jason
Full Member
Posts: 114
|
Post by jason on Jan 13, 2015 13:27:04 GMT -5
As you can see..my conclusion has changed over the course of this. ..
My experience with backlash in industry lead me to a false conclusion or two.. But the machines I deal with would not put the leadscrew or even rack outside the guide rails. And without a significant rework this isn't going to change..
|
|
jason
Full Member
Posts: 114
|
Post by jason on Jan 13, 2015 13:38:52 GMT -5
Btw bdureau...looks like you have already remade the x carriage...and obviously still have backlash...I really would like to know if anyone gets good results out of their printer... And did say the belt drive version suffer less?
Either the rack gets moved or maybe pulley/ belt system to balance up the pull around the guide rod furthest from the rack.
|
|
|
Post by bdureau on Jan 13, 2015 14:54:29 GMT -5
I had no choice but to redo the carriage because I have changed the hotend. so the backlash is due to that? I know that it needs to be improved and I was thinking of printing a new one. Could we compensate it programatically?
|
|
|
Post by bdureau on Jan 13, 2015 15:33:18 GMT -5
I wonder if we could adjust it here in the Marlin firmware
#define DEFAULT_AXIS_STEPS_PER_UNIT {84.88264,84.88264,800.00,60.0}
|
|
jason
Full Member
Posts: 114
|
Post by jason on Jan 13, 2015 15:38:23 GMT -5
I think the backlash is in part due to where the rack is relative to the guide rods.
Imagine the rack end of the carriage moves ( due to stepper action)..the beam moves..the next linear bearing moves a bit ( not quite so much.since it has to be loose to move)..then the rest of the beam moves..nit quite so much..until we get to the furthest linear bearing....which doesn't move at all but rocks on the play in the bearing..... This continues until finally the far end does move ..but now behind the rack end of the slide... So..the carriage is now moving in sync with the stepper. .. All is well now ..until the carriage needs to move the other way.... Now in this direction the error is double what we saw before... You ask if backlash can be taken out in programme..haven't seen any params in marlin.or slicers..or host software... Manual programming might be possible to minimise such but fancy editing 30 layers worth of g code?
|
|
jason
Full Member
Posts: 114
|
Post by jason on Jan 13, 2015 16:19:53 GMT -5
|
|
|
Post by bdureau on Jan 14, 2015 15:16:40 GMT -5
Thanks Jason I will take a look at the patched firmware to see if it can help. So far I have changed the DEFAULT_AXIS_STEPS_PER_UNIT for the X axis and I can now print perfect cubes. However circles are far from being perfect and I really need to get those right because I need to print model rocket part such as noise cone.
|
|
jason
Full Member
Posts: 114
|
Post by jason on Jan 14, 2015 16:17:51 GMT -5
A possible solution offered.. ( not my idea)..is to measure the error in x..and scale the object to increase the size in x yo compensate. .. Might work well enough for simple circular shells.. ..
|
|
jason
Full Member
Posts: 114
|
Post by jason on Jul 20, 2015 7:46:18 GMT -5
Ok. News such as is My mdf sumpod used polymer linear bearing Often a good solution..but always cased in ridgid housing. NOT IN A HALVED HOUSING like the mdf carriages on the sumpod..the housing "crushes" the polymer bearing to the correct size to match the rod. Halved housings fail to support the bearing correctly ( particularly in the plain of say x-y in the x direction)
Thus my firmly ( near stiff) clamped bearings rock on the guide rails. This leads to massive backlash
Solution Bin the mdf bearing holders x and y Replacwewith recirculating ball linear beatings ( and hard chromed rods..the mild steel rods supplied would not last with ball bearings rolling over them...).
Consequence.....gain 12 mm on z travel .. 10 mm actually usable. .. Circles are now round ( to .05 mm). Small features are printed not just blobbed
.this printer actually works.. Oh btw..auto levelling. ..won.der .full. Do it if you can..well worth it.
|
|
jason
Full Member
Posts: 114
|
Post by jason on Jul 20, 2015 18:26:25 GMT -5
Ok. News such as is My mdf sumpod used polymer linear bearing Often a good solution..but always cased in ridgid housing. NOT IN A HALVED HOUSING like the mdf carriages on the sumpod..the housing "crushes" the polymer bearing to the correct size to match the rod. Halved housings fail to support the bearing correctly ( particularly in the plain of say x-y in the x direction)
Thus my firmly ( near stiff) clamped bearings rock on the guide rails. This leads to massive backlash
Solution Bin the mdf bearing holders x and y Replacwewith recirculating ball linear beatings ( and hard chromed rods..the mild steel rods supplied would not last with ball bearings rolling over them...).
Consequence.....gain 12 mm on z travel .. 10 mm actually usable. .. Circles are now round ( to .05 mm). Small features are printed not just blobbed
.this printer actually works.. Oh btw..auto levelling. ..won.der .full. Do it if you can..well worth it.
|
|