Messages

21

General Mach Discussion / Re: Bridgeport Knee Mill Conversion?

« on: October 08, 2016, 05:56:42 AM »

You generally need a lot of damping to run closed loop with a VFD, as they are usually very slow to respond in comparison to a servo drive. Try dropping P to near zero, and I&D to zero as a starting point. Personally, as you're running direct drive, I'd be inclined to leave it running open loop, as the VFD itself will maintain near constant speed.

Regarding your G28 issue, have you got an example of the code?
If the post processor is inserting a co-ordinate as part of the G28, it could well be causing a move outwith machine limits. Have a read of this article - http://www.cnccookbook.com/CCCNCGCodeG28ReturntoReference.htm

22

General Mach Discussion / Re: Why won't they send me my license?

« on: October 06, 2016, 04:36:42 PM »

IIRC Mach 3 licenses are generated manually, so they take a day or two to be processed.

23

Promote and discuss your product / Re: Step-Direction Servo PID Closer - Pixi revival

« on: September 23, 2016, 03:39:06 PM »

Old machines with original motors generally don't have high resolution encoders, so it's not really an issue.

And just how many high-power AC servo drives exist that don't support step/dir?
Any I've looked at recently all have multiple control input options.

24

General Mach Discussion / Re: Bridgeport Knee Mill Conversion?

« on: September 23, 2016, 03:29:37 PM »

I'm sure Hood will give a more definitive answer, but I can't see any real reason why an index pulse would be essential on a mill that doesn't need any sort of spindle indexing.

If you've not looked at it already, US digital (www.usdigital.com) is probably the best place to see lots of encoder options.
Maybe also worth checking British Encoder (www.encoder.co.uk), although they're mostly complete units i.e. you can't get sensors and discs separate like you can at USdigital.

25

Promote and discuss your product / Re: Step-Direction Servo PID Closer - Pixi revival

« on: September 23, 2016, 03:18:08 PM »

In all honesty, I doubt there would be very much demand for such a device.

Nearly all modern drives will accept step/dir as an option.
On machines with more basic analogue drives/amplifiers, by the time you factor in the cost of two/three of these, will it be cheaper than just buying a controller that supports +/-10V?

To put it in context, a CS-Labs IP-A unit is €600, which is only €40 more expensive than the IP-S (step/dir) version.
A Dynomotion Kanalog add-on for the KFlop is $250.
But both those options are also giving you high quality cnc controller with 24V compatible IO.
Even in the LinuxCNC world, MESA provide +/-10V cards for not that much more than step/dir cards.

Thinking you can convert an industrial machine cheaper to a high standard is a false economy. You still need to factor in the cost/hassle of converting/interfacing any logic to low voltage. Admittedly on a basic machine there won't be much in the way of logic, but there is still a cost of converting it.

26

General Mach Discussion / Re: Bridgeport Knee Mill Conversion?

« on: September 23, 2016, 03:02:04 PM »

Hard part is mounting the encoder - will have to go on motor shaft as the draw-bar means i can't use the spindle shaft - guess that will mean i am limited to 1:1 ratio??

Got any photos of what space is available?
There is nearly always some way to mount an encoder.
Reflective optical encoder just needs a stripped wheel. Large bore encoder discs are available, which you could possibly mount on the toothed spindle pulley. Third pulley running on the belt between the main pulleys, or on the motor shaft, but both those options would need to allow for a gearing ratio. If you do need a spindle index pulse (Would it actually be used?) you could then use a slot sensor via a notch on the spindle pulley flange.

27

General Mach Discussion / Re: External EStop Requested - These random estop's are killing me!

« on: September 18, 2016, 12:29:04 PM »

Erm, Dave, did you miss the bit about VFD and 2.2KW, so I'm going to hazard a guess there are no brushes involved...

Certainly does sound like a noise problem though, and running a job with the VFD disabled would help narrow down the cause.
Does your e-stop wiring use shielded wiring?
Also, is the E-stop circuit switching the ESS input to Gnd or 5V?

28

General Mach Discussion / Re: Bridgeport Knee Mill Conversion?

« on: September 03, 2016, 12:00:48 PM »

Has Mach4 got CV yet or S-Curve acceleration, I hear thats pretty useful??

From the various discussions I can remember on the yahoo group, the trajectory planner is the main thing that has had minimal changes from Mach 3. However Brian has said due to the way Mach 4 has been designed/written, the TP is a module, so there is the opportunity for it be re-written later without any major alterations to Mach 4, unlike Mach 3 where the TP was a core part of the program.

29

CS-Lab / Re: Encoder error accumulation

« on: August 21, 2016, 04:15:39 PM »

How long is the cable before the connector/line driver convertor?

I'm going to guess by connecting the shielding, you've introduced a ground loop, which is causing interference on the encoder wires before they reach the line driver. First thing I'd do is move the line driver to as near the encoder as possible, as then you minimise he length of wire where the single ended signals can be affected, and get the maximum benefit from the differential pairs/line driver, which is to greatly reduce interference problems.
Alternatively, you could upgrade the encoders to line driver versions, and remove the need for any converter.

30

General Mach Discussion / Re: Machine setup help - urgent if possible :)

« on: July 25, 2016, 05:06:12 PM »

The problem with your current code, is there is nothing to stop motion. It simply displays whatever message, then continues on it's way.
Within the If Else statements, you'll need to call a feedhold prior to displaying the message, so that motion is physically stopped. Once you get that working, you could then expand it to stop the spindle if needed, and then restart in whatever direction you need.

What I'd suggest before you try editing/adding to the code, is create a flow chart of the process, so that you know exactly what the process is, before you start coding anything. That way you're simply converting from flow chart to code, and not trying to figure out the process at the same time as trying to code it.