Messages

721

General Mach Discussion / Mouse interfacing

« on: September 16, 2012, 08:09:23 AM »

It seems to me that an optical mouse would make a good basis for an "MPG".  One could mount the mouse electronics in a little housing, looking at the surface of a disc rotated by the jog wheel.  The easiest approach would be to cannibalise an off the shelf mouse rather than try to use the bare sensing chip.  However most of the mice you get these days have a USB input and this is handled on the chip itself so you would have to interface using USB.  Has anyone interfaced a mouse to Mach, does anyone know please?

John.

722

General Mach Discussion / Re: Myford Super 7 - adding ball screws

« on: September 09, 2012, 09:48:48 AM »

Thanks to all.  Hood, I've had a couple of tries at buying a "Connect" CNC lathe on ebay that is ML10 based but the seller doesn't seem interested in realistic pricing.  This was a conversion that was sold as a CNC training lathe competing with Denford I think.  How are you protecting the leadscrew - is that a telescopic cover?

723

General Mach Discussion / Myford Super 7 - adding ball screws

« on: September 06, 2012, 10:40:56 AM »

Hi folks, after a lot of thought I have decided that using plain screws on my lathe conversion is never going to give good results, and as far as I can see after a little experimentation I'm not going to be able to apply backlash compensation because the play in at least the Z leadscrew is too large.  So, has anyone got any experience of replacing the myford screws with ballscrews please?  Or could point me to info on the web?

I think I can see how to do it on the crossslide without too much trouble, but getting a long Z screw aligned with the nut on the saddle, and mounting the nut, has me stumped for now.  Another problem will be protecting the screws from chips.

Thanks in advance!

John.

724

General Mach Discussion / Re: PWM spindle - Circuit advice

« on: August 24, 2012, 07:00:22 AM »

I think your problem may be the TL082.  Looking at the data sheet its common-mode input range doesn't include the negative rail, in fact it doesn't go to less than a volt or 2 of the negative rail.  CA3140 is a much better bet as it is designed for both inputs and output to swing to the negative rail.  I suspect basically when one input goes too low the op-amp just gives up!

725

General Mach Discussion / Re: PWM spindle - Circuit advice

« on: August 22, 2012, 07:01:37 AM »

That should do it!

726

General Mach Discussion / Re: PWM spindle - Circuit advice

« on: August 22, 2012, 05:34:12 AM »

When you say base frequency, is this for PWM or Mach kernel?  I assume the former, it isn't clear from the manual that you can set it that high - they recommend 5 - 10 Hz.  But of course setting it to 1500 Hz would solve the problem with the smoothing circuit time constant, though reducing the number of speeds steps available.

As you pointed out in a previous post, an emitter follower is not an ideal buffer since it gives a -0.6 volt offset between input and output, which means that when the speed setting in Mach gives a smoothed voltage of 0.6v (6% of maximum, 180 rpm?) the VFD will see zero volts and give zero rpm; and it will also cut 180 rpm off the maximum speed.  An op-amp voltage follower would be better - but then the problem is that opamps often can't swing rail to rail anyway.  The CA3140 can swing to the negative rail if supplied from a single rail but only to about 2v below the positive rail.  So you would need both the opamp connected as a voltage follower and an auxiliary power supply giving say +12v.

Overall, if the VFD input impedance is very high, you may be better off not having any sort of buffer at all but just connecting the smoothing cap straight to the VFD control input.

727

General Mach Discussion / Re: PWM spindle - Circuit advice

« on: August 21, 2012, 08:44:14 AM »

Doh!  A bit of investigation shows that the PWM control pulses have a default base frequency of 5 Hz - for some bizarre reason this is set in the "spindle" tab of "ports and pins".  So the 0.1 second smoothing time constant in your circuit  is actually shorter than the pulse period, rather than 10x bigger!  This will likely account for the problem.

An easy way to fix this is to increase R5 and/or C1 by quite a large factor.  You can also increase the PWM base frequency - this reduces the number of speeds available but even if you increased it to say 20 Hz the number of discrete speeds will be of the order of 1250 at a kernel speed of 25 kHz.

So revised suggestion would be...

Increase PWM base frequency to 20 Hz in above config menu.

Increase R5 to 47 microfarads.

Trim speed adjust to compensate for the 10% loss caused by the potential divider effect of R5 and the Lynx input impedance.

728

General Mach Discussion / Re: What would you like to see in a new Mach3 User's Guide?

« on: August 21, 2012, 08:23:20 AM »

Why not make it a Twiki?  Then users can contribute as they solve problems...

729

General Mach Discussion / Re: PWM spindle - Circuit advice

« on: August 21, 2012, 08:15:52 AM »

I'm not sure what the nominal output frequency of the PWM signal from M3 is, but the time constant of C1/R5 should be at least 10 and prefereably 100x the period to give good smoothing.  The values shown have a TC of 10 microfarad x 10k ohm = 0.1 seconds, which should be OK for a PWM prf of 500Hz or more.  Also the effective resistance driving the capacitor is 10k in parallel with the 100k input impedance of the Lynx and the effective output voltage is dropped by 10% as a result, also reducing the max speed by 10%.  Finally the output resistance of the 4011 driver will also reduce the effective output voltage as it is in series with the 10k.  So you could reduce the 10k and increase the capacitor by the same factor.  

If there is adjustment in the lynx for maximum speed (it sounds like there is) then I would just do the following: increase C1 to 22 microF; reduce R5 to 4.7k; trim the max speed adjustment in the Lynx to get 3000 rpm on maximum spindle speed setting through Mach3.  You should not need to worry about separate power supplies or buffers.  

Note that some thyristor/triac DC motor drives (certainly my KB Electronics one and possibly the Lynx one) do NOT have control electronics isolated from mains, which means that you could have severe (or even fatal) problems if you try connecting a DC supply directly which has one side (usually the negative) earthed.  This would normally be the case for your breakout board or controller.

730

General Mach Discussion / Re: Has anyone actually got a C11 to control a KB speed control via Mach?

« on: July 06, 2012, 08:58:11 AM »

Normally, two wires from the KB supply +12 v and 0 v to the ends of the pot track; and the slider picks up a variable voltage between 0 and 12 v depending on its position.  What the KB needs is a voltage at something between 0 and +12 v relative to its own 0 v supply, which is what the C11 supplies.  The problem is that the 0v in the KB may be waggling around relative to the ground of the C11 which is tied to "mains" ground, by up to +/- 250 volts, because of the way the KB controller works (I think).

So the question I would ask is whether your separate 12 volt supply shares an earth (or common) connection with the C11?  Assuming that it's connected to the appropriate pins on the C11, it may be that its -ve ouput is connected to mains earth which will also be connected to the common connection of the 5v supply.  I think that to work correctly the 12 volt supply may need to be completely isolated from the other supplies.  In my reading of the manual it is very vague on all of this.