Absolutely, first things first. I am curious how consistent the timing is of the current task Maslow is using. If not consistent, it might make the encoder feedback loop suboptimal. But I’m going to skip that rabbit hole unless there is a good reason (other than my curiousity) to jump in there.
1 Like
It should be pretty consistent but only because we’re polling the encoders at about 1/10th the speed that the main loop is running at. I think we could read from the encoders faster if we improved how it worked
Yes please
1 Like
It looks like it is dead to me as well.
I’ve just looked at the v3 - it is well organised, they’ve dropped gulp and all of its problems, but they are running webpack, which is still maintained, but definitely showing its age.
2 Likes
Maslow is my first foray into FluidNC. I run pretty much everything else with Klipper, which does have a cable winch kinematics.
It would seem valuable to have servos available in a general purpose cnc geared firmware I would think…
1 Like
speaking of speeds, is there a reason the I2C is only running at 200k? According to the multiplexer’s datasheet, it can support 400k, and the encoders data sheet indicates 1000k.
It looks like all the I2C pullups are very strong. Have you sampled a waveform to see their tesponse?
I’ve been running at 400k in testing this aftetnoon with no yet observed ill effects, and maybe a slightly crispier feel.
Also noticed there’s a 1.5k resistor between tmc uart tx and rx, but I’m pretty sure the trinamic datasheet specifies 1k might be helpful for bom consolidation
1 Like
I think I bumped that down when we thought we might be having EMF issues (which turned out to be dust in the ethernet connection issues). I’m fully on board with bumping that back up.
1 Like
Will Puckett wrote:
It would seem valuable to have servos available in a general purpose cnc geared firmware I would think…
It is, but servos are far more complex to run and tune than steppers, so many
firmwares resist adding the first one.
David Lang
2 Likes
Curious about the 1.5k resistor too? I don’t know enough to have an opinion, am genuinely curious after seeing 1k in some board designs and the DataSheet. Tried asking AI for ideas/reasons why non 1k resistor would be used… ChatGPT - UART Resistor Value Guide
Am asking after stumbling across git/Boards/Schematic_Five-Motor-Control-Board.svg while I’m learning how to drive couple of steppers for a pet project I’m goofing around with. Am learning by digging around schematics of misc projects I like. fwiw, am skimming through TMC2209 datasheet (which only shows 1k between TXD and RXD)…
Cheers!
Nudge… Am still curious about this one, but understand if Bar and others are super busy and unable to respond, or, may not recall even why 1.5k instead of 1k was chosen.
Semi related, noticed V2.0 of GitHub - creapunk/CLN-ClosedLoopNemaDriver: Compact closed-loop stepper motor driver designed for NEMA motors moved away from TMC2209 to DRV8844.
Also, TMC2226 seems better than TMC2209 for dissipating heat (learned this from Ryan), however IC package/pin layout is different, they seem to be functionally same.
1 Like
Sorry for the slow response! I wanted to dig into it, and then I got swamped. I don’t recall an exact reason, it might be that 1k is better there!
2 Likes