Messages

161

General Mach Discussion / Re: Mach3 and second parallel port card

« on: December 29, 2013, 02:30:08 PM »

Thank you, Dickeybird.   

162

General Mach Discussion / Re: Mach3 and second parallel port card

« on: December 29, 2013, 02:08:29 AM »

jude:

I am sure the address ends up in the XML, but I never bothered to look, or found it necessary as everything works OK.
However, going into the XML, and playing around is not advisable, and is not necessary.
I don't know a lot of details regarding many "undocumented" things in Mach3, but there are a few here in the forum who know loads of odd little things. Mach3 is a really complex, huge program that has evolved over the years, as you probaly know.

Did the Windows installer recognize the second port, and install the drivers, as I described my port install.

One option may be to un-install the PCI parallel port, and download the drivers from the board manufacturer. Then re-start the computer, and point the Windows installer to the location of the driver, and do a new PCI install per Windows.

Then you can look in the Resources tab, and get the correct address.

Your program apparently just can't find the PCI board. Sort of like driving around in a new city, looking for a place without knowing the address. If the address's are both correct, Mach3 will find the boards. Likewise, if the address's are incorrect in the Resources tab, for some unknown reason, Mach3 will never find the board.

Also you said the first PP was not recognized, and then after trying the second port, and then going back to PP #1, it worked, and then on a restart it did not work again. I don't know the details about computer address's, but perhaps if the #2 PP is incorrect, it affects getting the correct address of PP#1 in the Mach3 program.


Another option is to get the same board I have, as it does work perfectly with XP, the company has all the drivers on their site for every version of Windows PC made, and has the best chip for the Parallel port. If you go to Newegg, or the Syba website you can read about that chip. I have read some PP boards use a different chip. However, I would think CNC4PC would know all of this anyhow.

John

163

General Mach Discussion / Re: worm gear cad design

« on: December 28, 2013, 03:47:39 PM »

Hi Tony:
Basic may be old, but for generating certain kinds of data, like your increasing pitch problem it works fine.  I made cam programs in basic that you could input the starting degrees, ending degrees, starting radius, ending radius, desired angular step size(like .1 degree per step) for different types of cam motion; Cycloidal, Harmonic, Parabolic etc., and it would crank out the data in seconds. The X,Y coordinates in the case of a flat plate cam, or X, A if the cam was machined on a rotary table. This is sort of "brute force" generating every point of motion by actual G-code X, and A steps. You can compile the Basic program, and have a bunch of little programs for each specific machining need. I should add, the code generated is to the "center of the cutter", not the cam surface. Depending on the size of the part, the program could be several thousand lines long. There are formulas in the math books for such motions, as well a curve fitting.

In the case of my increment example earlier in this post, making a loop, using G91 takes only a few lines of code. Mach3 does all the work adding numbers to the constantly changing X axis motion per the loop, and moving the X and A axis in sync accordingly.

It certainly is possible to write a Basic program that would read user input of starting pitch, ending pitch, and total length of shaft, that would do all the math and output a working G-code, G91 incremental program. But, it is most likely a one time job.

Sometimes just thinking about a given machining problem for a while will let you see a simple solution.

Regarding the 24" screw with variable pitch, there were enough "known" values to derive the "unknown" variable (pitch increase per 1 REV).

One place the Basic has been very useful to me, was writing code to calculate the offset from a NON-Linear line "Normal" to the line. There are cases where G41 or G42 get goofy, and will wreck the part, thus being able to create a "brute force" machine pass offset from a line in which there may be X,Y,A all moving together. Then following that line with a G40 (no offset) cutter will be a perfect offset. I am sure there are CAD programs that will do that, but they can run into the multiple thousands of dollars.

At the end of the day, simple is better. However, there is really no easy way. It all takes time; programming, setting up, test cutting in air (sometimes), special cutters, fixtures, setting  Zero's ....etc., and editing you code to fine tune the actual machining.

John

164

General Mach Discussion / Re: Mach3 and second parallel port card

« on: December 27, 2013, 11:38:45 PM »

jude:

I will go ahead and give my answer, as upon a re-read you have only 1 PCI card in place.
My PCI card was purchased from Newegg. Syba also makes a PCI card with 2 parallel ports.

My XP computer has the Syba PCI PP card, #SP-LP-MCS1P. Everything works fine. When doing a total format and fresh install of XP recently, it could not install, "No Drivers found" "Your device may not work properly" I opened the Hardware, and under ports, the port showed the yellow question marks.

I download the drivers from the Syba website. I may have had a disc with the original install, but could not find it. After downloading the drivers, and directing the wizard
to the driver location, everything installed OK. "Found new hardware", and all of that.

The second PP installed as LPT3, and the original built-in PP is LPT1 No more yellow question marks.

Opening the new LPT3 port, and opening the tab - Resources, the following showed:
I/O Range  EC00 - EC07
I/0 Range   E880 - E887

I closed the tab, and opened up the Mach3 config, Ports and pins, and checked the Port#2 box, Port enabled. in the Port Address box I entered 0xec00.
Also checked box "Pins 2-9 as inputs. The address is proceeded by 0x, indicating Hex. I see you do not have the Ox in your Port Address. That might be
a problem.

Port #1 has never been changed, and that Port Address is: 0x378.

Both boards work A-OK.
My system is an Intel Atom, CPU D525 @1.80GHz, Foxconn Barebones computer, from Newegg.

John

165

General Mach Discussion / Re: Mach3 and second parallel port card

« on: December 27, 2013, 04:18:28 PM »

jude:

Are both parallel ports on PCI cards, or is only the second port on a PCI card?

John

166

General Mach Discussion / Re: worm gear cad design

« on: December 27, 2013, 04:04:43 PM »

Tony:
Really not off subject, as the original post was about cutting a variable pitch screw 24" long, and then Khaled wanted info how to cut a 27mm pitch conventional thread, then things evolved into how to set STEPS PER with his gearing, ....... and so it goes. I suggested starting a new topic, but khaled is new to Mach3 and the forum, perhaps did not get my drift.
 
Never heard anymore about the original post. That was a really interesting problem.

I spent over 20 years doing prototype 4-axis CNC machining including all kinds of cams, so that post really got my attention.
 
John

167

General Mach Discussion / Re: worm gear cad design

« on: December 27, 2013, 01:02:46 PM »

Hi Tony:
We cannot use a non-whole number, even though as you said it will stop on 267, or 266 because each revolution will be either too little or too much. I know in the case of 6 revolutions it doesn't  mean that much, but the original post which is not even the same topic was about cutting an increasing pitch in 24" of travel.

In the case of my explanation of how to make a repeating subroutine, each pass though the loop had to be an exact amount. If the loop was executed multiple times, the error would be accumulative.

If the error in 44.4444 degrees per step was used, each 360 degrees of rotation would be off about a half step for each degree, which way I do not know. The would equate to about 8 steps for each revolution. This is not totally bad depending upon what the end use is, but I tend to think of how to get all the numbers correct.

Now that you have me thinking, we could always lie to the program, and instead of writing the G-code to go say 3600 degrees (10 revolutions), just write 3602 or 3603, and now it all works.

John

168

General Mach Discussion / Re: worm gear cad design

« on: December 27, 2013, 12:42:22 PM »

Hi Tony:
I thought I made this point in my post about steps and degees of rotation. I did not say to use a fraction like 44.44444. I fully understand a stepping motor will go to either 44, or 45, not some interval in between, because we are dealing with a magnetic lock on 1 step out of 200 per revolution. I thought this was clear.

My reply was to explain how the STEP PER number is derived, and why it must be a whole number.

I am trying to explain things that are pretty fundamental to a non-English speaking person, who although new to Mach3, is serious about doing a fairly difficult machine job. His English is 100 times better than my ability to speak any foreign language.

Regards,
John

169

PoKeys / Re: Pokeys - Eagle Cad

« on: December 26, 2013, 05:22:40 PM »

txpto:

Here is the link.
http://www.poscope.com/PoKeysCNCaddon

Also, while there, you can download the pdf manual. I thought it was a pretty impressive addition to the Pokeys boards.

John

170

General Mach Discussion / Re: worm gear cad design

« on: December 26, 2013, 03:54:18 PM »

Khaled:

The 254mm table, 90:1 ratio is a much better choice. However I  would still use a reduction from stepping motor to worm screw of table. Since the table worm shaft is much higher from the base, a 60t pulley or even an 80T woulld be better. If you used 20t to 80t timing belt reduction, the steps per would be exactly 200. This would make 1 turn of the motor = 1 degree of rotation. This will have much better torque, and the motor will be operating in its better range. The the cutting action will not be fast anyhow, as a 10mm ball cutter at full depth in the finish pass cannot be pushed fast at all. It also will require a quite LOW spindle RPM, as a great deal of cutter surface is in contact; it must be a slow, steady cutting action with coolant. So, the greater reduction all works for the positive.

As far as backlash, there is usually a means of adjusting the worm screw to ring gear to account for general wear. In addition to getting less backlash, if you put some kind of slight drag on the table with a friction material like a piece of leather, it will stay tight against one side of the ring gear; and the cutting force is a constant load in one direction only. You are not making a cut in a forward/reverse situation, like an X/Y axis anyhow. If you get it apart, you will probably find it needs lubrication if it has sat for a long time. The backlash will not be a problem if all things are setup correct. In rotary axis machining, having everything rigid is very important.

Testing the cutting action with a standard cutter of smaller diameter as I suggested, you can test how the cutting action works best by programming the cut from left to right, and then try cutting from right to left X axis motion. One direction may prove better due to the cutter forces, and backlash. When starting the cut, you can also do an MDI, moving the table and X axis backwards from the start position about 10mm, and then back to the start point, and all backlash will be out of both the X axis, and the rotary table. You can get all the correct settings, feedrate, best cutting direction on a test blank.

Once your mechanical motor setup is done, spend time getting the A axis motor tuning working smooth. There are some good trial and error tutorials in the forum how to get the best performance with stepping motors. The Velocity and Accel settings for the Rotory table will be quite a bit higher than the X or Y axis settings. Just keep increasing the numbers, and listen to the motor sound, and smoothness. It will eventually stall, or sound really bad. Back off the settings about 10 or 15%. You will get a feel after testing, what works best. It is all a trial and error unless you have the high-end drivers that have computerized analysis feedback.

John