RE: New VFA Tuning Method rocks!
No matter how I adjust the screws, with the "new" method the lower belt is always 6hz higher than the upper belt. Example, I adjust the upper belt to 98hz then the lower belt becomes 104hz. I adjust the lower belt to 92hz and upper belt is now 86hz. Also, have to be careful not to have to much tension on one side or the gantry becomes skewed on the opposite side. With the tensioners fully slack, the gantry is square.
RE: New VFA Tuning Method rocks!
Also to add, I switched back to the "old" method and easily retuned both belts to 85hz.
RE: New VFA Tuning Method rocks!
I try today, but imposible for me to read something if the axis is in the front right and pluck two front belt. But if i move the axis to the back and i pluck the belt on each side, i can reach 98/92. But when i'm back to the frond and finally read a value, it to much :/
RE: New VFA Tuning Method rocks!
Here are all the details, maybe it helps out.
https://www.reddit.com/r/prusa3d/comments/1m995lw/psa_vfa_update_3/
RE: New VFA Tuning Method rocks!
If I tune the upper belt to 98 Hz and the lower one to 92 Hz, the entire gantry becomes warped. I also have to adjust the tensioners for them extremely differently.With equal tension—85 Hz using the old method—the gantry is perfectly aligned (even without any tension on the belts). My understanding is that with different frequencies (meaning different tension), the gantry can’t be aligned.Does anyone have a tip on what I’m doing wrong? Or is mechanical distortion intentionally accepted in order to minimize VFAs?
RE: New VFA Tuning Method rocks!
If tightening warps the gantry, you're doing it wrong.
Loosen the belts entirely. Ensure that the gantry is square.
Then, tighten each side equally. Get the two sides to the point where the belts are starting to show some tension, then tighten each side in sequence.
Initially one full turn on the left, one on the right. When the belts start to show some response when plucked, reduce the adjustment to a half turn per side. Once you get over 80 Hz or so, drop to quarter turn. Alternate side-to-side. There will be some interaction, and as you approach the right frequency you should be making very small adjustments.
I also move the extruder around to the four corners once in a while as I start to get to the right range. I don't know if that really is necessary, but it won't hurt.
If you crank down one side to get to 98 Hz then go set the other to 92 Hz it's going to bend the gantry.
If you've got 85 Hz with the "old" method, what readings do you get with the extruder in the front/right side as in the new method?
I really think someone could make some coin building hardened replacements for the gantry brackets that don't bend. Or a simple stiffener to stop the bending. This shouldn't be so prone to being pulled out of square.
RE: New VFA Tuning Method rocks!
My understanding is that with different frequencies (meaning different tension)....
The different frequencies when measured in the 'new' position actually mean that the belts have EQUAL tension. Their frequencies are different only because of their unequal lengths (due to the idlers in the left block being at different distances from the extruder) when in that new position.
As @k1mu says, you need to go carefully with the tensioning - don't let one side get too far ahead of the other, and try to start building the 6Hz difference from early on.
The plain fact is that if your gantry is square when there's zero tension, and you then carefully crank on EQUAL tension (those UNEQUAL frequency measurements), your gantry will remain square. Unless....
1) Something is broken - so that, e.g. something flexes more than it should when under the correct tension
2) Something is not assembled correctly. The most likely culprit now seems to be the pulleys on the motors - one should be one way up, the other is the other way up. This is quite a common mistake to make during the build because it seems reasonable, after doing the first one, to assume that the second one is the same. Although the gantry motion will appear to be correct, the belt path won't be correct and belt tension will cause this error to show up. Double check your assembly against the photos in the manual: https://help.prusa3d.com/guide/5-corexy-assembly_835522#843284
RE: New VFA Tuning Method rocks!
I try today, but imposible for me to read something if the axis is in the front right and pluck two front belt. But if i move the axis to the back and i pluck the belt on each side, i can reach 98/92. But when i'm back to the frond and finally read a value, it to much :/
Do you mean that your phone can't get a reading at all at the front, or that you can't get to the right tension?
If your phone only reads correctly when inside the enclosure, is there maybe some ambient noise, something with an electric motor nearby?
If you have to move to the 'old' position to make any successful readings, don't aim for 98/92 on the side belts - they should have EQUAL frequencies when measured at the sides because those lengths of the belt have equal length. I think somewhere around ~104Hz on both side belts will be the equivalent of 98/92Hz at the front.
RE: New VFA Tuning Method rocks!
@k1mu @chris hill: thank you for your reply checked:
- The gantry is perfectly square without any tension on the belts.
- The idler pulleys are correctly installed.
- The belts are exactly the same length and mounted in the extruder plate with the same offset of 5 teeth.
- To get both belts to the same frequency using the old method, I need to make exactly the same number of turns on the tensioners.
- When tuned to 85 Hz using the old method, and then moving the gantry to the front right, I measure 75 Hz on the top belt and 74 Hz on the bottom belt.
.Further tightening of the belts does not achieve the 6 Hz difference, unless I tension them so unevenly that the gantry twists and is no longer square.
RE: New VFA Tuning Method rocks!
Here are pictures to clearify
how I meassure the new belt values.
RE: New VFA Tuning Method rocks!
It's ok, i succeeded to have a reading on the front. But if i go for 98/92, it make a lot of bounce when homing. If i go for 97/93, just 2 bounces (never see less)
RE: New VFA Tuning Method rocks!
@k1mu @chris hill: thank you for your reply checked:
- The gantry is perfectly square without any tension on the belts.
- The idler pulleys are correctly installed.
- The belts are exactly the same length and mounted in the extruder plate with the same offset of 5 teeth.
- To get both belts to the same frequency using the old method, I need to make exactly the same number of turns on the tensioners.
- When tuned to 85 Hz using the old method, and then moving the gantry to the front right, I measure 75 Hz on the top belt and 74 Hz on the bottom belt.
.Further tightening of the belts does not achieve the 6 Hz difference, unless I tension them so unevenly that the gantry twists and is no longer square.
If you simply aim for 104Hz on both sides (old position), what frequencies do you see on the front (new position)?
I would add that you should maybe hesitate to continue experimenting just to satisfy my curiosity, unless you've at least lubricated the tension screw threads. If there's something pulling the gantry off square it seems possible that your tension screws might be experiencing greater tension than they ought to, increasing the risk of galling.
RE: New VFA Tuning Method rocks!
It's ok, i succeeded to have a reading on the front. But if i go for 98/92, it make a lot of bounce when homing. If i go for 97/93, just 2 bounces (never see less)
Sounds like a good place to stop!
RE: New VFA Tuning Method rocks!
This new 'strobe' tuning method can't come soon enough.
It really is a complete PITA right now using the current method.
Prusa's priority right now should be a simple, foolproof way to get the gantry perfectly square, followed by a simple foolproof way to get the tension set....using the current methods there are just too many variables that the user can't control that can throw the readings off.
RE: New VFA Tuning Method rocks!
@k1mu @chris hill: thank you for your reply checked:
- The gantry is perfectly square without any tension on the belts.
- The idler pulleys are correctly installed.
- The belts are exactly the same length and mounted in the extruder plate with the same offset of 5 teeth.
- To get both belts to the same frequency using the old method, I need to make exactly the same number of turns on the tensioners.
- When tuned to 85 Hz using the old method, and then moving the gantry to the front right, I measure 75 Hz on the top belt and 74 Hz on the bottom belt.
.Further tightening of the belts does not achieve the 6 Hz difference, unless I tension them so unevenly that the gantry twists and is no longer square.
This is my experience as well. I've double checked the same and there is no issues with the X/Y pulley direction, belt lengths etc. However there is no way to get a 6hz difference without severely skewing the gantry.
RE: New VFA Tuning Method rocks!
If I tune both belts to 104hz then move the gantry forward and head to the right, the top belt is 82hz and the bottom belt is 96hz. So something is obviously not right. As in they are reversed in some way.
RE: New VFA Tuning Method rocks!
Wait... Which side should be "plucked" to measure the frequency? Is it the left pair, the right pair, or the front pair that go across the linear rail when the gantry is forward and the head is to the right.
RE: New VFA Tuning Method rocks!
Wait... Which side should be "plucked" to measure the frequency? Is it the left pair, the right pair, or the front pair that go across the linear rail when the gantry is forward and the head is to the right.
Old method, old values
Move the print head to the rear centred position - the left and right belts should both measure 85Hz. Note that this refers to the 'inner' belts on both sides - on the left side the 'inner' belt is the right hand one of the pair, and vice versa.
New method:
Move the print head to the front right position - top belt should measure 98Hz, bottom belt should measure 92Hz
Old method, new values:
If the top and bottom belts are tuned to 98Hz and 92Hz respectively at the new location, then when the print head is moved to the old position (rear centred) the left and right belts should both measure around 104Hz (according to my measurements - there's no official guidance on this as far as I'm aware).
RE:
Wait... Which side should be "plucked" to measure the frequency? Is it the left pair, the right pair, or the front pair that go across the linear rail when the gantry is forward and the head is to the right.
Old method, old values
Move the print head to the rear centred position - the left and right belts should both measure 85Hz. Note that this refers to the 'inner' belts on both sides - on the left side the 'inner' belt is the right hand one of the pair, and vice versa.New method:
Move the print head to the front right position - top belt should measure 98Hz, bottom belt should measure 92HzOld method, new values:
If the top and bottom belts are tuned to 98Hz and 92Hz respectively at the new location, then when the print head is moved to the old position (rear centred) the left and right belts should both measure around 104Hz (according to my measurements - there's no official guidance on this as far as I'm aware).
Yes, I get that.
What I'm trying to say is, first I tune both belts to 86hz (roughly). Then slide the gantry to the front-right. Then, in my case, the top belt on the left side is 6-10hz lower than the left bottom belt. On the right side it reversed, i.e. the top belt is 6-10hz higher than the bottom belt, same as along the linear rail.
Anyway, I'm not going to mess with it anymore. I'm printing just fine and have wasted too much time on this for little to no gain.
I've had this printer for 5 weeks now and I've been endlessly tweaking to get to a point the where it's almost as good as the MK4S it descended from.
RE: New VFA Tuning Method rocks!
What I'm trying to say is, first I tune both belts to 86hz (roughly). Then slide the gantry to the front-right. Then, in my case, the top belt on the left side is 6-10hz lower than the left bottom belt. On the right side it reversed, i.e. the top belt is 6-10hz higher than the bottom belt, same as along the linear rail.
There is no "left side" and "right side". This may be the cause of the confusion. Apologies for the following, but I'm going to be as precise as I can with directions.
Move the extruder. For clarity, this means that facing the door of the Core One, move it to the right hand edge then pull it towards the front of the printer, closest to you. This is "Right/Front".
To the left of the extruder is the linear rail with two belts, one on top of the other. These run from the left edge of the printer to the right edge. The one higher up is the TOP belt, the one lower down is the BOTTOM belt. Those are the belts that you are strumming to be able to tune things.
Don't reach in and strum the belts on the left and right side, which run from the front of the printer to the back. Only the two belts running left to right. Right there in the front of the Core One.
"Left" and "Right" do not enter into this. Again, to be completely clear, this means you're plucking in a completely different place than you used for the 85 Hz, back-center extruder position tuning setup.
I'm hoping it's not really this simple, but it sounds like you're trying for 98/92 by plucking the belts you plucked for the 85 Hz process, which is absolutely incorrect.