Wikipedia:Reference desk/Archives/Science/2017 August 29

= August 29 =

Frequency dependent amplification?
In case of amplification of alternating current by a transistor, is the amplification dependent on the frequency of the source?Sayan19ghosh99 (talk) 01:13, 29 August 2017 (UTC)


 * Yes. Transistors have frequency limits. If you are working with audio or power, you can pretty much ignore this, but get into radio frequencies and it becomes important, get into microwave frequencies and everything changes, and by the time you light frequencies even the wires that connect to the transistor cannot transmit a signal. --Guy Macon (talk) 02:16, 29 August 2017 (UTC)
 * Although there's also the issue of Q, considered very simply as a ratio between the signal bandwidth and the carrier frequency.
 * For audio work, these are comparable frequencies - the bandwidth might go from near zero to a few 10s of kHz. Old analogue phones with carbon mics were just 300Hz–3.3kHz. For AM speech radio though, you might have a similar bandwidth but on a signal of around 1MHz.  So although at higher frequencies the transistor's bandwidth might become seriously limited as a fraction of the carrier, that could also be far more than the bandwidth that the signal requires, so the design problem becomes relatively simple (if you have MHz transistors, then modulating an audio signal onto that isn't a big deal).
 * Back in the early days (1960s) of transistors with physically large junctions, even good audio design could be a problem as the transistor gain was rolling off noticeably within the bandwidth of the signal. Amplifiers were designed to require less gain than this, and to use negative feedback as a control mechanism instead of being simply gain limited, but even so they might run out of gain at the top end, because the raw transistor could just no longer achieve it.
 * Spread spectrum radio changes this simplification. RF circuits now need massive linear bandwidth as a proportion of their carrier, because the bandwidth used becomes so broad - far in excess of what the simple modulating signal might suggest. So they can become a difficult design problem.
 * In radio there's also the related, but distinct, issue of the linear amplifier. Is the transfer function of the amplifier linear across a range of signal powers? Now that's not needed for FM modulation, as transmitter power remains constant as the signal is modulated in frequency (but it needs your original point to be stable). But for AM (or SSB) modulation, the amplifier must also remain linear in response as the signal amplitude changes. This makes a crude FM transmitter often easier to design than a crude AM transmitter. Andy Dingley (talk) 09:48, 29 August 2017 (UTC)
 * In a bipolar junction transistor, the internal capacitance of the junctions causes gain to decrease with increasing frequency. See Gain–bandwidth_product. A transistor's current gain between the cutoff frequency at which the current gain drops by 3 decibels and the transition frequency fT can be estimated by dividing fT by the frequency. Blooteuth (talk) 10:45, 30 August 2017 (UTC)

While shepherds didn't watch their flocks by night
Absent humans who might induce them to stay awake, will sheep generally graze at night? Iberia Airlines Flight 062 crashed at 10:02 PM and killed a flock of grazing sheep; I don't suppose shepherds were that common in England by the 1960s, so the idea of a flock autonomously feeding at 10PM sounds rather odd. Nyttend (talk) 04:55, 29 August 2017 (UTC)


 * Sheep grazing at night seems unlikely to me, this is confirmed in the sheep article. My understanding of flock animals such as sheep is that they congregate at night (possibly for added protection) and this fact would account for the large number of sheep killed. My assumption is that grazing sheep would be more separated and the strike rate would have been lower. Richard Avery (talk) 07:15, 29 August 2017 (UTC)
 * Agreed - the cited source uses the phrase "grazing sheep" but I suspect this is just shorthand for "sheep on a hill", who would have been grazing it during the day. Sheep also have a tendency to move uphill at night (see here), another strategy against predation. AndrewWTaylor (talk) 07:53, 29 August 2017 (UTC)
 * We don't need to repeat the Fernhurst Society's error in our article, so I'll remove the word.   D b f i r s   06:06, 30 August 2017 (UTC)

Aural boxing?
What exactly is the "aural boxing" technique described in Cessna 188 Pacific rescue? 2601:646:8E01:7E0B:84FD:477A:71AB:7A1E (talk) 08:33, 29 August 2017 (UTC)
 * It's described here. Basically, it involves flying just on the edge of VHF radio reception. Because VHF radio is high frequency, it's more or less line of sight - this means that when the plane slips over the horizon, you lose contact. With a bit of geometry, you can then work out how far away the other plane is. The rescue plane flew in a kind of ⋉-pattern, and by tracking where it made and lost contact, it was able to triangulate where the Cessna was. Smurrayinchester 09:07, 29 August 2017 (UTC)
 * (It's called aural boxing because a) you're drawing a box around the target plane, and b) you're using voice radio and you have to keep talking constantly to avoid radio silence) Smurrayinchester 09:08, 29 August 2017 (UTC)
 * Thanks! 2601:646:8E01:7E0B:84FD:477A:71AB:7A1E (talk) 05:05, 30 August 2017 (UTC)
 * Sounds like a good red link opportunity for someone so inclined... shoy (reactions) 14:37, 31 August 2017 (UTC)

LM317 - conceptual difficulty
"the device is conceptually an op amp..." is it just me or is this bit (and the explanation that follows it) a bit misleading? First, the equation is different for a real non-inverting opamp amplifier (the feedback resistor is in the numerator), and second, an opamp does not work by "adjusting (...) the voltage of the output pin to be a fixed amount (!), the reference voltage, above that of the adjustment pin" [=inverting input in this analogy], it adjusts the current through the feedback resistor to minimize Vdiff between both inputs, in other words both resistors carry the same current and the output voltage is Vout=Vin*(1+Rf/R1), thus the output is Vout-Vin=Vin*(Rf/R1) above (either) input. I see how it's tempting to conceptualize the LM317 as a high-ampacity opamp that just amplifies the 1.25 reference voltage applied to a (non-externalized) non-inverting input, but the analogy seems faulty (why are the equations different?). At the same time, as a hobbyist, I don't know what is wrong with the analogy or where exactly it breaks down 78.53.240.75 (talk) 11:11, 29 August 2017 (UTC)
 * The LM317 sense/feedback input is a preconfigured OP amp. Configuring its output voltage is calculating a voltage divider of two resistors, matching the prefered output divided ratio the internal reference in addition to the power consumption of the LM317 as it does not touch the ground reference potential. The higher the voltage divider is, the more will the output modulated by the internal power consumption, rising the voltage drop of the lower resistor of the voltage divider. For recommended values and calculations, refer the manufacturer's datasheet. -- Hans Haase (有问题吗) 11:33, 29 August 2017 (UTC)
 * the equations are different, though. There's something wrong with the analogy. I'm not arguing that the LM317 wasn't, at some level, an opamp, - most circuits that implement some kind of servo loop are one or contain one. 78.53.240.75 (talk) 11:41, 29 August 2017 (UTC)
 * here's how to get 5V from an opamp and LM317:

720                                                                                     ___                                                      _____                                |-|___|--|                                                 |     |                               |        |                                   12V --|LM317|---|- 240   |        |                                                 |_____|   |                  ___    |  |\    |                                                   |       |              ---|___|---|--|-\   |                                                   |       |              |             |  >--|                                                   |      .-.            ===         ---|+/                                                       |      | |            GND         |  |/                                                        |      | | 240 |                                                           |      '-'                     + ---  1.25V                                                    |       | -                                                           |---|                            |                                                                    |                         |                                                                   .-.                       ===                                                                  | |                       GND                                                                  | | 720 '-'                                                                                            |                                                                                             ===                                                                                            GND


 * see what I mean? 78.53.240.75 (talk) 12:07, 29 August 2017 (UTC)


 * See DS LM317 Chapter 8.2 internal block diagram, which draws a voltage follower which is the classical full coupled servo loop. Is think: YES! -- Hans Haase (有问题吗) 12:14, 29 August 2017 (UTC)
 * This discussion belongs at Talk:LM317 where I shall express agreement with User 78.53.240.75. Blooteuth (talk) 13:13, 29 August 2017 (UTC)
 * Thank you! When it occurred to me the equations were different, I thought "huh?!" 78.53.240.75 (talk) 17:38, 29 August 2017 (UTC)
 * A servo loop is coupling the OP amp output to its negative input. The voltage follower is an OP amp with output direct fed into the negative input. It is also the basic circuit of linear voltage regulators. The data sheets of LM317 draw such circuits to make engineers understand how the device works. Such diagrams not not really differ for the real internal circuit. Maybe to save transistors or compensate some unwanted effects, the real circuit differs for the published diagram. When archived, please have the article discussion a link to this section. -- Hans Haase (有问题吗) 13:04, 1 September 2017 (UTC)