azonenberg_work has quit [Ping timeout: 244 seconds]
<rqou>
offtopic question for people who like weird power hacks (azonenberg :P ): do you know of the magic search terms necessary to find an inverter that outputs 120vac and takes as input around 190-250vdc?
<rqou>
i was thinking around 2000w and _without_ MPPT
<rqou>
for context: I want to see if solar inverters have dropped in price enough to make a "tap power from a hybrid car traction battery" hack (http://techno-fandom.org/~hobbit/cars/plugout/) cost less $$$
<whitequark>
rqou: any inverter that has a rectifier stage?
<whitequark>
i.e. literally any inverter
<rqou>
most are rated only for 12v/24v/48v
<whitequark>
no, I mean, an UPS or something
<rqou>
not all the way up to ~250v
<whitequark>
get an UPS that has 240V input and 120V output
<rqou>
that's even more hacky :P
<whitequark>
really?
<rqou>
especially since you don't need any of the other parts of the UPS other than the inverter
<whitequark>
I think that powering SMPSes with DC is a perfectly legitimate way to use them
<whitequark>
oh, in that sense
<cr1901_modern>
A UPS with 240V=>120V out isn't going to have a step-down transformer for AC?
<rqou>
probably not
<rqou>
step down transformers are heavy :P
<whitequark>
expensive
<whitequark>
takes place and copper
<whitequark>
and a buck converter is rather easy to build, especially if you're staying under like 5kW
<rqou>
the normal way to do this is to take the AC in and immediately rectify it into DC
<rqou>
then use a DC-DC SMPS
<whitequark>
then the ghettiest UPSes will just output a square wave
<whitequark>
if everything you power from it is just another SMPSes then that will work, because they don't care
<whitequark>
if you try and run a fridge from it, lol
<whitequark>
your fridge is fucked
<cr1901_modern>
My UPS outputs modified sine wave
<rqou>
apparently some active PFC power supplies have problems too
<cr1901_modern>
Which is jargon for "square wave with a dead zone"
<cr1901_modern>
And sure you can go without. Just I'm used to seeing a transformer in a power supply lol.
<whitequark>
you can see that it's basically a boost that got interrupted by a trafo
<rqou>
how do you know so much about smps design?
<cr1901_modern>
The 20V output needs to be AC. So if you boost it, you'll need an inverter
<cr1901_modern>
Inductor isn't useful for DC
<whitequark>
rqou: I built a few shitty SMPSes and then drilled the brain of a friend EE until I understood why they didn't work
<whitequark>
I actually don't know that much about SMPSe
<whitequark>
*es
<whitequark>
I still have a loooot to learn
<rqou>
ah, much more effective than my "stare at TI appnotes" :P
<whitequark>
SMPSes aren't inherently hard to understand (but are hard to master)
<cr1901_modern>
The 20V from the first boost is AC when it reaches the inductor, correct?
<whitequark>
but you do need a change in point of view, so to say
<cr1901_modern>
whitequark: I understand. I asked b/c I'm not very good with isolation as a concept period. I know what it means (voltage at output terminals cannot be easily referenced to input terminals), but aside from a few (life saving) applications, i don't know much else about it.
<whitequark>
ok, I think I'm not explaining this very well, because there are so many topologies to talk about
<whitequark>
let me try again
<whitequark>
so let's say we have a naive buck converter
<cr1901_modern>
That website is REALLY cool, btw
<whitequark>
we take 240V from mains and output 5V
<whitequark>
at low loads, this converter will have to run at *extremely* low duty cycles
azonenberg_work has joined ##openfpga
<whitequark>
for various reasons that you're better off reading the appnotes for, SMPS controllers don't do them well
<cr1901_modern>
Understood/makes sense. I can imagine/do my own calculations later :P
<whitequark>
so, what you do instead, is you use a transformer to take high voltage to much lower voltage
<whitequark>
you want to run at 90-240V, so you design for 90V, and then gradually lower your duty cycle as your input voltage rises
<whitequark>
given constant load
<whitequark>
and you take the turns ratio such that on the entire input range, your duty cycle stays in reasonable limits
<cr1901_modern>
But *with* the transformer, the duty cycle is still better than if you were controlling directly even with 240V input
<cr1901_modern>
?*
<whitequark>
um, what?
<cr1901_modern>
"then gradually lower your duty cycle as your input voltage rises" <-- by the time we reach 240V input, do we still have shitty duty cycle?
<cr1901_modern>
From what you're saying, as long as you use a transformer to step down the voltage, the duty cycle over your entire input voltage range should be okay
<cr1901_modern>
whitequark: This site you linked to is absolutely amazing :o