Re: [CAD_CAM_EDM_DRO] T-Tech 5000 PCB Milling machine

Alan,

If I had the time to invest, I'd go for reverse engineering the RS-232
interface and command sequences. It's possible a Google search might
even find someone who has already done so.

I'd write a very simple program with the demo software -- something like
"Cut 1 inch in the X-direction". See what you get in RS-232 commands.
Then I'd change it to "cut 2 inches in the X direction", and see what
changes. If you do enough trials, you should be able to figure out what
is different between the command strings when only the length of travel
on one axis changes. Characters preceding or surrounding the changing
bytes will probably be associated with the X-axis command. Then I'd
write two consecutive identical X-axis commands. The repeating parts
would probably represent the entire X-move command syntax.

After that I'd change to a Y axis move, while checking that the
previously identified X-axis command bytes have not changed, and then on
to any other axis moves as well. Then I'd see what parts of the string
change with commands to change the spindle speed, and so on.

If you can isolate the individual command sequences, then it should be
feasible to write your own control program in a BASIC dialect (I like
Liberty Basic [LB] myself), or possibly modify EMC (Free Linux software)
to run your machine. There is a simplified dialect of LB Called
JustBASIC that is free for the downloading, and even the very powerful
LB is something under $50.00 for a full license.

Granted it's not a trivial process, but it looks like something you
could certainly do. I'm in the process of deciphering old NASA mission
control display panels that used RS232 communications, and came with no
docs. I'm half tempted to pull the ROMS that run the 6502
microprocessor inside and figure out what the command protocol was, but
I'm playing with reasonable ASCII sequences for a while longer before I
resort to that approach. It's a lot more productive than Sudoku!

Dave

Alan Rothenbush wrote:

> Being firstly an electronics kinda guy, and secondarily a machining person, for some years now I've been dreaming about a PCB milling machine, either building it from scratch or modifying something else.
> My lack of progress has been stunning, but the desire lives (lived) on.
>
> Well, yesterday I was at an auction where they were selling a T-Tech 5000 PCB milling machine. I somehow convinced myself that I
>
> a. could afford it
> b. actually had a need for it
> c. would save me money in the long run making my own prototype boards
>
>
> None of the above are of course true, but it's mine now.
>
> I also got a real deal on blank PCB, a zillion cutters for free (close enough), and the computer that ran it all, with the T-Tech software (ISOPRO) .. ERASED.
> OK, no problem, contact T-Tech .. yes, they'll sell me the software, for about a month's wages.
>
> Well, I'm into this too deep already to spend much more money at all, and certainly not that kind of money.
>
> They do have a sort of demo version, and I was able to hook that up and make sure that things move. I then looked harder at the computer to controller interface; standard RS232 serial. So, I installed a serial port monitor to look at the data travelling back and forth and HOPED for standard GCode.
>
> No such luck. What was being sent was something that looked a LITTLE bit like HP plotter language, but only a little bit. Some was obviously decipherable, some of it gibberish.
>
>
> As I see it, I have a few options, and I'm not at all sure where to go.
>
>
> 1. Just buy the _very_ nice looking T-Tech SW, plug in the dongle (hire a divorce lawyer) and get to work.
>
> 2. TRY to locate an old copy of the software. (I don't think I'd be violating any copyrights here, as the SW _only_ works with T-Tech machines, and the original purchaser of this machine DID have a copy; it comes free with the machine)
>
> 3. Buy a copy of DeskPCB (which I can just afford), then reverse engineer the protocol from PC to controller and write some sort of translation program from GCode to T-Tech.
>
> 4. Buy a copy of DeskPCB, build my 2.5 axis controller, run things off CNCPro.
>
>
>
> Option 4 is certainly do-able, but it hurts to replicate that very nice controller that's already there. Also, the controller is able to control the spindle speed, which is apparently a useful thing to do (although it can be done manually) Finally, I'd have to send out to get some boards made up, a bit of a cruel irony.
>
> Option 1 is where I'm headed, when I can afford it, as this software really is sorted out, and it all integrates very nicely. I'll just have to wait for a (long) bit.
>
> Option 3 MIGHT be doable, is certainly cheapest, but would also be the slowest, and I've got boards I need (OK, want) fabbed TODAY. Plus, I would probably need a copy of the real software so as to spit out and capture all the various options, but if I had a copy of the software, I wouldn't need to do this at all.
>
> Option 2 does not seem likely, as I can't seem to source this old stuff.
>
>
>
> Anyway, if anyone has any thoughts, comments, opinions or help to offer, I'd love to hear from them.
>
>
>
> Alan
>
>
>
>