HAAS VF1 retrofit

sspeed33317 wrote: And is sensorless vector good enough for CNC? I can make it sensor vector also if i have a encoder hooked up.
This is with the WJ200,

It depends on what you need to do. Many people CNC with manual on/off control of the spindle.
Sensorless vector will help at very low spindle speeds, and might be enough for spindle orient with a plunger/switch arrangement.
If you want spindle-synchronised motion then you need a spindle encoder, which might or might not also be suitable as an encoder for the drive depending on if the spindle motor is direct-drive or not.

A typical arrangement is probably a sensorless vector drive controlling the motor and an encoder feeding back to LinuxCNC on the spindle for rigid tapping and/or tool change orient. But with a spindle encoder you could rigid tap with a simple non-vector drive, LinuxCNC makes the axes follow the spindle encoder, rather than trying to control the spindle as a closed-loop axis.

Well I received an email reply from Lee Osborne

He is very busy but gave me a quick explanation on how he managed
his Cincinatti tool changer, operates pretty much the same as a HAAS.
He did it using GCode subroutines.
Commanding a o<tool1> instead of a T1 M6.

Also he says he is keeping the machine running 12 hours a day...

I'm starting to feel a little more confident I can get this thing going.

sspeed33317 wrote: He did it using GCode subroutines.
Commanding a o<tool1> instead of a T1 M6.

The (yesterday) released 2.6 version of LinuxCNC allows you to remap M6 to call a G-code subroutine.
All it takes is a line in the INI like

REMAP = M6 ngc=mygcoderoutine

Then (with a switch, also in the INI) the G-code can use

#1 = #<hal[iocontrol.0.tool-pocket-prepare]>

to read the HAL pin driven by the T-word directly.
Or, if you prefer, you can have the REMAP routine run a (provided by LinuxCNC) Python routine that reads the T-word into a parameter called #<tool> for the G-code

These two approaches may not be entirely identical. I think that the latter reads the T-word on the same line as the M6. The former reads a T-value from the last invocation of the T-word.

OK, Mesa boards arrived today. So I can get to work this weekend.

I already mounted the motherboard in the same space as the old controller.
Taking pics as I go.

Tonight I will go back over to the shop and grab the input/output panel out of the enclosure
to mount the 7i77.


I have one question how did you guys get so familiar with programming?
Do you have coding experience? You make it sound so easy but it is still over
my head.

I have read a lot of the documentation but it reads like I need a little more to understand it.

What do I need to read/learn in addition to the LInuxcnc docs ?

Thanks
TIm

sspeed33317 wrote: I have one question how did you guys get so familiar with programming?
Do you have coding experience?

Yes, quite a bit, in my case.

But don't mistake HAL for programming. It looks a bit like a programming language, but it isn't one.

Think of HAL more as building an electrical circuit. You load the "components" you are going to need, then you add them to a realtime thread (I tried to find a convincing metaphor for that part, but gave up) then wire them together with "net" statements.
A net statement creates a signal, or connects a previously created one (think of it possibly as a wire) to the input and output "pins" of the components.
So, rather than a programming language, HAL is more like a list of connections.

G-code wasn't meant to be a programming language either, but it can be used as one now. However it is perfectly OK to produce "linear" G-code with no loops or conditions or variables, in fact that is what most CAM packages create.

Well tonight is a bust, did not realize the HAAS I/O board was attached straight to the enclosure panel.
So I will be stopping by my friends sheet metal shop in the AM to have a stand off plate made.
The I will mount the new Mesa I/O to it.

The original board is pretty simple, has a bunch of optoisolated transistors on it and 12 volt coil relays.