of anything else.
I don't think it's wrong I think it is more likely the problem lies on the other side unless your left side is missing the bushing.If I understood you correctly, the right side of the carriage being able to go slightly up and down is not the case either because mine doesn't and you can adjust yours too...
It isn't so much that it can go up and down as much as it can twist and therefore cause binding as those Teflon blocks pinch the rod. That carriage should run on 3 points not 2.
We know exactly where the symptoms lie, has anyone pushed them? When I get home, I'm going to try and exacerbate the error we are seeing. I plan to do this by drawing a line across the platform from right to left with a degree offset of 1 degree over the length of the platform and see if the line printed forms steps instead of a straight line. If the printer is capable of performing this test, it could just be a software problem.
The STL files I attached will be my test. One test is with 1 degree, the other is half a degree.
I like the idea of this test but am confused as to how you think this is software/firmware related. We are all running open source Marlin. Most of us are using Slic3r and R-H and I am concerned that this will bring other SD4 users down a rabbit hole.
Pretty soon I am going to just buy an SD4 just so I can fix the problem for you guys
Sorry wardjr, I don't get it. Too exhausted. I will reread what you guys wrote tomorrow. I wish you luck and hope to see some good news when I come back.
Just a random thought.... Have any of you tried rotating your y-axis rods to see if they are bent?
I have. Mine is not bent.
I'm back home so I'm working on this now.
Thee post hat trick. Put software concerns to rest. I know that isn't where this is. Software is what I do. All day. I've investigated that deeply and I promise you software is not the problem and you will get nowhere looking for software problems.
Just a random thought.... Have any of you tried rotating your y-axis rods to see if they are bent?
Correct me if I'm wrong but, if a rod was bent, the problem would happen consistently in one location and not in other locations on the bed. I've tested several locations of my bed and it's always showing the same problem.
Regardless, it's worth testing so I'm going to pull my rods out tonight or tomorrow (when I get home from work) and see if the rods are bent. Such a daunting task.
Rabbit holes are sometimes required when all other ideas are not working. Yes, we are still waiting for the results of the 2 axis of motion but we can still perform tests while we wait. I'll admit, my reasoning for software/firmware being the issue is a stretch, but at this point, nothing gets ruled off the table.
Thee post hat trick. Put software concerns to rest. I know that isn't where this is. Software is what I do. All day. I've investigated that deeply and I promise you software is not the problem and you will get nowhere looking for software problems.
Noted, I'm still going to perform the angle tests as I'm stubborn and it might unveil slip in the steppers or a friction/sticking or something.
@all,
According to the solidoodle tech response, a multimeter test was possible. I did this last night and noticed that upon testing location E (X,Y,Z,E), the voltage was 1v when solidoodle claims it should be .250v. This could be due to old documetnation on solidoodle's part but worth mentioning. I don't even know what E is... LOL
If your Extruder is running at 1v your prints would look terrible, and I mean god awful.. Make sure you measured correctly. I wouldn't mess with it. They are finicky and you don't want to fight that variable as well.
wardjr wrote:Just a random thought.... Have any of you tried rotating your y-axis rods to see if they are bent?
Correct me if I'm wrong but, if a rod was bent, the problem would happen consistently in one location and not in other locations on the bed. I've tested several locations of my bed and it's always showing the same problem.
Regardless, it's worth testing so I'm going to pull my rods out tonight or tomorrow (when I get home from work) and see if the rods are bent. Such a daunting task.
Rabbit holes are sometimes required when all other ideas are not working. Yes, we are still waiting for the results of the 2 axis of motion but we can still perform tests while we wait. I'll admit, my reasoning for software/firmware being the issue is a stretch, but at this point, nothing gets ruled off the table.
You're not wrong but also not entirely correct. A bent rod could cause binding and if the bend is in the horizontal plane wouldn't really show any other symptoms. Especially if it were one of the
Y-axis rods.
So I'm working on this. I removed the bottom Teflon block and replaced it with something like in the picture below. It slides fine, but no real progress yet. Imperfect circles. HOWEVER... this is only one idea and as you can see it's held in with cardboard. Printing another and then going to find out where I need to put a screw hole. This will reduce motion in the thing I made to replace the block. So for now. Keep standing by. Wardjr's idea is to copy the contours of the carriage roughly enough to get a snug fit. If this doesn't work, I'll give that a shot.
I had a different design on there earlier that clipped to the carriage, and it made my circles MUCH worse, so I am confident that this is the source of our problems. Even if this fixes it, I am going to print the parts to upgrade to a new carriage soon. It's so cheap, that I don't see why I shouldn't just do it, because this design is clearly flawed.
But.. going to do my best to see if a free printable modification can fix it.
edit: Alright. it doesn't work. so copying the contours of the carriage it is. I may not finish this tonight, I have to sleep (work sucks right?)
I'll have to take off the top teflon block be able to put it in, but I will post results when I have them.
No simplify3d is its own slicer.
Depending on the source code license, Simplify3D could use the Slic3r source code, in whole or in part ( since apparently Slic3r is open source ) and not have to mention it, so you wouldn't even know it.
I didn't pay $200 for software without knowing something about it. It is not slic3r. Not even close. The GCode between the two is very different even for simple parts. I had experience with S3D before I even purchased my Solidoodle. We use it at my university because it is a high quality slicer well worth the cash.
Lately I've been enjoying the newest version of slic3r, but I still prefer S3D when I need supports. Its supports are far superior to slic3r's in my experience so far.
It's not slic3r. It's a good slicer though. It offers a lot of options slic3r does not. But I can't really use it because my backlash prevents me from taking full advantage of the variable speeds s3d offers . If slic3r behaved the same way as s3d my backlash would prevent me from printing at any reasonable speed. Luckily slic3r is relatively constant and I can configure it to get decent prints with my janked carriage.
I am going to try to make a part to fit the stock carriage and increase the contact points. But full disclosure--I am printing parts for a new carriage tomorrow... So that will be the ultimate route I take. I need to print and move on. It's sad that free solutions are better than the ones I paid money for, but that's why this community is so great .
SoliForum - 3D Printing Community → Solidoodle Discussion → Odd circle test results
Powered by PunBB, supported by Informer Technologies, Inc.