#robotics Logs

Jan 31 2018

12:24 AM rue_: so, fixed temp control of the reflow oven is working, it overshoots by about 10c
12:27 AM rue_: should be able to do better
12:34 AM mrdata: mhm
12:34 AM mrdata: some PID adjustment should eliminate overshoot, methinks
12:44 AM rue_: not good at tuning pid
12:44 AM rue_: esp when the timer constant is in the order of 8 seconds
12:44 AM rue_: time constant
12:44 AM rue_: the pwm is 1Hz
12:50 AM mrdata: 1 Hz period?
12:50 AM rue_: yup
12:50 AM mrdata: okay
12:50 AM rue_: heater, ya know.
12:51 AM mrdata: mhm
12:51 AM rue_: it takes 16 seconds for my code to go from 0% to 100%, which is too fast
12:52 AM mrdata: and within that period, what is smallest increment of on-time
12:52 AM rue_: 1/127
12:52 AM rue_: 7bit, 1Hz pwm
12:52 AM mrdata: ok
12:52 AM mrdata: what is the desired time to go from 0 to 100%
12:53 AM rue_: not sure, it takes 8 seconds for the thermal peak after the elements have been turned off
12:55 AM mrdata: so if you did 100% on for 8 seconds, then off for the remainder, what would be the peak
12:56 AM rue_: about 10c over the temp when the elements were turned off (between 40c and 130c)
12:56 AM rue_: at 40c it might be 12c over, at 130c it might be 9c over
12:59 AM rue_: I'm not sure what the temp gain would be over the 8 seconds
12:59 AM mrdata: but if you then do 50% duty for each of those 16 seconds, you'd get about the same temperature, because you added the same heat in the same time
01:00 AM rue_: but it would be 12Kj
01:00 AM rue_: kJ
01:00 AM rue_: hmm, so you think I should use the error to bias a 50% pwm cycle?
01:01 AM rue_: I could also work out the energy loss rate
01:01 AM rue_: put 12kJ in and see how long it all takes to leak off
01:01 AM rue_: tho I suppose its not linear
01:01 AM rue_: but I could chart it
01:02 AM mrdata: higher temperature difference means higher heat loss
01:02 AM mrdata: sure
01:02 AM rue_: there is a working usb port with serial
01:02 AM rue_: currently, all it does it say "hello" when you boot the controller
01:03 AM rue_: I'll do that, it sounds handy
01:03 AM rue_: I might be abel to makea less compansation chart
01:03 AM rue_: loss
01:04 AM rue_: hmm, and if I make the target settable via usb, I can have PC side profiling software
01:04 AM mrdata: it would supply good info to tune the controller, yeah
01:05 AM rue_: its a direction, I'll take it
01:05 AM rue_: gnight
01:09 AM mrdata: nn
03:44 AM OpenSorc_ is now known as OpenSorceress
05:48 PM veverak: https://www.aliexpress.com/item/10pcs-Digispark-kickstarter-development-board-ATTINY85-module-for-Arduino-usb/32580777597.html
05:48 PM veverak: this looks interesting
08:13 PM rue_: hahaha $21!!!!!
08:14 PM rue_: oh 10pcs!
08:14 PM rue_: ok, better
08:43 PM rue_: this is not a good chair to doze off in, you wake up with one hell of a stifff
08:43 PM rue_: neck
08:43 PM Tom_L: so why do it?
08:45 PM Tom_L: was gonna work on truing the column a tiny bit more but decided to take the night off
08:45 PM Tom_L: rue_
08:46 PM Tom_L: will voltage or current increase stepper speed?
08:46 PM Tom_L: or both or neither?
09:27 PM Tom_L: rue_ you should really wake up now
09:28 PM Tom_L: as a sparkie, is there a standard way to orient a fwd/rev switch?
09:28 PM Tom_L: as far as codes go
09:29 PM rue_: forward is up
09:30 PM rue_: and yes
09:30 PM Tom_L: what about left and right?
09:30 PM Tom_L: on a rotary switch
09:31 PM Tom_L: for fwd/stop/rev
11:20 PM rue_bed: forward on the left...