Retrofitting a 1986 Maho MH400E

I thing it’s correct, but we haven’t find solution to “chicken/ egg problem”.
Good work
John

I think we need to disconnect the 7A1 from the 19K1 chain and use it to feed the axis.0.amp-fault-in pin in the hal file.
Something like net Xfault-in hm2_5i25.0.7i77.0.0.input-12=> axis.0.amp-fault-in. but reversed. (Reverse the output)

So when the amplifier is in trouble it will tell the pin axis.0.amp-fault-in to stop everything and put on estop the machine.

I need a day off work to sit down and study the drawings, to propose a solution to this.

I'd like to keep the MAHO E-Stop and electrics as original as possible. How about this idea...

Button SH-1 was the Phillips start switch, for me this is "E-Stop release". I will probably hard wire it in the top RH position of my control panel.

Have LinuxCNC set TB2-17 - "CNC-Ready" as soon as it starts, and hold it on as long as there is no LinuxCNC fault. I think this is what the Phillips did. Is there an existing HAL signal for "LinuxNC is running without fault"?
net CNC_ready.LinuxCNC is now running => hm2_5i25.0.7i77.0.0.output.08 # sorry, time I start learning HAL syntax properly
sets CNC_ready.LinuxCNC is now running 1
So 19K1 is primed to close, waiting for "7A1" Bb1-Bb2 to close.

Then, when the SH-1 is pushed the Indramat gets power via the 7K1/7K2/7K3 E-Stop relays, and closes its "7A1" Bb1-Bb2 contact in the "NC-Ready" chain, thus 19k1 can close and the system is on and should remain latched even after SH-1 is released.

If the indramat fails, it opens "7A1" Bb1-Bb2, which releases 19K1. If the LinuxCNC has a fault or fails, 19K1 is released directly. Either way the machine fails into E-Stop per MAHO's original design, without software intervention.

This way, the "drive enable" is only doing that single function, and we don't need to change anythig in the E-Stop wiring.
Mark

If we leave the wiring as is, then the indramat must energize with start button, not separately. Or as soon we release the start button then everything is deactivated again.