Software bundle 3.4.0 beta 4 released
-
@dc42
Super excited about the forward feed heater parameters.Can you go into a little more detail on how the forward feed parameters were (/are) derived?
As much as I hate fluid dynamics, I can totally see some dimensionless numbers involved here. There will be a direct correlation to the following:
Thermal capacity of the filament
Temperature rise of the filament
Velocity of the filament extrusion
Thermal capacity of the heater block
Temperature latency to the Fremont sensorAs the last two are captured during Hester tuning, the extrusion velocity is known in the look-ahead and temperature rise can be assumed delta to ambient, it comes down to thermal capacity of the filament and a measure of reaction times / overshoot. Overshoot could probably be captured from the tuning.
Care to share a few more details? This is rather important for the items I work with.
-
@oliof said in Software bundle 3.4.0 beta 4 released:
Question about the new M309 command: What are expected/useful/sane values for the heater PWM coefficient?
In my tests with 1.75mm PETG I found 0.01 was about the right amount. 2.85mm filament will need more than 1.75mm, and high power heaters will need less than low power heaters. The units are PWM fraction per mm/sec extrusion rate.
The test I did was to monitor the temperature for 20 seconds, then extrude filament at close to the maximum rate for 20 seconds, then stop and monitor the temperature for another 20 seconds. Without feedforward there is a temperature drop shortly after extrusion starts, which recovers after several seconds. Then when extrusion stops there is a temperature rise, which again reverts to normal temperature after a few seconds.
This feature is intended for high flow rate hot ends. It's not needed on regular hot ends with a 0.4mm or similar size nozzle where the temperature drop caused by extrusion is less than 1C.
-
@jens55 if running in SBC mode, you need to be on the unstable branch and use
sudo apt update & sudo apt upgrade
-
@jay_s_uk, I run in stand-alone mode. Is DSF only for SBC mode ?
-
@jens55 yes. Stands for DuetSoftwareFramework and is only ran on an SBC
-
@jay_s_uk, thanks ! (I was aware it stands for DuetSoftwareFramework but that didn't clue me in to the fact that it is only for the SBC mode)
-
@dc42 You could add in the DWC changelog that the new gcode viwer 2.0 from @Sindarius is included!
-
@pcr Thanks, I forgot to add that note - now updated.
-
@dc42
Can you expand a bit on the expected usage for new datetime() function?
A little experimentation suggests that the datum is 1970-01-01T00:00:00, but I see no way of determining the number of seconds since then.
If I calculate it externally, I can confirm it converts back to a DateTime (see below) and doing math based on the time isn't so hard using variables, but I'm at a loss as to the use of the integer value part of it. -
@owend
datetime()
can be used to convert a string back to a datetime value, like e.g. state.time. Something like this gives you the difference between two times in seconds:echo state.time - datetime("2021-09-28T12:11:16")
-
I've been testing ramping up my print speed over the weekend, noticed some extruder temp drops, and thought feed forward would be good for that, as if by magic it's appeared!
Will have a play at the weekend, great work @dc42, as always!
@dc42 said in Software bundle 3.4.0 beta 4 released:
@oliof said in Software bundle 3.4.0 beta 4 released:
Question about the new M309 command: What are expected/useful/sane values for the heater PWM coefficient?
In my tests with 1.75mm PETG I found 0.01 was about the right amount. 2.85mm filament will need more than 1.75mm, and high power heaters will need less than low power heaters. The units are PWM fraction per mm/sec extrusion rate.
The test I did was to monitor the temperature for 20 seconds, then extrude filament at close to the maximum rate for 20 seconds, then stop and monitor the temperature for another 20 seconds. Without feedforward there is a temperature drop shortly after extrusion starts, which recovers after several seconds. Then when extrusion stops there is a temperature rise, which again reverts to normal temperature after a few seconds.
This feature is intended for high flow rate hot ends. It's not needed on regular hot ends with a 0.4mm or similar size nozzle where the temperature drop caused by extrusion is less than 1C.
-
@chrishamm
Thanks!
Simple when you put it like that.
I'll give myself an uppercut -
@owend the main purpose of the datetime function is that it allows you to make datetime values persistent by constructing a command to reload a previously saved datetime value into a variable.
The datum is currently 1 Jan 1970 but we will likely change it because year 2038 is looming.
-
@dc42 - our chamber heaters are still faulting when heating up the first time with the following error. We have verified this is does not happen when rolling back to 3.3.
Error: Heater 2 fault: temperature rising much more slowly than the expected 2.4°C/sec
-
@oozebot 2.4C/sec sounds very high for a chamber heater. I suspect you haven't tuned it or set manual M307 parameters.
-
@dc42 I started another thread about this. No, we have never been able to get tuning to complete for our chamber heaters, and our M307 is not optimized, but it worked previously. If you have a minute and could weigh in on the other thread with some mock tuning suggestions, we'd be all ears! Thanks
https://forum.duet3d.com/topic/25099/6hc-3-4b3-chamber-heater-faults
-
@oozebot what M307 parameters are you using for your chamber heater?
You may need to reduce the R parameter and/or increase the D parameter.
-
@dc42 Thanks - whatever was fixed with heater faults in 3.4 exposed the issue and we've suspected it was a configuration issue. After a quick review, I think the missing R value is the culprit. It has been added and we are now performing additional testing.
-
@dc42 - we are still consistently getting chamber heater faults with 3.4 beta when heating from room temp. I've lowered R to 0.1, jacked the C value up to 30 minutes, set D to the maximum of 350 and it still happens!
Error: Heater 2 fault: temperature rising much more slowly than the expected 0.1°C/sec
I think what's happening is the new fault protection that was added in 3.4 is not allowing the temp to swing negative - which can and does happen when the chamber initially heats.
Are there any other M307 variations you'd like us to test? We really need to get this resolved before the release of 3.4 stable. Thanks
config.g
;== Enclosure Heater ========================= M308 S2 P"temp1" Y"thermistor" T100000 B3950 M950 H2 C"0.out2" T2 M307 H2 B1 C1800 D350 R0.1 M141 H2
Edit - we are still working on this to see if it's our settings. This just happened - we have R set to .5 but got a fault that it was rising slower than .2c?
9/29/2021, 11:37:03 AM Error: Heater 2 fault: temperature rising much more slowly than the expected 0.2°C/sec 9/29/2021, 11:34:44 AM m307 h2 Heater 2 model: heating rate 0.500, cooling time constant 140.0, dead time 5.50, max PWM 1.00, calibration voltage 0.0, mode bang-bang
-
@oozebot your M307 D parameter looks far to low for a chamber heater to me. Try 20 or 30.