#garfield Logs

Jan 02 2021

01:27 PM penloy: pog
01:34 PM Tom_L: peg
05:17 PM polprog: thermite track welding: https://www.youtube.com/watch?v=035DOWfo2Xs
05:18 PM polprog: looks like the 70s
06:08 PM rue_mohr: I worked on the math for the servo till 6am, and woke up at 3pm
06:38 PM rue_mohr: polprog, did you take a look at any of that bode plot stuff ?
06:38 PM polprog: no, i was doing school stuff
06:38 PM rue_mohr: oh :(
06:39 PM rue_mohr: looks like I'm alone on this one
06:39 PM polprog: no ;)
06:39 PM rue_mohr: but I made some progress
06:39 PM rue_mohr: there isn't a missing zero, its just got a pole and some overshoot cause its underdamped
06:40 PM polprog: so where is the pole?
06:40 PM rue_mohr: about 1.8Hz
06:40 PM rue_mohr: the bandwidth is about 2.7Hz
06:40 PM rue_mohr: damping ratio is about 0.5
06:41 PM rue_mohr: but if I look at it like a series RLC, I can see how to compensate it out
06:41 PM polprog: ok
06:41 PM rue_mohr: but I dont know any more
06:42 PM rue_mohr: and I'm not sure if the start of the overshoot technically counts as a zero
06:42 PM rue_mohr: and I cant add a zero a cancel out the pole, filters can only add poles?
06:43 PM polprog: hmm
06:43 PM polprog: i have to read up
06:43 PM polprog: my internet is spotty today
06:44 PM rue_mohr: I dont know how to get from the frequency domain chart I have to an S plane chart
06:44 PM rue_mohr: and almost everything I'm reading has been taken to s plane
06:46 PM rue_mohr: I have the type of transfer function for it
06:46 PM rue_mohr: I may have to sit down with just an RLC network and a filter, see if I can understand how things interact
06:46 PM polprog: http://users.wpi.edu/~mcneill/handouts/s-plane_handout.pdf
06:48 PM polprog: https://www.maximintegrated.com/en/design/technical-documents/tutorials/7/733.html
06:48 PM polprog: these look good
06:48 PM rue_mohr: its funny, cause a lot of docs expect you to engineer things from the ground up
06:48 PM Tom_L: ever created a .deb file?
06:48 PM rue_mohr: but I already hav a system and I need to adjust it
06:48 PM rue_mohr: Tom_L, you created 3!?
06:48 PM Tom_L: i did?
06:49 PM Tom_L: well those don't count
06:49 PM rue_mohr: yea, you ask me every time and I dont know
06:49 PM Tom_L: like for a python app
06:49 PM Tom_L: and i forget every time i ask so i'll keep asking :)
06:49 PM rue_mohr: you like 300% more versed than I am
06:49 PM rue_mohr: I thought you kep a notebook?
06:49 PM Tom_L: i dunno where i put it
06:49 PM Tom_L: on one of the ssd here but which one?
06:50 PM rue_mohr: find a notebok for documenting this, and I bet it'll be the same one
06:50 PM Tom_L: carry on.
06:51 PM rue_mohr: really, everything was written twice, but one with jw instead of s ?
06:52 PM polprog: i think so
06:52 PM polprog: i dont know laplace transform
06:55 PM rue_mohr: oh maybe you cant add a pole without adding a zero
06:58 PM rue_mohr: oh, the throretical integrator has no pole because its voltage source is unlimited
06:59 PM Tom_L: gawd i hate when noobs become experts overnite
06:59 PM rue_mohr: subtle difference between an integrator and a low pass filter
07:00 PM Tom_L: if only i were an op
07:00 PM polprog: :)
07:00 PM Tom_L: hey i recognize that jargon! i use lowpass on my mpg
07:00 PM rue_mohr: on your stereo too
07:01 PM rue_mohr: called bass
07:01 PM Tom_L: yeah i got one in a box in the garage
07:01 PM Tom_L: smells like it's been hot
07:01 PM Tom_L: wouldn't know why
07:08 PM rue_mohr: polprog, I like the 3d graphs here, but I'm not sure its helping much
07:08 PM polprog: i dont really get it either
07:08 PM polprog: i think we should learn laplace transform
07:09 PM rue_mohr: it would seem to imply that what I want to do is pull the highs up by adding another pole, and then adding a zero to push down the overshoot a bit
07:09 PM rue_mohr: you want a cross section of the tent thats quite level till it falls off
07:09 PM rue_mohr: I suppose the application limit is the supply available
07:10 PM rue_mohr: because the output of the motor driver saturates, I can only add so much gain till its useless
07:10 PM rue_mohr: because the driver cant apply it
07:11 PM rue_mohr: so I would like a filter with 1 pole
07:11 PM rue_mohr: and then I place that pole after the drop-off of my systems natural pole
07:14 PM rue_mohr: but their chart for a low-pass shows that by varrying soemthing, you can get a peak in the freq response
07:14 PM rue_mohr: but their chart also shows ...
07:15 PM rue_mohr: oh, you cant utilize the negative freqs
07:15 PM rue_mohr: ok
07:15 PM rue_mohr: so..
07:17 PM rue_mohr: something is off about this maxim page
07:50 PM rue_mohr: but I kinda already know I need to add a high pass fitler
07:51 PM rue_mohr: its funny to think that they are using integrators that need antiwindup instead of a high pass that cant go insane in the first place
07:51 PM rue_mohr: k, so...
07:51 PM rue_mohr: I have two issues, adding a high pas to pick up the responce
07:52 PM rue_mohr: and adding damping to keep the response at/under 0db
07:52 PM rue_mohr: https://www.maximintegrated.com/content/dam/images/design/tech-docs/733/733Fig04b.gif
07:52 PM rue_mohr: this says a bit tho
07:53 PM rue_mohr: if I consider the first pole to be the pole I have now, then I can look at the second pole as the one I want to add
07:53 PM rue_mohr: so, the question is where do you put the high pass
07:54 PM rue_mohr: thanks Rue
07:54 PM rue_mohr: np
07:54 PM rue_mohr: its based on driver saturation
07:55 PM rue_mohr: so we dont want to push the driver over +-255 after the high pass
07:55 PM rue_mohr: and that should bring back where we put the pole
07:55 PM rue_mohr: right, cause the amplitude would have to be too high to make the correction
08:13 PM Tom_L: uh oh
08:13 PM Tom_L: rue was left alone to play
08:17 PM rue_mohr: }:]
08:56 PM aandrew_ is now known as aandrew