[–] [S] 1 point 2 weeks ago

I forgot to put this in my other post, but I think that the traditional discrete solution is the Voltage-reference (or LDO) into Current reference circuit. Basically you can turn any LDO or plain old voltage regulator into a current reference by setting their feedback in a different manner.

That doesn't work for me here because I do have a 1V minimum voltage requirement. But most people probably are working closer to 3V or 5V.

You can mirror a current generated by a LDO over somewhere else though. The current mirror methodology is truly quite flexible.

  • source
  • parent
  • context
  • [–] [S] 1 point 3 weeks ago

    Yeah, OpAmps are easier for sure.

    I'd expect this design to be used "seriously" only with discrete components. Its slightly cheaper to use a dual-transistor + 3 resistors (BCM847B) than an Voltage Reference (probably a diode) + OpAmp + BJT + resistor...

    Even then, the OpAmp circuit probably has lower quiescent current, probably has better output stability. A simple diode-as-voltage-reference doesn't seem especially good though: the Peaking Current Mirror will have lower PSRR than diode-as-voltage-reference (because Peaking Current Mirror has a mild amount of negative feedback).

  • source
  • parent
  • context
  • [–] 1 point 3 weeks ago (2 children)

    cool writeup. i'd like to add that at the board level discretes aren't really used unless you're shooting for certain performance characteristics (like low noise) or application or doing it for funsies. the op amp current source you showed remains king -- so king that it's even used at the chip level, where it would make most economical sense to just use two to four transistors, to provide a stable master current reference.

  • source
  • [–] 4 points 1 month ago

    I think the word you are looking for in English is "Design", not "Conception". A "Conception Engineer" sounds like you are helping someone make babies. ( ͡° ͜ʖ ͡°)

    My personal experience is largely from before AI ate everyone's jobs, so may be outdated. But my quick reaction is that an IC designer in the right area might make more than an RF designer, but IC work is highly specialized while RF design principles are applicable much more broadly.

    Also, to me, "Design" corresponds to the high-level concepts. Someone still has to translate that design into a physical structure. "Layout" is still a thing that humans do, heavily aided by tools (and probably AI by now) but a human still has to oversee it all. You probably go to school to learn all the same basic Engineering concepts, but then go in a different direction post-University. I know several people who make a good living as independant PCB and RF Layout consultants, after spending several years doing the job for large electronics companies.

  • source
  • [–] [S] 1 point 1 month ago*

    I'm trying to learn electrical circuit design and programming and want/need an oscilloscope for obvious reasons (I already have most of the basic stuff such as a soldering station and a multimeter). I know I want to work with various ICs, FPGAs, and of course stuff like USB to plug stuff. Though budget isn't that extravagant, I should have decent wiggle room (I suppose totals of below or around 700$ is fairly manageable at the moment).

  • source
  • parent
  • context
  • [–] [S] 2 points 2 months ago (2 children)

    Been doing some digging now knowing about logic analyzers and I found that RIGOL has some oscilloscopes with digital channels that are within my budget range (needs the RPL1116 probe to function). Though they're bandwidth is 100MHz for the one's I could find that are under 500-1,000$. Would this be enough at least for USB 2.0?

  • source
  • parent
  • context
  • [–] 2 points 2 months ago (3 children)

    I have the same scope, but it doesn’t have USB decode. I watched the video and he’s initially doing the decoding very manually. Fun, but not something you’re going to want to do when someone’s paying for your time.

    He later switches to a Keysight DSO-X 4024A. It costs a fair bit more. :)

  • source
  • parent
  • context
  • view more: next ›