system won't movie once first alarm is cleared

I’m using Labview, sometimes jumping to Coolterm.

One of the things that is messing me up is that although I’m in txt mode I sometimes get a json replay so my text matching gets messed up.

Here is my config fileconfig file

[You do have G55 and G56 offsets, but $$ agrees with you that G54s are 0.0.
Could be a Chilipeppr bug, or….]
I have the TinyG v8 board, latest firmware according to your website (.9?)
What is Chilipepper?

[This is a unique machine setup, X,Y,A1,A2.
Are A1 and A2 motors driving independent moving elements?]
I must have left A2 hanging because I am only using X, Y, and A1.
The Y is mounted on the X stage but the rotary stage is independent. This is not a CNC rig; the setup is being used in a lab to automate image taking of a tooth (graduate student project).

[I’d suggest you use $_pm=2, in-cycle. That way, motion on any axis will energize all motors, thereby holding position while other axes are moving.]
I can understand your suggestion although both my XY stages have acme screws and a minuscule load so it probably is not an issue. The rotary stage is heavily geared and horizontal so no issue with that.

[Are X and Y belt drives or something else?
At $_tr=40 mm, I doubt they are screws.
I ask because some of your jerk values seem strange, but I doubt that is your issue.]
I might have played with the jerk trying to speed up or smooth out something but it is not important in this case. The movements are strictly travel only and I may have been trying to minimize the noise and/or travel time. Also not important.

[You have soft limits turned on, why?
What was the Alarm that was cleared]
Somebody else is going to take over this project after I make a set of low-level subroutines for her. I’ve been trying to make the subroutines robust because she has no experience with motion control. After discovering the limit switches required a board reset which might screw up her situation, I decided the soft limits could be used to protect her from mistakes without bollixing up whatever she has running (she will be programming a Labview interface that she will then use for her graduate project. As typical for Labview, the development and usage are likely to be intermixed)
Her usage will be custom step-and-repeat sequences and probably make mistakes).
[One of the things that is messing me up is that although I’m in txt mode I sometimes get a json replay so my text matching gets messed up
Is this when using Chilipeppr?
CP(SPJS) forces tinyG into JSON for a lot of transactions, because SPJS expects JSON in return]

[Describe how Labview interfaces – does it interface to the Chilipeppr console, or what?]
Labview is talking directly through the serial port with $ and G commands.

I’ve tried to simulate what is going on through Coolterm but it’s hard to do; the interface routines so stat checks and stuff. I’ve built a logger so that I can repeat exactly what the Labview code is doing up to the hangup but it’s still difficult to do. That is where I discovered a few problems of my own.

One of my difficulties with this board has been an unclear picture of the state of the board in different situations. Many commands that don’t expect a reply get one anyway, so I inserted a serial port read command to clear the buffer before sending a command to the board that DOES expect a command.
I experimented a little with the alarm and certainly am confused. The alarm does not seem to be system wide. After getting an alarm on X axis, a did a successful Y move and $stat says no alarm. If I try the X axis again even with a valid move, I get another alarm unless I use $clear. How does the flowchart actually look in this system with regards to alarms? My general approach has been to do a move, use $stat to see if the move was ok. If Alarmed, I send a message to the user interface. Something in my code is assuming something that is not true but after massive experimentation I have yet to figure out what.

There is actually only one command that returns a json text but it is not a command that expects a reply so not a problem right now.

The only problem I really have right now IS the soft limits. In the end I can disable those but not happily so.

BTW, I fixed my homing problem as described in my original post; I was doing a text search in the reply for ‘homed’, not realizing the false reply was ‘not homed’.

• This reply was modified 7 years ago by matthewdc. Reason: removed redundant text