#garfield Logs

Jun 25 2021

12:07 AM rue_mohr: hey kicad might be able to generate outputs for my plotter
11:43 AM aandrew: I have a xenon flash that I trigger just to test with for now - I see a VERY fast rise which is great, but the tail is like 1000x longer than what the xenon flash datasheet says it provides
11:43 AM aandrew: now the photodiode has 140pF of capacitance and since the output of the TIA is so fast on the rising edge, I don't think that THAT is actually the issue
11:43 AM aandrew: I think it's the photodiode taking forever for the energy from the light to drain out
11:44 AM polprog: GBW means gain*bandwidth = 15MHx
11:44 AM polprog: MHz
11:45 AM polprog: so for gain 10 you have 1.5MHz, for 100 you get 150kHz and so on
11:46 AM polprog: so in your case you have 15MHz / 500,000 which is roughly about
11:46 AM polprog: 30Hz.. not a lot
11:46 AM polprog: photodiodes are usually used with transimpedance amps
11:47 AM polprog: which is like a regular inverting amplifier but the photodiode is used instead of the first resistor
11:47 AM aandrew: right
11:47 AM aandrew: this is a TIA
11:48 AM polprog: strange to need this much gain though
11:49 AM aandrew: well no, it'll be measuring very low amoutns of light eventually
11:49 AM aandrew: so PD -> TIA -> level shift+minor amplify -> ADC
11:49 AM aandrew: I'm trying to do all the gain on the TIA to keep the high impedance, noisy stuff very contained and probably under a metal can
11:50 AM aandrew: but I can possily just give the TIA a gain of 10 or 100 (not k) and then it'll be uV signal going ot the second amp which can gain it a lot more
11:50 AM aandrew: and since the uV signal will be coming from a decently low impedance source (the TIA) then it should be a lot "safer" from impedance
11:50 AM aandrew: er interference
11:52 AM aandrew: I've got a guard voltage of a hundred mV or so driven from an opamp around the TIA input/feedback
12:18 PM aandrew: polprog: I understand the concept of GBW, but my question is why would I able to see such a fast signal change if my gain is theoretically like 26Hz?
12:20 PM aandrew: https://ibb.co/p1x87Mk is the ADC output
12:20 PM aandrew: horizontal is sample number (sample rate = 32kHz, so 30.125usec between samples
12:20 PM aandrew: vertical is voltage
12:20 PM aandrew: that's not a small signal change, yet I went from -0.4 to 2.5V in like 60usec
01:01 PM Tom_itx is now known as Tom_L
01:56 PM polprog: i dont think you can catch a signal that fast with this GBW
01:56 PM polprog: :(
03:11 PM aandrew: polprog: I should not be able to catch that rising edge that fast though should I?
03:12 PM aandrew: that's 2V in 60us, that's only .3V/us
03:15 PM aandrew: let me verify gain, one second
03:26 PM aandrew: nope it's worse
03:26 PM aandrew: gain resistor is 1M in parallel with 220pF, removing that cap too but I don't think that's the issue
03:26 PM aandrew: that means my bandwidth is like 13Hz
07:44 PM rue_bed: noise amplifier
08:00 PM Tom_L: finally got a big piece to the puzzle
08:02 PM Tom_L: now i can get the mill alot closer to where it needs to be
08:05 PM Tom_L: https://i.ebayimg.com/images/g/INYAAOSw9rVgCT3S/s-l300.jpg
08:35 PM aandrew: hm interesting. screwing around in ltspice (which is probably lying through its teeth becaue I don't know how to properly run spice), I should be able to get reasonable settling time both on charge and discharge and it's my filtering capacitor which is the likely culprit
08:37 PM aandrew: that means I *really* don't understand GBW
09:26 PM aandrew: polprog / rue_bed https://ibb.co/thvvMXh
09:26 PM aandrew: that is PLENTY of bandwidth for my needs, and using the photodiode equivalent with 140pF of parasitic capacitance
11:02 PM rue_mohr: hi
11:02 PM rue_mohr: Tom_L, :)
11:02 PM rue_mohr: I made one for the lathe, but it doens't have a web
11:03 PM rue_mohr: aandrew, yea, sorry I'm alive now
11:04 PM Tom_L: it needs to be trammed again
11:04 PM rue_mohr: ok, I'm a bit worried about the 1Gohm resistor
11:04 PM Tom_L: but making adjustments wasn't well thought out in the design
11:04 PM rue_mohr: and I'm worried about the 590 ohm in parallel with the 0 ohm
11:05 PM rue_mohr: Tom_L, which way?
11:05 PM Tom_L: i need to do it all
11:05 PM Tom_L: i didn't have a good way to check everything before
11:06 PM Tom_L: the column isn't perfectly square with X for one
11:06 PM rue_mohr: k
11:06 PM Tom_L: it's square with the vise because i adjusted the vise
11:09 PM Tom_L: i gotta figure out a way to make the mounting holes bigger on the column mount
11:10 PM Tom_L: so i can adjust it
11:10 PM rue_mohr: aandrew, whats the photodiode for?
11:13 PM Tom_L: i need an idea to make a 1/2" hole bigger with hand tools
11:13 PM Tom_L: i've got a burr tool but kinda hate to use it
11:13 PM Tom_L: i'm afraid i may have to
11:17 PM rue_shop3: whats the hole in?
11:17 PM Tom_L: 6 holes in the Z column mount
11:17 PM rue_shop3: ah
11:17 PM Tom_L: i need to turn it so Y square with X
11:18 PM rue_shop3: aah
11:18 PM rue_shop3: I suggest tipping a drillbit for the first pass
11:18 PM Tom_L: it's not off much but it's a real pita to do
11:18 PM rue_shop3: then paralleling it up with the burr
11:18 PM Tom_L: no bit that will work in my hand drill
11:18 PM Tom_L: i tried one that had a step but it kept coming loose
11:19 PM Tom_L: may see if i can get an ovesize bit stepped to a 3/8 shank
11:21 PM rue_shop3: I'm half way thru an upgarde on a z motor
11:21 PM rue_shop3: your hand drill wont take a 1/2" bit with reduced shank?
11:21 PM Tom_L: also need to check the spindle parallel to the Z column
11:22 PM Tom_L: in X and Y
11:22 PM Tom_L: i don't have one
11:22 PM Tom_L: and if i get one it will be bigger than 1/2"
11:22 PM Tom_L: i'll just drill them a bit oversize
11:23 PM Tom_L: once it's tight it shouldn't move any
11:24 PM Tom_L: my hand drill sucks anyway. it's got a hand tighten chuck
11:24 PM Tom_L: i got a big arm buster i could use
11:26 PM rue_mohr: yes, the z axis motor upgrade is probably a good thing for me to do while I work out where to start
11:33 PM Tom_L: not too excited to take the column off the base to work on it but i think i'm gonna have to
11:35 PM Tom_L: storm movin thru...