I tired to respond to a old thread (http://www.artsoftcontrols.com/forum/index.php?topic=2570.0) on this same subject but it does not seem to show up properly so I thought I would start a new thread.  The subject is retrofitting a Bridgeport Discovery 308.  While it's basically in good shape the control is shot and no parts are available, etc.  I would like to keep the drives etc, but the drives have analog inputs.

I have the same mill in my lab that we are giving to another department and I am going to retrofit it for Mach III for them.  From reading the above mentioned post and other posts about the Galil 18x2 boards and Grex, it looks like the Galil might be the way to go. The Grex analog outputs are only 8 bit resolution so I would have to us it's SD outputs and something like Pixie's to control the analog drives.  Considering that fact that both the Pixie's and Grex would want acccess to the encoder signals, I can't see how that would work; please correct me if I'm wrong. (I did email Gecko to see what they say  ).

Here is the cost breakdown of both options as I see it:

Gecko:        $400                                  Galil: DMC-1842         $1200
Pixe's: (4)      320                                  64 bit I/O expansion     200     
                   ------                                                               ---------
                   720                                                                  1400

So from a initail price the Galil is about twice as much but the Grex might not be able to do the job, any input?

Thanks,

The grex doesn't care about the encoder inputs. I doubt it ever will. The Pixe's and a grex should work fine. ( I haven't tried it yet.)

Brian would be the one to ask about the Galil. He is the one working on that.

Chad

I ordered a galil to have here for testing and I HOPE to have it in a few weeks I want to add rigid tapping with the Galil!!! that is about the only reason that have ordered it

The grex doesn't care about the encoder inputs. I doubt it ever will. The Pixe's and a grex should work fine. ( I haven't tried it yet.)

Hmm...that's curious as the G-Rex description PDF says...(the inputs are also described in teh programming manual)

2) 6 quadrature encoder inputs including index channel. Maximum encoder count rate is 1 MHz. Encoder inputs are filtered and have indicator LEDs on each input.

So, am I to understand that this functionality is not available/utilized with Mach III at this time (or ever)?  You still set accel/decel and velocity withing the G-Rex, correct?  And, I guess if the G-Rex does not handle the encoder inputs it is runnign in open loop mode making the Galil boards the only current choice for closed loop operation?  I would also like to hear from anyone who is succesfully running a Galil board.

Thanks for all teh help guys, just trying to get this all striaght in my head.

Perhaps I am confused but if the servos/drives are ok all you would need to do is install the pixies to the drives and hook up to the lpt ports?? The LPT's are hard to beat in form and function.  The servos are closed loop at the servo drive level and can be setup to stop the program if the error level is tripped. True closed loop error correction at the controller level is very complex and is rarely needed.

If the drives are bad then all you need is the gecko's/powersupply and hook up to the LPT ports

Keep it simple  (;-) TP

I'm just rying to consider all the options.  Mach works great with a single LPT on my Dyana 2400.  The Bridgeport has MANY more requiments such as, tool changer, more limits, enclosure safety switches, spindle control, probing, MPG and etc.  Now if I went the LPT route I would need at least two LPT's, two breakout boards, 4 Pixies, and likely a PLC to handle the tool changer or other auxillery functions.  Add up the cost and complexity of such a system and your in the same ballpark as a Galil, the G-Rex however seems to have 90% of the lower cost Galil board functions, for 1/3 the price.  I would only slightly disagree with you on the necessity of closed loop operation, I would not say it is 'rarely needed', but rather 'it is not always needed', it all depends on what the system is designed to do.  (But then I could be picking nits  )

So, if anyone is really using a Galil board how is it working for you?  Are you closing the loop?  Are the encoder inputs really used on the G-Rex? I have emailed Gecko but not heard back yet, guess I'll try calling tomorrow.

Thanks again to everyone for all the help,

Hi Jeff,
I am running a series 1 boss Bridgeport , 4axis,mpg, probe, spindle control,modbus, on 2 lpt ports with a lot of extra inputs/outputs leftover. I do have a closed loop at the drives, if it exceeds the tolerance it signals Mach to stop. It just does not do online corrections with mach. I don't think any other low end controllers do either. I really think it could handle a tool changer with no problem, but the qc30 spindle is not easily adapted to tool changing. Keep us informed on your project, it is always exciting to see a new project starting up. I usally learn something following a project such as yours.
 (;-) TP

after a bit more research I find that the Discovery 308 requires the following I/O:

19 GP inputs (18, digital 1 analog)
15 axis specific inputs (encoders and limits)
13 GP outputs
4 Axis specific analog outputs (drive command signals, or could be generated with S/D converter)

So it looks like I could get by with a G100, I am still unclear on how/if it works with S/D converters but am attempting to find out.

I still like the idea of the Galil PCI card, but will have to do a bit more reading on it.  Which ever route I go I'm sure that that the Mach III crew will be there to help out.

The G100 is working but only in Exact Stop mode CV is not working at all on my mill... I know Steve will get this fixed... I just don't know how long it is going to take him.
The Galil can do all the IO and will not have a problem with the CV but I would need to know more about your application and the parts that you would like to make before I can recommend anything

The type of parts, that's a tough one....the machine is going into a university machine shop setting.  It will be used by a limited number of students to make parts for student design competitions.  So, stuff like hubs for the solar car, parts for robots, etc.  They have no CNC machines presently so we can live without features that may not be working yet like, rigid tapping or CV mode.  Clear as mud?

What I'm most concerned about is system complexity.  Although the G100/Pixies may work fine the route of Mach III -> G100 -> Pixies -> Drives is more convoluted than Mach III -> Galil -> Drives.  Simpler = Better