7' X 4' Combi

[Ricardoco]

Hard to tell from a video, but it looks to be running smoothly without whipping, which is good.



If the step/mm value before, which caused it to move 100cm when commanded 100mm was accurate, then you just need to multiply it by 10.

There's a bit more to it though. You need to make sure that you don't hit the kernel frequency limit in Mach3. It's best if you can to use the lowest frequency, which is 25kHz. This means the port will output a maximum of 25000 pulses per second, so if you divide that by the step/mm value you get the software limited maximum feedrate. For example if it's set to 2000 step/mm, then that's 25000/2000=12.5mm/s or 750mm/min. If that feed-rate is too low, then you either need to reduce the microstepping resolution (e.g quater stepping instead of 1/8th) or raise the kernel frequency.

Ok i went to the workshop and in motor tuning ive set steps per to 20- Velocity to 3000 accl to 200, Oh im using metric, now when you type in X1000 the gantry travels 1M so that much is correct but im not sure it is the only way to do it the drivers are on the lowest setting which is 200 per rev. with four people sitting on the gantry it dosnt evn know you are there..

Rick

[Jonathan]

ive set steps per to 20 Velocity to 3000 accl to 200, Oh im using metric, now when you type in X1000 the gantry travels 1M so that much is correct but im not sure it is the only way to do it the drivers are on the lowest setting which is 200 per rev.

200mm/s^2 for the acelleration is very low. You should be able to get around 1 to 2 m/s^2.
The motion will be a lot 'smoother' if you use microstepping. I reckon you'll get around 8m/min, so that's 133mm/s. 200 step/rev is 20 step/mm, so 25000/(133*20)=9.4. So you can have anything up to 9.4 times the number of step/rev ... so try 1/8th microstepping.

with four people sitting on the gantry it dosnt evn know you are there..

The overall mass of the gantry doesn't make much difference to the feedrate you can get, it mainly affects the acceleration obtainable (Newton's 2nd law). Increase the acceleration and it will 'notice' plus the people will probably fall off which is an added bonus.

[Ricardoco]

200mm/s^2 for the acelleration is very low. You should be able to get around 1 to 2 m/s^2.
The motion will be a lot 'smoother' if you use microstepping. I reckon you'll get around 8m/min, so that's 133mm/s. 200 step/rev is 20 step/mm, so 25000/(133*20)=9.4. So you can have anything up to 9.4 times the number of step/rev ... so try 1/8th microstepping..

I set the accelleration slow while i was setting things up as i was affraid of missing steps in one of the twin steppers and it all getting out of shape but as you suggest i will speed that up, Ok so how do i set it up to use microstepping then, (I thought i was LOL)

The overall mass of the gantry doesn't make much difference to the feedrate you can get, it mainly affects the acceleration obtainable (Newton's 2nd law). Increase the acceleration and it will 'notice' plus the people will probably fall off which is an added bonus.

LOL that would be funny, if not for the fact i was one of them...

Was it good to see your ballnuts working then? and where are these extra points you said i could have.

Rick

[Ricardoco]

...... Ok so how do i set it up to use microstepping then, (I thought i was LOL)
Rick

Ok i think i have got itLOL

To get 1/8 microstepping.

I have 200 full steps per full revolution of motor as it is a 1.8deg motor.
360degrees (1 full revolution of motor)/200steps=1.8degrees per step
8 Microsteps per full step (1/8 microstep)
200x8=1600 steps per rev (1600 set on the driver)
10mm travel on gantry for 1 revolution of the stepper,
1600 steps/10mm=160 Microsteps per mm set in "steps per" in motor tuning
1mm/160microsteps=.00625mm gantry travel per microstep given 1:1 gear ratio motor to screw.

YES??

Rick

[Jonathan]

8 Microsteps per full step (1/8 microstep)
200x8=1600 steps per rev (1600 set on the driver)
10mm travel on gantry for 1 revolution of the stepper,
1600 steps/10mm=160 Microsteps per mm set in "steps per" in motor tuning
1mm/160microsteps=.00625mm gantry travel per microstep given 1:1 gear ratio motor to screw.

Perfect :)

where are these extra points you said i could have.

[Ricardoco]

Perfect :)

Ok Put all the settings in and im Over the moon the motors sound great it manages 8M p-min as Jonathan said it would (not tried higher) so Ive backed it of to 7 and thats sweet. The steppers sing with those settings. I will do a Video maybe tomorrow. I wrote a quick bit of G-Code (my first bit of G-Code writing.. (Thanks for the tuition i2i)) just to watch the gantry going back and forth and stopping and starting with small travel and big travel just to see if everthing is how it should be, then copied and pasted it several times, then just watched it run for 45 mins, I then tested the wireless pendant and that works great as well... Im a happy bunny now .

Now the Y axis (oh and lots more playing with the X as well..

Rick

[GEOFFREY]

Congratulations. G.

[Lee Roberts]

Nice one Rick, looking forward to continued success for you and following religiously.

.Me

[Jonathan]

I wrote a quick bit of G-Code (my first bit of G-Code writing.. (Thanks for the tuition i2i)) just to watch the gantry going back and forth and stopping and starting with small travel and big travel just to see if everthing is how it should be, then copied and pasted it several times, then just watched it run for 45 min

Try this spreadsheet to generate similar G-Code. Especially if you just test one axis, the machine can seem fine from just jogging around at full speed, or repeated code like you tried, but in reality fail a test with random code. The reason is if you just repeat the same commands and run it for a long time, that's not much better than just running those lines a few times since it's doing the same thing every time, so the currents will be virtually the same every time. If they're randomized then over time you can expect to find if there's a problem.

Edit: also what acceleration have you set it to?

[Ricardoco]

Try this spreadsheet to generate similar G-Code. Especially if you just test one axis, the machine can seem fine from just jogging around at full speed, or repeated code like you tried, but in reality fail a test with random code. The reason is if you just repeat the same commands and run it for a long time, that's not much better than just running those lines a few times since it's doing the same thing every time, so the currents will be virtually the same every time. If they're randomized then over time you can expect to find if there's a problem.

Edit: also what acceleration have you set it to?

Cheers Jonathan, I will give that a go, i think i set it to 1000mm sec i will have a check tomorrow.

Rick