I’m just starting to look at a project to convert a manually operated “Surface Grinder” into a hydraulic one with a numerical control system, and would value the experience and ideas of others as I’m not 100% how to accomplish some of the details.

Basically the machine is a small Pallas/Trident horizontal mill, on which the bearings and spindle were improved by a previous owner to allow it to run at speeds sufficient for grinding, along with fitting a high speed motor. The current feed is via a rack for the X axis and Acme Screw for Y&Z axes, which is suboptimal for my needs, but suited the man who re-manufactured it very well.

My basic plan is to start by replacing the t-slotted table designed for milling, with a purpose built carriage consisting of a cast iron dovetailed slideway to which a welded metal drip tray and an assembly to allow quick-change between mounting of both an 18″*6″ magnetic chuck and a dedicated machine vice. This would not be driven by the rack, but by a hydraulic ram attached to the saddle, rather than using limit switches to reverse the valves during operation, I was planning to have a linear scale output to a PLC (likely a 32-IO Mitsubishi unit I have sitting about) allowing control of where the machine begins and ends it’s stroke to minimise wasted time during the cycle.

This leaves me with enough IO’s left over to also automatically actuate the other two axes (y&z), using the same linear scale self measurement system with either more hydraulics, or via servos acting on a leadscrew.

I have a couple of stumbling blocks, firstly actuating the knee by leadscrew gives a issue as to both replacement of the crude acme with a much higher tolerance ball screw in order to realise the kind of accuracy a surface grinder should deliver, and then the issue of developing sufficient torque to move it… But equally, using a ram as an actuator, opens up scope for it to move slightly during operation if the system isn’t perfectly sealed in order to resist any downward force, that’s after accurately quantifying the behaviour of the cylinder and valves to allow very precise movement initially

The second is how to set and display the parameters of the machine, a shared numeric keypad, parameters selection switch and LCD for each is logical, but would take up quite a lot of IO’s; the alternative approach is to have a separate logic controller to deal with the input of settings, which then simply outputs the setting for each to the PLC.

The Y-axis would need to remain on a leadscrew due to space constraints, and is likely to be sufficiently precise retaining the existing acme, torque is also less of an issue, compared to raising the knee.

In an ideal world, I’m aiming to finish up building a system where I can chuck a part; set the final dimension to grind to, Y-axis & z-axis step distance, x-axis limit postions and feed-speed; touch it off using a jog-function; press run and leave it to cycle.

You’re entirely correct, an existing hydraulic grinder would be a lot easier to start from and I had an opportunity to get a Norton grinder, which simply needed the hydraulic cylinder reconditioned, and some new piping…

The issue that stopped me is twofold, this ision loan in a communal workshop by agreement with the other users, so to take it back to a mill and bring in a purpose built for see would need quite a lot of non-engineering effort to build consensus (and I don’t quite have it in me to do all that right now). Secondly we’re tight on both space and floor load (with a Bridgeport Mill, Myford Super 7, Warco Engine Lathe, jig borer, Another small lather, Denford Novaturn, CNC mill, donkey saw, Vertical and horizontal bandsaws, plus three welders and a Layout table, all on a concrete slab support ed by RSJ’s over a basement we’re really wary of additional floor loading), so the aim is to make the Pallas conversion both smaller and lighter than a conventional grinder would be,accepting that it involves possibly sacrificing some accuracy when using the full 18″ of the chuck.

LabVIEW looks and sounds ideal, but looking at their I/O hardware, made my wallet make some funny hissing noises and snap shut…

I think need to fully document the intended process flow and then start building up specifics until I really understand where the big challenges are, and probably end up buying LabVIEW because the savings in time are enough to justify the cost of getting an interface board.

Unless of course I’ve misunderstood and I can use the software to directly access the PLC’s I/O via serial.