QTplasmac, Estop Mode 2 and estop-latch - a safety issue

rodw wrote:

phillc54 wrote:

I don't think a GUI should have an Estop button.

Technically it is an estop reset button.
It should not display an invalid state if the estop reset fails. It should stay red.

[/quote]
I wanted to get rid of the button function completely but there were complaints. IMHO all estop should be done external to the GUI, that is why I had the only options originally of an estop indicator or no estop indicator.

I also find it a bit odd that you mention safety and yet have an Estop over wireless.

One of the problems with LinuxCNC, everyone has an opinion and there are a lots of them.

Even if I was a volunteer developer, I don't think I'd like to knowingly allow a non standard safety circuit situation to develop in the GUI.

Well in that case I will remove the button function and go back to Estop modes 0 and 1 as they were once:
Mode 0 enable the Estop indicator
Mode 1 hides the Estop indicator
All Estop functions are handled external to the GUI.

There are also outputs on a safety relay, so why not setup your estop circuit with the safety relay, and when you set it with an input that clears estop. Take one of the outputs of the safety relay as an estop input into Lcnc and use the estop in QTP as an indicator. Use a second output to remove power to motors. This will turn off the machine in the gui and wont be able to turn it back on until your estop circuit is clear. 

bevins wrote:

There are also outputs on a safety relay, so why not setup your estop circuit with the safety relay, and when you set it with an input that clears estop. Take one of the outputs of the safety relay as an estop input into Lcnc and use the estop in QTP as an indicator. Use a second output to remove power to motors. This will turn off the machine in the gui and wont be able to turn it back on until your estop circuit is clear. 

Thanks Bevins, I don't have the estop relay installed on my machine but do have on one a desktop 7i76e testbed.

Looking at some other QTvcp screens I see QTdragon gets the terminology  right and uses "estop reset" in its GUI. It appears to follow the same methods as axis and gmocappy regarding estop.

QTplasmac has some conditional hal code in machine.tcl that overrides behaviour to support its estop modes. I don't think MODE 2 fully replicates the default behavoiur of axis. Maybe it can be fixed by unlinking some pins and redoing the hal connections. I'll try that when I get a chance.

You can demonstrate the issue in the sim if you change the estop mode to be type 2 in the SIM ini file.

This shows the state on power up.
 

This shows the external estop is off  but the screen button is enabled. This is the bug. The state of the estop button should be reset to red in this case.
 

This is with the machine on
 

So linuxcnc behaviour is correctly respected but the state of the GUI button is allowed to be inconsistent with the machine state.

It would be good if this was reviewed.