Forum Replies Created
-
AuthorPosts
-
mcgyvrMember
and more info (don’t forget to read what I’ve already wrote in the last 2 posts above)
Just switched my solenoid to its own power supply on its own independent circuit. So I have the 24V supply (Meanwell SP320-24) for tinyg just running the board/steppers, the 24V supply is plugged into a UPS. The solenoid now has its own 24V supply (adjustable Protek lab power supply) plugged into its own circuit. I have snap on ferrite beads around each pair of limit switch wires as well as on the DC input wires to the tinyg board. I also have .1uF caps from each limit switch to ground. Using NC “industrial” ($70/each) honeywell limit switches. Any “disabled” limit switches do NOT have caps on them. My limit switch wiring is all shielded cables with the shields tied directly to the ground pins for each limit switch and the other end of the shields (far end from tinyg) not connected.
It made it through 4 complete gcode cycles then..boom psycho speed mode again.. It “attempted” to complete 3 G01 moves during that psycho time and then tinyg went to the flashing light/reset type mode..
mcgyvrMemberMore info.. Finally got my UPS in the mail so I hooked it up to that..It ran just fine 4 times through the code then boom again..
This time coolterm had finished sending the gcode and I was actually able to see an error Its all right here. Is it “telling” of something that the system shutdown error came then a movement then again ? Or how about a -0.000 velocity..? pulling at straws..Note it did hit a limit switch (never should have been close but when it went psycho speed mode the y axis slammed into the non-homing side limit switch. It really seems like its still trying to execute the code but when the axis goes to full speed obviously tinyg would be lost as to where it really is.
Here is what I was able to get in coolterm..
tinyg [inch] ok>
qr:1
tinyg [inch] ok>
tinyg [inch] ok>
qr:1
tinyg [inch] ok>
qr:2
tinyg [inch] ok>
tinyg [inch] ok>
tinyg [inch] ok>
posx:6.675,posy:6.750,vel:-0.000
qr:3
posx:6.675,posy:6.750,vel:0.416
posx:6.684,posy:6.773,vel:41.623
posx:6.744,posy:6.924,vel:149.636
posx:6.858,posy:7.206,vel:199.584
posx:6.975,posy:7.500,vel:176.587
posx:7.053,posy:7.694,vel:81.582
posx:7.075,posy:7.749,vel:6.660
qr:4
qr:5
qr:7
posx:7.075,posy:7.750,vel:-0.000
qr:8
posx:7.075,posy:7.750,vel:0.416
posx:7.097,posy:7.738,vel:41.623
posx:7.228,posy:7.667,vel:149.636
posx:7.493,posy:7.524,vel:199.063
posx:7.783,posy:7.367,vel:200.000
posx:8.073,posy:7.211
posx:8.362,posy:7.054
posx:8.635,posy:6.899,vel:185.016
posx:8.853,posy:6.789,vel:100.000
posx:8.922,posy:6.752,vel:12.591
qr:9
qr:10
qr:12
posx:8.925,posy:6.750,vel:-0.000
qr:13
posx:8.925,posy:6.750,vel:0.416
posx:8.936,posy:6.777,vel:45.890
posx:8.994,posy:6.924,vel:149.636
posx:9.108,posy:7.206,vel:199.584
posx:9.225,posy:7.500,vel:176.587
posx:9.305,posy:7.700,vel:75.858
posx:9.325,posy:7.749,vel:6.660
qr:14
qr:15
qr:17
posx:9.325,posy:7.750,vel:-0.000
qr:18
{“er”:{“fb”:370.08,”st”:27,”msg”:”System shutdown”,”val”:1}}
posx:5.000,posy:8.000,vel:0.000,stat:3
qr:24
{“er”:{“fb”:370.08,”st”:27,”msg”:”System shutdown”,”val”:1}}- This reply was modified 11 years, 4 months ago by mcgyvr.
mcgyvrMemberJust tried sending the file with coolterm and it went into “psycho fast mode” right after the second G01 move.. 🙁
Ran it again 3 times and it completed without any issues. Ran it again and it went crazy again.mcgyvrMemberAlden, (and Riley)
Thanks..
I know these types of problems are virtually impossible to “internet diagnose” so I really do appreciate any help you can provide.mcgyvrMembersorry just frustrated.. It will run through the code a few times np too..but then boom.
The settings I have for max velocities are fine for these. They are Haydon Kerk linear actuators rated to 20 in/sec… its just when Tinyg randomly takes them over that I’m having the problems 🙂 🙂
File I’m now running is here.. all comments stripped.. just the M8/M9 bug workaround and some positioning steps. When the issue happens is right around line 180 on the g01 moves (cylinders always been in the up position..with the M9)
https://dl.dropboxusercontent.com/u/9726802/TinyG/testgcode.txtMatt said he was going to write a little “black box” to log all the communication from tinyg/android app and hopefully we will see something odd there ..I will post that when I get it.
- This reply was modified 11 years, 4 months ago by mcgyvr.
mcgyvrMemberOk.. update.. Removed all the ; comments..problem still occurs, removed ALL comments..problems still occured.. Rolled back to the master build firmware and the skipping of G code lines is 100% gone.. HOWEVER I still have the problem that I talked about in another post with the stepper motors going to an “insane crazy” velocity.. WAY faster than what I have set as the velocity or feedrate maximum. It only happens on occasion and the only help I can give at this point is that it happens at the exact same time that (or a split second after) the android app finishes sending all the gcode to tinyg. (note its also happened with tgfx so its not just a Matt Stock app problem). At that time it “seems” to be attempting to finish the program but the stepper speed is so fast that its too much for my leadscrews/steppers to deal with so its a grinding noise,etc… Note that the android app has also stopped reporting velocity at that time but the positional DRO is still updating. I’ve had this exact same issue with tgfx.. Tomorrow I will run the gcode program multiple times from coolterm to see if it shows the same issue.
Really need some help to deal with this..
Note I’m only using simple X/Y axis G01 commands BUT I’m also using M8/M9 commands with G4 Px delays which I would expect very few if any other TinyG users are. So please keep that in mind when you think about a possible issue.
Again ..PLEASE help. I’m getting close to the edge of sanity now..mcgyvrMemberI can’t really run the master build due to another bug 🙁
or maybe I don’t understand how git branches work?Tomorrow I have a new apk from Matt to test and try a separate power supply for the solenoids. And I’ll see what happens with that first then look to downgrading the firmware for some testing I guess. Any chance I can get a master hex file with the m8/m9 bug fixed?
mcgyvrMemberoh ok found out that Matts app was turning it off.. But does TgFx use/need it on?
mcgyvrMemberJust let me know what I can do to assist in this… I really need something besides coolterm to reliably send code.
Should $ex be set to 1 as default.. Does TgFx use it?
Prior to trying coolterm per Riley’s suggestion $ex was = 0 (loaded like that by default which I thought your wiki says it should be 1 by default)mcgyvrMemberFYI..
I am using NC limit/homing switches..
My issues so far have been
#1-blinking board mid job
#2-“supposed” tripped limit switch mid job but not even close to the switches.
When #2 has happened TgFx has popped up the “a limit switch has been triggered click here to reset.blah blah dialog.
BUT at that time it also went “crazy” in that the steppers went FULL speed (as in WAY faster than my max velocity is set for) and “sounded” like it was stripping all the threads in my leadscrews. The axis doesn’t move as speeds that fast cause my leadscrews to bind (for lack of a better term) but I just don’t know why it would go into that “crazy..way over max velocity speed cycle”I don’t have a spindle to attribute noise too but rather a pneumatic cylinder that is triggered with the coolant pin.. But its running to a SSR and I have flyback diodes across the solenoid valves. I am powering the solenoids from the same 24V supply as TinyG uses though and am going to try to separate it on its own 24V supply.. However seeing as the gcode runs perfectly with coolterm Xon and I’m not seeing those issues I suspect its more of a bug somewhere vs a noise issue.
I do have shielded limit switch wiring (grounded shield at tinyG) and NC switches and have placed a .1uF cap across each switch line. But the problems still occur.. But only when running TgFx or Matt Stocks android app
And I did notice $ex was set to 0 in the tinyG config by default.. I now have it at 1 and will try again..
I assume that its ok to have $ex=1 when using TgFx or the android app??? right?- This reply was modified 11 years, 4 months ago by mcgyvr.
mcgyvrMember$ex was set to “0” would that matter? (and I thought default was supposed to be “1”) I didn’t change it to 0..Does TgFx use it? or change it?
I’m running the “tgfx_windows-x64 v.95 – build 2012.exe” from that dropbox download site you have set up. I’ve never had a netbeans before? are they tasty? (bad attempt at humor)
Coolterm + XON worked prefectly..well so far (I ran the gcode 10 times anyways with no issues at all)
Gcode is here. (note my machine is X/Y axis only with a pneumatic cylinder being triggered with the coolant pin)
https://dl.dropboxusercontent.com/u/9726802/TinyG/7000300005C_A_test.GCODE
TinyG Config is here.
https://dl.dropboxusercontent.com/u/9726802/TinyG/TinyG%20Configuration.txt
This is what comes back when I reset the board with coolterm connected.
https://dl.dropboxusercontent.com/u/9726802/TinyG/coolterm_reset.txtAnd my main goal is to run TinyG with Matt Stocks Android app and I have been talking with him and having him send apk files for testing,etc.. and I had similar problems with that and thought I’d give TgFx a try to see if the problems were with the app or TinyG and it “seems” that since the issues are showing up with both the app and TgFx that something about TinyG is to blame… Just trying to help as much as a non-java programmer can.
mcgyvrMemberI don’t use linux but.. on my winblows machine I just loaded the exe and get an icon to run the program..
mcgyvrMemberThere is a 32 bit linux build of TgFX..
https://www.dropbox.com/sh/huiupgemipv8f4q/X2l_1EH-gxmcgyvrMemberI’ve been having LOTS of crashes with the app. Mostly when connecting/disconnecting. I have been submitting a few of the crash reports.
(java unhandled exceptions or something like that)Also the G28.2 button doesn’t seem to work (I only have X and Y axis on my machine) and when I press the button it shows a “seek” in the mode section but nothing happens and the app becomes totally unresponsive to the point I have to restart my tablet for it to come back. Even stopping the app and clearing the cache doesn’t bring it back.
But issuing a G28.2 x0y0 in the gcode command line works just fine.As for bugs/additions #1, #4 and #5 from above would be a priority for me.
Heck for #1 just having a Start AND a Stop button would be fine if you can’t workout the bug.. And of course addressing the connect/disconnect issues.I’d also LOVE to have the ability to “lock out” the machine/axis tabs to prevent “users” from messing with the settings. (an admin password to allow/activate those tabs would be perfect or something like that)
This is being used as an assembly machine in a factory and I just think some level of “security” is a good idea.I basically want my “users” to walk up to the machine, plug in their tablet to the USB, start the app, home the machine, load a file and run it..over and over again.
mcgyvrMemberwell I rewired all homing/limit switches with shielded 1 pair cables and grounded the shield at one end and now it seems to be working perfectly.
I’m really surprised that noise effected NC switches but it did.. Half the time it would “think” it was homed and stop before it even got to the switches, the other half of the time it would work just fine.. But now with the shielded cables its working perfectly all the time (so far.. homed it 25 times so far).Oh and it sure would be REALLY..REALLY nice going forward if you either switch those pin headers to screw down terminals or at least included a proper mating connector and a handful of crimp contacts with each board. (even just a few 2 position connectors with a 6″ pigtail of wire on each would be fine) It was a pain to scrounge my junk draw till I found some mating connectors that would fit then splice into them.
-
AuthorPosts