Advanced Search

Bonjour,
Il est bien possible de configurer une broche asservie en vitesse avec une carte Mesa 7i96s en utilisant le codeur de broche comme retour de vitesse?
Merci

> downside to the ethercat is its ONLY ethercat. i wish it has pulse as well. would be more versatile.

That's right. Now I remember that this is what pushed me to choose the M version. It seemed to me at the time that the pulse/dir was the simplest method, so it was less risky for me (this machine was my first build), and I paid some extra to get another option for the future. Turns out that CanOpen does not seem to be a viable option with LinuxCNC, but well...

If EtherCAT works well, wouldn't it make the pulse option look rather outdated ?

(Still need to do a lot of reading before I can consider an upgrade to EtherCat)

Good evening,

G96 constant surface speed on my lathe works fine with CW spindle movement. Now I want to use an upside down parting tool with CCW spindle rotation.

But the x-axis does not move. It works with M3 instead M4. Gmoccapy shows correct -350 rpm (minus!) during M4. I assume, that feed does not start because it waits for positive spindle speed.

Do I realy have to use a separate HAL file with negated spindle feedback?

Best regards,

Hartmut

That manual self-centering 4 jaw chuck came yesterday, and that brand seems pretty well made and works really well.

i have some material ordered and coming for a attempt at a rotary union, the seals should be here today, i hope this material is not under sized or i will have issues. for whatever reason thick-walled tube in this size range is expensive and i only had one supplier that i have used that was fair on price, but sizes were limited. i checked on bar stock prices and that was even more expensive yet it's amazing how you get screwed so bad on prices for small orders.

these unions are hard to make on the cheap. i did see one import that had 120 mm bore but was around $400 each. what i am trying is 140 mm bore. not big enough for a 4" square tube so it is what it is i guess. if i can make these i can make 2 of them for around half the price of one of the smaller imports.

i hope i can make it work would be a neat build and if nothing else i am eager to try it at least. nothing ventured nothing gained

You say LinuxCNC is way way better than Mach 3.

How so?

As far as I could research, Mach3 is not able to do fully closed loop operation with feedback between motor axis and rotary or linear encoders.

• feed rate based on spindle power (cutter breakage detection)
• EDM with reverse movements or slow down based on electrical values
• a hobbing machine that slows down workpiece spindle and tool spindle in case something reaches the limits.
• Turning excenters or augers on a lathe
• cam shaft grinding
  
  These are examples I can think of. Things that Mach3 could not provide, according to my understanding of the available information.
  Of course those items go bejond the average DIY user and will not be easy to do. Not easy to do in any system, unless available as turnkey in exchange for $$$
  
  Linux CNC seems to be in the gap between, "very basic good enough" and "just pay money to one of the usual suspects"
  
  I see a lot of misconception that Linux CNC ist "Parallelport technology" for 3-axis open loop stepper machines. Which is actually only the lower end of the spectrum.

cool.

downside to the ethercat is its ONLY ethercat. i wish it has pulse as well. would be more versatile.

The spindle driver is some weird (undocumented) BLDC unit with a motor using hall sensor feedback.
This benchtop mill has the tool spinde parallel to the motor and a short 1:1 toothed belt transmission. Amplifier works I can spin it up connecting 5V to signal wires.
How to connect wires, power supply and HAL components is another story.
Probably similar to a frequency inverter.

I am using the old transformer, which includes a AC to DC rectifier board.

Hello,

Got the drive working again today, the parameters for Control Mode and Pulse Type were indeed wrong (so you were both right

Turns out that my backup parameters file was not good at all, I ended up just copying the working parameters from one of the X drives to the new Y drive. Both axis have the same mechanical setup, gear reduction ratio etc...

@ihavenofish : you are right, these drives have CAN bus capability. I made this choice at the time as I was not quite clear about the various options. Now, I wish I had chosen the E version that offers EtherCat...

I'm thinking about upgrading the whole setup to EtherCat, possibly running LinuxCNC from a RaspBerry. Pi. That would allow me to get rid of the PC/Parallel Port limitations, move all the hardware inside the control cabinet, and have more I/O to finally add some extras to the machine.

Thank you all for your help, great to see this community thriving !

Is there a GitHub tracking thread (or something else) for those patches?

I would be really interested in following that thread.

fully_defined wrote:

 
"Help" is a strong word for what's happening here.

You seem to be getting frustrated.  Maybe I can help.

Correct me if I'm wrong, but your situation is that you have a Mesa 7i96s and Clearpath SDSK motors.  You want to:

• Set some rational values for max velocity and accel
• Connect the HLFB pins to the 7i96s
• Make LCNC do something with the HLFB signals

Have I got it right?

langdons wrote:

You say LinuxCNC is way way better than Mach 3.

How so?

In short:
Mach3 = extremely limited
LinuxCNC = extremely versatile, like the T1000 from The Terminator, it changes shape to fit almost everything as far as machines go.

Perhaps I'm still misunderstanding, but the actuator you showed looks like it has a leadscrew with a pitch more like 5mm, not 50mm.

Just to make sure we're talking about the same thing, 'pitch' to me means the spacing between the screw threads. 50mm is an abnormally steep pitch unless it's a multi-start screw.

In other words, does the carriage really move 50mm for a single turn of the screw?

Yes, if the channel in enabled in the hal file, it should be recognized.

You can see the 7I74 pinout a bit better with:

mesaflash --device 7i97 --addr [card_ip_addr] --dbname3 7i74

Try out the transformers you have with the motors

Make sure the transformers don't overheat.

If they start to get warm, turn the machine off ASAP.

Remember that the wire inside the transformer heats up before the iron will.

The wire will burn before the iron laminations do.

Why not just isolate the logic with optoisolators?