TL-Smoother for Stepper?
-
I can see why you would avoid these if you had 256 microstepping available as it works by smoothing the ripple out, and in doing so on a 256 step you'll introduce an additional lag. So rotor lags the fields, which lag their drive signal, which lag the signal fed into the smoother which will fractionally lag the pulse train into the driver.
On another thread dc42 talked about they used larger caps to control the electromagnetic interference. Any idea if these have a similar purpose or are they merely to prevent power dips on the drivers as the MOSFETs trigger?
-
TL-smoothers are useful with some stepper drivers, in particular the DRV8825 for the reasons described at http://cabristor.blogspot.com. I have seen some evidence that under some conditions the same issue can happen to a lesser extent using A4988 or A4982 drivers. I've never heard of it happening using Trinamic drivers, which use spreadCycle mode (or sometimes stealthchop mode) instead of constant off-time mode.
-
Cheers for the link I'll have a read.
Looking on you tube it's hit and miss whether it would help:
For: https://www.youtube.com/watch?v=ipnwq_3GD3o
Against: https://www.youtube.com/watch?v=Vt-7xZUsPJ0
Edit: I do appreciate that this issue is largely resolved with the more recent Duet boards. Working on tackling the low fruit and educating myself before I feel both I and my machines are ready for Duet Ethernet/Maestro/ or maybe 3!
-
Reading between the lines is this suggesting 24V maybe worth delaying until a swap to the newer boards? Could the swap end up reducing print quality?
-
I ran my delta on 24V using a Duet 085, without this type of problem. I think it may depend on how the R_OSC pin of the A4982 or A4988 driver is wired. But the problem is usually seen only with DRV8825 drivers anyway.
-
@doctrucker said in TL-Smoother for Stepper?:
Any pictures of the salmon skin effect?
Not sure why it was referred to as salmon skin, it looks more like salmon meat. Note the diagonal lines on the cabin and the ridges on the bow.
-
Here's a close up of the recent test following a right battle with my custom cantilever. Pushing my luck here with a 0.4mm nozzle using 0.1mm layers on the perimeter and 0.24ish on the infill. There are two diferent sets of the lines on the bow and where they intersect on the bow the surface finish goes to pot.
-
@doctrucker The overhang angle on that part of the bow is pretty extreme. Are you sure it's not just a result of a little curl up?
-
Works fine at larger layer thicknesses, but then again there is lokely to be a higher back pressure with the lower thickness that may be dumping out of the side of the nizzle as it goes around the corner.
All that said, it does have a very regular diagonal ripple pattern running through it.
-
I have noticed with that print, that too high of bed temps can creep up, this makes the material softer and more prone to curling up then squishing out.
Judging by the base, it looks like it's a little too hot, or your layer height is a little too close to the bed.
The ripple you talk of could be over-extrusion, filament variation or Z axis rounding errors or slight binding. What stepper/lead screw do you use?
Also how many shells did you use?
-
Three vertical shells, three top and bottom. Base layer accepted as not too accurate. Bed temp 50. Extrusion accurate for 100mm at lowest rate DWC would dish out and volumetric extrusion set by vernier caliper 90 degree opposed for diameter. Extrusion multiplier about 0.93 and temp at 180.
