God-like semiconductor gurus stand in judgement
I saw the first article on the SemTest semiconductor test set in the February 2012 issue. So why is Jim Rowe spouting that silly line that BJTs are “current operated” and FETs are “voltage operated”? That is soooo last century.
What does it really mean? Is it really true? I don’t believe so and I cite as expert witnesses Peter J. Baxandall and Douglas Self. PJB is, like, God, so you know I’m right. Horowitz and Hill only ever talk about either in terms of transconductance.
The input variable for BJTs, FETs, and IGBTs is the base-emitter or gate-source voltage, and the output variable is the collector or drain current. In fact, the same model should adequately describe thermionic valves too. What distinguishes all these devices is the magnitude of their various parameters.
BJTs would better be described as low-input-impedance high-gain transconductance devices and FETs as high-input-impedance low-gain transconductance devices. The primary gain parameter of a BJT is its
“gm” and that of a FET is its “gfs”.
But wait! What’s this? On page 45 under “Testing Mosfets” you write that BJTs are “transadmittance device
”? And what is transadmittance? It’s the complex form of transconductance, ie, the ratio of output current to input voltage. In what sense can that be called “current-controlled”? Maybe you’re a bit confused. I certainly think this mode of thinking is confusing.
Does this mean the SemTest doesn’t do a proper job of measuring BJT gain? No. The fantastic thing about BJT transconductance is that it is determined by physics and its collector current. It is so well described by the Schockley diode equation, that they form the basis of log and antilog amplifiers that operate with precision over six to eight decades of current. There is simply no need to measure BJT transconductance.
FETs are another matter entirely, since their transconductance can be
affected during manufacture. Curiously, diodes don’t follow the Schockley diode equation as well as BJTs.
The SemTest looks like a thoroughly comprehensive instrument. What took you so long?
Jim Rowe remains defiant: you can quote all of your God-like gurus if you like but I still stand by my simplified explanation of BJTs as a current-controlled device, as most people still think of them this way in practice. What’s more, most people still use current gain or hFE as the most useful “quick measure” of a BJT’s health and its suitability for the majority of practical applications.
Time for 12V lighting in houses?
Is this a crazy idea with an obvious flaw? Your 10W LED floodlight project in the February 2012 issue got me thinking. I thought a couple of those would be great to replace the two 48W CFLs in my shed. Just plug a computer type 12V DC power supply into a power point and chain two or three 10W LED floodlights off it. Then I thought, why stop at the shed? Maybe power-efficient LEDs have gotten cheap enough that its no longer necessary and appropriate for lighting to be based on 240VAC.
All the lights in your house are on a separate circuit, right? So you disconnect the lights circuit in your house from its circuit breaker in the meter box, connect a 12VDC power supply to that circuit instead and replace all the 240VAC light bulbs in your house with LED lights like the 3.6W K318s that Oatley Electronics sell for $6 each.
Sure, you’d have to disconnect the circuit safely but it’d mean you’d have very power-efficient lighting with long-life globes using all the existing wiring and switches already in the house. And you could maybe go one step further and run the light circuit off a battery that’s charged off a solar panel – they’re getting cheaper – with an automatic switch-over to the 240VAC to 12VDC power supply if the battery charge level dropped too low. No rocket surgery in any of that. It’s all safe low-voltage stuff only connected to the mains through a professionally-designed 12VDC power supply.
I don’t know whether it’s still possible to get those bayonet light socket plugs so the LED “globes” could be built onto and plugged into the existing socket. That’s the thing I’d be uncomfortable with – a 15VDC maximum “globe” built on a plug that could be plugged into a socket that could have 240VAC connected to it. But by doing it the way I propose you could be sure all the bayonet-type light fittings in your house, at least the fixed ones, would only have 12V DC in them.
Comment: this is an interesting idea and is along similar lines to an idea suggested some years ago in a Publisher’s Letter whereby, since so many electronic appliances can be run from 12V, why not have a 12V grid running through the house? Now that LEDs are getting much cheaper, your idea is more practical.
It also has a big advantage in that you could back it up with a 12V battery bank and then you would have less problems during a blackout. However, there are two problems with using existing switching on 12V DC rather than 230V AC. First, switch and wiring corrosion is bound to more of a problem (it’s bad enough in seaside areas with 230VAC switches and power points). Second, in the event of a short-circuit, a standard 230VAC switch may not be able to break the circuit because an arc may be established across the contacts. This could be worked around by having a suitably-rated DC circuit breaker in the system.