|
TWEAKIE my CNC machine under construction. |
|
CNC is only limited by our imagination. |
|
Construction 5. |
|
(25) Cable routing again. The neater the wiring, the more reliable it will be.
(26) Connecting to the electronics. The control board is a '3 axis StepMaster' from USA and the PSU is a switched mode 24 Volt 6 Amp OMRON unit.
(27) An axial fan has been incorporated into the control box rear cover to provide forced air cooling. In practice this has proved not to be really necessary, as nothing gets hot. Stepper motor current has been set to 2 Amps per phase, and this works extremely well.
(28) All the wiring finished, tested and working OK. The fan, which is 12 volt is operated from the 24 volt psu via a 7812 series regulator.
(29) Wiring secured to frame with small 'P' clips. Tweakie is ready to 'Rock n Roll'.
(30) I will almost certainly be fitting a vacuum table at a later stage and this temporary, sacrificial, cutting table is just to try everything out. |
|
I have included my schematic diagram here and it can be seen that it is all pretty simple stuff really. It looks more complicated than it is in the above pics because each limit switch has its own dedicated cable and all cables (outside the control box) are of course screened.
I did initially fit pull up resistors for each limit switch (and axis disable switch) connection but in practise found this to be unnecessary as the Stepmaster p.c.b. already incorporates this function.
As of Jan 2008, I understand that the Stepmaster p.c.b is no longer available (following problems with their long delays in shipment and poor or non-existent back-up service).
Personally I have had no problems either with supply or usage of this product (perhaps I am the lucky one).
|
|
When setting up Mach3 or connecting the limit switches, estop etc to a controller board it is useful to know and monitor the state of the pins on the LPT1 port. A small utility which does just this is Parmon, written by Fred Bulback at GeekHideout.com. This utility can be downloaded from here Parmon. It is incredibly easy to use and as well as monitoring the individual pin states it also enables direct switching of pins by clicking on the pin you wish to toggle state. |
|
A cautionary tale of a lesson to be learnt.
I recently tried 3 of the Routout 2.5 Amp bipolar stepper driver boards and separate breakout board (on a new machine) with very unexpected results. I was getting extra (unwanted) steps and these could be heard as an occasional tick or stutter in the usual note produced by the motor when jogging an axis from one end to the other. All three axis were affected and the problem appeared to be occurring totally at random. Extra steps (as opposed to missing steps) were identified by fitting a pointer to one of the stepper shafts and driving it exactly 50 revolutions (in my case this corresponds to an axis movement of 200mm). The pointer should have stopped at exactly the same point from which it had started but it had sometimes travelled just a little bit extra (never less). The amount of over travel was extremely small but nonetheless it was an error in position which was unacceptable. Using an oscilloscope the problem was identified as noise spikes occurring on the ground side of the LPT connection between the computer and the breakout board. The only spikes which were causing trouble were in the interval between the positive going step pulses and when they were of sufficient amplitude they were being treated as an additional (unwanted) step or steps. The direction pulses (which are either on or off) were unaffected by the spikes (as far as I could ascertain).
Obviously the passive breakout board supplied by Routout would be just fine if there were no signal problems but in my case it was certainly not suitable. Curiously, this breakout board connects LPT screen to GND which it should not do as this will create a ground loop via the domestic supply - no connection should be made between the computer GND and the machine GND. Although the computer GND and machine GND are supposed to be at the same potential they rarely are. With regard to the machine GND - there should be only one GND point and all grounds should be run to this point - do not daisy chain grounds - do not produce ground loops.
Problems such as this are avoided by using an active breakout board with opto isolators between all inputs and outputs (which allow perfect signal transfers between sources of different potential) and a Schmitt trigger circuit in series with each step input to condition the pulse shape and remove any false triggering. Connecting a new breakout board, incorporating these features, completely cured the problem and the new machine started working perfectly.
It is perhaps interesting to note, that profile cutting a complete shape, as an example, has an equal number of +X steps as it has -X steps and the +Y steps also equal the -Y steps (this is true because the toolpath starts and finishes at the same place). Assuming that extra steps occur at random they will be more or less equally distributed around the profile (law of averages) and an extra step whilst travelling in the + direction automatically becomes a lost step when travelling in the - direction (and visa versa) therefore extra steps can and do cancel each other out. It is for this reason that setting a dial test indicator and moving a table or gantry there and back again can show no positional errors on it's return and it is therefore not a good test to demonstrate the accuracy or otherwise of a machine.
Try jogging an axis, any axis, from one end to the other whilst carefully listening to the sound of it's stepper motor. Does it occasionally stutter or produce a tick, interrupting it's usual flow of sound ?. If it does then it could just be a bit of grit in the ballscrew nut or in the linear bearing perhaps but could it be an extra, unwanted step ?.
|