Advanced Search

Novice here with zero knowledge and experience of industrial control, please forgive the lengthy post.  I have read and searched numerous threads here as well as YouTube videos to ad nauseam and been at it since 0430 this morning with only frustration to show for it.  Here is my setup of a DIY PrintNC machine:

• "Axis" 3-axis configuration, X, Y1, Y2, Z.
• Mesa 7i76E.
• 120VAC 2.2kW water-cooled spindle (Z-mounted, but not connected to VFD).
• Huanyang VFD (powers up).
• PNP proximity sensors at both ends of X, Y1, and Y2.  Z only has software limits defined.
  • Prox sensors wired to 3-wire PLC input module, 6-way sensor terminal block with the signals wired to Mesa digital-in.  HAL "sees" the states of the prox sensors which appropriately change when triggered.  Home and min/max limits have not been defined - yet.  That is for another day.  At the moment only software limits defined for all axis.
• Main multi-branch power circuit enters emergency "oh sh*t" paddle switch mounted to the enclosure, much like a table saw or router table.  Pushing the paddle terminates power to the entire enclosure, everything turns off - motion and control.
• From paddle switch, multi-branch main power circuit enters enclosure to door mounted two-position lock-out type main power rotary switch. One branch is power for typical 2-button 120VAC 24VDC coil contactor holding circuit, attached to it is all motion (VFD and D542 drivers).  Power to motion side can be killed entirely as expected with the holding circuit being turned "off".  The other phase provides power to the control side (Mesa, exhaust fan, and other peripheries).  The control side power is turned off when the main power rotary switch is turned off.

What additional hardware is needed to implement a hardware e-stop and a reset function?  I do not wish to route main 120V through e-stop buttons, rather like to keep it inside the enclosure. I only have e-stop NO/NC mushroom buttons of various styles and hole diameters.  Nothing else.  I've read the "External Power On/Off + EStop" and "Emergency STOP on mesa card 7i76e" threads several times.  Please interject and correct my understanding of how to proceed:

• Connect a Mesa TB1 field power output (pin 0-3) to the NO in terminal of the e-stop button.  Using more than one e-stop button requires wiring the NO circuitry in parallel so that a possible malfunctioning switch along the chain does not cause an unwanted e-stop signal - the switch must be intentionally depressed by human action.
• Connect the e-stop NO out terminal to Mesa digital-in pin (TB5 or TB6, pin number to determined for coding in HAL).  This is the e-stop pressed "event notification" signal.  In my case it is 24VDC.
• Connect Mesa digital-out (TB5 or TB6, pin number to be determined for coding in HAL) to +VDC terminal of solid-state relay.
  Q: This and the GND connection will cause the relay to energize its coil and open (if NC relay) thereby removing the +5V signal from Mesa to the stepper drivers ENA+, thereby "disabling" them so they no longer output Step/Dir pulses to the stepper motors?  Is that its intent? 
  Q: Does not removing power to ENA+ cause the stepper motors to stop moving?  Power to them can be left on?
  Q: Do I need a SPDT or DPDT and what kind NO only, NO/NC, NC only?  I ordered two of these based on Rod's recommendation in one of the aforementioned threads, is this correct?  Ultrathin DIN Rail Relay Module, MRD-060D2, Input 4-32VDC NO
• Connect Mesa GND to -VDC terminal of solid-state relay.I think I am a-OK to this point.  My conundrum is the "load" side of the relay.  Stepper drivers ENA+ can accept +5V~24V and they and accompanying ENA- are NOT connected on my machine.
  Q: Does that signal pulse have to come from the +5V terminal of one of the unused Mesa Step/Dir output connectors on TB2 or TB3?  Does this Step/Dir output need to be added to the system configuration if I did not select 5 steppers in configuration wizard?  It isn't being used as step/dir controller, how does the +5V terminal actually go high?
  Q: Could the signal instead come from one of the Mesa digital-out terminals of TB5 or TB6 with a buck/step-down converter reducing the 24VDC to 10VDC max and programmed in HAL appropriately?  Only need an "on" signal, right?  I suppose the buck converter is another point of possible hardware failure to contend with.
• Connect Mesa GND to D542 ENA- terminal.
  Q: Does it actually matter the source of the Mesa ground?  Could it be a "field ground" (terminal block connected to Mesa TB1 or to TB3 supply power ground) or does it have to also come from the same unused/spare Step/Dir output connector supplying the +5V?
• Daisy-chain connect each of the D542 stepper driver's +ENA and -ENA.

Now the part I need the most help with.  What to do with the NC branch of the E-Stop button in order to to stop I guess the power to the VFD?  Should power to it be removed entirely or should the output speed, direction, etc. to the spindle be turned "off" - and how?  Another relay and routing through Mesa?

Thank you.
Darryl

Can someone share working config.ini and friends for Linuxcnc (on PC)3axis+ 10V spindle + 3 *2 limits?
I am getting stuck somewhere between working 3axis steppers Nema23, Estop and missing spindle control in GUI but it seem to be enabling signal spindle, latency value update
This way I could work out my values and update them.

pippin88 wrote:

Could be Z axis stalling - stepper stalling, or coupler slipping.
 

x 2
Either your velocities and acceleration are too high and it misses steps when retracting against gravity or the Z axis coupler is slipping.
Either way your z axis is not fully retracting  due to lost steps or mechanical slippage.

You need to post the gcode files (to make sure it's not in the gcode)

Could be Z axis stalling - stepper stalling, or coupler slipping.

Hello Everyone,

It has been a year or 3 since I needed to change my machine, I've reading on the forum and so it seems linuxCNC has again evolved a lot since then. Congrats and thanks to all developers!!!

I was finally able to buy an ATC spindle and thus I have to rewire and extend my machine. First steps are of course the CAD designs of spindle and carousel, and open source wouldn't be open source if people didn't share their work, so please follow along this thread to find all parametrical cad designs I will draw to build your own carousel. I will upload them as I go along in the design. (Solidworks and .step format)

Of course I will be having some questions of my own, I have looked into the probe basic GUI and that looks awesome. For the moment I thus only have 1 question, how much pockets can the GUI of probe basic handle. I saw in most screenshots a 12 tool carrousel but I would like to build a little bigger. Up to 36 tools. Can probe basic screen show this graphically or is there a certain max of toolholders for the graphics screen? I looked to the documentation and forum but couldn't find an answer to this. Hopefully someone knows this?.

While reading through the forum past days, I think I could conclude that a stepper/Geneva drive would be best suited to keep the implementation as simple as possible. Am I correct in my assumption? As this will determine the type of design I will.make and share for others

After the designs i will step by step update how I build the entire ATC, so anybody else in need can follow this as a tutorial as I didn't find any complete tutorial to build and ATC with cad designs on the forum. I will try to do my best and try to keep a standard as high as the pro's around here.

Happy milling everyone!!!!!!

 

A bit of an update...

The XYZ steppers are working well (using the original drivers).

There are homing switches for each axis - also now working.

The spindle is working (including stop, variable speed and direction changes).

I've connected up the lamp relay and made a custom pyvcp control for it.

This has been enough to do a couple of little milling/engraving tests under linuxcnc control. (The original manual buttons are currently disabled).

The next thing to tackle is probably the spindle encoder. The machine currently has a single slotted switch (TT OPB866T55) but has a wheel with both holes and an index slot. I guess it was intended to function only as a tachometer in this version but the manufacturer offered upgrade options.

The existing slotted switch plugs into the original main control board (which is going to be removed). So this means that the encoder will lose whatever circuit it currently has to work. I'm not sure if this switch can function directly with the Mesa card or what circuit it would need.

I'm considering replacing that single slotted switch with 3 of the TT OPB900 series as I think these can give TTL level outputs that might have a better chance of working with the Mesa card. Two would be for the quadrature signals, the third for index. I might need to make a little bracket to mount them (although it looks like the existing bracket could already mount 2 slotted switches).

Obviously it would be nice if the mill could tap holes but the spindle control seems fairly basic. It has a relay to switch the direction polarity, an enable/brake relay and a 0-10v analogue signal for speed. It probably won't ever be able to change direction quickly - it has to be stopped before it can engage the other direction.

What do you guys recommend?

I'm leaning towards slotted switches over a stand alone encoder as that will allow me to use the existing drilled wheel. Are there particular models recommended here? Also, are there any ready made pcb's that I can use to wire the slotted switches to and get a more standard encoder output?

Hi all,

This is some feedback on my first cutting job using LinuxCNC – QTDragon GUI.

Firstly, I would like to thank those who made this GUI – it’s great, and I’m really warming to this interface.

I don’t fully know if the issues referred to in these notes are because I haven’t set it up correctly or, they are short comings or, bugs…

My machine was a CNCEST 6040, with a USB interface and Mach3. My initial step was to design and build a parallel port interface, which I did, and engraved it with Mach3. I installed it into my CNCEST Controller, replacing the USB interface.

I then installed LinuxCNC via the linuxcnc-2.8.4-buster.iso (as a dual boot). I then ran stepconf:-

Setting up the stepper motors was pretty easy, mainly because of the Test button. Also, having the computation results from the max speed, acceleration, pulse frequency etc was very helpful.

When it came to do the spindle – without a Test button, I went through that loop about 10+ times…

Looking at several GUIs, I thought that QTDragon would suit me best. I couldn’t find a configuration template for it, so I just manually added it into my INI file. Without adequate documentation, it was a trial-and-error process.

One thing that I really found annoying was, stepconf overwriting my changes. There should be more options about that in stepconf or, a custom INI file with additional settings…

My first cutting job – engrave a spare parallel interface board.

The NC uses probing to generate Z offsets, with 90 (10 x 9 grid) probe points. When running, I’d randomly get this Error - “Probe tripped during non-probe move.”, and the machine couldn’t complete the 90 probes. QTDragon didn’t help – the highlighted Gcode line is wrong, it would rewind back to the start on this error.

Sample GCode
...
G1 X41.16252 Y2.095 F1000
G38.2 Z-1 F5
G0 Z2
…

I changed the GUI to AXIS which showed the error was occurring with G0 Z2

I’m using the engraving tool as the probe (not spinning), and PCB is sitting on 1.0mm double-sided foam tape.

In this case, it should NOT produce this error. Also, it should NOT rewind the program – it should go to the Pause state allowing the Resume command…

Drilling the PCB – The software that generates the drill code embeds manual tool change coding.

I don’t have an automated tool changer, I need to do this manually. The code raises the spindle to a height for me to change the tool, but now I need to zero Z. QTDragon won’t allow me to move the Z axis while in Auto mode. When I change it to Manual mode, it rewinds the program!!!

The finished board is really good, and I’m very happy with it. Much better than the Mach3 board.

I hope the above info can be used to improve this product. And perhaps improve me too!

Kind regards
Pedro
 

PCW
Okay, I have wired the spindle up that way and have put my VFD back into the 0-10v range. Will test this when I get replacement cable and box installed along with testing everything else. Thank you. 

MesaFlash: Okay, I tried the command that you suggested, as well as a bunch of permeations of said command as well as using sudo, and a I kept getting "No 7I96S board found". Also tried with just 7I96.  The 7I96S was plugged into my comp via ethernet. Maybe I need to update mesaflash? maybe there is some other issue? However, based on the output you provided, on the 7i85S table 2 - pins 11,12 are step5 pins, and table 2 - 13,14 are dir5 pins. This jibes with where I thought they would be based on the manual.  However, it raises the question, does this mean that the - and + sign on step and dir pins can be ignored when comes to wiring stepper drivers?   

Aciera
Appreciate the spoiler. Well that worked in the sense that I was able to launch the GUI and see 3 rotary axis.  However, the TCP is all messed up when the machine is in "bridgemill" mode.  In Vismach, the A axis controls what should be the B axis, the B axis controls what should be the C axis. And the C axis isn't moving anything on Vismach.  When it is in TCP mode Vismach shows all the axis moving when I jog one of the A or B rotary axis, so it is still trying to do the kinematics but they are completely wrong and its draws weird archs instead of remaining at the same point.  Good news is that in the preview and "identity" mode, the A and C axis are spinning round the same axis, which is good as there are supposed to be two axis aligned with each other, but they should be the B and A axis.  So it seems like if we could redirect the axis to the appropriate joints on the machine, everything should just work I would think. Do you happen to know how to do this? 

Thank you. 

Aciera:
I tried altering the ini file as you suggested.  There were a few places where I added to the xyzabcw config that weren't where you suggested.  As well as moved a few other things around in the axis configuration file, mostly with setting up the joints and axis.  But I ended up getting an error when I tried to launch that configuration. I have attached a picture of the error below as well as the full ini file.  Note: I did not add the first line of code you specified as I did not know where in the initiation file to place it, perhaps this is the issue. 

I do not consider myself a shy person, but I also have been working on this machine for so long and really just want to see it running right now if thats possible, even if it is a little hacked together.  I have plans to build a more substantial machine in the future and would like to get deeper into writing kinematics then. 


Wiring issues:
I have most of the board wired up.  I have 5 of the 6 axis wired into the 7I96S.  Have all my sensor wired into the inputs of the 7I96S (and written down so I know what goes where). I ordered the wrong cable to attach the 7I96S to the 7I85S so will get a replacement on Monday.  

I do not know where to wire in the last axis into the 7I85S.  The manual on this is very sparse. Its states "Note that actual signal functions depend on FPGA configuration" but gives no data on the different config or how to change the config. I need to know how to change the output of these pins to Step/dir and which pins are the ones I need to use.  Also, I assume something needs to be done to the config to make it talk nicely to the 7I96S.  Note: the only thing that I am plugging into the 7I85S is one stepper driver. 

Lastly, I need to know how to wire my Huanyang VFD into the 7I96S. First off, I have set my VFD to receive between 1-5 volts input for spindle speed control (factory setting is supposed to be 0-10v but since 7I96S is a 5v board I figured this needed to be modified) by adjusting parameter PD070 according to my manual.  From my last attempt at getting everything running I had the FOR (forward run pin), VI (analog frequency reference input), and ACM (GND) pins of the VFD wired into 5v, spindle output, and GND of my old board respectively. However, the 7I96S manual is being characteristically vague and only provides spindle , spindle out, spindle - as descriptors for the pins.  I can take a guess on these and see how they work, but does anyone know how this should be wired? Also, there is a jumper that needs to be adjusted to change speed control from the onboard potentiometer of the VFD to the inputs I referenced, so I am aware of that. 


Thank you. 

Pendants:

The Vista pendant is USB - it plugs in to your PC's USB port and there are configuration files from Vista which let the pendant communicate with LCNC.  They are solid, and the features on them are really nice.

My only gripe with them, and USB devices in general, is that they aren't 'real time'.  If you jog slowly (turn the handwheel) it's fine.  If you move it faster it's fine...until you reverse directions.  At that point the USB lag or latency becomes painfully obvious as the machine doesn't stop moving immediately.  If you move the handwheel back and forth fast you can quickly get the MPG out of synch with the axis in question.

Some folks have reported intermittent connection issues, possibly related to EMI interference bugging up the USB connection & data.

I've noticed the same latency issue on xbox controllers connected as CNC pendants.

Other USB pendants are, I presume, similar.  Easy to connect, have lots of features, and likley have the same latency issues as the Vista.

Wireless USB pendants are probably in the same situation, only with the additional issue of wireless interference potentially causing problems.

The above may be a non-issue for some users, but my use-case and risk-aversion led me to switch to a hardware MPG.  My Vista has been sitting in a box for a year or so waiting for me to sell it off.

The 'hardware' pendant/MPG I mentioned is simply a 100ppr encoder hooked up directly to a Mesa board's inputs.  In this case, there are no latency issues - it's a direct input.  They require two inputs (one each for A/B signals).

In addition, some method of selecting an axis as well as a jog increment is needed.  The axis selection & increment could be done through an on-screen control, but the typical scheme is to use two multi-position selector switches also connected to a Mesa board.

Oddly enough, the non-USB MPG's on Scamazon are not bad.  I also discovered that the only difference between a "4-axis", "5-axis" and "6-axis" MPG is on the sticker.  They all use the same 12-position selector switches and they all use the same coiled cable.  The extra wires are left un-terminated and the selector switches are set to 4 positions for the 4-axis (cheaper) versions.

I've looked at two different brands, and they're both as I've described above.  Crack it open, solder some wires, set the switches, and print a new legend sticker.  There's probably some that don't follow the scheme I described - but I doubt it.  YMMV.

IO:
There may or may not be 'plenty' of IO on a 7i96s + 7i85s.  There are 11 general-purpose inputs on the 7i96s.  5 or 6 of them are taken up by axis home switches (A-axis is optional but nice).  Should probably have an estop input (from a latching button or safety relay).  Might want to have one connected to the VFD in case it faults.  Maybe another one for a water flow-meter so you don't burn your spindle up... but your VFD might have an input you can use for cooling so no need to use up a Mesa input.  Looks like you've got open-loop steppers, so no drive alarms to worry about.

That leaves you three to five inputs to use. You could:

• Connect a hard-wired MPG encoder to two inputs, and use on-screen controls to do axis selection and jog increments
  • 5 inputs for axis selection (for 6 axes)
  • 2 inputs for jog increment (for 3 increments - 0.1, 0.01, 0.001)
  • You don't have enough inputs to do the above plus everything I mentioned...thus the 'on-screen' selection
• Buy a Mesa daughter-board that has more inputs (connects to 7i96s or 7i85s)
  • A 7i73 has four encoder inputs (8 pins), as well as the ability to use a matrix keypad for up to 64 programmable input buttons for user controls.  Downside is you can't really use it for 12/24v field IO (like homing switches)
  • Other smart-serial boards, like the 7i84, have lots of IO.  Some have 48 inputs and 48 outputs - that's quite a lot.
• Use a USB pendant
  • no IO required and probably has additional controls like a cycle start/stop button, feed-hold, feed/speed overrides.
• OR....skip the pendant for now and just use on-screen jogging controls until you've got your machine moving properly.  Then figure out what you want to do about a pendant.  All of the options I mentioned can be done later with no loss of money or time or brain-investment.

VFD:
If you're going with the 7i96s, there's no reason to bother with modbus right now.  Later on you can still use it to get information from the VFD and display/use it in LCNC - while continuing to use the 7i96s output to control the VFD.

tommy

I added a photo of the histogram.  Honestly not sure how to interpret this. Let me know what you think. 

So, are the step/dir signals coming to the mesa board through the ethernet cable? Because I don't have a parallel port of my computer, just a vega connector.  


spumco

I added a screenshot of the cad of the 4th/6th axis (kinda works as a 4th axis but will technically be the 6th axis) since it is not assembled yet.  Again, only 5 axis will be moving at any one time and there will be two different configurations.  And a picture of the full machine which you can tell is a head-head I built on your standard gantry style machine. 

Good call on removing power to the driver, I do plan to removing the 4th axis from the table when not in use.  Interesting idea about using a 6 pin connector, I'll def think about that. 

Okay I'll plan on purchasing a 7i96s and 7i85s. Hopefully on Monday.  Lets assume I envy your handwheel and want the vista one and not a amazon one, I assume there would be plenty of I/O for that with the 7i96s and 7i85s?

So it looks like I need a parallel port to run the VFD? As I mentioned above I only have a vega port on my computer (plus usb, ethernet, aux..).  Is there a way to run the VFD off of the Mesa? It looks like there is a way to do it off the usb, I know you just said you don't like usbs but...  I definitely will not be doing rigid tapping with this machine. It would be very cool to get temp and particularly load (I think I am running this machine significantly under its potential power) off the spindle, but honestly, it will just be nice not to have to crawl behind my machine and manually set the rpm like I was doing before this build.  The name of this machine is the "Hackbot" and I consider it a means to an end and also a learning experience.  I plan to make a significantly stiffer, more powerful, more accurate machine in a few years and I want that one to be perfect.  But, if this one is a little janky its okay.  Just to set expectations accordingly. Obviously I still want to be able to machine percise parts..  Let me know what you think. 


I was also poking around on LCNC and I think I can use two sample configurations as the basis for my machine.  The "5axiskins - 5 Axis bridge mill" should work for when I am using the machine in head-head configuration.  Though I will have to flip the y and z axis so that I don't have to change the kinematics (this was a good suggestion by rodw).  This will probably mean that I can never get the A axis to 0 degrees (where zero is in the preconfigured file anyway) though as that would crash the spindle into my A stepper motor.  Will this be a problem?  By the way, does anyone know how to change that machine configuration to TCP control? I would think it is either the 'bridgemill' or 'identity' buttons but is unclear.  I plan to try posting some code and testing that configuration on the laptop soon. 

For the head-table configuration. I think I can use the max5kins or the max5triv to run the machine, it will just be that the home (0) position of what is currently my b axis will have to be 90 degrees off from where 0 is on those configurations,  this moves what I am calling my A table axis to a C axis. Does that make sense? 

Thanks for the clarification on your number of joints/axes.  A pic or two of your motion hardware would be helpful (beyond just the head)

Regarding your number of joints:  You might get confused, and your CAD/CAM might get confused... but LCNC wont.

Mesa board selection: The example I gave will get you everything you need, is currently available (Mesa is having the same supply chain issues as the rest of the world), and is easily expandable.  You may get suggestions from other folks here and there's nothing wrong with a different setup.

If you go with the 7i96s and 7i85s combo you won't need to plug/unplug your table-mounted rotary axis.  You'll have 9 step/dir outputs and can leave it plugged in permanently.  However, if you plan to unplug it at a bulkhead connector (maybe to take it off the table) you'll probably want to have a separate relay or contactor to remove bus power from that motor's drive.  Stepper drives don't like to be energized with no motor connected.

It wouldn't be too difficult to use a 6-pin panel connector for that motor.  Four would be the standard stepper A+/A-, B+/B-... and the last two would be used to signal LCNC that the motor is unplugged.  If LCNC see the motor is unplugged, it won't energize that drive - and maybe signal you through the GUI that LCNC knows the axis is unavailable.

VFD: you have two options to (easily) control that VFD.  0-10v (plus FWD/REV signals) and Modbus.

• Some people on the forum aren't too keen on using modbus for spindle control, and others love it.
• Modbus can have some latency which isn't good for tight spindle control (rigid tapping).
• However, it saves a bit on wiring, and you can use modbus data from the VFD back to LCNC for spindle speed, spindle load, temp, etc.
• As usual, read the manual - there's a modbus driver specifically for that brand VFD.
• If you don't know what modbus is, search the forum and user manual.  There are also some good YT vids on LCNC and modbus for VFD's.

Your PC: aside from the spike, your latency numbers look ok for use with a Mesa board setup.  Read the forum latency threads, and see if there are some PC/bios settings you can edit to get rid of that spike.  Worst case you pick up another USFF PC and use that instead of a laptop.

On my mill, I've got an Odroid H2+, and it gave me fits for a while due to network instability and some self-inflicted installation troubles.  Sorted out now.

On my current (in-progress) project I'm running an ebay HP EliteDesk 800 G2 mini I got for very little money and it's rock solid.  Add a 15 or 17" touchscreen and good to go.  Not as tiny as the Odroid, but it's got a fan, was cheap, and zero hardware-related headaches.

Handwheel: some handwheels are USB-based, and some are 'hardware' based - those connect directly to IO ports on a Mesa board (or Pico, or parallel port).  Certain Mesa boards have some inputs specifically designed to use handwheel (MPG) encoders.  Hardware versions are more fiddly to wire up than USB but not too bad.

Keyboards aside, I am not a fan of USB-based anything on CNC controls.  I have a VistaCNC MPG and while it works exactly as designed, I found that the latency inherent to USB is unacceptable.  I got used to 'industrial' CNC controls - they are instant when you push a button or move a handwheel.

Mesa 7i84 does only inputs and outputs, nothing else, and although it uses an RJ45 connector, it is not an Ethernet card, it is a Smart Serial card. Can not be used alone.
-
Mesa 7i92 can do 10 stepper drives and still have some inputs and outputs left, but it is harder to wire and very sensitive to miss-wiring, total of 34 pins.
Mesa 7i96S does 5 steppers and has headers for easy wiring and a spindle analog output. It also has an IDC26 header that can do up to 5 more steppers or other stuff, again wiring does require a bit of care.
-
Also, 2 parallel ports can work, but will limit the speed of steppers, depending on microstepping might not limit it much or limit it way to much. Cheap and useful for testing and getting stuff going, you can upgrade later.
Friendly warning, everyone everywhere will advise on using BOB (break out boards), i resorted to never using them, ever, cheap and cause more trouble, but offer a bit of protection (some none, some quite good).
And if you do go the parallel port route, upgrading to Mesa 7i92 is very easy as there are ready made cables for it, even if you use BOB's.

tommy

So what is the 7i84 doing exactly? It is essentially just a breakout board, it can take the inputs from the ethernet and output to the steppers? 

The only thing that I am running off the machine is 6 stepper drivers and output for spindle speed control. the stepper drivers are just normal old stepper drivers.  3 of them are from steppermotor online and 3 of them are some other company that I don't know the name of. I currently have 4 inputs going into each driver (DIR+, DIR-, PUL+, PUL-) I believe that the PUL pin is also called the step pin on some drivers. So 24 output pins minimum I guess. Though I know that these can sometimes run off of 1 step and 1 dir pin too. I may want to add an output to control a solenoid at some point so 25.  None of my motors or spindle have encoder feedback.

Then I have homing switches on each axis. 2 of them are your old school limit switches. I believe those are just wired to 5 volts and then have their input going to the board. So 2 inputs there. I have a probe that is just a simple switch. So 3 now. Then I have 4 hall effect endstop sensors.  Those have 3 cables, but I believe only one output pin that goes high when is in proximity to metal.  So 7 inputs total.  Lastly, I wouldn't mind adding a handwheel at some point, but that isn't pressing or necessary.  Idk, how many inputs they require. 

Looks like the 7i96S is just 5 axis, and I will have 6 in total. So I am looking for another mesa board but there are like a million of them.  Is it better to find a board that has step & dir in the name? 


Spumco

Just to be clear the main reason I am worrying about post processing is that that is the main reason I can't seem to use my current Mach4 board.  I cannot find a post processor that will output gcode to handle all the kinematics of my machine (even though it can do the simulation of the machine in fusion 360 just fine).  Perhaps I could do more research on post processors and find or modify something that could get that Mach4 stuff running? But, having a controller that cannot handle TPC is obviously not ideal since I cannot really probe in my tool heights effectively. And my spindle simply has a an er collet, so repeat precise location of each tool really presents a challenge.  I'm sure I could figure something out but.. Anyway, I just want to make sure there is a clear path to success here, since not being thorough last time cost me a bunch of money and time. 

But, can you elaborate on making sure my laptop is stable? That is concerning.  I have a Debian Bookworm running right now on my lenovo thinkpad T440 and was able to get LCNC 2.9.0 loaded and opened.  But do I need to do some stress testing to make sure that it is stable? Can you recommend how to do this? 

Do you happen to have a link to the LCNC manual? I am not seeing anything that is clearly a manual when I do a search for it. I will definitely read it if you do. 

I don't know how to modify the GUI right now, do you have any information on where I should start on this?  If axis is the default, yes it is pretty damn rough looking, but I care much much more that it works than if looks good.  I don't know how to get vismach opened up in the GUI though.  I would like to look at a machine simulation and see what 5 axis machine setups are available. 

I added all the I/Os up above where I was responding to tommy.  Would love to hear if you have any suggestions on a board that would accommodate everything that I have listed.

Your last sentence, by control panel are you talking about the GUI? 


One other general question.  Does LinuxCNC have backlash adjustment? I figure it does but wanted to make sure.  I made the gear reductions for my rotary axis, they are not back drivable, but do have a little backlash. 

tuxcnc wrote:

scotta wrote:

I'll ask you now to read the docs!
 
You don't understand me.
Did you test the encoder?
If you did tests, you must have the config.
Why don't you share the config instead "read the docs" advice?
For example the docs says "loadrt PRUencoder names=encoder.0", but I have no PRUencoder.so file, I don't know where from can I get the PRUencoder.so file, so this is useless for me.
Second example is PWM. There is wrote in the header of spindlePwm.cpp file "Module for NVMEM spindle RPM on pin PA_0", but I don't know where is the PA_0 pin in my EC500, so I don't know where I should connect the spindle driver...
I suspect you know, but you don't share this knowledge...

I've tested the functionality of the encoder module to ensure that the encoder count is available in LinuxCNC. I can't test your application of it.

If you would like to see the application of the encoder feedback for a closed loop stepper application, you can look at the freely shared and available example configuration.

github.com/scottalford75/Remora/tree/mai...s/remora-closed-loop

I spend time to research and understand problems and scratch my head to solve them. And yes I share my learnings and the results. Maybe look back over this exact thread where the journey to get to where we are now has been shared every step of the way.

When you read the Digital Dream documentation for your controller board, you will find that there is only one pin for the 0-10V output for spindle control. It's labelled VSO.