Advanced Search

The 7i96S board has an analog spindle out which should work for that purpose - it's implemented as an isolated digital potentiometer.

Different question: if I want to control my spindle's speed, I do this through software, right? My Sherline mill's speed control uses a 0-7V range to control its speed using a potentiometer.

I have something very similar and would love to pick your brain on the ins and outs of your retrofit. I just acquired a machine myself and would like to bring this into the modern world now!

my machine was "only" a three axis machine too, but I just kept adding to it...
Mesa and Wifi should be OK; my machine has two ethernet ports. You just don't want mesa packets on the same wire as other packets.

My machine is only a 3-axis machine. I'm not sure of how much more I/O I really need. I plan on using the old Dell for LinuxCNC, so I really don't need to worry about having a free Ethernet port. Can you use a Mesa card and a WiFi card at the same time? If not, I'm fine with using the old sneakernet method.

If you're using one of the ethernet-connected Mesa cards, you'll want a dedicated ethernet port on your pc to access it - don't share it with a connection to your home router (if you're not using wifi). PnCconf is otherwise sufficient to get you going, although I think you'll eventually find yourself hand-modifying your config files to do custom things

One consideration for the Mesa cards: if you think you'll need more field I/O, you might as well pay a little extra and get a card with that built-in. I.e. 7i76EU instead of 7i96S

I feel like making foolish decisions, so I think I'm going to go with the Mesa card. How hard is the setup? I know you can use PnCconf, but is that enough to get you going, or do you need to do anything extra?

I am not a regular on the linuxcnc forum so please bear with me if I am not following the protocol/etiquette.  The Remora RP2040-w5500 looks like a great approach to a cost effective linuxcnc setup.  It is hard for me to understand how stable the current software/design is.  Are people using it routinely or is it still being developed?   About a year ago I started down the Remora path with one of the 3D printer stepper boards only to find out that the on-line documentation was not up to date.  I appreciate how complex this project is and can understand if it is not ready for prime time.  TIA.
 

Logthor wrote:

GRBL or LinuxCNC

GRBL is a good and cheap driver, but only with basic capabilities.
The problem is in communication, the PC sends a full line of g-code and waits for confirmation of its execution.
It is not possible to create loops, conditions or use variables.
(In FluidNC they try to bypass this, but the effect is far from expected.)
LinuxCNC does not have these limitations that will start to irritate you sooner or later.

So if you only want to use some specific CAD/CAM program to generate g-code, then you can use GRBL.
LinuxCNC, on the other hand, offers practically unlimited possibilities.

I converted a manual KBC GRIP-1340 lathe to CNC, kept the handwheels connected via encoders. I can manually machine or NC with it, It is more like a 'teaching' lathe, or tool room cnc lathe. Gives you the ability to quickly face or turn a diameter without having to write code or install a gui/cam thing to do it for you.

When i spin the handwheels, then stop, the machine stops. Encoder over-run is something you can disable. Sure you dont get feedback like a manual crank, but for the odd time doing something it doesnt matter. I put microswitches in the feed sticks and spindle fwd/rev lever, right now im using on screen dials to set the pitches and speeds, but its very similar to a manual lathe when you want it to be, and exactly a CNC lathe when it has to be.

I ended up using a 7i76E and 360ppr encoders connected with a small timing belt to 3d printed gears secured to the hand wheels inside an apron box i made from some rectangle tubing. My X ballscrew isnt in the apron it is beside the cross slide. I still have the cross slide on the lathe incase a quick taper job is required. Ill probably end up ditching it eventually for a more solid mount as its a weak point.

Anyways, the machine has made its money back and then a small fountain of cash to fund the other toys. It might not be everyones cup of tea, but it works very well.
 

Thanks for enlightening us ignorant folks. I'll call Siemens and Heidenhain right away and tell them to remove that feature from DMT Kern, DMG, and the rest. I'm sure they'll be thrilled that someone has finally pointed out they've been doing it wrong ever since the 1980s.

I started with a parallel port for a lathe conversion, it was robust enough to do the job, but eventually went with a 7i76E for it, then a VMC i retrofit with 7i76E+7i84S, then now im doing another mill with a 7i95T, the Mesa hardware is fantastic- so good that china is now knocking it off!

Parallel port to start, then work your way through things. Lots of help here on this forum!!

The log file produced by the polygon pocketing program  produces g-code linear moves as its output.  If you stick a few m2 and percents and a few setup commands in that file and change its name to a .ngc file you can  mill simple things. 
(maybe)   (I'm assuming you have some other machine that you can run that program on)

If you have access to a PC with a parallel port, it's a cheap and easy way to get started in CNC using knockoff breakout boards from Amazon and some cheap stepper/driver kits. That's what I did. However, this is not a robust solution. I had many problems with port pins getting fried, for example. Getting any one of the Mesa cards (I used the 7i76E) is a bit more of an investment but the electronics are much more robust, and you get to enter the worlds of "industrial lego". At that point, you'd need LinuxCNC anyway, so might as well start there.

Also, it looks like GRBL is designed to run on an embedded board, not a PC, so if you want the lowest cost-of-entry, go LinuxCNC.

The 7I97T enable outputs should be able to drive that directly.

If you worry, you can always measure the current with a DVM