transistor

113 Articles

Frequency Tells Absolute Temperature

January 18, 2023 by 21 Comments

It is no secret that semiconductor junctions change their behavior with temperature, and you can use this fact to make a temperature sensor. The problem is that you have to calibrate each device for any particular transistor you want to use as a sensor, even if they have the same part number. Back in 2011  1991, the famous [Jim Williams] noted that while the voltage wasn’t known, the difference between two readings at different current levels would track with temperature in a known way. He exploited this in an application note and, recently, [Stephen Woodward] used the same principle in an oscillator that can read the temperature.

The circuit uses an integrator and a comparator. A FET switches between two values of collector current. A comparator drives the FET and also serves as the output.  Rather than try to puzzle out the circuit just from the schematic, you can easily simulate it with LT Spice or Falstad. The Falstad simulator doesn’t have a way to change the temperature, but you can see it operating. The model isn’t good enough to really read a temperature, but you can see how the oscillation works

You can think of this as a temperature-to-frequency converter. It would be easy to read with, say, a microcontroller and convert the period to temperature.  Every 10 microseconds is equal to a degree Kelvin. Not bad for something you don’t have to calibrate.

Thermistors are another way to measure temperature. Sometimes, you don’t need a sensor at all.

It’s Not Easy Counting Transistors In The 8086 Processor

January 16, 2023 by 15 Comments

For any given processor it’s generally easy to find a statistic on the number of transistors used to construct it, with the famous Intel 8086 CPU generally said to contain 29,000 transistors. This is where [Ken Shirriff] ran into an issue when he sat down one day and started counting individual transistors in die shots of this processor. To his dismay, he came to a total of 19,618, meaning that 9,382 transistors are somehow unaccounted for. What is going on here?

The first point here is that the given number includes so-called ‘potential transistors’. Within a section of read-only memory (ROM), a ‘0’ would be a missing transistor, but depending on the programming of the mask ROM (e.g. for microcode as with a CISC x86 CPU), there can  be a transistor there. When adding up the potential but vacant transistor locations in ROM and PLA (programmable logic array) sections, the final count came to 29,277 potential transistors. This is much closer to the no doubt nicely rounded number of 29,000 that is generally used.

[Ken] also notes that further complications here are features such as driver transistors that are commonly found near bond wire pads. In order to increase the current that can be provided or sunk by a pad, multiple transistors can be grouped together to form a singular driver as in the above image. Meanwhile yet other transistors are used as (input protection) diodes or even resistors. All of which makes the transistor count along with the process node used useful primarily as indication for the physical size and complexity of a processor.

A Transistor? Memory? Wait, It’s Both!

December 18, 2022 by 13 Comments

What do you get if you cross graphene, hexagonal boron nitride, and tungsten diselenide? Well, according to researchers at Hunan University, you get a field effect transistor that can act as both a switching element or a memory cell. The partial floating-gate field-effect transistor or PFGFET uses 2D van der Waals heterostructures to deal with isolated atomic layers. The paper in Nature is unfortunately behind a pay wall, but you can read a summary over on [TechExplore].

The graphene acts as the gate, and the transistor can be switched between n-type behavior and p-type behavior. It can also be configured as a switching element or as a memory element similar to an EEPROM cell.

One advantage of having configurable transistor types is that a single transistor structure can produce CMOS or complementary circuits. Traditionally, a CMOS IC has two different transistor structures and often producing one of them requires extra effort.

The configuration takes place by applying a control voltage pulse. A negative control voltage produces a p-type FET and a positive voltage configures the same transistor as an n-type. If you don’t have access to the paper, the figures available online offer a good bit of insight into the device’s design.

If you want to learn more about ordinary MOSFETs, we talk about them often. You can also get the skinny on CMOS from [Bil Herd].

When Only A TO92 Will Do

November 5, 2022 by 17 Comments

As through-hole components are supplanted by their surface-mount equivalents, we’re beginning to see the departure of once-common component form factors. Many such as the metal can transistors became rare years ago, while others still hang on albeit in fewer and fewer places. One of these is the once-ubiquitous TO92 moulded plastic transistor, which we don’t see very much of at all in 2022. [Sam Ettinger] is a fan of the D-shaped plastic blobs, and has gone as far as to recreate them for a new generation to enjoy.

Though a TO92 was a relatively miniature package in its day, it’s still large enough to easily fit a SOT23 or similar SMD packaged device on a small PCB. So the tiny board with just enough space for the part and the three wires was fabricated, ready for encapsulating. Epoxy moulding a TO92 gave very poor results, so instead an SLA print of a T092 shell was made. It fits neatly over the PCB, producing a perfect TO92 package. We’re sure a translucent pink package would have raised a few eyebrows back in the 1960s though.

There will come a time when restorers of old electronics will use and refine this technique to replace dead components. We’ve seen the technique before, after all.

Practical Transistors: JFETs

June 2, 2022 by 16 Comments

Transistors come in different flavors. Tubes used an electric field to regulate current flow, and researchers wanted to find something that worked the same way without the drawbacks like vacuum and filament voltages. However, what they first found — the bipolar transistor — doesn’t work the same way. It uses a small current to modulate a larger current, acting as a switch. What they were looking for was actually the FET — the field effect transistor. These come in two flavors. One uses a gate separated from the channel by a thin layer of oxide (MOSFETs), and the other — a junction or JFET — uses the property of semiconductors to deplete or enhance carriers in the channel. [JohnAudioTech] takes a decidedly practical approach to JFETs in a recent video that you can watch below.

The idea for the FET is rather old, with patents appearing in 1925 and 1934, but there were no practical devices at either time. William Shockley tried and failed to make a working FET in 1947, the same year the first point-contact transistor appeared, which was invented while trying to create practical FETs. In 1948, the bipolar junction transistor hit the scene and changed everything. While there were a couple of working FETs created between 1945 and 1950, the first practical devices didn’t appear until 1953. They had problems, so interest waned in the technology while the industry focused on bipolar transistors. However, FETs eventually got better, boasting both very high input impedance and simplified biasing compared to bipolar technology.

Continue reading “Practical Transistors: JFETs” →

suspended carbon nanotube

Falling Down The Carbon Rabbit Hole

January 11, 2022 by 18 Comments

Research projects have a funny way of getting blown out of proportion by the non-experts, over-promising the often relatively small success that the dedicated folks doing the science have managed to eke out. Scaling-up cost-effectively is one of the biggest killers for commercializing research, which is why recent developments in creating carbon nanotube transistors have us hopeful.

Currently, most cutting-edge processes use FETs (Field Effect Transistors). As they’ve gotten smaller, we’ve added fins and other tricks to get around the fact that things get weird when they’re small. The industry is looking to move to GAAFETs (Gate All Around FET) as Intel and Samsung have declared their 3 nm processes (or equivalent) will use the new type of gate. As transistors have shrunk, the “off-state” leakage current has grown. GAAFETs are multi-gate devices, allowing better control of that leakage, among other things.

As usual, we’re already looking at what is past 3 nm towards 2 nm, and the concern is that GAAFET won’t scale past 3 nm. Carbon Nanotubes are an up-and-coming technology as they offer a few critical advantages. They conduct heat exceptionally well, exhibit higher transconductance, and conduct large amounts of power. In addition, they show higher electron mobility than conventional MOSFETs and often outperform them with less power even while being at larger sizes. This is all to say that they’re an awesome piece of tech with a few caveats. Continue reading “Falling Down The Carbon Rabbit Hole” →

Miller (Effect) Time

July 21, 2021 by 10 Comments

While the Miller effect might sound like fun, it is actually the effect of parasitic capacitance in amplifiers. What do you do about it? Watch the video below the break from [All Electronics] and find out. We like how the test circuit it uses has a switch to put the mitigation circuitry in and out of the test for comparison purposes.

Actually, the Miller effect can refer to any impedance but in practice that is most often parasitic capacitance because of the construction used for tubes and transistors. The sometimes tiny capacitance gets multiplied by the inverting gain of the stage and increases the amplifier’s input impedance. This, in turn, reduces the bandwidth of the stage.

Continue reading “Miller (Effect) Time” →