2642: Meta-Alternating Current

Explain xkcd: It's 'cause you're dumb.
Revision as of 14:20, 18 March 2024 by (talk) (Explanation already consistent, marked complete.)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search
Meta-Alternating Current
It's always bothered me that you can't cancel out an inverter by putting a second inverter after it.
Title text: It's always bothered me that you can't cancel out an inverter by putting a second inverter after it.


This is the sixth installment in the series of Cursed Connectors and presents Cursed Connectors #120: Meta-alternating current extension cord.

Direct current is a unidirectional flow of electrons from a power source to something being powered, through one or more conductors, before returning to the power source via one or more other conductors, thus completing the circuit. Batteries produce direct current. It is commonly used in electronics applications, including computers. Alternating current, on the other hand, frequently reverses the direction of electron flow, and is commonly used for longer-distance transmission (such as from the power plant to an outlet).

This comic proposes a humorous Meta-Alternating Current, which uses a series of adapters to "alternate" between DC and AC current along the length of a connector. This is absurd in part because typical power inverter efficiency is 90%, and maximum bridge rectifier efficiency is about 99% for 120 V,[actual citation needed] so an extension cord made in this manner would lose about 11% power (compounded) per such pair. For the wire shown in the comic, with seven pairs, the efficiency would be 0.897, which is 0.45, that is, 45%.

The title text bemoans that an inverter, which converts direct current to alternating current, does not work in the other direction, as a layman's interpretation of the word "inverter" might assume. Rather, a separate device, a rectifier, also pictured in the comic, must be used for this second conversion. (However, a similar circuit to an inverter may be used to rectify in a process called active rectification.)


[Above the drawing there is a header. Below the header is a chain of rectifiers and inverters. There are seven pairs of rectifiers (the smaller ones) and inverters (the big chunky ones). The first five inverters have arrows pointing to them from a label above the chain, whereas the last five rectifiers are labeled below the chain, but also with arrows pointing to them from that label. The chain starts on the left with a mains plug (type B), followed by the first rectifier. It ends with the last inverter, and a female mains plug (type B socket) to the right. Beneath the chain there is label for the connector and an explanation in brackets below that.]
Cursed Connectors #120
Meta-alternating current extension cord
(alternates between AC and DC)


  • This comic replaced No One Was Hurt as comic 2642 after it was taken down from xkcd's website.
    • See why on that comic's explanation.
  • This comic initially did not have a high-resolution (2x) image for high-DPI displays, making it look more blurry than usual on such devices (such as smartphones). This was probably because it was uploaded hastily to replace No One Was Hurt quickly. It has since been updated to include a high-resolution version.

comment.png add a comment! ⋅ comment.png add a topic (use sparingly)! ⋅ Icons-mini-action refresh blue.gif refresh comments!


And today, we are reminded that Randall used to be a physicist (or at least has a physics degree). Not worth mentioning in the article, but while inverters can't reverse each other, transformers can. (Has Randall done the transformer/Transformer pun yet as an excuse to mock the movies?) Nitpicking (talk) 11:10, 7 July 2022 (UTC)

I haven't picked up the physics reference yet. I see electrical engineering here. Randall strikes me as somebody who would study physics given the opportunit, though. It's notable that this webcomic started while Randall was in college, if I recall right. 11:58, 7 July 2022 (UTC)

Long distance links, especially those between separate unsynchronized grids, use high voltage DC. There is a 2,000-mile link in China running at 1 MV. Arachrah (talk) 11:32, 7 July 2022 (UTC)

That's because at great distances, relatively high frequency AC loses a lot of ("active" = actually useful) power as ... reactive power, I think (didn't learn the terminology in English, unit seems right though). A typical grid has a lot of generators and load. A long distance connection results in a phase shift according to the transmission time (speed of light in medium x distance) in about the order of magnitude of the AC period (usually somewhere between 1/10 to 1/60 seconds) wastes a portion equal to the sine of the phase shift angle (up to 90° = all of it) as reactive power. DC isn't quite as easy to use but on long distances there is no power loss to reactive power. 627235 (talk) 12:25, 7 July 2022 (UTC)
Do you have references? That doesn't seem right to me, speed-of-light lags ought to simply place widely separated power stations at different phase angles which appear the same, without causing a clash. And even in cases where there is an irreconcilable mismatch (say, if there's a loop flow), it's possible to use a "phase-angle regulating transformer" to match the phases. -- 06:27, 14 January 2024 (UTC)

Let's face it, the thing should be called an alternator. Of course that name's taken as a redundant word for (electrical) generator. 627235 (talk) 12:26, 7 July 2022 (UTC)

Alternator and Generator shouldn't be used interchangeably. At least, in automotive, aerospace and industrial discussions, a generator is normally thought of as a DC device while an alternator is AC, even if we usually rectify it's output to 12 or 24VDC... SwervingLemon (talk) 19:33, 9 July 2022 (UTC)

It always bothered me that UPS battery backups take the wall AC and convert it to DC to charge the battery, but then have to turn it back to AC to send it to the computer, so the power supply can convert it to DC to run the thing. I picture some connector that goes directly from the UPS to the power supply so that if power is lost it can just pull 12V directly from the battery. Andyd273 (talk) 12:47, 7 July 2022 (UTC)

some UPSs do this. They normally power the computer directly from the input AC, but if there is a power failure, they use the battery to power the inverters and switch the output to the inverter. Other UPSs always power the computer from the inverter. They have the advantage that there is not even a milisecond time to start powering the computer. That can be better for some equipment, and that kind of UPS often costs more. It is also worth noting that in some data centers, they bypass the AC step and have one big DC power supply that directly powers the computers. WhiteDragon (talk) 16:49, 7 July 2022 (UTC)

NOT logic gates are also often known as inverters. An even number of those would indeed produce the same output as the (true/false) input. 16:03, 7 July 2022 (UTC)

“Further chaining this into more inverters/rectifiers would normally not be considered.” Well, if you take a DCC controlled model railway for camping, you get a second stage of inverter/rectifier. The power supply of the DCC control station usually expects AC input, so you invert the DC of the car battery. The PSU then makes DC for the control station processor, which is then made AC to generate a DCC signal. The locomotives always have a rectifier to get a DC power supply from the DCC signal (which is confusingly AC).

You get a third stage with another plausible trick: put the battery on a car battery charger, which converts AC from the camping site power grid to DC. Then use a locomotive with a (rarely used) BLDC motor, which confusingly needs an inverter generating AC. -- 18:01, 7 July 2022 (UTC)

The efficiency calculation is bogus. For the rectifier, the "efficiency" of 81% relates to voltage, not power. (See https://en.wikipedia.org/wiki/Rectifier). I don't know what the power efficiency is, but I do note that my computer's power supply is not glowing white hot.

Where do you propose the extra current to make up for such difference would come from? 21:38, 7 July 2022 (UTC)
When you convert AC to DC or DC to AC, you can't qualify either with single number for voltage OR current. You need to examine whole graph, because both voltage and current are changing with frequency of (original) AC. I'm pretty sure that 81% figure is related to the different way how voltage is computed for AC and DC. That said, regarding the gloving power supply ... usually, power supply contains fan, and while it's partially used to cool the case interior, the power supply might not like it being stopped either. -- Hkmaly (talk) 22:24, 7 July 2022 (UTC)
Your computer's power supply has switching H-bridge MOSFET transistors that sense and match the AC phase, not an ordinary diode rectifier. 23:25, 7 July 2022 (UTC)
The mains input bridge rectifier in every PSU I've seen is just a normal 4 diode bridge, it's at worst a 2% loss. The output from the transformer probably has synchronous rectification due to the much lower voltage, but not an H-bridge, instead a center-tapped transformer winding and 2 MOSFETs. The only H-bridge in a PSU is the inverter. 21:43, 8 July 2022 (UTC)
Please see https://www.physics-and-radio-electronics.com/electronic-devices-and-circuits/rectifier/bridgerectifier.html "The maximum rectifier efficiency of a bridge rectifier is 81.2% which is same as the center tapped full wave rectifier." See also https://techweb.rohm.com/knowledge/acdc/acdc_pwm/acdc_pwm06/8786 19:36, 16 July 2022 (UTC)
Sigh, that's for 5V DC output. A rectifier isn't some magic device that wastes a fixed percentage of power, it's just some bloody diodes. Your first "source" is rubbish, and your second source is for the output rectifier for a 5V DC power supply. 22:56, 24 July 2022 (UTC)

I interpreted Randall imagining "inversion" as "reciprocal" (or maybe the inversion of a function) rather than physically turning something upside-down, since mathematical inversion is typically reversible... 03:48, 8 July 2022 (UTC)

Should the description also comment on the choice of number for this "connector". Obviously related to the fact that wall outlet voltage in the US is frequently (pun intended) 120 Volts. MAP (talk) 14:41, 8 July 2022 (UTC)

Two inverters *can* cancel each other out if they are the simplest type (a commutator, a.k.a. square wave inverter) *and* they happen to be synchronous. Expect glitches at the commutation points though! 16:39, 8 July 2022 (UTC)

Totally irrelevant to the comic. 20:58, 8 July 2022 (UTC)
How is it irrelevant?
The comic is about "meta alternating" pairs of consumer inverters and rectifiers, not idealized chains of the same circuit. 00:58, 9 July 2022 (UTC)
The title text is about pairing two inverters and having them cancel each other out. It's relevant. SwervingLemon (talk) 19:33, 9 July 2022 (UTC)

And another one marked "Done" (by removing the Incomplete bit) and then the same editor making further changes (and then correct their change). Wrong way round, mate! Make sure it's correct, make changes you really feel need making and then subsequentally (ideally after checking your addition) make your bold proclaimation of it being Complete. 22:01, 24 August 2022 (UTC)