#garfield Logs

Feb 11 2024

12:00 AM aandrew: rue_mohr: I wonder if you're sampling fast enough to see the velocity jitter caused by brushes skipping over the commutator
12:13 AM rue_mohr: I'm catching every line just fine
12:13 AM aandrew: not what I mean
12:13 AM rue_mohr: its definitly not missing any of its targets
12:13 AM rue_mohr: oh I count time between lines
12:13 AM rue_mohr: and I just use the inverted units for the velocity
12:13 AM rue_mohr: I do a cutoff and snap to 60000 if the timer rolls over
12:13 AM aandrew: yes, but if there aren't enough slots you won't have the resolution
12:13 AM rue_mohr: I only need to know velocity as the time between encoder lines
12:13 AM rue_mohr: ?
12:13 AM aandrew: yes I understand that, I am just wondering if the velocity changes line up with how many commutator positions there are
04:52 PM polprog: https://polprog.net/rozne1/ircjunk/rack3u/3railPSU_rev2_wip.pdf
04:52 PM polprog: thoughts?
04:56 PM rue_mohr: what you up to?
04:57 PM rue_mohr: your from/to are backwards
04:58 PM rue_mohr: you have 12 in and 10 out
04:58 PM rue_mohr: 8 to 10V from 12V
04:59 PM rue_mohr: -8 to -11V from 12V
04:59 PM rue_mohr: -24 to -27V from 12V
04:59 PM rue_mohr: take the 5V from the 8V rail...
04:59 PM rue_mohr: less dissipation
05:00 PM rue_mohr: er the 10V
05:01 PM rue_mohr: never seen the feedback tapped between the filters like that, but I think I know why
05:02 PM rue_mohr: if you want to get fancy, you might be able to pull all the voltages from a single transformer that generates the -24V
05:04 PM rue_mohr: actaully I think the flyback from the -8V generates the +8...
05:04 PM rue_mohr: maybe that needs a second coil... yea, cause only one polarity of the pulse is regulated
05:04 PM rue_mohr: but its still an 8V delta
05:05 PM rue_mohr: Rue, your talking to a wall
05:05 PM rue_mohr: yea, I kinda just realized
05:05 PM rue_mohr: EGMK?
05:05 PM rue_mohr: ok, I'm taking the music with me tho
05:06 PM rue_mohr: I want to crank it, think he'll take the oncall on time?
05:06 PM rue_mohr: watch for the message is about all we can do
05:34 PM polprog: its a power supply for the microwave generators
05:34 PM polprog: 8, -8, -24 are analog voltages
05:34 PM polprog: i dont want to take from them for the atmega
05:34 PM polprog: -24 needs to be very clean, it goes to the VCO
05:35 PM polprog: "+8 from +10" means that i will be using the linear reg (didnt draw yet) to take the 10V from the DC-DC (dirty) to make a clean 8V
05:35 PM polprog: and so on
05:36 PM polprog: hm i could take the 5V from the 10V rail, there will be a lin reg after it so should not hurt it
05:36 PM polprog: thats a good diea
05:36 PM polprog: idea*
05:36 PM polprog: i was thinking about using some smart transformer setup to use 1 DC-DC to make +8 and -8 at the same time
05:36 PM polprog: er
05:36 PM polprog: +10 and -10
05:38 PM polprog: thansk rue_mohr!
05:38 PM polprog: did you ever experiment with using ethernet transformers to make symmetric power supplies?
05:38 PM polprog: gotta sleep now
06:35 PM rue_shop3: ah
06:35 PM rue_shop3: polprog, no, but a 1:1 should be able to do it
06:35 PM rue_shop3: in forward mode....
10:28 PM rue_mohr: https://github.com/ruenahcmohr/EGMK
10:28 PM rue_mohr: ^^ peer review?