Comment by chermi | Hacker News Reader

Comment by chermi | Hacker News Reader

Hey guys, I learned electronics from a nobel laureate!

Throughout my physics career including PhD, analog electronics was the most difficult but probably also the most rewarding class to me. I fondly remember staying until 2am in broida at ucsb trying to get a filter to work, getting a few hours sleep, then being back in the lab before sunrise. Of course, this was mostly the result of procrastination, but damn were those good times.

One thing that really bothered me then was the idea of a current source. I was perfectly happy with a voltage source, perhaps naively(1). But a current source seemed magical. I was asking Martinis about this and he seemed dumbfounded that I didn't understand. Of course, the answer is feedback. And, of course, good voltage sources also require feedback. But he was so familiar with feedback control he didn't even consider saying that's whats happening, while I never even heard of controls.

Long story short, sometime later I asked to join his lab as an undergrad researcher. He said no, and to this day I think it's because I didn't understand current sources. Or maybe I was too late, or maybe the A- (see the aforementioned procrastination). That led me to asking a biophysicist, and therefore I became a biophysicist instead of condensed matter/QI/QC. In hindsight, I think this was fortunate. I would've never considered biophysics, which has been one of the loves of my life since then. Who knows, maybe I would've been just as happy with quantum stuff. I'm working through Mike and ike now and find it fascinating.

Funny enough, after my PhD, I co-founded a startup in industrial control & automation. Now I understand feedback quite well, and thus current sources, albeit many years too late.

(1) Of course, good voltage sources vary their resistance just like good current sources vary their voltage. My best guess as to the reason I was more bothered by the current sources is that I was so familiar with voltage sources with confidently claimed constant voltages (batteries). Not a very good reason, I should've questioned it more. In practice, it's much easier to make a near ideal voltage source (very high resistance) than a near ideal current source (0 resistance).

Ideal current sources would also be pretty terrifying. Set it to even a "measly" 50mA and go poke someone with the output...

No need to poke somebody. Just turning it on would cause it to ramp up until air breakdown at the output electrodes, where you'd have a 50mA plasma!

"Of course, good voltage sources vary their resistance just like good current sources vary their voltage. "

Did you mean to say "good voltage sources vary their current just like good current sources vary their voltage"?

(I know nobody really cares, and I promise I'll seek help for whatever neurological condition I seem to have.)

There are two different ways to produce a voltage source. The first is to put a variable resistor in series with the power source, which you can adjust to control the voltage. This works, but dissipates a lot of heat. The second is to put a switch in series with the power source and a capacitor in parallel with the load, then switch the power source on and off very rapidly and use the duty cycle to control the voltage. This is how modern switching power supplies work. In actual practice, there are also inductors in the circuit which cause resonance, and allow the switching to happen when there is no current flowing through the switch. This is how modern switching power supplies can be so efficient.

Ok, thanks. But I was calling attention to a point in the previous comment that makes it difficult to see a kind of "dualism" between idealized current sources to idealized voltage sources. Idealized current or voltage sources don't necessarily have any series resistance, and it doesn't matter how they're realized.

Your point is well taken. My guess was that the OP had implementation in mind when they wrote "good voltage sources vary their resistance".

Or that might have just been a mistake.

I'm certainly no electrical engineer, and I see how the nice duality was lost in my description. Yes, I was thinking about the ones I'd actually worked with, where the resistance was the control knob. I haven't done much electronics since, so my recollection isn't perfect. I have recently been doing some esp-32 control projects where I just used some power supply I bought. I should look into how it works!

Edit- I just looked up switching power supplies and remembered that I did actually know about those!

It's best not to assume the motives behind a rejection unless they align with things you already see as shortcomings. Doing so might cause you to over-correct on things that weren't problems. Being incorrect in a discussion or getting an A- in a class doesn't seem like things that would completely turn a professor off from accepting you.

correlation is not causation

Interesting, this constant current source notion.

For a design without feedback, and in an energetically inefficient way, maybe this can work too:

1. Determine what will be the maximum resistance of the "current consumer" part of your circuit throughout its operation.

2. Prepare a resistor several magnitudes larger than the resistance above.

3. Connect to the resistor above a (huge) voltage source so that the resulting current is the one you target for your current source

4. Put the "current consumer" part of your circuit in series with the large resistor.

It works. In most applications it would be wasteful of electrical power (and money). You have to generate this very large voltage that (it turns out) you don't really need.

On the other hand that circuit is very easy to understand and build and test.

This is pretty much what we do to apply small bias currents to our superconducting circuits. The signals are small (<1 uA), and the power is dissipated outside of the cryostat, so this method is very simple and effective. The voltage and resistors don’t even need to be that huge, ~10 MOhm or below, and correspondingly, <10 V.

Oops, meant "understand feedback quite well*", can't edit for some reason?

FYI, I think the time limit for editing is 1hr. It might change if someone replies? I'm not sure that these rules are written down anywhere.

@Dang, if you see this can we add to the FAQ?

There's also a delay setting in your profile that delays the message but has a max of 10 minutes. There's also the noprocrast setting to help reduce procrastination. The default values let you in for 20 minutes at a time and don't allow you back in for 180 minutes.

I did a s/while/well in there. Hope that's ok.

All: you are welcome to email hn@ycombinator.com if you need something like this after the editing window has passed

Reply on news.ycombinator.com