#robotics Logs

May 11 2019

01:43 AM Jak_o_Shadows: afternoon
01:48 AM rue_bed: my filter analysis tool is becoming quite a project
01:49 AM Jak_o_Shadows: I have lots of the stuff with holes and pads
01:49 AM Jak_o_Shadows: but no strips
01:49 AM Jak_o_Shadows: which is nice sometimes, but also annoying
01:49 AM rue_bed: yup
01:50 AM rue_bed: it looks like me talking to myself had been adding quite a base-level to the channel
01:51 AM rue_bed: 57 of you should contribute more
01:51 AM Jak_o_Shadows: For sure
01:52 AM rue_bed: I'm annoyed tho, the circuit is going to require a +-15V supply for sure
01:53 AM Jak_o_Shadows: How much power? Just shove an ebay module on
01:54 AM rue_bed: do they make dual output ones?
01:54 AM rue_bed: I dont think a max232 will do it, or I'd just use that
01:55 AM Jak_o_Shadows: Just get two modules?
01:55 AM Jak_o_Shadows: up/down converters are like 70+% efficient
01:57 AM rue_bed: or, just power it from an AC transformer anyhow
01:58 AM rue_bed: it wouldn't be an unreasonable thing
01:58 AM Jak_o_Shadows: Yeah, i was assuming plug in pack
01:58 AM rue_bed: I suppose I need to choose a shape for this project
01:58 AM Jak_o_Shadows: My comment about efficiency was more about heat
01:59 AM rue_bed: totally reasonable, but I'v got a wide one on this
01:59 AM rue_bed: its for analizing filters
01:59 AM rue_bed: so, I'lll output a sine signal, and you give it your filters output
02:00 AM rue_bed: it works out the gain/phase per freq
02:00 AM rue_bed: sweeps it
02:00 AM rue_bed: I suppose two BNC
02:00 AM rue_bed: USB for comm.
02:01 AM rue_bed: and a IEC for power
02:01 AM rue_bed: or just an AC adapter
02:01 AM rue_bed: w/AC
02:01 AM Jak_o_Shadows: What sorta frequency range is this good for?
02:01 AM rue_bed: the goal is 2Hz to 20kHz
02:01 AM rue_bed: 1Hz to 40kHz would be nice
02:01 AM rue_bed: something like that
02:02 AM rue_bed: I'd like to use it on servos to audio filters
02:02 AM Jak_o_Shadows: ayup, cool
02:02 AM rue_bed: it stem'd from looking at PID as high/low pass filters
02:03 AM rue_bed: which puts PID in a new light
02:03 AM rue_bed: that of an EQ for system compensation
02:04 AM rue_bed: if I can sweep a servo, I can know what the compensation should look like
02:04 AM rue_bed: 1Hz is a bit high, hard to go to the ~0.2Hz that would be more reasonable
02:04 AM rue_bed: whatever
02:05 AM rue_bed: I want to play with digital filters too
02:05 AM rue_bed: so, being able to compare an analog w/ digital would be nice
02:06 AM rue_bed: for checking math
02:06 AM Jak_o_Shadows: Yeah
06:05 AM Tom_itx is now known as Tom_L
10:48 AM rue_mohr: Tom_L,
10:48 AM rue_mohr: so far my code looks kinda like this:
10:48 AM rue_mohr: set the t1 divider to /256 and time an input cycle
10:49 AM rue_mohr: if the count is less than 8, go use the cycle-counter-for-250ms and remeasure
10:49 AM rue_mohr: if the count is less than 256, switch to /1 on t1 and remeasure
10:49 AM rue_mohr: if the count is less than 2048, switch to /8 on t1 and remeasure
10:50 AM rue_mohr: if the count is less than 16384, switch to /64 on t1 and remeasure
10:50 AM rue_mohr: otherwise go with what we have
10:50 AM rue_mohr: if using the period measurement, f = 16Mhz/(clockdiv*count)
10:51 AM Tom_L: what �C?
10:51 AM rue_mohr: if using cycle count f = count *4
10:51 AM rue_mohr: �C ???
10:51 AM Tom_L: arm?
10:51 AM rue_mohr: no, M88
10:52 AM rue_mohr: tho, as far as counter resolution you might have a point
10:52 AM rue_mohr: hmm, but 3.3V and ugh
10:53 AM rue_mohr: 5V is enough of a restriction for reading the oscillator and filter return
10:54 AM rue_mohr: I may need to adjust some of those measurement switching points tho
10:54 AM Tom_L: somebody wrote a frequency counter for the saleae you could swap out the firmware
10:55 AM rue_mohr: the uC will be measuring freq, measureing phase offset, and 4 voltages that become the Av of the filter
10:55 AM rue_mohr: and resetting the peak detectors and controlling a sine wave osc
10:56 AM rue_mohr: I'm showing 6 seconds irc lag
10:57 AM rue_mohr: and its 0.1... huh
10:57 AM rue_mohr: anyhow, I can make it take wild input for now
10:57 AM rue_mohr: so I'm starting with that
10:57 AM rue_mohr: I'm still having a time designing an adjustable sine wave osc
09:22 PM rue_mohr: --
09:23 PM ace4016: ++
09:24 PM rue_mohr: trying to spend some time outside, all my projects are in door stuff
09:24 PM rue_mohr: well, there is some grinding i can do