having looked at the code, I think the encoders register the position and convert that to speed and check against a feed rate. If the feed rate is too high or low, the rate changes to get it to the right speed.
Then 24v will not speed up the shield unless the firmware is modified accordingly. Would increasing the feed rate values/limits be enough to solve this?
This is going to sound overly pessimistic, but: I’m not sure what we are solving. The OP’s question was could using 24 V be done. The answer is yes but the real question is what will we get from it? The higher voltage would give more instantaneous torque to move the sled… more force on the chain, and at some point, more speed, but speed is more about the cutting bit and the ability to maintain stable control as the sled moves, so I don’t think it would really change anything other than quicker acceleration for starting movement and a greater ability to snap bits.
Since the It is effectively hanging with gravity holding it in place, when one chain goes slack because the cutting isn’t progressing as fast as the chain is being let out, it becomes a pendulum and we get weird cuts. Ever tried to push a rope? A faster chain let-out doesn’t help the sled go faster. I had a rough surfaced piece I was cutting on Saturday and the sled started to hop while cutting. I was expecting a snap, but it didn’t. More torque from the motor would mean more force to the chain and that would have likely just broken the bit. A motor with more force only helps when pulling the chain, not pushing it or letting it out. I could be on the wrong thought train here though. I’ve never experimented to see how fast gravity will cut. Maybe I’ll run a test and see how fast I can cut a straight vertical line down without going slack on the chains. I presume this will be very much bit and material dependent. I think the gantry setups have one thing going for them. They cannot deflect. For Maslow, this is s blessing and a curse. I imaging we would all be cursing broken bits more if we didn’t have the invisible bungee that is gravity pulling on the sled. It makes me also wonder if whenever CERN runs the LHC and gravity changes a bit, we all have to recalibrate. - not serious.
I did not want to throw in that we don’t have acceleration/deceleration planing (unless Webcontrol has it), ergo ‘overshooting’, but now I did. Forgive me. Went from 1000 back to 800.
I did temper with the max feedrate that is/or was hardcoded in the firmware.
@Gero How high of a feed rate did you try and what depth did you cut at the high rate?
You answered before I asked the question… I thought acceleration planning would be done in the firmware, but either the Mega doesn’t have the horsepower to do it or the firmware just isn’t there yet. I’m thinking the second.
I don’t remember if it was very shallow or with no bit in the sheet at all. I do remember stopping my attempt due to some sort of observation. After 10 months hibernation I’m back and might try it again to give better answers.
Edit: The 2 posts I could find are related to AMP and that is not allowed in this post (joke)
I have 10A and guess I did not try after that.
That’s what i meant in this previous post
However we went a bit off topic discussing about how the motors are controlled I think it could be beneficial for users to have the ability to control a bit more the feed rate depending on their individual needs, some users are just cutting foam with it.
The motors are 24v capable and the TLE will hold up to 5A of constant current (6A peak)
I went from a 17kg sled to a 7.5 kg sled. That would also allow me more speed I guess. Keep it up!
Not with the v1 board. I made a patch to the firmware and webcontrol so I could see the PWM being used. If it went purple (PWM>240 / 255) we’re going as much power as possible. On my machine, that was $15=1100 mm/min; I wanted to go faster, so I need moar power!
I expect that it’s going to get dicey without acceleration planning, so I’m thinking of a small patch to do something about that (in hack-tastic fashion) while thinking about trying to put the “brains” of Marlin into the Maslow (>3x readahead so you can do accel planning and can go “hey, this smells like an arc”) [larger patch].
I hope you are successful! Higher speeds could be good for PVC, foam, balsa woods, coroplast, ridgid insulation, HDPE plastic, etc. I know most cut wood, I think people would be impressed how well and easily these less dense materials cut.
I think I finally understand what you were saying now… it just took a couple days. Rereading your post and then mine, I really just said the same thing you said above. Sorry to plagiarize and not give credit.
@Thormj: the webcontrol / groundcontrol feeds the file line by line to the firmware, so I think the small patch in firmware will end up being be several in number to both webcontrol and the firmware to get the two coordinated to buffer correctly. In Webcontrol, there is a buffer option, but my system doesn’t like it and does strange things when it is enabled. I’m willing to test. I have a second pi set up for this purpose with a second controller with no motors. Let me know if/when I can help out with testing.
grbl has acceleration planning running on the same CPU, but the maslow firmware
the issue is teaching marlin how to deal with two things
how to control the DC motor/encoder instead of a stepper
how to do the kinematics calculations
just like grbl
I totally empathize with you, I don’t even understand my own posts sometimes
I measured the amperage on motors. It picks 0.5A on no load, 0.9A on heavy load and 1.3A on blocked. So L298N 2A bridges are enough to run. 12V PC power supply.
Max feedrate is determined by how much your motors can forece move. If there is not enough powh there You may get Keep-up error. In GC is setting about max feedrate available. Rise it and try to go faster. PID controller will force PWM to go higher till it reach desired feedrate.
There is no problem with higher voltage. Read datasheet of H bridge, or use one of my DIY shields for external H-bridges. Typical H bridge is opticaly/specially separated and its signals are TTL compatible. So go and have some fun with rising voltage even to 50V
First try to redesign your Pulley system to reduce friction, gravity pulling etc. Then rise GC feedrate in options. Fusion360 got Gcode turn(acceleration control) optimization so if you making sharp turns there wont be much problem with this. Program Your code in Easel/Fusion to run faster.
There are some normal limitation for feedrate… try to cut somthing by hand with Your router and see that even if go faster, the result are not so healthy for bit and router. Typical result will be that your bit will come out from chuck or get stuck in wood due friction.
- Optymize puley system
- Rise feedrate in settings
- Rise voltage
- Get keeping-UP error and start form 1st step.
For Z axis there was special option for allowing it to move faster (Z axis PID is prbbly controlled on diffrent method and by default its max speed is set on ~30%). Look in options
Resistance on motor coil will rise depending on friction. Less friction = less current draw. H birdge got stall protection and will cut current depending on that setting.
1st Arduino send signal to motor, if it moves then it move encoder. Shield reads encoder then ajust the PWM to desired speed. Motor has to start move as fast as passible so get 100% power in first few moments.
Ahh. Check what regulator your shield got. If its not capable to hang more than 12V it will blow up
It look like Arduino is powered by maslow CNC shield(i don’t own any normal shield xD) so Arduino power controller can be the factor what woltage can You plug in.
In my case i build my own shield. I take power form USB port for arduino and 12V from PC power supply to power only H bridges. I can swap that 12V with any voltage i want becouse it wont go to my arduino regulator.
Thats why there should be optical separation between H Bridge and arduino shield. Diffrent voltage sources can destroy PC mobo
I have one thing to say. Until you research and understand motor drivers you can’t apply basic electronic principles to them. Drivers are not linear in function and performance drop off is normally sudden and drastic. 25% and 75% output can be unrelated proportionately to each other. I say grab a material, take a controller, run it at 12v then run the same job at 24v and see what you get, compare the output. Empirical testing to find the answer.
I no longer have any spare time for this, but as they say "proof is in the pudding. "
@Gero I knew you added more power, I didn’t remember how.
Just my 2 Cents.
IIRC he found that his power supply as dropping voltage under load, so he went
with a beefier 12v power supply and it solved his problem
PC Power supply got over current protection so it will turn off before frying your Maslow CNC shield
Go for something 100–>150W. (FUN FACT)Another problem is low temperature. They wont work if there is -10oC so pick hair dryier or use vacum cleaner fromy cyclone to heat it up
You may need somthing that can output 6Amps, so 6A*12V= 72W. So 100W PC power supply can fit. They are cheap and easy to get.