Excellent post, John.

John Prentice wrote:

>Greetings
>
><head_above_parapet>
>
>From: "Andre' Blanchard" <[EMAIL PROTECTED]>
>  
>
>>At 10:39 AM 8/20/2007, you wrote:
>>    
>>
>>>I beg to differ.
>>>      
>>>
><snip>
>  
>
>>The work offset numbers are available in EMC but, how to access the tool
>>table values in EMC?
>>  These move around a lot on different machines but are usually around
>>#2000 or #10000 depends on if you have type I or type II offsets.
>>    
>>
><snip>
>  
>
>>Machine position information.
>>  Preceding block endpoint, #5001,#5002, etc..
>>  Current machine coordinate, #5021,#5022, etc..
>>  Work coordinate, #5021,#5022, etc..
>>    
>>
><snip>
>  
>
>>And while EMC can do some of this the syntax is often very different 
>>making
>>running a program written for a Fanuc on an EMC machine a challenge.  Also
>>gives part programmers one more reason not to like EMC based machines,
>>which makes them less likely to get purchased in the first place.
>>    
>>
>
>If one could make a control that would run code unaltered from "a Fanuc" 
>then it might be useful. I suspect the truth is the best that is possible is 
>to run code for a "Fanuc xxE Rev d.q Option 3" and that is hard and has 
>little general applicability.
>  
>
Additionally, it's impossible to make something that's compatible with 
*both* Fanuc *and* Haas, for example.  So although users of a specific 
machine we might emulate will be happy, the users of the other 200 
models/brands won't.

>As soon as one conteplates modifying legacy programs full of # parameters 
>one is doomed.  Is anyone left of this list who literaly programmed a "real" 
>program in binary/octal/hex machine code without the benefit of a symbolic 
>assembler? The nearest I came was keying the standard PDP8 boot loader.
>
>I submit one should not contemplate extension of the # parameter mechanism.
>  
>
I agree wholeheartedly.  Unfortunately, it's often easier to make small 
changes to what's there than it is to design something new.

>So what is needed? Probably many things are useful but I suggest two that 
>are different in kind.
>
>(a) Interactive CAM. The distinguishing feature is that the result of using 
>such a system is a "G-code" program or program fragment. There are plenty of 
>exemplars both in commercial machines and PC controllers. In essence one 
>defines the desired cut(s) by choosing from standard operations like 
>pockets, facing, PCD circles, standard panel apertures etc. and/or teaching 
>by recording jogs.  Such a system have very little interaction with the 
>control beyond being easy to initiate, have a neat way of loading the 
>generated code and perhaps picking up DROs for teaching.
>The author of this type of program needs access to tools to implement a GUI 
>for his/her users and the ability to write to a part program "file".
>
>(b) Interactive machining. Here the distinguishing feature is that the 
>thread of code run by the user actually controls the machine. This allows 
>two things. Firstly the control flow of the user program depends not only on 
>user input but on the machine state (e.g. the tripping of a probe, the input 
>power to the spindle). Secondly it can take place whilst a "G-code" program 
>is loaded - so is useful for tasks like semi-automatic stock preparation.
>The author of an Interactive program needs access to tools to present the 
>GUI at the user side, the ability to issue commands to the machine (say in 
>the form of G-code strings) and to read the current state of hardware and 
>software (offset systems, modes, speeds etc). This is a tougher call.
>  
>
There's a (c) as well:  non-machining applications.  EMC2 has at its 
core a very nice motion controller, and an excellent I/O interface 
layer.  There are a number of ways of using motion and I/O that aren't 
related to machining:  robotics, packaging, cell-to-cell part moving, 
etc.  There are a number of generic motion controllers out there, some 
even have pluggable kinematics like EMC (so you can use a serial robot 
but specify waypoints in cartesian coordinates instead of joint 
positions).  Most start at the $2000 range, and don't have a lot of 
I/O.  They're programmed in various ways, but there's usually some 
scripting language (see Yaskawa's SMC4040 for an example of a script 
language I don't much like :) )  I can see EMC2 with a similar scripting 
language as a front end instead of the G-code interpreter.  Ideally, I'd 
like to be able to have the motion controller capable of servicing 
multiple clients simultaneously, but that's a very radical design 
change.  Note:  I mean multiple clients, like a G-code interpreter that 
outputs commands for 4 axes, plus a robot controller that outputs 
commands for a separate two or 3 axes, each being able to do its thing 
without the other necessarily knowing about it.  This is distinct from 
several user interfaces all controlling the same "motion client":  the 
G-code interpreter and its associated stack of protocols and processes.

><head_even_higher>
>
>These interfaces feel to me to be much like interactive web page designs. 
>Flash (and its support tools) seems to be a strong contender (very powerful 
>graphics - static and dynamic, a form of object orientated programming, 
>surprisingly efficient) although the Linux world seems currently rather 
>deprived on implementation.
>
></head_even_higher></head_above...>
>  
>
<head out of a$$>

I'd say that flash, although it's more or less fine for media, has no 
place in the requirements list for EMC2 :)  There are any number of fine 
programming languages and environments to use for the UI.  I'm not sure 
what you'd use to make a flash presentation on Linux anyway, and I 
wouldn't want to be beholden to Adobe to make updated versions for my OS 
(which they don't - I use a 64-bit version of Linux, and they don't seem 
to like supporting 64-bit OSes on anything but PowerPC macs AFAICS).  On 
a technical level, I'm not sure what facilities Flash has for actually 
doing things that aren't "media" or web-related anyway.

</head back in a ... oh, wait a minute>

>John Prentice 
>  
>
- Steve


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >>  http://get.splunk.com/
_______________________________________________
Emc-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/emc-users

Reply via email to