Hello Guest it is April 26, 2024, 02:05:40 AM

Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - Jeff_Birt

881
General Mach Discussion / Re: Noob is mystified
« on: April 24, 2009, 04:40:43 PM »
Quote
"Mach 3 just homes by moving into the home switch and then slowly backs out until the switch changes and stops.
Backlash must be determined with a dial indicator. Then there is screw mapping which is another whole topic.
Your plan to install without the driver on the office machine is a good one. Enabling the axes is needed so the program has a reference to track on the screen."

Yowza.... really?

It's such a simple thing to do as part of the homing... and can be done for each axis in a matter of seconds.

The controller that comes with the Dyna machine measures and compensates for backlash every time you initialize the machine, and it displays the actual backlash values.

The Dyna controller tries to measure backlash by sensing the number of pulses between the closing an opening of the switch. That is not real accurate at all it does wind up giving you some idea if your switches are sticking though. To really measure backlash you really need a more accurate way to measure it, like a dial indicator.

Backlash compensation on a mill is somewhat of a myth. The cutting forces (bit on material) will have a tendency of pushing the table around taking up any backlash opposite the direction of the force. If you enable backlash compensation, set it right and, move one axis at a time it will look like you accomplished something. In actual practice you'll find that it did not help at all.

A more accurate homing solution is to use an encoder with an index pulse on each axis. The home switch then becomes a 'hey home is near' switch and 'Home' is seen as when both the mechanical 'home' switch and index pulse are seen.

Quote
My feeling is... I shouldn't have to change jack-squat... plenty of online stores are able to work without me lowering my security settings... this is bad.

The last thing the ArtSoft people should want is for people that are ready and willing to buy... not being able to because of the way SecurePay's order processing works.

VERY FRUSTRATING!

I've seen that other merchants sell Mach3 licenses.... can anybody suggest a reputable dealer that has a better online store?

I can understand being frustrated but I can see the other side of the coin too. As an online merchant I can check and double check things and then my shopping cart software, some browser, etc will change on little thing and something breaks for someone. It is frustrating for all involved. Try it again and if you have no luck you can pick up a license at my site: http://soigeneris.com/shop/Mach_3_97626.aspx (I guess I would have to count myself as being reputable  ;) )

882
Sorry, I thought I had included it on this thread? Anyhow here is the link: http://www.machsupport.com/forum/index.php/topic,11108.0.html , basically with the motherboard that Rusty had not even another PCI card would work although the driver test looked great. Something in the chip-set that handles the PCI bus (perhipials) did not like being driven that fast. He had tried several things to try and make it work though so I thought some of them might apply in your case as well. If you have a PCI parallel port card I would try that first to see if it helps.

883
Sometimes it is just the parallel port or South-Bridge chip-set that prevents the driver from driving the parallel port properly. The link I gave a few posts above this one had that very problem.

884
That would be a really bad idea. The voltage drop across a resistor is directly proportional to the current through it. So when your servo driver is drawing less current it will have more voltage across it. Not a good thing.

885
Check the docs for your drivers to what edge they trigger on and set Mach accordingly. If that is not set properly than any other changes you make won't produce meaningful results.

Also keep in mind that WinCNC uses dedicated external HW to generate the stepper pulses where as Mach uses the parallel port. As great as the Mach LPT driver is it can't match having a dedicated HW built for the purpose. Also, some motherboards just can't actually drive the LPT that fast. The driver test will come out great but the actual LPT hardware is lacking. Sometimes an add-on PCI parallel port card will make all the difference, sometimes not. Take a look at this thread as it may have a few other ideas that you can try. For a more realistic comparison you would need to use something like a SmoothStepper: http://soigeneris.com/shop/Warp9_97624.aspx  or other motion control card.

886
Have you done the mach driver test? That is a good way to see how smooth the output wave form will be. You can also turn on 'Sherline' mode. This really stretches the step pulse out. I had to do this one one machine to get good results. Also be sure that you have the step/directions signals set up to the correct state. Some driver switch on the rising edge of the signal (Active Hi) some on falling edge (Active Low). If you get it backwards the drive won't work properly. You can see anything from lost steps to rough running.

887
General Mach Discussion / Re: Noob is mystified
« on: April 22, 2009, 04:52:12 PM »
I'm not sure Hood, that could be. Maybe he can post the file here so we can try it. I also vaguely remember Brian saying something about the simulation mode using a completely different 'driver' (in other words not the LPT driver) perhaps this simulation 'driver' is getting confused?

By the description of the problem it sure sounds like crop circles though.

888
General Mach Discussion / Re: Noob is mystified
« on: April 22, 2009, 04:34:06 PM »
I converted a Dyna 2400 a few years ago to Mach control. It is just a slightly 'newer' machine than the 2200. I pulled out the distribution board and made my own opto-isolation board to fit the same geometry and that the stock wires would plug into. This worked well but the phase jitter from the LPT port driver combined with the stock stepper drivers left with with a max traverse speed of 10 IPM. Just last week I tired it with a SmoothStepper instead and got the rapids up to 30 IPM.

If I had it to do over again today I would use the SmoothStepper and a Gecko G540 driver. Neither were available when I did the machine. The Gecko G540 is nice as it has the optoisolation built in, among other things. It would have made the whole job much easier.

The 'crop circles' you describe are a classic symptom of the IJ mode being configured wrong (as others have stated.)

889
SmoothStepper USB / Re: Help ! From Incremental Jog to a runaway !
« on: April 22, 2009, 12:07:29 PM »
Have you tried a different keyboard? I remember hearing of some folks having lots of problems withhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh

Wow! That was strange. While I was typing the batteries in my wireless keyboard dies and the 'h' key stuck! OK, what I was trying to say is that some folks have had problems with keyboards in the past. It would be worth swapping it out with another one to see if the problem goes away.

890
SmoothStepper USB / Re: Smooth Stepper & PoKeys55
« on: April 18, 2009, 09:46:51 PM »
Long parallel cables can cause problems as well. As the length of the cable increased the signals will interact with each other more and more. This has the effect of some signals being delayed slightly more than others. At fast pulse rates it can cause problems.

IMHO, it is better to keep the PC, SS, and drives in close proximity and extend the cables to the monitor, keyboard, etc. These are far less critical. Someone just posted a neat KVM extension that works over a single Ethernet cable on the Warp9 forum.