Den ons 9 dec. 2020 kl 13:33 skrev Gene Heskett <ghesk...@shentel.net>:
> On Wednesday 09 December 2020 03:52:16 Sven Wesley wrote: > > > Den ons 9 dec. 2020 kl 02:27 skrev Gregg Eshelman via Emc-users < > > > > emc-users@lists.sourceforge.net>: > > > The way to wire an emergency stop circuit with 2 or more things that > > > can trigger it, is to have it a normally closed circuit and every > > > stop switch, mechanical or electronic, in series. > > > That way when anything triggers a stop it opens the circuit, the > > > controller senses that and stops sending motion commands, shuts off > > > the spindle(s), pump(s) etc. > > > What does this controller mean by "high" if "low" is grounded? If > > > "high" is open circuit then you should be able to connect every > > > fault sensor and a normally open E-Stop button in parallel. You'll > > > also need to have the new control hardware and software able to be > > > set to use a closed circuit instead of open circuit as the E-Stop > > > condition. > > > If "high" is something other than open circuit then things get > > > more complicated. One may curse the engineer(s) who thought that > > > making things more complex than open or closed was a neat idea. ;) > > > A NC E-Stop circuit is quite fail safe because if a switch fails or > > > a wire breaks, the circuit goes open and the machine stops. If a > > > wire breaks in a NO circuit then the E-Stop button or sensor the > > > broken wire goes to cannot initiate a stop. > > > > > > On Tuesday, December 8, 2020, 04:19:16 PM MST, Sven Wesley < > > > > > > svenne.d...@gmail.com> wrote: > > > > Friends, > > > > > > > > I am in the process of refreshing a CNC that has been running in > > > > > > "temporary > > > > > > > state" the last 12 years, time to give it an overhaul. I've > > > > found the > > > > > > magic > > > > > > > little piece we all call 7i96. It will do magic. > > > > But this isn't an E-stop circuit though, it's an error signal from the > > drive that can/should be reacted upon. It makes total sense to have an > > active low on such wire. If the drive dies completely the signal will > > disappear as well. > > > This idea came from my r-pi4 driven Sheldon lathe. I had recently > replaced the x & z drives, which were junk box occupants when I did the > conversion to cnc originally, with one of the new 3 phase stepper driven > servo's a, 3NM for z and because it was shorter and its length would fit > on the back of a new apron, a 2NM for x, both nema 23 sized. Because I'm > using the spi driven 7i90HD and buffering it with a triplet of 7i42TA's > which make the wiring nice and gives me 72 i/o's, I had gpio's to throw > away. So I brought the 2 error signals in on separate pins and that > allows me to issue a pyvcp based video advisory. Because all the r-pi4 > stuff is on the rear of the lathe, leds in that location, includingthe > fault leds on the drivers themselves, are out of sight, so I did the > video ouyput thing. > > Basically, a fault means a loss of home because it shuts down the drive, > the machine will need to be rehomed anyway. So a fault bangs off the F2 > status, which all machine power to be removed as I have a couple 40 amp > Solid State Relays in series with the 250 volt feed. 10 seconds of > powerdown resets the faults, so I remove the tool, wait for the vfd > display to quit, press F2 to re-enable machine power, which is taalied > in the video, rehome the machine, replace the tool, and decide what to > do so it doesn't fault again. Since I write my own gcode, thats the > fitst thing I inspect, usually finding a mistake in the gcode. Fix that, > reload the gcode again and hit the r key. > > I have not had an x caused fault yet because a 2NM motor driving that > tiny x screw with a 2/1 reduction, is enough to damage the 8mm screw > that moves my x, so since that screw has no replacements, I take no > chances of stripping it. I can't even find new balls to re-stuff it > today. And there simply is not enough space for a bigger, commonly > available 12mm screw. > > These 3 phase drives have no currant setting switches, use very little > current in normal operation as they control the current by using enough > to attain the position and fault when they cannot do that because > something has blocked its movement, the driver being its own PID, they > issue the fault and shut the maqchine off if they can't follow orders. > > There are NO math intensive PID's in that r-pi4 install, so it either > does what motion tells it to, or faults. With a vfd controlling the > spindle, the spindle encoder only has one job. making G33.1 and G76 > work, the left end of that lathe has been totally stripped. Linuxcnc is > the compound so there is a block of cast under the toolpost, giving new > meaning to its rigidity. The old compound was broken beyond any hope of > repair when I bought it about 5 years back because I needed something > bigger than my 7x12, alhough I made most of the parts to do that > conversion on that 7x12, which itself bears very little semblance of its > original assembly. It also has lost its sponge rubber compound, and > gained a 5" chuck, tapered gibs, and now has about a 2 horse treadmill > spindle motor. And they both can cut any thread I specify including many > that aren't in the machinists handbook, metric or inch. A 7/8" 50 tpi? I > think there is gcode to cut it already written. :) > > > Regards, > > Sven > > Take care and stay well Sven. > > Cheers, Gene Heskett > Nice read Gene. Actually, I think one of the first replies I've got in this list was from you like 16 years ago or so. :) Maybe my intentions have been unclear and now I'm a bit confused. The drives provide a high signal on one pin when they are working. If they fail the pin goes low. My initial thought was to connect all four pins to one input on the 7i96 and configure it to trigger a software stop. Like Gene proposed, shouldn't it work with all wires to one input with a diode in series on each pin? Best regards, Confusedius _______________________________________________ Emc-users mailing list Emc-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/emc-users
