[Brian L]

OK, I've loaded this program 3 different times into the machine, at over 45 minutes per time, and it still crashes in various places. It will say illegal character, for example one line had no X value, just a -1.575 type of thing.... so all I can assume is the program is getting scrambled on the way over.

So... I check the program in my NC Link text editor (from One CNC) and all looks good, at least on every line it crashed in the machine on, there are a tad over 102,000 lines of code... I didn't check every line. Did some searching here on DNC and I think I have things set correctly, just have questions about this buffer size, where to set it, and such.

So... I tried setting the machine into DNC, then Terminal, and then send the program from the laptop to the machine.... works for a bit, varies as to when it craps out, but it always craps out before I get too far. So maybe those smarter than I can take a peek and tell me what I'm doing wrong.

A picture of a rough idea of what the part will look like:



My roughed out blank of MDF:



The alarm I'm getting now:



What shows up on the screen a few minutes after the alarm:



and what my comm software settings screen looks like:

[ZZZZ]

On the machine, cursor to the bottom of the Parms-Ctrl list to find the communication parameters.

[Brian L]

That's where I got the Com1, 9600 baud, Even, 2 stop bits and 7 data bits..... So now, given the com software screen, I have XON/OFF checked, which is what my Milltronics paper on RS232 Communications says... but do I check start at machine, or start at computer.... pretty sure I've tried both.

Do I need anything in there for First lines and Last lines? I put 255 in the line delay, but I think that's between each line, where/how do I set a buffer value? Read in an old post here that it should be 255. It would seem I'm overflowing the control, at least that is what I assume based on the "Insufficient RAM memory" error I'm getting... that shuts the whole machine down, servos and all, have to go in back and flip the power switch to clear it.

It shouldn't be this hard to run DNC, tons of guys do it, I must be missing something easy.... like where on the machine do I set the buffer amount, or do I set that on the PC, and if so, how?

[moldcore]

Hey Brian, does your machine have a floppy? I DNC from the floppy on my Cent. 5. Been doing it for 15 years without a problem. No need to wait for it to load, just put in the floppy hit a few keys on the controller and off you go. I have a cable from my computer to the machine but haven’t used it for so long I can’t remember how it was set up. I was able to set it up with help here on this forum about 10 years ago but it looks like yours is set up like mine was (just checked, it is). I do a lot of large programs but have found ways over the years to break them up to fit on the floppy drive. Besides my old Centurion 5 can’t hold anything over 100k I believe, so I gave up loading long ago unless it was a production job. If the machine wasn’t so old and slow I would update it.

[Brian L]

Well, it does have the floppy drive, but that's limited to 1.44mb and my program is larger than that. I upgraded the computer and have 2 gb of memory now, so once the program is in the computer, I can DNC right from there... if the program is right. Unfortunately I suspect the program is getting garbled slightly on it's way into the machine, and because of the DOS limit, I can't open/edit the program while it's in the machine to double check it.

Hence my attempt at DNC from the laptop. I did some more searching on this forum last night and found some reference to adding a 200 ms delay and LF to each line, so I can try that and see if it either transfers in without errors, or will run from the laptop.

[Brian L]

As an update.... I don't think I have my DNC process incorrect.... I think something is wrong with the machine. I worked on it late yesterday and after reading a ton of posts here, set a 200ms pause at the end of each line, made sure my file was a .txt file and checked the option to turn the screen on during the feed on my pc, I was able to start DNC, watch the computer feed in approximately 20-30 extra lines of code as expected.

Then the feed stops until the machine uses some code, and feed of the program resumes. Long linear moves weren't eating code very fast, so easy to watch, then as I got into more 3D moves the code started feeding in faster. The machine never seemed to outrun the computer, even running 140% (42ipm) of my programmed 30ipm feed rate.

So, it runs long fine for an hour or so, part is 8" square and it's moving over .010" X minus per pass, cycling back and forth with the YZ axis doing the up/down and in/out. Seems to be working just fine and then about 2500 lines into the program it stops and throws the 901 Parity alarm... the line of code it stops on is fine, it has executed that line at least 100 times already as it moves across the part.... So I am at a loss as to what is really going on here.

[ZZZZ]

Could you have some kind of a Windows timing problem?

Do you just throw away the first 2500 lines and start again?

[Brian L]

Could be a timing issue, like the laptop trying to hibernate or windows off doing some BS. I will try a couple things.... we did fish a new ethernet cord from the back of the machine out to the control pendant, thought it would be easier to hook the pc up in the front. I wonder is the proximity to the other cables in there, including a 110V power line, could be scrambling the input.

Figured I'd pop open the back door and attach my short ethernet cord back there directly to my laptop, turn off all the power saving features on that, and try to get a clean copy of the program into the control.... at least it is something to try.

[Brian L]

Another update.... I did get the program to reliably go into the Mill's computer. I opened the back door, hooked my ethernet cable (I have a DB9 to ethernet adapter plug on the computer, not using the DB25 on the side of the cabinet) to the computer and used a short 6 foot cable hooked to my laptop sitting on a stool behind the machine.

I also added a 200ms pause to each line, not sure what fixed the problem, the cable or the pause, but the program did go in.... albeit slower than molasses... took over 4 hours for a 1.3mb file.

So, my next question, is anybody using a pause at the end of each line, and if so, what sort of value, could a 5 ms pause be as effective as 200 ms? Or do I even need it and the trouble was in the cable? I'll test it both ways again and see what is causing the problem, but just wondered what others that download large files or run DNC are doing....

[sportybob]

Brian, I belive that your problem is in used newer Windows IE; newer then Win98, in that the orginal IO used back then did not have the newer I/O chips. Go into your advanced settings on your PC and turn off the FIFO ( first in first out) buffers. this will make the newer I/O from the PC act like the older I/O chips.

Sportybob

[Brian L]

Hi Sportybob,

I found that all the issues appeared to be coming from an unshielded RJ45 (ethernet) cable I had run from the back of the machine to the front. It ran inside the tube with the power cable for the monitors and such. I have since replaced it with a shielded cable and am doing tests now to see if I have any more trouble, so far, so good.... although I haven't tried a DNC from the computer again, since I have plenty of memory on the mill's computer (2 gb) I have just uploaded and ran DNC direct from that computer.

[Nathina]

Hi Brian L, I have the same question. I've also put this application onto the system three times now, each time taking over 45 minutes, and it still crashes in various locations. Thank you for your cookie clicker and great post.