Grounded gate --or grounded base-- amplifier stages are very useful!
A cascode is a grounded source stage followed by a grounded gate stage. That greatly reduces its input capacitance, and so helps to extend its useful bandwidth. I use cascodes extensively in beam instrumentation amplifiers in particle accelerators.
I also like Norton amplifiers, which are a grounded base stage with transformer feedback. Those can be designed to have a real 50 Ohm input resistance, excellent 3rd order intercept, low noise and a flat gain over a very large bandwidth. High quality radio receivers invariably have some version of this concept in the first IF stage. There are examples in the ARRL handbook.
The theory makes assumptions that are sometimes unphysical, specifically that the source can be accurately described by a Thevenin equivalent circuit of an infinitely-stiff voltage source in series with a fixed source impedance. In situations where that's more or less true, the theory works fine, but in real life that's not how you design circuits when efficiency matters.
The problem with that assumption is maybe easier to see if you consider an op amp. Near DC, a unity-gain follower probably has an output impedance of a milliohm or so, and might have an output swing of 12 V, depending on the supplies. With the prof's assumptions, the maximum available power would be
P_matched = (12 V)**2 / 0.004 ohms = 36 kW.
Maybe not. ;)
Knowing a bit of symbolic logic is good for the mind, anyway, and can help in all sorts of situations. Last year I designed a temperature controller that uses wire-NOR to save parts in a four-quadrant current limiter. (The four quadrants had to have unequal limits, for TE cooler datasheet reasons.)
Cascodes are super useful, I agree. A barefoot common-gate design requires a bit more nuance, though. (BJTs generally make much better cascodes than FETs, anyway.)
Because the input current goes into the output circuit, the device loads down the input far more than it would in a nice unilateralized common-source configuration. In CS, a matching network can multiply the input voltage by a factor Q**2+1, and furthermore the noise in the drain circuit doesn't get dumped into the input. So with a low-level signal in a quiet background, a narrowband CS amp is going to beat CG.
In communications service, especially at HF and below, you don't have a quiet input at all, and you want your front end to have fairly low gain and match to your antenna over some reasonable bandwidth without adjusting the tuning. For that, a FET with a 50-ohm source impedance (g_M ~ 20 mS) can be good medicine.
In instruments, common-gate amps are pretty rare IME, though I've often used common-base input stages to improve op amp TIAs. That's more like the cascode situation, though, since your average photodiode looks like a current source with a lot of shunt capacitance.
Noiseless feedback sure is attractive sometimes, especially when there's no antenna attached. In my world that usually matters most in very high-Z circuits, where transformers aren't much use. I usually wind up doing it by using capacitive feedback in the front end and fixing the frequency response in the second stage.
The antenna really does look like a Thevenin source, so you want the line matched to the antenna, right. That helps in both Rx and Tx. It's especially important for transmitters with solid-state final amps, which can get blown up by a sufficiently bad impedance mismatch.
The transmitter does not look like a Thevenin source, though--you want it to have a very low output impedance to avoid dissipating power in the transmitter. Otherwise no transmitter (or audio amp, or motor driver, or what have you) could be more than 50% efficient.
I had Spice in mind - which is way more complex internally.
I find that *very* hard to believe. Did they sleep through their lectures?
Sounds to me much more like what EE's do to software.
And some very big software has latent issues that only ever get found when some unusual error occurs. Often such latent faults lurk in dark corners of error recovery or reporting code.
We found one in VAX VMS once due to someone (a beginner) writing a programme that opened new IO channels every time around a loop. Eventually it ran out of them at system wide level and what was the first thing the error handler did?
Try to open an IO channel to the system console and report the failure.
I note that US domestic flights are down today due to a probable software SNAFU. It is being reported as a computer failure.
I don't entirely trust Spice. Once I found a crash bug LT Spice that Mike fixed in hours.
formatting link
I like that idea, using a simulator to train your instincts. I do disagree with Mike about some of his other no-nos.
Don't know. I never took any computer courses.
(Nor digital design, nor control theory)
I've wondered how awful it is to restart a giant system, credit card or air traffic or Amazon or something, after it crashes. Some programmers vacations and sleep cycles are probably affected.
At the classic maximum power transfer point, the generator dissipates as much power as the load. So efficiency is at best 50%. You can do that in small systems, like an RF MMIC driving a cable or something.
One can systhesize a fake source impedance in a switcher amp, at high efficiency. Carefully.
IC distributed amplifiers are, I think, always cascodes. Fun and expensive parts. I used a lot of HMC659's in a recent project. I modestly note that the results made international press.
Nortons were a fad for a while. I never understood why. Do people still make them?
Darlington MMICS are great. Some have NFs below 1 dB.
Those stupid noisy crappy LM3900/LM359 things were the fastest you could get, for a brief while. Then (apart from their zombie spawn in the National Linear Apps book) they pretty well sank from sight.
On a sunny day (Wed, 11 Jan 2023 08:26:56 -0500) it happened Phil Hobbs snipped-for-privacy@electrooptical.net wrote in snipped-for-privacy@electrooptical.net:
It all depends. A class C output stage for FM A class B output stage (linear amplifier) for for example SSB etc etc
Most of the time matching the load impedance to the electronic optimum designed output stage is done with a so called PI filter (look it up). It has the virtue of filtering out harmonics too. Simple transformer works too.
I had a 1 kV tube linear (2000 Vpp at 250 mA) matched to a few hundred Ohm antenna with a simple few turns around the output tuning coil.
The Pi filter matching network you find in almost every transmitter these days, matching it to a 50 Ohm cable..
Then there are 'antenna tuners' to make the antenna look more like a given impedance (say 50 Ohm) so that you can use almost anything as antenna.
RSGB radio handbook ... Or just look at some transmitter circuit diagrams.
And then there are those HUGE ones with tubes in towers...
And very small very high frequency ones in for example cellphones. And wave guides and dishes and cavities....
Sometimes it's quite bad. In the worst cases, _very_ bad, up to company-wrecking bad.
One sort of problem I've observed amounts to "circular dependencies" creeping into an infrastructure over time. As software is upgraded (or, at least, made more featureful and complex) and slotted into an already-running infrastructure, you can develop situations in which a bunch of critical pieces of the system are all dependent on one another; you can't start one unless the others are up and running and stable. This means that you can restart any one of them safely, but if they're all (or mostly) down, it may be impossible to find a startup sequence which actually works and won't deadlock.
It's a bit like the "start up with limited battery power" portrayed in "Apollo 13". If you're lucky, there _may_ be a cold-start sequence which actually works, and doesn't overload or freeze up. If not, you're screwed.
This sort of problem can creep into an infrastructure silently. An apparently-harmless addition of a feature, or a change to the implementation of a library may add a new dependency between modules which works OK during a normal upgrade, but which deadlocks during an infrastructure cold-start.
Preventing this requires understanding (and documenting, and validating) the actual dependencies, and careful planning. More than occasionally, it makes the addition of new features much more expensive than was planned for, because you end up having to re-factor a whole bunch of infrastructure in order to avoid introducing a new circular-dependency loop and creating deadlocks. Since most software these days consists of dozens of layers of complex libraries and toolkits layered on one another, it's often hard even to understand the possible problems and risks adequately.
Anecdotally, there were press reports to the effect that (at some point after Musk took control) Twitter's server infrastructure came very close to a meltdown... and Twitter's fleet engineers were really concerned that it would be impossible to restart the fleet after a meltdown/shutdown. This was (if I recall properly) due both to the complexity of the system, possible deadlocks, and the "We've lost too many critical people, and those who are left don't really understand how it all works" problem.
Bingo. I told him that the cooling pond of the nearby shortwave transmitter would evaporate within a few hours if the TX effcicieny were that low. And it would be less than 50% with his idea. Not to even mention the cost of electricity being almost double.
In Germany we always used their expressions of plug and jack for connectors. One document was in English and it mentioned a female connector. The guy reading that passage aloud blushed a bit ...
That's part of the aspect. Industry buy-in wasn't great, especially not in small business which is what I always perferred. This is one of the teachable situations, to find such a temporary assignment even while the odds are against you. University isn't just to teach science, they are also there to teach perseverance and toughness. IMNSHO, and I personally know ivory-tower academians who absolutely don't believe that.
Excellent! Every engineering candidate should do that.
One of my jobs came with paperwork that contained a passage "In case of your death or permanent disability an insurance policy has been ..." Unfortunately my mom saw that when I was visiting my folks and had the stuff with me. She just about freaked out and asked me to reconsider. I didn't and during one storm several semi-submersible oil rigs got into major trouble. Ours almost did. Since that made the evening news my mom became really worried but because the antennas were down nobody could message their relatives that we were ok. No satellite stuff back then.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.