Re: [CAD_CAM_EDM_DRO] CAM to CNC interface?
Posted by
stan
on 2008-02-01 04:20:41 UTC
Sounds like you have the expensive parts already, if you have time to put into
it I would say go for it as the interface you would need to make up would
give a lot of usefull knowledge.
On the other hand the geckos suggested are an out of the box solution that are
tried and tested and accepted as being value for money with little or no
faults.
I'd guess around 60 hours to get it all together allowing for debugging and
having to go back and do bits again. Put that against $500 for the geckos and
they may start to look cheap.
Not sure what output your getting from the encoders, if your +-10v is
controlling position then I would go with the geckos. Pure analogue is mostly
dead and buried with good reason, it would be a pain in the neck to find all
measurements are out by 0.2mm because the temperature in the control box has
gone up by 5 degrees or the Z axis drops at full speed due to the phone
ringing.
If the encoders give digital output then I would recommend getting an I2C bus
set up on the PC and choose from the many different type of chips available.
TDA8444's will give 8 6bit analog outputs controlled by writing the value to
a text file on the PC end and PCF8574's will 8bit in (or out) for parallel
output from the encoders. The 8574's may be a bit expensive as you will
probably need at least 2 for each axis, however....
Ideally you want start with a signal at the parallel or serial port and end up
with the machine at the desired position. The I2C stuff is great for testing
and prototyping but what you really need is a microcontroller. If you can get
your head around these things then you will never ask a question even
remotely like this again. For hardware a PIC with 3 PWM outputs (or 3 PICS
with 1 PWM output each, but that's a lot more complicated) and 3 amplifiers
to take the (normally) 0 to 5v PWM output and turn it into +-10v (if MAX252's
give enough current they would be ideal).
Input would be better coming from the serial port as most microcontrollers
have a built-in serial line, the parallel port can be used but it needs a lot
of pins on the controller and the added complication of software flow
control.
Software needs a component to read the desired value for each axis from the
serial line input and a component to take each axis to the desired position
at the desired speed and hold it when it gets there (makes a lot more sense
after getting a controller and trying a few example programs).
This really isn't a big job, especially if the MAX232's have enough power to
drive you controllers, and the coding needed is quite simple and an ideal
introduction to microcontrollers.
If that sounds like way to much work then go with the gecko's. To get set up
and connected with the microcontrollers can be done for less than $50 but a
budget of $150 to $200 would be more realistic to allow for a programing and
development board and all the caps, resistors, plugs, cable, magic smoke,
etc. Also, the 60 hours isn't taking into account the time needed to get
comfortable with programming in assembler.
PIC's are probably the easiest microcontroller to understand but there are
many others available. AVR's are a common option (more powerfull in my
opinion) but are more complicated to get to grips with and are really
intended for C programing rather than assembler. AVR assembler is a lot more
like assembler used by PC processors and is more or less unreadable without
quite a bit of experience, while PIC assembler makes sense quite quickly.
Good luck with it.
cheers
it I would say go for it as the interface you would need to make up would
give a lot of usefull knowledge.
On the other hand the geckos suggested are an out of the box solution that are
tried and tested and accepted as being value for money with little or no
faults.
I'd guess around 60 hours to get it all together allowing for debugging and
having to go back and do bits again. Put that against $500 for the geckos and
they may start to look cheap.
Not sure what output your getting from the encoders, if your +-10v is
controlling position then I would go with the geckos. Pure analogue is mostly
dead and buried with good reason, it would be a pain in the neck to find all
measurements are out by 0.2mm because the temperature in the control box has
gone up by 5 degrees or the Z axis drops at full speed due to the phone
ringing.
If the encoders give digital output then I would recommend getting an I2C bus
set up on the PC and choose from the many different type of chips available.
TDA8444's will give 8 6bit analog outputs controlled by writing the value to
a text file on the PC end and PCF8574's will 8bit in (or out) for parallel
output from the encoders. The 8574's may be a bit expensive as you will
probably need at least 2 for each axis, however....
Ideally you want start with a signal at the parallel or serial port and end up
with the machine at the desired position. The I2C stuff is great for testing
and prototyping but what you really need is a microcontroller. If you can get
your head around these things then you will never ask a question even
remotely like this again. For hardware a PIC with 3 PWM outputs (or 3 PICS
with 1 PWM output each, but that's a lot more complicated) and 3 amplifiers
to take the (normally) 0 to 5v PWM output and turn it into +-10v (if MAX252's
give enough current they would be ideal).
Input would be better coming from the serial port as most microcontrollers
have a built-in serial line, the parallel port can be used but it needs a lot
of pins on the controller and the added complication of software flow
control.
Software needs a component to read the desired value for each axis from the
serial line input and a component to take each axis to the desired position
at the desired speed and hold it when it gets there (makes a lot more sense
after getting a controller and trying a few example programs).
This really isn't a big job, especially if the MAX232's have enough power to
drive you controllers, and the coding needed is quite simple and an ideal
introduction to microcontrollers.
If that sounds like way to much work then go with the gecko's. To get set up
and connected with the microcontrollers can be done for less than $50 but a
budget of $150 to $200 would be more realistic to allow for a programing and
development board and all the caps, resistors, plugs, cable, magic smoke,
etc. Also, the 60 hours isn't taking into account the time needed to get
comfortable with programming in assembler.
PIC's are probably the easiest microcontroller to understand but there are
many others available. AVR's are a common option (more powerfull in my
opinion) but are more complicated to get to grips with and are really
intended for C programing rather than assembler. AVR assembler is a lot more
like assembler used by PC processors and is more or less unreadable without
quite a bit of experience, while PIC assembler makes sense quite quickly.
Good luck with it.
cheers
Discussion Thread
you get one guess
2008-01-31 22:29:07 UTC
CAM to CNC interface?
Michael Fagan
2008-01-31 22:46:21 UTC
Re: [CAD_CAM_EDM_DRO] CAM to CNC interface?
hannu
2008-01-31 23:54:54 UTC
Re: [CAD_CAM_EDM_DRO] CAM to CNC interface?
stan
2008-02-01 04:20:41 UTC
Re: [CAD_CAM_EDM_DRO] CAM to CNC interface?
Jon Elson
2008-02-01 11:58:36 UTC
Re: [CAD_CAM_EDM_DRO] CAM to CNC interface?
Danny Miller
2008-10-14 14:32:55 UTC
Polished acrylic machining
NEVILLE WEBSTER
2008-10-14 14:38:16 UTC
Re: [CAD_CAM_EDM_DRO] Polished acrylic machining
aclausing2003
2008-10-14 15:33:04 UTC
Re: Polished acrylic machining
kefale1
2008-10-15 07:53:37 UTC
Re: Polished acrylic machining
Danny Miller
2008-10-15 09:56:35 UTC
Re: [CAD_CAM_EDM_DRO] Re: Polished acrylic machining
caudlet
2008-10-15 10:49:45 UTC
Re: Polished acrylic machining - OFF TOPIC
Mark Hubrich
2008-10-22 16:55:58 UTC
RE: [CAD_CAM_EDM_DRO] Re: Polished acrylic machining