LikWidChzyeah watch the video on the teardown of the video camera pill, ill find a link
kaoshavocI think one of the big reasons why I like greatscott is he does soo much for no real reason other than learning or want
kaoshavochis is not really a channel of practicality but of fun and learning and experimenting
kaoshavocthen you got AvE, that bumblefuck is just awesome and if I could learn anything on a physical trade, it would have to be machining, but he even knows how the angry pixies work for the most part
LikWidChzhah yeah its good stuff! HEY nice chatting, ive got to be going :)
kaoshavockrasnow ?
kaoshavocor kraznow ?
kaoshavockk, see ya
wolfmanjamesdo people chat here? I have a question
pwillard ;-P
IcePic"do what I want" can fail on subtle things like buying an enc28j60 which basically can put "bytes on the wire" and you wanted to run tcp/ip, which means you get to invent and implement the whole tcp and ip layers yourself.
pwillardIf you try hard enough you can attach anything
deskwizardlol considering our mcu's are more powerful and faster than the ones they were originally ran on .. :P
catphishIcePic: yeah, won't be writing my own tcp stack for this :)
pwillardI have an RC2014 Z80 CPU at home and struggle to keep it pure 80's technology.
pwillardConstantly tempted to add a ESP8266 or Arduino.
IcePiccatphish: which is fine, just make sure you know what "I want" actually is when looking, because its easy to get tricked into "oh, this ethernet blah is cheap" when "ethernet" doesn't really mean tcp/ip to such chips
pwillardexactly. I have some extra unused Enc28J60 for that reason
YotsonLet alone wanting some fancy buttons web page. :)
catphishIcePic: indeed, i'll probably want a TCP/IP ASIC to keep things simple, unless i just use UDP
catphishno http needed :)
YotsonAh ok.
pwillardarduino is really horrible for web pages... when for the same price of arduino and wiz ethernet board you can have a Pi and do this.
IcePicI think this is why esp8266 became so popular
IcePicit just does a huge amount for you
catphishesp8266 is cool for this stuff, shame no ethernet
IcePicexactly, someone (not me) should invent the wired version of esp8266
YotsonI just don't like wifi. Especially not on the 2.4GHz band.
Yotson-sigh- Colder weather, old car -> battery died, rusted fastener broke off getting it out.
Yotsonafk... :)
deskwizardYotson: arff GL
sibiriai like brisk walks in chilly weather
catphishso i just need a couple of and a
sibiriaas long as it's not a couple of miles
pwillardcatphish: probably
deskwizardcatphish: yeah the PCF would do, but I'd suggest the MCP23017, you get a lot more flexibility and 16 instead of 8 I/Os as we mentionned
pwillardIf you are not treading new ground on your own... at least follow the most worn path...
catphishdeskwizard: ah yes, even better, thanks
deskwizardcatphish: np, I have a lib for those where everything is implemented in case you're interested I can push a bit early
catphishdeskwizard: thanks, will probably be back once i have some hardware to play with
deskwizardcatphish: aight, by that time I should have everything checked out
catphishcool, thanks
catphish1.8v to 5.5v awesome
deskwizardand another one done :)
deskwizardindeed, I'm making the damn thing for myself anyway, might as well push it for everyone else
coinomy usb-to-serial has 5/3.3V, TX,RX,GND, but board has GND, 3.3V, CLK, I/O, pls advise
timemagecoino, "board" ?
coinotimemage, the mcu board
pwillardUnless your USB has an onboard 3.3V regulator... only connect TX RX and GND
coinothe usb-to-serial has one pin 5V and another pin 3.3V
coinobut the TX RX vs CLK IO i don't know
Entroacceptorcoino: It'll _probably_ work if you connect the 3.3V from your usb-serial to 3.3V on the mcu. In theory that could fry the mcu, but so far I haven't had it happen.
EntroacceptorBut do have a look if you've got a jumper on the USB-to-serial board to select the voltage
timemagecoino, not sure how anyone can know how to hook your cable to an arbitrary board.
coinoit's a USB-to-TTL and a STM32F103C8T6 ARM board
coinosorry I'm new to this
timemagecoino, a nucleo?
coinoI think they call it blue pill
timemagecoino, okay, you've been looking at the SWD port, it seems
coinothe board has lots of pins on the sides this is a separate 4 pin connector
timemagecoino, you wouldn't be hooking that to a usb-ttl table, unless you have one that does bitbanging and you are trying to jtag/swd program through something that knows how to do that.
_ami_timemage: he has usb to ttl so i think he wants to program blue pill via serial bootloader?
_ami_there are two ways to program blue pills
timemage_ami_, maybe, if so he has the wrong port.
coinotimemage, that's the one
coinoI saw people using the arduino IDe with it
_ami_1. via STLink v2 using SWD
_ami_2. via usb to ttl using serial port of blue pill
coinoso this 4 pin is the SWD port
timemagecoino, assuming they're using te serial device as actually serial, it seems one of the RX TX pairs, maybe the one on PB6, PB7
_ami_To program it via usart, you need to set BOOT0 to 1 and Boot1 to 0.
timemagecoino, yup. as in "you've been looking at the SWD port, it seems"
_ami_and connect Rx/Tx pairs which timemage already described.
coinoUSART being RX TX
_ami_coino: using linux or windows?
_ami_coino: btw, refer to this guide:
_ami_also after flashing make sure you set BOOT0 to 0
_ami_its better if you flash blue pill with dfu bootloader
_ami_but i heard it has problems because of wrong usb D+ pullup
_ami_it should be 1.5k but it is generally 10k thanks to great chinese designers
timemage_ami_, that has it as pa9 and pa10 rather than my guess. same signal labels though. maybe a remappable function.
_ami_let me check it in libopencm3 codebase
_ami_yes, PA10/9 - Rx/Tx
coinoweird i can't get a read on R10
coinotried all scales
coinoit's 10k
_ami_coino: for usb to work, you need to replace it with 1.5k 0603 size Resistor
coinoso Im stuck with USART
_ami_coino: getting a STLinkv2 programmer would be nice.
_ami_then you don;t need to change boot0/1 jumpers position
_ami_coino: you could buy a black pill instead
_ami_which has correct pullup
_ami_coino: - very good seller.
_ami_and a very good product.
_ami_it has serial bootloader flashed but you need to pay 1$ extra. :)
_ami_usb to serial bootloader*
coinogreat i bought 5 of these
coinoit's more the hassle i guess
_ami_well, get those repaired - hardly 5 mins job per board for replacing R10
_ami_i heard stm32duino is pretty nice.. i have not used it though.
timemagecoino, you might be able to find a usb hub that doesn't care about the resistor. maybe an old usb 1.1 hub.
_ami_coino: you can do cool stuffs with this board. recently i made a usbdisplay with this board: stm32f103 + ili9488 & custom framebuffer kernel driver.
_ami_timemage is right. it might work for you. AFAIK, usb bootloader does some software hack to take care of this wrong pullup at D+
_ami_timemage is right. it might work for you. AFAIK, usb bootloader does some software hack to take care of this wrong pullup
timemageit's always nice to be right twice in a row.
coino_ami_, looks good
coinotwo rows even
coinoi got blinky
coinoI mean, just power up
_ami_coino: woah! congratulations!
coinobut it's default program i guess
_ami_oh, thats the default program seller load up :)
coinoyes, soon
coinowindows says best driver already installed
timemagecoino, best driver is a bad thing?
coinowell there's a warning
coinoand no COM available in IDE
coinois that the resistor problem?
timemagecoino, i see. what is the actual warning on the device in device manager?
coinois this what the pullup resistor problem does
timemagecoino, well, from what i see in searches this is due to a counterfeit pl2303
coinowhat does that mean
timemagecoino, pl2303 is the chip used in your usb to serial ttl adapter. i'm guessing you knoww what counterfeit means.
_ami_i think its a problem with a your ubs to ttl device not blue pill board
coinoso the usb device I got from china is conterfeit
pwillardGosh... ya think?
timemagecoino, seems likely.
coinoi figure all these videos on youtube are doing the same thing
coinoI got unlucky?
pwillardChina will rule the world as a result of "death by a thousand tiny cuts"
timemagecoino, indications are it worked with older drivers.
_ami_or use linux! :)
coinoah, maybe in another life
timemagecoino, the only pl2303 i have (that i know of) is from maybe 2006 (perhaps earlier) and came from somewhere like digikey, mouser, and not ebay or someplace like it. so far no problems. i've used it under windows 10 sometime in the last four months.
coinotimemage, any idea how I can get old drivers
_ami_by installing the old drivers ?
_ami_lol -->
timemagecoino, yeah, if you go that route you might consider getting them from someone reputable that made a product using that chip. in other words, dont get it from somewhere random.
timemagecoino, or... get another cable from a trusted source.
timemagecoino, and return your other one. even if you break even on shipping you'll have wasted a bit of their time.
coinofor some odd reason i bought 10 lol
coinoI must've misclicked or read
timemagecoino, one for every month of the old roman calendar maybe.
coinono idea really
timemagecoino, they all do this, yeah?
coinoyeah i tried a bunch
coinoit werkd
coinoI followed this
coinonow...has anyone done OOP UMl with this thing
pwillarda really good reason to not use those.
deskwizard*sigh* not working, don't know why, comms are fine :| damn it.
DacoTacohello hello! o/
DacoTacothis channel is euh...alot more crowded then i had thought xD
pwillardyes, this is one of the busier "basic" knowledge channels
DacoTaco"basic"? :P
Rickta59maybe basic+
Rickta59when is it that china goes on holiday? january?
DacoTacono idea tbh, why?
Rickta59i want to plan my aliexpress purchasing
Rickta59ah . . Feb 16th, 2018
DacoTacoyou got time :P
DacoTacoare general related avr questions ok? :P
pwillard usually... though there is an #avr channel as well
DacoTacopwillard: looked dead before :P
pwillardhas its moments... sae as here
DacoTacowell, i guess this problem can also happen with an arduino i think
DacoTacoi am connecting a few data lines to a port, set as input, pull down disabled, and the device outputting to those lines has 5v. but asa i connect those data lines to the port, the voltage drops and therefor the pin never is pulled high (or very shortly so its not detected maybe)
DacoTacoany ideas? :/
pwillardNote: there are no pulldowns in arduino
pwillardData lines... what are they?
DacoTacopwillard: i mustve read wrong that they used the internal pullups. anyway, the data lines are just pins set high or low. the multimeter confirms the right pins are set to high. but when i connect those pins to the avr, the voltage drops and cpu doesn't detect the input as being set high.
DacoTacomy code & setup is actually a copy of some dude who did it with arduino too xD
Snert_you sure a pullup should not be used?
DacoTacoSnert_ : no idea, the original project didn't use them, and when i use them, the pins wont go low when the data line is pulled low by the device (a Gameboy cartridge btw)
DacoTaco(the pull up overrides the low pull from the cart)
kaoshavochmmm, I see 2 mistakes I made on the boards I ordered,
raphain, am i right to believe that R6 changes U2's slope, and, if R3 were made a trimmer, it would adjust U2's gain?
kaoshavocI have not even got them yet LOL
kaoshavocand I thought that when you turned a schematic into a pcb, it force saved the schematic, so, now I dont' even have the schematic I made for it on easy eda LOL
raphakaoshavoc: are you talking about EasyEDA?
kaoshavocI really did think it force saved,
kaoshavocboth errors I see are easy fixes when the boards come in
kaoshavocbut, since all the stuff is lables R1 ...... R20
kaoshavocand I don't have the correct schematic, its hard to verify traces LOL
raphai've started automatically hitting Ctrl-S in it
kaoshavocbut so far that I can see, its not too bad, I expected there to be mistakes, its my first board ever ya know
raphabtw have you figured out a /reliable/ way of making connections between wires and things, or wires and wires? i mean that when you move things around the connections move with them?
kaoshavocno, I don't care much for the environment yet and thats part of the problem
kaoshavocthe red dots seem to mean multiple attachments? but thes sometimes it seems like they mean no attachment, I dont' know
raphai actually sort of like it, except for that particular annoyance
raphaand i thought big red dot meant "permanent attachment"
raphadoesn't seem to be so
kaoshavocyeah, I think I see more errors, I wish I had the schematic
kaoshavocthese are probably junk boards tbh
raphasorry to hear that :(
kaoshavocoh lord
raphayou're right, red dot is multiple attachments
kaoshavocit looks soooo messed up , Idon't even understand how
kaoshavoc Ihate that it attaches to anything around instead of what you point it at,
kaoshavoclike if you accidentally cross paths with more than one componant, it will absolutly attach to it
kaoshavocjust gonna take some getting used oto
kaoshavoctoo 8
raphabut it does not automatically mean that if you have a red dot it will be fixed
raphahmm interesting, that didnt happen to me yet, accidental connections
timemagekaoshavoc, i've had new boards being made while others were in flight. =). if you already familiar with them, check out the software's ERC and DRC features. most schematic capture and layout software has them. it can save quite a bit of trouble.
kaoshavocI am not frustrated that I have probably made a useless board, I am a bit frustrated that it did not even save the file I used to make the board, like, you would think that would be standard procedure
kaoshavoctimemage: you just spoke greek to me LOL
kaoshavocerc ? drc ?
timemagekaoshavoc, you're using something called easyeda ?
raphakaoshavoc: what does work is to select (Ctrl, not Shift!) a component and a "multiple attachment point" on a wire and then move both together
raphatimemage: it's actually quite nice, i'm liking it so far. more fun to use for me than either fritzing or qcad.
kaoshavocso, I 100% understand why people say to never use auto routing
kaoshavocit may be easy, but freak does it make tracking down mistakes harder
timemagekaoshavoc, often autorouters simply don't understand the constraints you understand. if you could feed it enough knowledge about what your care about it would probably make better decisions than you.
raphatimemage: do you have a minute to take a look at ?
timemagerapha, hmm, what am i looking for?
raphatimemage: i'm wondering whether R6 controls U2's slope or gain, but I'm guessing slope? And then I'm wondering how to control gain, and I'm guessing by making R3 a potentiometer?
kaoshavocoh man, these boards are so junk LOL
raphakaoshavoc :(
timemagerapha, this is an opamp?
kaoshavocoh well, next month I will try again
timemagerapha, okay, says opamp. looks like gain control
raphatimemage: yes, both U1 and U2 are opamps. U1's function is to amplify the signal coming in through the BNC connector. So if it sees 0V there, ofc its output to U2 will be 0V as well. That's the reason I'm thinking to use R3 for gain adjustment. And U2's function is to offset the signal so that an input of 0V falls into the middle (~512, 2.5V) of an ATtiny's ADC range.
timemagerapha, not sure my brain is functioning well enough for this. not that i'm well suited to answer the question anyway. but... you're expecting a dc offset out of this circuit?
raphaoh do you need caffeine?
raphatimemage: well, first an amplification and then an offset, yes.
raphaand yes, DC.
timemagerapha, i'm not seeing how a dc offset would come about.
raphatimemage: i thought by way of a portion of VCC being fed into U2's positive input?
raphabut this is the first time i'm trying to understand something like this
timemagerapha, ehh, i think that will wind up clamping the signal. usually if you're adding a dc offset you'll have a dc blocking cap between the input signal and divider that you feed to the opamp.
raphafor what it's worth, my multimeter says that if input is around -100mV then output is around 2V. so it is doing that.
kaoshavocok, this is not looking as bad as it seems, going to have to cut 2 traces, add a couple jumpers, and it should be good to go
raphatimemage: "to clamp" means "to hold tight"?
timemagerapha, to make sure it can go outside a certain bounds.
kaoshavocsomehow my power wire went to pin2 instead of pin3, and pins 1&2 are connected instead of going to seperate LEDs, Weird as hell, but whatever
raphathe input signal can be between -350mV and +350mV timemage, so it's sort of keeping itself within bounds.
raphaall i can really say is that it does work as advertised. i'm just missing a way of adjusting both gain and slope and i'm not sure whether R6 is the gain or is the slope adjustment.
timemagerapha, i'm so used to dealing with single supply systems that i'm probaly missing something really obvious about how this works with the negative rail.
raphaah i see
raphaperhaps pwillard will peek by sometime later.
raphai'd be surprised if he wouldn't have experience with negative rails.
raphathanks for taking a look :)
timemagerapha, well, most of us have some experience with them, if for no other reason than dealing with serial. but it would help to have someone that deals with analog signals.
raphamaybe it's a good idea then if i go ask over in #electronics
timemagerapha, it wasn't that long ago that i was reading about how headphone amplifier ic's deal with "dc pop" that happens when you connect the load.
raphalet me guess, they use negative rails? :)
timemagerapha, traditionally positive and negative rails weer more common.
timemagerapha, often yeah. some of them just generate a negative rail and use that.
timemagerapha, sometimes they do something like slow-start the dc offset.
raphaso what's bad about that, why did they change that?
timemagerapha, just more complexity we usually don't need.
timemagerapha, it's great if you're working with analog signal though.
raphai always felt more connected to analog stuff
raphait's kind of funny that i haven't yet drawn the ADC into that picture :)
raphaas a kid, before i was even 4 years old, i wanted to know how a radio worked. wasn't yet strong enough to use a screwdriver on it so my solution was to throw it out of the window as many times as necessary to get it open. that's an analog method!
timemagerapha, heh.
timemagerapha, a friend of mine and i developed our own definition of 'solid state'. which was something like: a thing is solid state if hitting it with a hammer has no chance of fixing it.
raphai like that :)
CheatermanBliblibli I hope y'all doing goodie ^__^
CheatermanApparently using Serial on a Nano at 115200 bps is a bad idea?
CheatermanIt's said it can have 3.5% error rate
Cheaterman TIL
CheatermanMight be a good idea for me to go down to 76.8k
timemagerapha, made some approximation of your circuit in a simulator.
raphai went to ##electronics and became part of a discussion about when and how the US will implode o_O
timemagerapha, heh, i see.
raphadying to see that simulator approximation, though :-D
timemagerapha, it's not named what i thought in the trace, so i need to check that.
timemagerapha, the output is not unlike what i expected to see though.
raphait's completely unlikey what i expected to see ^^
raphawhy is it going up and down like that?
raphahmm also for what input signal strength is that output signal?
raphaoh wait
raphaso you put a sine signal as the input?
timemagerapha, a 440hz 0.5V signal
raphathe real input can be assumed to be a non-changing DC voltage of between -350mV and +350mV
raphaso you just look at the point in time that interests you
raphadoesn't make sense then though ... the output signal should be between 0 and 5V. it's between 4.48 and 4.96V instead.
timemagerapha, i'm not sure you even need to worry about that exactly. look at oa_neg, the negative input to the opamp.
raphatimemage: so _your_ U1 is _my_ U1? or is it _my_ U2 rly? *confused*
timemagerapha, your U2, iirc.
timemagerapha, just the part that tries to add DC offset. because this is what doesn't make sense to me about your circuit.
timemagerapha, so instead of a trim pot i put 2.5V there through the 5k/5k divider.
raphathe guys in ##electronics said the same thing btw, the second part doesn't rly made sense to them.
timemagerapha, the opamp tries to drive the output to make the input to the positive pin look like the input to the negative pin.
timemagerapha, erm. i think i just said that incorrect
raphaand it's giving my brain trouble parsing :)
timemagerapha, i know where i'm trying to go with the explanation, i just don't want to get there via incorrect statments.
raphaif my U1's amplification is 3.5times, could you make the input signal go from -1.25 to +1.25V? i'm curious to see whether that would bring the output into the range it's having in reality...
timemagerapha, i guess i should put it this way. you expect oa_neg to go above 2.5V at some point, right?
raphai'm still looking at it wondering about oa_neg and how it differs from input signal
raphaR4's purpose can only be to limit the op-amp's input voltage to something that won't destroy it, right?
timemagerapha, if you wanted oa_neg to get above 2.5V, wouldn't you need more than 5V at the output?
raphai don't know at all what i want from oa_neg
raphai know that the pH probe produces -350mV to +350mV and i know that the ADC needs 0-5V input to produce a number between 0 and 1024.
raphaor rather, 1023
rapha(i also know that to be a gross oversimplification because the pH scale is log)
raphawait, oa_neg is not fed by R4 only, it is also fed by R5, which is its own output
raphaor a portion of its own output
raphaokay i need to see what's inside an opamp because i'm completely confused by it having a positive and a negative input
Mikeeerapha 5v positive and 5v negative is exactly the same as 10v and ground.
raphaMikeee: i don't mean the rails. i mean the inputs.
timemagerapha, mm, you may find that more confusing than helpful. the internals of an opamp are more complex than you'd probably guess.
raphaholy shit
raphaindeed they are
Mikeeeinverting and noninverting input
raphabut why, Mikeee ?
MikeeeTo make it universally useful
timemagerapha, high gain / high bandwidth, low temperature drift, high linearity.
Mikeeeand one has to be compared to the other. One has to be the one comparing and the other the one being compared TO
timemagerapha, is your ph sensor a 2 terminal device?
raphaMikeee: the schematic we're talking about seems to be an "inverting amplifier" according to why is the resistor going from inverting input to output not causing the op-amp to produce an ever, ever, ever increasing output signal?
raphatimemage: yes.
Mikeeerapha because it's a divider
timemagerapha, have a look at the schematic on page 4
raphatimemage: if you mean GND and signal by two-terminal?
Mikeeethere's a divider between Vin and Vout, and that division shows up on the inverting input
raphawow TI has seriously cool application notes
timemagerapha, they're using an opamp to make a strong divider for the lm4040 (1v reference)
Mikeeeso the opam will ALWAYS do whatever it can to make sure that point in the divider is exactly equal to the positive input (which is likely ground most situation)
timemagerapha, well 1.024v reference.
rapha*head begins to steam*
timemagerapha, so as it says, 512mv offset. i was thinking of suggesting diodes, thought they'd be pretty crude and subject to temperature problems
raphatemperature is not really a concern in the temperature-controlled environment of a fishtank
raphaokay, because of steaming head
raphaMikeee: the original question was, in, is R6 controlling gain or slope of U2?
timemagerapha, i'm not really referring to the temperature part of the circuit. i was only speaking of temperature because of how ti would affect what i'd be considering as an alterative to their opamp design.
Mikeeeits a falstad simulation of an inverting amplifier
raphatimemage: i see ... but diodes can not amplify?
timemagerapha, no, i was considering them only as as crude way to add an offset voltage.
raphaMikeee: everything you can read about opamps uses frequencies as input. my pH probe inputs a steady DC voltage.
timemagerapha, what ti is doing here make more reasonable.
raphatimemage: i was wondering why not use a transistor voltage for the offset voltage?
timemagerapha, i'm not sure how that would work (well).
Mikeeerapha there's no AC component to its operation at low frequencies. You can just see the sine wave as temperature going up and down
raphatimemage: not well, just work
MikeeeI can make a circuit with a DC slider if it would help you understand it
raphai think it may, yes ...
MikeeeSlider is on the right, labeled "voltage"
raphatimemage: btw the TI paper says "[the probe] produces a voltage that is /linearly/ dependant upon the pH of the solution being measured". i wasn't clear on that part before. it's cool that that is not log as well.
raphathank you Mikeee
raphaACTION plays
timemagerapha, your ph sensor is of the same sort of design as theirs?
MikeeeNotice no matter what you do to the input voltage, the meter reading the inverting input's voltage stays pretty much 0 volts (it shows up as microvolts because there's a TINY amount of current going into the opamp's input.) For all intents and purposes, you can consider that input current to be 0, and the voltage at the negative input to be zero.
raphatimemage: my probe? yes, just lower quality (not made from glas)
raphatimemage: theirs is much closer to an "ideal" probe
raphaMikeee: why is there no requirement for a supply voltage here? much less two of them...
raphaand why are we not using opamps to generate negative supply voltages?
raphathat microvolt thingy between the 3k resistor and gnd
raphathat's just a voltmeter
raphanot an opening in the circuit
timemagerapha, i'm not suing their nice voltage reference, just buffering a 1/10th divider to get .5 V. the "sensor" (V3) sits on top of that. amplified by U1
raphaMikeee: i tried to change it to a non-inverting amp, i failed horribly :-P
timemagerapha, this is using "ideal" opamps
raphaACTION looks
timemagerapha, usually you don't want the input that close to the rail. though some opamps do okay under those conditions. you could just add a bit more dc offset and a bit less gain in the second part to deal with that.
raphathe triangles i'm assuming are ground
timemagerapha, yup
timemagerapha, usually you'd have caps on this. often the positive is connected to negative through a resistor rather than just direct like that.
timemagerapha, erm, output is connected to negative input through ... rather.
raphafunny that now looks a little bit like the 1st and 2nd staging are swapped, compared to the design i built
raphaperhaps it would be a good idea to replicate the real thing in falstad
timemagerapha, i didn't show "strong_div", but it would be a flat 500mv
timemagerapha, if you're working with an mcu, you could take the strong_div and feed it to an adc channel so you know what the offset was.
raphai still have two spare opamps
raphacould also try to build TI's/your design on the breadboard, why not
raphatimemage: it's just i spent so much time and effort to build the original schematic that i'd rly like to fix it instead of having to tear it down again :(
timemagerapha, what they're doing to make a voltage reference is a better idea. i just did the divider/buffer because it was simple
raphayeah i couldn't build theirs without buying more parts, but i could build yours
timemagerapha, well, someone that uses opamps enough to just be fluent with them may be able to tell you how to alter it.
raphayeah best to have some patience
timemagerapha, somehow your opamp would need feedback that is proportional (linear basically) and yet provide an offset (which puts it in the realm of an "affine" transformation). maybe you can do that, just not coming to mind. it occured to me that if your device has only two terminals that it needn't be ground referenced. that's part of why i just went searching for design notes. and sure enough, what we end up with is something that isn't
timemageground referenced.
raphadoesn't it sort of generate its own reference somewhere between where the device is placed and where the probe is placed? (i.e. table and fish tank)?
timemagerapha, ehh, what?
raphahmm but then the fish tank stands on top of the table
raphaso they have the same potential anyways
raphaare we talking about two different things?
timemagerapha, maybe.
raphawith reference i thought you meant the base potential of the whole system
timemagerapha, you might want to look at "summing amplifier" and your original circuit. might lead to a modification that you could do.
raphabut you meant what's pH 7 / 0V?
timemagerapha, nah, i meant in ti's circuit they have an lm4140. that's a 1V reference. roughly like a voltage regulator. i'm not doing that, i'm using divider but then buffering it with an opamp. so it's only 500mv if the resistors are perfect and the "5V" line is really 5V.
raphatimemage: the only thing i need is a way to control normalization (lower or raise U2's output at 0V input) and the amount of offset applied by U2.
timemagerapha, the lm4140 is a precision voltage reference. iirc, something like .5% accuracy. i'd need to look it up to be sure. more than i've needed in any case.
raphai have a different thing: in the ATtiny's sketch I specify the voltage as i measure it :-)
raphalike, okay, it fluctuates a bit, but it doesn't seem to be a real problem
timemage0.1% i guess.
raphait's amazing how you just look at all this stuff like it were LEGO and you build something from it because you understand the different LEGO blocks' properties and behaviours
timemagerapha, well, the people that design this stuff kind of made it to behave like that, as much as they could.
timemagerapha, i mean legos are fun, but after the fact it's kind of obvious that they stick together =) i mean, if they didn't there wouldn't be much point.
timemageexcept maybe ruining childhoods.
raphaofc just legos and their interactions are not as complex as these components and theirs
raphayet you seem to have quite some level of clarity as to that
timemageseamonkeys have that market, i think.
raphaseamonkeys? o_O :-P
raphaoh, ruining childhoods
raphawell, i only got them to live beyond a week of age once during my childhood so you may be right