#avr Logs

Nov 13 2021

02:30 AM josuah: on the back of these websites live ERPs that manages all the parts, talking with web frameworks APIs, sometimes the website takes the central role of synchronising everything
02:38 AM josuah: and it is a big mess as the website takes development age: sometimes they are build in a portable way, but sometime they get broken from the inside to fit the client's need
02:38 AM josuah: maybe Digi-Key is having its own range of website cruft to sort-out
05:51 AM exp: there's something about electronics suppliers
05:51 AM exp: they are contractually obligated not to have a good website
05:52 AM exp: i've seen a few attempts now at "this equivalent part" feature but it's usually the worst
05:52 AM exp: "hey this 25v 200mohm P-MOS will be a great substitute for your 80v 25mohm N-MOS!"
05:56 AM exp: i suppose they've probably analysed my circuit with an advanced AI and determined its best usage is as a fire starter, and so the two parts are equivalent :D
06:45 AM twnqx: i've yesterday got a full lab power supply as a compatible replacement for my PMIC...
06:46 AM exp: technically correct :)
06:47 AM exp: to be quite honest, that could make a good youtube show
06:47 AM exp: set out to build a project, but use only parts explicitly recommended to you
06:47 AM exp: i'd love to see the LED light bulb that ends up being made of a single red LED but powered by a 10kva genset
06:52 AM cehteh: hey back in times when i did rock'n'roll technican we had this US musican with a hammond organ who needed 110V 60Hz ... and we have 220V 50Hz ...
06:52 AM cehteh: turned out to be powered by a lab signal generator and a fet Bass Amp :D
06:53 AM cehteh: prolly the first tuneable hammond around
06:56 AM Thrashbarg: wow haha
06:56 AM Thrashbarg: cehteh: a current project of mine is a VFD for a brushless DC motor, driven by an Arduino and MIDI, to controll the speed of a motor in a 1/4" tape machine
06:56 AM twnqx: actually it wasn't, my pmic has 4 channels, the lab supply didn't
06:57 AM exp: buy 4?
06:57 AM exp: :D
06:57 AM exp: they've got you covered
07:36 AM specing_ is now known as specing
12:18 PM exp: anybody around and have some spare brain power i can borrow to rubber duck a problem?
12:22 PM qu1j0t3: exp: sure!
12:22 PM qu1j0t3: exp: i'm procrastinating so your timing is perfect!
12:26 PM exp: qu1j0t3: my boss messaged lol
12:26 PM exp: give me two seconds and i will explain :)
12:28 PM * qu1j0t3 nods
12:28 PM exp: qu1j0t3: so I have a device which produces 12v out to power some internal components. I will be connecting many of these together on a DIN rail or equivalent
12:29 PM exp: so smart person me thinks: I know this, I'll add a bus bar to the bottom and that way I can share their output
12:29 PM exp: they have internal batteries, and don't always run at the same duty cycle, so it would be advantageous to be able to balance them intelligently
12:30 PM exp: so, i have a CAN bus already available and use UCC29002s for balancing, so I can inject enough current to do this and everything is great
12:30 PM exp: however, I now need to think about protecting this interface, as someone shorting the battery, or shorting the bus is absolutely not out of the question
12:30 PM exp: I have a decent stock of LTC4368-1s which are almost ideal in that they have -40 to 100v protection and bidirectional current limiting
12:31 PM exp: however, it has a strict requirement that upon a reverse overcurrent fault, a 100mV difference must develop between the sides before it will re-enable the fets
12:32 PM exp: in my use case, this isn't strictly possible as i'd have to generate a voltage 100mV above my internal 12v supply to accomplish this
12:32 PM exp: but it strikes me that... i can connect SENSE and VOUT in reverse
12:32 PM exp: i retain all features and all protection, but now Vin must drop 100mV below Vout
12:33 PM exp: as a corollary, when this is the only device powered and it needs to enable the bus initially, i plan to use a simple diode + power resistor to bring it up within the UV/OV range which will wake up the 4368
12:33 PM exp: this will also work to my advantage in the case of an overcurrent fault, as it will naturally charge the bus to > 100mV below my internal rail
12:33 PM exp: my question to you: am I dumb? :)
12:33 PM * qu1j0t3 reads
12:34 PM exp: muchos gracias
12:35 PM qu1j0t3: ok it's a bit above my pay grade but sounds feasible? checking the datasheet on that thing
12:36 PM exp: i'm no qualified engineer either :D
12:48 PM qu1j0t3: exp: is it a concern that you'd be swapping the very asymmetrical forward/reverse behaviours? would the behaviour be safe for a forward overcurrent trip/reset loop?
12:49 PM exp: qu1j0t3: they're only asymmetric in the -2
12:49 PM exp: in the -1 they're 50mV either way
12:50 PM exp: it isn't quite ideal in that i'd prefer it be 50/25mV but 50/3 is way outside my load profile
12:50 PM qu1j0t3: ok, but what about the cooldown? that's absent from the reverse recovery at least int he Rev C
12:50 PM qu1j0t3: i probably should be looking at the right -1 datasheet :)
12:50 PM exp: that should be fine, everything internally has overcurrent limiting directly built in
12:50 PM qu1j0t3: ok.
12:50 PM exp: so even if it reconnected immediately it'd just be re-disconnecting I think
12:50 PM exp: I will double check that though thank you
12:50 PM qu1j0t3: seems like something you could test empirically
12:50 PM qu1j0t3: i mean, your idea
12:50 PM qu1j0t3: it's a pretty clever idea
12:51 PM exp: I may have accidentally done the CAD for the chassis and now some crazy boss of mine has decided to make it
12:51 PM exp: so i'm out of time and space and using what I have :D
12:51 PM qu1j0t3: exp: yeah i might be concerned about a high frequency trip/reset loop
12:51 PM qu1j0t3: exp: since the 1.2sec cooldown just isn't present in that direction
12:51 PM exp: yeah i'm combating motorboating on a few parts of this circuit
12:52 PM exp: the one saving grace i have is, i'm going to add a fairly big RC to the gate line of the dual nmos, as there still could be around 4v difference between the buses when it enables
12:52 PM * qu1j0t3 nods
12:52 PM exp: I need to keep the slew rate low enough that the sharing bus control loop can keep up
12:53 PM exp: so even if it does constantly re-enable, it will take some time to cycle, not 1.2s though
12:53 PM exp: this part I definitely do need to test empirically
12:53 PM exp: I will have to max my load tester out too which will be fun
11:06 PM rue_mohr: hmm
11:44 PM Thrashbarg: hmm
11:46 PM rue_mohr: I"m not doing well with inductors, I suspect if I'd been raised in an age where I were rewinding slot-car motors to beat the other guy, I'd be better with magnetics
11:47 PM Thrashbarg: that or making your own radios
11:48 PM rue_mohr: so
11:48 PM rue_mohr: I'm working on a flyback converter
11:49 PM Thrashbarg: ah yea nice
11:49 PM rue_mohr: 110V to 12V DC
11:49 PM rue_mohr: its being a lot of trouble
11:49 PM Thrashbarg: my advice is to get a prewound transformer :P
11:49 PM rue_mohr: maybe its the 20W
11:49 PM Thrashbarg: been there done that
11:49 PM rue_mohr: I should know how to make one tho
11:50 PM Thrashbarg: yeah fortunately you only need a few dozen turns for a flyback. If it was linear, you'd need thousands... done that too, wasn't fun
11:50 PM rue_mohr: a forward converter is just a boost converter really
11:50 PM qu1j0t3: Thrashbarg: well that's the thing, not everything is on the shelf
11:50 PM rue_mohr: your just using a different winding for charging the inductor than discharging it
11:51 PM rue_mohr: so, I'm starting with out the primary
11:51 PM Thrashbarg: qu1j0t3: https://au.rs-online.com/web/p/switch-mode-power-supply-smps-transformers/4185543 :P
11:51 PM rue_mohr: the idea being to build, basically a voltage doubler, but I'm taking the output from vcc, so, its just a voltage offseter
11:51 PM rue_mohr: I have a lot of documents, I need to fix whats in my head tho
11:51 PM qu1j0t3: Thrashbarg: exactly. very limited specs out there.
11:52 PM Thrashbarg: oh I don't know. If you just want 12V at a few amps, that'd do fine
11:52 PM qu1j0t3: Thrashbarg: for example, a possible need i had recently - say a 1:1,1:2 or 1:3 stepdown, say 20v to 6.3v
11:52 PM rue_mohr: so , I should be able to take an inductor, and at 50% duty, convert the input to an output with the same voltage
11:52 PM rue_mohr: +- a bit of loss
11:52 PM qu1j0t3: Thrashbarg: nothng exists, especially when you add my need for 6kV isolation...
11:53 PM Thrashbarg: oh yea if you're talking outside of a narrow band, yea
11:53 PM rue_mohr: I'v fought a lot of breadboard and winding loss, but
11:53 PM qu1j0t3: Thrashbarg: yeah, that's all i mean.
11:53 PM qu1j0t3: anyway i will continue to spectate rue_mohr since i might learn something
11:53 PM rue_mohr: the equations I have from my documents talk about the duty as a ratio of input and output voltages
11:53 PM rue_mohr: but its not quite right
11:54 PM rue_mohr: an inductor isn't outputting a voltage
11:54 PM rue_mohr: its outputting a current, with a joules limit
11:54 PM Thrashbarg: got a schematic?
11:54 PM Thrashbarg: if you're only needing 20W, you'd not need a forward converter. They're better for >200W
11:54 PM rue_mohr: http://ruemohr.org/~ircjunk/tempimage/p1210751.jpg
11:54 PM rue_mohr: if it helps
11:55 PM rue_mohr: the idea is to be dumping ~12V into the load
11:55 PM Thrashbarg: looks like you're making a boost converter?
11:55 PM rue_mohr: which is conducted thru the inductor
11:55 PM rue_mohr: it does, but its a flyback where there is only 1 winding
11:55 PM rue_mohr: and that winding is the secondary
11:55 PM Thrashbarg: ahhh
11:56 PM rue_mohr: if I can get this to work, I simply need to create a new primary for the available voltage to charge it with
11:56 PM rue_mohr: here, I'm simplifying by making the input the same as the output I want
11:57 PM rue_mohr: as soon as I know that winding can produce the right output, I can move on
11:57 PM rue_mohr: I got close
11:57 PM rue_mohr: I put a HUGE multifillar inductor on, and it got 1V short of the target
11:57 PM rue_mohr: but, its breadboarded, so thats ok
11:58 PM Thrashbarg: are you using a gappe or ungapped transformer
11:58 PM Thrashbarg: *gapped
11:58 PM rue_mohr: its gapped
11:58 PM rue_mohr: http://ruemohr.org/~ircjunk/tempimage/p1210756.jpg
11:58 PM rue_mohr: its a oversized one to do away with needing to design one
11:59 PM Thrashbarg: did you wind the transformer yourself? If you made two 12V windings to test it you could rewind it with a 110V winding afterwards
11:59 PM rue_mohr: this one came prewound
11:59 PM Thrashbarg: okay
11:59 PM rue_mohr: ;)
11:59 PM Thrashbarg: dare I ask if there's a datasheet on it? :P
11:59 PM rue_mohr: the point is, I know it has ~0V winding loss, and it has more than enough storage capability