|
|
Page 4 of 5
1 2 3 4 5
First test!
22:03 1/12/2004 |
| There comes a point in any build where you finally have to do "The Test", so you check and recheck all your cables, make sure
you didn't cross the polarities or wire live to ground; sacrifice a small goat or other deity-approved offering. And you still
sweat it! The worse is when everything powers up and nothing happens.. not a failure - just the system waiting, all the time you're
terrifed that in a second that ethereal magical smoke which makes modern technology work will escape.
Since I was quite worried something would go insane, I built my E-stop circuit first, complete with a massive big red emergency stop button.
My first test was to power it all on for a few seconds - then power it down again. This confirmed that it wasn't an immediate-mode
pyrotechnic, but it didn't ease the next task - connecting my PC to this as-of-yet appliance of dubious electrical certification! |
Okay, so now it's time we talk about what's inside the chassis of the mill's controller. The "box" itself is from an OLD Dell 486
which I was given. In retrospect I should have just bought a NEW PC Case as they look a lot better, are very cheap nowadays, and
are a lot more accessible (I did this later on for my Sherline as I prepared it for sale and it looks really good).
So the components of the controller are:
Parallel port breakout board: basically a simple board with a parallel port plug on one side, and screw terminals on the other.
Man, that seriously beats soldering a dozen tiny wires. The normal board is flat, but I made a clever little bracket for it so it
could sit in the PC chassis like an ISA card would.
|
| Solid State Relays: Regular relays require a lot more power than a parallel port can drive and send back-EMF when you stop
supplying them power, SSR's have none of these disadvantages however, so you can power them straight from the parallel port. I use
two SSR's, one for controlling the spindle of the machine, and the other as the system enable/disable. The enabler basically controls
power to the entire controller. Unless the driving program drives the line high the mill will remain powered down. Further, the
enable circuit is chained to the E-stop: if a limit is hit, or the E-stop button is pressed, the system shuts down - the PC has no
say in this whatsoever: but it does have a sensing input which tells it whether the system is alive or dead (Estop sense in this case).
So there are two things which can shut down the mill: the software (by dropping the enable line) or the mill itself (by tripping an E-stop).
The spindle power is a lot more simple: it basically controls the AC power to the "original manual mill's" spindle motor. Since I didn't
want AC power inside my logic box, what the SSR actually does is simply put out 12VDC - which I then feed into a little box with a
DC relay that switches the AC power and has a plug and socket on it.
|
Gecko Servo Drives:These guys are little beauties. There are 3 of them here, one
for each axis, X,Y,Z. They act a lot like a stepper motor circuit in that they take a step and direction input (one signal tells them which
direction to move, and pulses move them a discrete amount in that distance) but instead of controlling steppers which are prone to
stall they run DC servo motors with encoders fitted. At rest they hardly use any power at all, idly bouncing the motor between
pulse edges, yet in an instant they can deliver a scary number of amps to the motors! Servo's cost a little bit more, but for a mill
they're highly worth it. It goes beyond the scope of what I can communicate here, so be rest assured: Servos are good, very good.
Big capacitor:A very big capacitor. It stores so much energy, that if you turn off the main PSU you can still run the mill for a minute before it dies...
PSU relay, discharge resistor:.. which is why you need this circuit. While being able to run the mill for a while when the
power is turned off is certainly efficient sounding, it's absolutely terrifying if the reason the power was turned off was
because you were screaming and hitting the E-stop button! The relay switches between two states: one where the PSU is connected
to the system and charging the cap. The other where the PSU is disconnected, and we're dumping the cap as fast as we can through
the massive resistor. Heat fins included.
|
|
Finally, the beginning..
22:03 21/12/2004 |
| The first test came, and passed. Then the second, more wires were layed, and then all the axis worked. But the mill controller was a horrible beast of
a monster with wires everywhere - poking through the frame in all places. It wasn't pretty - but in some bizarre way, it worked.
The original Dell casing used to contain an LED display which would show the clock speed of the processor, and a reset button, both
of which I hacked. I changed the original PCB to instead provide a reset button for the mill and 3 blue LED's - the fault sensors for each
axis. The Turbo and Hard-drive LED's were modified to instead indicate the state of "controller enabled" and "spindle on".
So there I was - a basically functioning CNC mill, no projects, and an ugly chassis to boot. What a fallacy! There was only one thing to it:
Doctor, Heal thyself! |
First project for mill with tangly cables? Create a backplate for the mill, so we don't have any more tangly cables. So I drew
up the CAD model, removed the old PC power-supply, milled some nickel sheet, installed some cable glands in the nicely milled holes
and voila. Done. This was pretty good - but the first hint of uncertainty had crept in... the holes were good, but not perfectly
round.. backlash! We'd done a good job.. but this would not do.. we'd need ballscrews to replace the poor linearity, and slack when we changed direction,
of the leadscrews. Well, not a bad start, at least...
|
|
Page 4 of 5
1 2 3 4 5
|
|
|