New experimental firmware 1.19beta11
-
FYI
Not sure if it matters for this purpose. The end stops I care about for W and A are 5 and 6. (X-0,Y-1, Z-2, U-3, V-4, W-5,A-6)
They're not triggered when carriage is not near them, however end stops 7,8,9 say they are. They're not configured as anything – those steppers will be extruders, so not sure if those end stops should be configured as something? -
The other bit, let me know if this is not right – I'm using 2 wires with my end stops -- I can add a third if it's necessary. 2 wires work with other axis. I don't have 3.3v connected -- I leave it unconnected, I assume you have a pullup. Let me know if that's not right in this case.
-
You are using normally open/active low endstops (S0 in M574). This is opposite to what most people use, i.e. normally closed switches, which is what I tested with. It shouldn't matter, but might possibly be relevant.
-
I've started using them this was since IMO it's safer as the switch is only job is to sync the signal line to ground, and carries no current, so if there is a pullup on the board, then the switch does it's job and all is fine. Since I'm not done wiring yet, I normally clip off the third pin on the switch – the NC pin. And I just connect 2 wires. With other applications with Marlin and Smoothieware, I just enable the pullup, I am assuming you have a pullup there already. If you think flipping it around to active high with 3 wires will get it working. I can try that when I get home (I thought I was done wiring, but adding a third wire to the end stops would be fairly easy.
-
NC endstop switches are normally used so that if a wire falls off, the carriage stops immediately when homing, instead of the motor trying to drive it through the endstop.
-
NC endstop switches are normally used so that if a wire falls off, the carriage stops immediately when homing, instead of the motor trying to drive it through the endstop.
Yep, I was thinking it over – I'll rewire my switches when I get home I keep 2 wires and just swap the NO to NC. (I'll have to replace the switches since I clipped off the NC end because I was trying to make everything neat) -- I had my wife check and I have at least 6 limit switches left over -- so I have enough to replace all 6 to use NC active high, I haven't setup Z homing yet -- I am considering using 2 Z limit switches and add a second driver for my 2 Z steppers. No sure if it's worth the effort.
My thinking was to do active low and only have the end stop trigger on homing. I go over the top and use silicone wire with heat shrink everywhere -- so my wires are as secure as it gets, but it makes sense to use NC to keep the machine from going crazy if there is an issue.
-
Installed today no issues with installation. Running 1.19 beta 11 and wifiserver beta 10. Only tried one print but got this error message:
Error: Failed to load height map from file heightmap.csv: bad header line or wrong version header
Should I have read the release notes?
-
I forgot to mention in the release notes that the height map format has changed, to a!low different X and Y spacings. You will need to run G29 again to generate a new height map file.
-
Okay thanks David.
-
Configuring end stops with S1 and using NC limit switches worked. Pretty cool homing now with all the extruders moving to their corners. Took some time to redo the end stops, but full steam ahead now.
-
I'll re-test using NO endstops, because that should have worked too.
-
While to a G29 I had to make an emergency stop because the sensor came out of bed. When you do the emergency stop, the firmware restarts. The only way to reprint was to remove tension from the duet. It seems that the firmware reset after the emergency stop does not work and you have to remove the voltage to remove the lock
-
Strange, I just tested emergency stop and software reset, and they did turn all the motors off.
-
Strange, I just tested emergency stop and software reset, and they did turn all the motors off.
The problem is when reestablishing. I have to remove tension to have control. If I reload the web page it returns and it appears that the firmware is being reset. The weird thing is that until the beta10 + 2 version did not happen to me
-
Just tried to upgrade to 1.19beta11 from 1.19beta8 on my Duet Ethernet through the web interface and I still get the same issues that I had trying to upgrade to beta 9. That is to say that after apparently updating, DWC reports that it could not connect. I have to cycle the power and when I do so, DWC reports that I'm still on 1.19 beta8.
-
Just tried to upgrade to 1.19beta11 from 1.19beta8 on my Duet Ethernet through the web interface and I still get the same issues that I had trying to upgrade to beta 9. That is to say that after apparently updating, DWC reports that it could not connect. I have to cycle the power and when I do so, DWC reports that I'm still on 1.19 beta8.
I think it is from 1.19beta10 + 2 when you let update from the web with no problems. I was able to upgrade to beta11
-
That's right, it's a bug in some versions of DuetEthernetFirmware 1.19beta, fixed in 10+2. If you are affected then you will need to update to a new version using SAM-BA or Bossa/bossac 1.8. Although it's possible that if you enable FTP on your old system and transfer a file, you might be able to update over the web after that.
-
BTW. I started seeing:
"Can't open 0:/sys/oem.json to read, error code 4" since I updated everything to latest and greatest. Is this an issue with DWC update? -
I guess you must have a USB host connected. You can ignore that message; but I'll discuss with chrishamm how we can get rid of it.
-
I guess you must have a USB host connected. You can ignore that message; but I'll discuss with chrishamm how we can get rid of it.
I don't – I did at one point when I was trying to debug liner motion. I actually couldn't leave USB connected as the board really didn't like it with ethernet. Basically the ethernet connection would drop out and I only had USB connectivity. I ordered a USB isolator as I want to have that option open to me. I am assuming ethernet issue was caused by a ground loop of some sort.
Y
