Fullscreen HTML5Med Res (??MB)HTML4

Improvised Electronics and DIY Components Part 2 by Dr David Maddison, VK3DSM Image source: www.pexels.com/photo/tools-on-a-desktop-7286026 Last month, we explained how components like capacitors, batteries, motors, diodes, resistors or even transistors can be made from scratch, even using scrap. Some people go much further, creating their own valves, integrated circuits and more. We’ll look into how that’s possible in this second and final instalment. A fter we’ve looked at fabricating more advanced components using equipment and techniques available to the general public, we’ll investigate some of the things that can be built using these components. But first, let’s look at making more advanced components and assemblies. Some complex devices or assemblies that can be made in the garage or shed include the following. Others videos that are worth watching are titled “Homemade Cathode Ray Tube, Transconductance Tube Tester, and other stuff” at https://youtu.be/ ZvZXtEbn1Zk and “Video Display On Homemade Cathode Ray Tube” at https://youtu.be/_PzoAReMXOE An old CRT TV can be converted to an oscilloscope, as shown in the GreatScott! video at https://youtu.be/ aScAZReGQc0 A simpler way to do it with no additional electronics is shown in Fig.36 and at https://youtu.be/vHvbCWVtPzY Cathode ray tubes Cathode ray tubes (CRTs) are a form of valve, so they can be made with similar techniques (more on valves later). They used to be common for televisions and oscilloscopes but are now mostly obsolete. A simple CRT is shown in Fig.34, including the basic principle of operation. YouTuber “jdflyback” has been quite successful in making CRTs (see Fig.35 and the associated video). Fig.34: a simplified diagram of a cathode ray tube as formerly used in TVs, oscilloscopes and radar displays. Source: https://w.wiki/Nv7D 12 Australia's electronics magazine Silicon Chip siliconchip.com.au but this is not for beginners due to the extremely high voltages present (typically over 10kV!). Electron beam lithography For the extremely dedicated DIYer, obsolete scanning electron microscopes (SEMs) can often be acquired very cheaply (for scrap value) or even free. They can be used for their original purpose and also modified to draw patterns for electron beam lithography, as shown in Fig.37. You need plenty of space and a means to move these large devices. By adding a pattern generator, such as a Raspberry Pi or FPGA controlling the scan coils via DACs, and disabling or bypassing the raster scan and using a resist-coated sample stage, the SEM can be turned into a basic direct-write electron beam lithography system. Many hobbyists and small labs have successfully patterned features down to about 50-200nm (sometimes smaller) using PMMA (Perspex) or HSQ (hydrogen silsesquioxane) photoresists. For more on this, see the video by Sam Zeloof titled “Making Tiny Things with Electron Microscope - E-beam Lithography” at https://youtu.be/ SB94rQtKlKI and Peter Bosch’s video on “Shooting electrons at plastic to make microscopic features” at https:// youtu.be/HA9p38AnByY Fig.35: an improvised CRT made by jdflyback. Source: https://youtu.be/ESwjVXjDtY Integrated circuit fabrication Home integrated circuit manufacturing is an extreme frontier level of improvised electronics, where dedicated hobbyists attempt to replicate professional semiconductor fabrication processes in garages, sheds or home labs. They are typically working to a late 1960s or early 1970s standard of technology. This is far beyond simple homemade diodes or crystal radios. The most prominent figure in this niche is Sam Zeloof (https://sam. zeloof.xyz), a self-taught engineer who has turned his family’s garage in New Jersey into a functional (though dated) chip fab. As a high school student in 2016-2017, he built his first working Mosfets and logic gates, then progressed to full integrated circuits. His milestones include: the Z1 (2018), a PMOS dual differential amplifier IC with six transistors fabricated using four photolithography masks (see Fig.38). Fig.36: using an old TV as an oscilloscope. Be careful of high voltages! Source: https://youtu.be/vHvbCWVtPzY ▪ siliconchip.com.au Fig.37: modifying a scanning electron microscope for electron beam lithography. Source: https://sam.zeloof.xyz/e-beam-lithography Australia's electronics magazine October 2026  13 Fig.38: Sam Zeelof’s Z1 chip, a PMOS dual differential amplifier with six transistors. Fig.39: Sam Zeelof’s Z1 and Z2 chips compared. There are twelve 100-transistor Z2 chips on the one piece of silicon for a total of 1200 transistors. ▪ the Z2 (2021) is an upgraded silicon chip with 100 transistors, with 12 chips on the same piece of silicon, so 1200 transistors total (see Fig.39). It is almost comparable in complexity to early microprocessors like the Intel 4004 (which had 2300 transistors). It features smaller, faster devices and better yields than the Z1. Zeloof’s work draws inspiration from vintage patents and textbooks, using salvaged or homemade equipment to perform processes like photolithography, oxidation, doping, etching and metal deposition, proving that 1960-70s IC fabrication techniques can be replicated on a tiny budget with persistence and ingenuity. Other notable hobbyists include Jeri Ellsworth (mentioned last month), who pioneered early home semiconductor work in the 2010s, fabricating thumb-sized transistors and basic logic gates using vinyl-cut masks, rust stain remover (for etching) and simple dopants. Her videos directly inspired Zeloof and many others in the community. Recent (2025-2026) developments include hobbyist-developed spin-on glass dopants using TEOS (tetraethyl orthosilicate)/thermal diffusion for safer phosphorus/boron dopant introduction into silicon and small-scale photolithography setups. There are often shared on the ProjectsInFlight website: www.projects­ inflight.com or on their YouTube at: www.youtube.com/<at>projectsinflight While no-one has yet matched Zeloof’s transistor count or IC complexity at home, these efforts show ongoing progress in accessible doping and lithography. Key equipment and materials 14 Silicon Chip needed for a garage-scale fab (based on Zeloof’s documented setup and similar projects) include: Substrates: silicon wafers sliced from ingots or bought pre-cut, often 2-4 inches (50-100mm) in diameter, p-type for PMOS. Photolithography: a UV light source (eg, a modified DLP projector + microscope optics for maskless exposure), photoresist (spin-coated), photomasks (printed or drawn) plus an alignment jig. Furnace/oxidation: a high-temperature tube furnace (up to 10001200°C) for growing gate oxide and diffusion doping. Deposition/etching: Sputtering system or thermal evaporator for metal layers (aluminium), wet etchants (HF-based for oxide, phosphoric acid mixes) and a plasma etcher (optional but helpful). Doping: Spin-on dopants (eg, homemade phosphorus/boron solutions) or gas sources in advanced setups. Inspection: a scanning electron microscope, an optical microscope ▪ ▪ ▪ ▪ ▪ ▪ and a probe station for testing. Safety/cleanliness: fume hood, acid cabinet, HEPA filtration (to minimise dust), gloves/goggles/respirator. HF acid and very high temperatures are extremely hazardous! Other: Vacuum pumps/chambers, chemicals (acids, solvents, photoresist) and wafer-handling tools. The costs for such a setup can be in the tens of thousands of dollars (mostly for the SEM and furnace) but some put together a lab for much less. The process is labour intensive (Zeloof’s runs took around 12 hours per chip) and low-yield, resulting in features in the 5-10 micron range (1000 times the feature size of modern fabs). This level of home fabrication remains a rare, high-commitment pursuit, far from everyday improvisation. Still, it embodies the ultimate in maker ingenuity, turning a garage into a micro-fab and creating working silicon chips from raw wafers. Some related videos are as follows: • “Photolithography on Silicon with PCB Chemicals”: https://youtu. be/Kx1TenvQXTg ▪ ▪ Fig.42: the structure of a TEA laser. Original Source: https://laserkids. sourceforge.net/eng_co2teaLaser.html Australia's electronics magazine siliconchip.com.au • “Making Spin-On-Dopant for DIY Semiconductor Fabrication”: https:// youtu.be/1dFj-tGn8DI • “Metallisation: Making Conductive Traces on Silicon Chips”: https:// youtu.be/Ddd_-4D76do The OpenSilicon Initiative (https:// github.com/RParkerE/OpenSilicon) aims to make it easier for individuals to design and build their own chips. It offers open-source guides, bills of materials, 3D-printable files, firmware and PDKs (process design kits) to help enthusiasts and researchers build semiconductor-related equipment and tools at home. Its current focus is on prototypes for spin coaters, direct-write lithography (405nm laser-based), metrology systems for film thickness measurement, oxidation furnaces and PVD (physical vapour deposition) systems targeting sub-micron resolution in small-batch processing. While this supports DIY experimentation in chip characterisation, testing and basic processing setups, it complements rather than replaces professional fabrication services like Tiny Tapeout or Wafer.Space shuttles (more on them later). The goal is to allow the maker to fabricate more advanced hardware without requiring a full cleanroom. Lasers A transversely excited atmospheric (TEA) laser can be improvised from some basic components (Fig.42) and a 10kV (or so) flyback power supply, as is common in CRTs. This type of laser can be extremely dangerous for a variety of reasons, so such a project is not recommended unless you understand and are trained to deal with the dangers. Laser safety glasses must be used. There are many online sources of Fig.43: one of the many steps in making a Nixie tube (similar to making any valve). Source: https:// youtu.be/wxL4ElboiuA siliconchip.com.au Fig.40: a home-built TEA laser. Source: https://youtu.be/Zv3DFTSs6NQ Fig.41: Nixie tube based artwork. Source: www.daliborfarny.com/project/ omnixie-clock instructions for building such a laser, such as Laser Kids (siliconchip.au/ link/acc5) and see the video at https:// youtu.be/Zv3DFTSs6NQ and Fig.40. Nixie tubes Nixie tubes are very popular ‘retro’ electronic display devices, but they became obsolete in the 1970s as LED and LCD displays became commercially available. Experimenters and restorers could only buy old stock or used devices until Dalibor Farny decided to make them himself and turn his hobby into a business (www.daliborfarny.com), making Nixie tubes. They are essentially artworks, with a price to match Fig.44: purple Nixie tubes made by jdflyback. Source https://youtu.be/ q7I61d27R3E Australia's electronics magazine – see Fig.41 and the video at https:// youtu.be/wxL4ElboiuA All the essential elements of making your own valves are shown in the video; one is shown in Fig.43. “jdflyback” made their own purple Nixie tubes too – see Fig.44 and the video at https://youtu.be/q7I61d27R3E Optical sensors An advanced DIY semiconductor fabricator used a copper oxide semiconductor on a silicon dioxide coated silicon wafer to make a simple optical sensor (Fig.45). Their video, titled “I tried to make a camera sensor”, is at https://youtu.be/O7xH9ZSp_B4 The fabricator used electron-beam Fig.45: an improvised copper-oxide optical sensor. Source: https://youtu. be/O7xH9ZSp_B4 October 2026  15 lithography to draw the pattern. This was accomplished with the aid of an obsolete and modified electron microscope. Valves (also called vacuum tubes) were common in electronic appliances until about the late 1960s or early 1970s, when they were almost exclusively replaced by transistors and ICs. Arguably the most basic type of valve is the incandescent light globe, which can function as a rudimentary thermionic diode due to electron emission from the heated filament under certain circumstances. A valve diode provides a similar function to a solid-state diode, while a triode is the equivalent of a transistor (it’s most similar to a field-effect transistor). Valves are typically fabricated by hobbyists for fun and experimentation, rarely out of necessity. Still, it is conceivable that one day they might need to be fabricated to restore antique electronic equipment if original valves can’t be sourced. The valve diode was invented by Sir John Ambrose Fleming in 1904 (see Fig.46), while the valve triode was invented by Lee de Forest in 1906 (see Fig.47). The latter can amplify signals or act as a switch by using a control grid to modulate the flow of electrons from cathode to anode. Incandescent globes can be used as very poor improvised diodes according to one report on Quora (siliconchip.au/ link/acba). After burning out one filament of a combined brake and indicator automotive lamp, one end becomes the anode and the other heated filament becomes the cathode. They can even be used as valve triode amplifiers, but likely with poor performance – see Fig.48 and the videos at https://youtu.be/LRLtiOMKIA4 and https://youtu.be/JkrrStcYZLk Improvised valves are typically made using a glass envelope, usually soda-lime glass blown and shaped with a torch, with tungsten wire for the filament (cathode), nickel or nickel-­ plated metal for grids and plates, plus Fig.46 & 47: the structure of a diode valve (left) and triode valve (right). Sources: https://w.wiki/Nv7M & https://w.wiki/Nv7N Fig.48: a dual-filament incandescent bulb can be used as a primitive triode by deliberately burning one filament out. Source: https://youtu.be/JkrrStcYZLk Vacuum pumps Vacuum pumps can be made from modified air compressors or refrigerator compressors, but these cannot achieve a sufficiently good vacuum for making valves. A modified refrigerator compressor can reach 10-50 torr, which might be OK for glow discharges and plasma experiments, but not nearly enough for valves, which require a minimum of 10-4 to 10-6 torr. At minimum, a two-stage mechanical vacuum pump with a gauge is needed. These can reach 10-3 to 10-4 torr, which is sufficient for simple diodes and triodes as long as a ‘getter’ is added to increase the vacuum further (more on this later). A torr is a unit of pressure that’s defined as 1/760th of an atmosphere. Such pumps can be purchased on eBay. Traditional quality brands are Welch, Edwards, Alcatel, Leybold, or Varian; and some recommended less costly Chinese brands, including the 2XZ series or similar. Valve fabrication 16 Silicon Chip Australia's electronics magazine getters, which are reactive materials (often barium or barium alloys) placed inside the valve to absorb residual gases after evacuation. Warning: some barium compounds are toxic. The characteristic silver-coloured coating seen on many valves is the getter flash, a thin layer of barium that reacts with stray gases to maintain the vacuum. It is usually a barium-­ aluminium or barium–magnesium alloy. Modern vacuum systems use zirconium alloys instead, which are also activated by heating but don’t vaporise and absorb gases continuously. Construction requires basic equipment like a vacuum pump, as mentioned above, a blowtorch or glass-working torch for shaping and sealing the envelope, and careful assembly of electrodes and leads. While challenging, hobbyists have successfully built simple diodes, triodes and even small amplifiers this way; see Fig.49 and https://youtu.be/-UEfqAWb3fE Here are some basic steps to build a DIY triode valve. Note that this is a rough guide only. The materials needed are: Envelope: Pyrex or soda-lime glass tubing for the envelope. Filament: thoriated tungsten wire (for high emission) or oxidecoated tungsten wire, about 0.6-1mm in diameter Grid: fine stainless steel or nickel wire, or mesh wrapped in a spiral Plate: nickel sheet or a stainless steel cylinder Feed-throughs: tungsten or Dumet wire (the latter is a special composition specifically for glass-to-metal seals) ▪ ▪ ▪ ▪ ▪ siliconchip.com.au ▪ ▪ Internal supports: mica sheets or spacers to centre and insulate elements Chemicals: barium or strontium carbonate slurry for the filament coating and potassium nitrate for cleaning the tungsten. There are restrictions on purchasing the latter in Australia; sodium hydroxide (NaOH) may be an alternative. Tools needed include an LPG/oxygen torch; carbon shaping tools for the glass; a glass lathe (useful but not essential); a vacuum system, as described above, with associated tubing; a spot welder for joining wires (these can be home-made); and a diamond saw or file to cut glass. The steps are: 1. Fabricate the stem – melt tungsten or Dumet wires into the glass to create airtight feed-throughs. 2. Prepare the electrodes – weld the filament, grid and plate to the stem’s internal leads and use mica to maintain spacing between them. 3. Coat the filament – apply a barium carbonate slurry to create an oxide layer on the filament for better electron emission. 4. Seal the envelope – put the assembly into the glass envelope and fuse the stem where the feed-throughs go through, but leave an exhaust port open. 5. Evacuation and bake-out – connect the exhaust port to a vacuum pump and torch the envelope or place it in an oven for a period to remove residual gases. 6. Once a high vacuum is obtained during the bake-out, melt and collapse the exhaust port to seal the unit. 7. Fire the ‘getter’ with an induction heater. 8. Test it to make sure it works. Hazards include heat from the blowtorch or furnace, harmful chemicals during fabrication, and high voltages and possible X-ray production during testing. You can watch an instructional video about making a triode valve at https://youtu.be/hLEYaV_Nl2Q (also see Fig.50). Charles Alexanian (siliconchip.au/ link/acbb) makes their own valves without the use of any especially harmful chemicals. If you decide to make your own, check the material safety data sheets (MSDS) for any substances you will use before starting. One prominent valve fabricator is “Glasslinger” (Ron Soyland), siliconchip.com.au Fig.49: a homemade amplifier by jdflyback with two DIY triode valves. The other components are vintage types. The grey boxes are repurposed doorbell transformers for a choke and the audio output transformer. Source: https://youtu.be/-UEfqAWb3fE Fig.50: a DIY triode valve. Source: https://youtu.be/ hLEYaV_Nl2Q Fig.51: a partially completed Audion valve made by Glasslinger. Source: https://youtu.be/ PxxLPrVbb-A a YouTuber (www.youtube.com/<at> glasslinger) focused on vintage electronics and valve manufacturing to an extremely high professional standard. One of his videos features the manufacture of the de Forest “Audion” spherical triode invented in 1906; see Fig.51 and video https://youtu.be/ mJgEjghVom0 Glasslinger also has a video of a 1912-style radio, which he made using mostly homemade components, including resistors, capacitors and an Audion valve; see https://youtu.be/U6ZVqr0fPo4 and Fig.51. Fig.52 shows a valve made by another constructor. Hobbyists have built X-ray tubes (see Fig.53 and the video at https://youtu.be/yL2RIzlo7W8) but these should be made with caution due to the possibility of radiation exposure. Australia's electronics magazine Other fabricators worth looking up are Claude Paillard (https://youtu.be/ EzyXMEpq4qw), Charles Alexanian (https://youtu.be/bZN735jtikA), Lea Barker (https://youtu.be/3_2n7fpW­ bXE) and Nick Poole (www.youtube. com/live/39-5WgcvaHk). There are several entries on improvised valves in the Nyle Steiner document (PDF) at siliconchip.au/link/ acbc Some additional videos to look at are titled “First Homemade Triode Vacuum Tube” (https:// youtu.be/ajo8SKXBbr8) and “Radio Built With Homemade Fig.52: a very neatly made, nearly complete DIY valve with the base yet to be installed. Source: https:// hackaday.com/2014/11/21/ artisanal-vacuum-tubeshackaday-shows-you-how October 2026  17 Fig.53 (left): an improvised X-ray tube; the light is from the filament. The emitted radiation can be detected with a Geiger counter. Source: https:// youtu.be/yL2RIzlo7W8 Fig.54 (above): the simplest possible crystal radio (left) and a betterperforming crystal radio circuit (right). Triodes” (https://youtu.be/BX7Oy9S7Kdw). The website https://diyvac­ uumtubes.com may also be helpful for would-be valve constructors, although it doesn’t appear to be particularly active. Radios Radios are important in times of crisis or isolation. They become lifelines for news, good morale and even survival. Long before smartphones or internet access, simple radios provided the only connection to the outside world; many of the most remarkable examples were built entirely from improvised or scavenged parts. The crystal radio, one of the earliest and simplest receivers, requires no batteries or external power as it harvests energy directly from the incoming radio waves. It can receive AM broadcast stations from many kilometres away. Its extreme simplicity made it perfect for improvisation. Crystal radios A crystal radio is perhaps one of the simplest possible electronic circuits of practical use you can build. Many have been built for survival by POWs (prisoners of war), or for fun and entertainment. If stranded, you could also listen to the news to see if rescuers are still looking for you! A crystal radio generally consists of a long wire antenna, a diode (which can be a homemade cat’s whisker) for rectification, one or two capacitors (fixed and variable) and a coil to form an LC tuned circuit and a high impedance earpiece – see the right side of Fig.54. Some or all of those components can be fabricated depending on one’s 18 Silicon Chip level of commitment or necessity as described earlier, including the diode, capacitor(s), inductor and earphone. The earliest crystal radios actually lacked a tuned circuit, so they were not selective and just picked up whatever station was strongest, although there were perhaps only one or two radio stations on air in the early days anyway. Such radios consisted only of an antenna, a diode, earphone(s) and Earth, as shown on the left side of Fig.54. This simplest possible radio is likely responsible for the rare examples of radio stations being heard from metal teeth fillings, which could act as an antenna, rectifier and transducer, and causing vibrations that could be felt or heard. Lucille Ball famously claimed to have picked up a radio station with her tooth. While considered scientifically plausible if near a strong transmitter, the TV show Mythbusters was unable to replicate the phenomenon. Regardless, radio signals can be accidentally detected by many electronic circuits. In 1922, the US Bureau of Standards published plans for building a broadcast band crystal set, which you can view at siliconchip.au/link/acbd You could still build one today from those plans using the same or similar components. Admittedly, it is not an absolute ‘minimalist’ design as it has a lot of unnecessary non-electronic hardware. An improvised crystal radio can also use small resistor-sized inductors instead of an air coil, although the performance will be inferior. As we discussed last month, many of the basic components of a crystal Australia's electronics magazine radio or other radios and devices can be made quite easily. Foxhole radios During World War II, Allied soldiers in the field built foxhole radios which were ultra-minimal crystal sets assembled from razor blades, safety pins, wire scraps and headphones, to listen to BBC or Armed Forces Radio broadcasts amid combat (see Figs.55 & 56). Soldiers were not allowed to use powered radios because unintended radio emissions could have given away their location to the enemy. Foxhole radios were built for self-entertainment by deployed soldiers. The terms foxhole radio and POW radio (see below) are sometimes used interchangeably, but POW radios carry the extra layer of secrecy and danger. Some POW radios also acted as transmitters, not just receivers. Improvised battlefield radios “Winnie the War Winner” was built by stranded Australian troops from the Battle of Timor in 1942. It was built using the power pack from a Dutch transmitter, 20m of aerial wire, a broken commercial medium-wave receiving set and a transmitter from another broken radio set. A generator was taken from an old car to charge the batteries. For more on this, see the video at https://youtu.be/r0PCqJyTjAs POW radios POW radios were clandestine crystal (and sometimes valve-based) receivers secretly built by Allied prisoners in German, Japanese and Italian prisoner camps during World War II. Unlike basic foxhole radios improvised in the field, these had to siliconchip.com.au Fig.55: a foxhole radio used by an American soldier on the Italian front. Source: https://w.wiki/Nv7X be completely concealed, often ingeniously hidden in soap dishes, canteens, brooms, fake books, gramophones, table legs, or camp furniture. So it seems the radio hidden in the coffee pot in Hogan’s Heroes isn’t so far-fetched after all. They were built in extreme secrecy, with scavenged or smuggled parts like bed-spring wire for coils/antennas, razor blades or burnt cinnamon bark/ foil for diodes, foil-and-paper capacitors, pilfered headphones, and occasionally smuggled crystals or valves. These sets were more elaborate than foxhole radios, featuring variable tuning (sliding coils or homemade capacitors), better selectivity, and shared listening among trusted prisoners. They received BBC and Allied broadcasts to track war progress, boost morale, distribute censored news summaries and occasionally coordinate escapes. In European camps (near strong signals), crystal sets sufficed; but in distant Japanese camps, valve receivers with batteries were more common. Most documented POW radios were receive-only devices (crystal sets in European camps near BBC signals, or one-valve regenerative receivers in distant Japanese camps powered by batteries or camp electricity). Transmission was highly dangerous because any radiated signal could be easily detected by direction-finding equipment. Borneo camp British Lieutenant Colonel R. G. Wells built a secret transceiver in a Japanese POW camp from scavenged materials like foil from tea chests and burnt cinnamon bark for resistors. He described his efforts in an interview at siliconchip.au/link/acbe Two Queensland brothers, Ernest siliconchip.com.au Fig.56: two diagrams of foxhole radios built by American soldiers on the European Front during World War 2. Source: https://w.wiki/Nv7Y and Charles Hildebrandt, built a POW receiver (strictly speaking they were interned, as it wasn’t a POW camp) which they operated in Java (now Indonesia), which was overrun by the Japanese in 1942. It was constructed from scavenged parts, built into a Dutch gas-mask container and hidden under a piece of concrete. For the last 16 months of the war, they secretly shared news from the ABC, BBC, USA and other stations. A photo of the radio can be seen at siliconchip.au/link/acbf These life-sustaining devices exemplify the ultimate in improvised electronics under duress, with pure ingenuity turning camp scraps into hope and intelligence. Tunnelling diode receivers Early radios that originally incorporated the negative differential resistance (NDR) tunnelling diode previously described were called Crystodynes. The term Crystodyne was coined by Hugo Gernsback, editor of Radio News, in 1924 to describe Oleg Losev’s invention. Gernsback published articles in September 1924’s Radio News detailing the Crystodyne principle, including circuits built in his labs to Losev’s specifications. These Crystodyne radios used the zinc oxide device for amplification, oscillation and detection, enabling regenerative receivers, simple transmitters and oscillators without valves. Tunnelling diode transmitters Using the NDR device described last month, Nyle Steiner demonstrated that he could create audio oscillators, RF oscillators up to 13MHz and even amateur radio transmitters on the 80 metre band (3.5-4MHz) or 40 metre band (7MHz) using crystal control for frequency stability. For example, he built a zinc-based 80m CW (continuous wave) transmitter that reached 8km, all from scavenged or household items, although it only had a power output of 1mW or even far less – see Figs.57 & 58. He also demonstrated the NDR effect with other materials like iron pyrite (fool’s gold). Another NDR device transmitter, by Ashish Derhgawen, is described in the video at https://youtu.be/LfUABN_ HGwU titled “A bizarre transmitter without transistors or tubes”. Improvised spark-gap transmitters Spark-gap transmitters are a very Fig.57: an NDRbased 80m CW (Morse code) transmitter with a range of 8km. Also see Fig.58 overleaf. Source: http:// sparkbangbuzz. com/zinc-osc-2/ zinc-osc3.htm Australia's electronics magazine October 2026  19 early type of radio transmitter that generate extremely broadband radiation. They were used from about 1887 to 1917. They can only transmit Morse code, not voice. A battery, a coil, wire and pieces of steel from wrecked equipment can be used to build a spark gap transmitter even more basic than the early designs. Such a transmitter generates electromagnetic interference that might be detectable by passing aircraft, ships or radio amateurs, although today no one is specifically equipped to detect this type of transmission. Spark-gap transmitters were phased out in 1934, but before that they were used for making distress calls, including from the Titanic; see the videos at https://youtu.be/izCV1WrPFds and https://youtu.be/CtqRwFQE2AM Commercially-made ICs Even if you have the equipment to make ICs and can use it successfully, where do you get designs from? One option is to learn how to design chips yourself, then get a commercial fab to make them. This is usually too expensive for individual but there are options for DIYers or small businesses to dip their toes into IC fabrication. Some possible options are: Tiny Tapeout (tinytapeout. com) provides access to professional silicon by letting individuals, hobbyists, students or small groups submit their digital (and sometimes analog/mixed-signal) designs to be fabricated at a commercial foundry through shared ‘shuttle’ runs. In a shuttle run, multiple user designs are combined onto a single multi-project wafer (MPW), drastically reducing costs compared to a dedicated run. It uses opensource tools such as Wokwi for graphical design/simulation, Verilog/VHDL/Amaranth for HDL (hardware description language) and GitHub-based workflows to reduce costs. No expensive proprietary software is required, nor do you need to sign any non-­ disclosure agreements or have a massive budget. For more details, see the video at https://youtu. be/qVWq_XZko-M which shows the process from idea to fabricated chip, 20 Silicon Chip Fig.58: the circuit diagram for the transmitter shown in Fig.57. including community sharing of designs and interactive demo boards. While Tiny Tapeout is the entry point for anyone to try chip design, wafer.space (https://wafer.space) provides the next step that bridges hobbyist prototyping to small-batch production without losing the open, collaborative spirit. An open-source Z80 clone The Zilog Z80 reached ‘end of life’ (EOL) in June 2024, with Zilog (now owned by Littelfuse) discontinuing production of the classic standalone Z84C00 family CPUs and peripherals after nearly 48 years of continuous manufacturing. In response, the open-source and Fig.59: the open-source Z80 tape-out (the final files that get sent to the IC foundry). Source: https://github.com/ rejunity/z80-open-silicon Australia's electronics magazine hardware preservation community has decided to develop a free and opensource silicon clone of this processor. Functional prototypes have already been made and confirmed working on multiple fabrication processes. Tiny Tapeout infrastructure has played a key role in its early development, enabling low-cost tapeouts on the SkyWater Technology (USA) 130nm process. Additional experimental runs have produced full or near-full pin-compatible chips that run Z80 code, communicate with external RAM, and pass major parts of test suites. Work continues toward a true drop-in 40-pin DIP replacement chip, including a Chip-on-Board (COB) variant on the GF180MCU 180nm process via Wafer.Space (Singapore) shuttles – see Fig.59 and the main project repository at https://github.com/rejunity/ z80-open-silicon Circuit boards Printed circuit board substrates can be fabricated by depositing layers of either insulating or conductive materials layer-by-layer using fused deposition modelling (FDM), a form of additive manufacturing. Insulating substrates include filaments like PLA, ABS, PETG and nylon, while conductive materials include PLA/ABS filaments incorporating graphite, copper or silver to create electrical traces. It should be noted that conductivity with this technique will be less than traditional copper traces. Simple circuits can be fabricated using conductive inks. An inkjet printer can be modified to use silver nanoparticle ink – see Fig.60. Components are held onto the substrate with copper tape. Such techniques are in the early phases of development. Conductive inks and paints can also be made using ingredients such as graphite powder, black iron oxide powder (Fe3O4, magnetite), clear nail polish, polyvinyl acetate (PVA) glue and isopropyl alcohol. Making an iron oxide (magnetite) based conductive ink is shown in the video at https://youtu.be/RjFeeptoilI Making another conductive ink is shown at siliconchip.au/link/acbg W. Wayt Gibbs made a conductive liquid metal that can be used to siliconchip.com.au Fig.60: conductive traces printed with a modified inkjet printer. Components are held on with copper tape. Source: www. instructables. com/PrintConductiveCircuits-WithAn-InkjetPrinter connect components without soldering, described at siliconchip.au/link/ acbh and siliconchip.au/link/acbi plus the video at https://youtu.be/ mF_Z8RUZjJE We don’t suggest you do this, as indium and gallium can be extremely toxic if inhaled or ingested. Light bulbs A DIY light bulb can be made using graphite from a pencil with CO2 as the inert gas; see Fig.61. Doing more with less The following sections demonstrate the use of normal construction techniques but in ways that are not normally considered practical. However, if you were stuck trying to ‘bootstrap’ technology; for example, after a devastating war, with no access to functional computers, some of these techniques might provide a good ‘shortcut’ to making the necessary tools and equipment. CPUs Discrete components can be used to make AND, OR and NOT logic gates to form a basic computer. So, if you could produce or obtain reasonable quantities of transistors and/or logic chips, you could make a relatively simple processing unit from scratch. A small 4-bit processor like the Intel 4004 from 1971 could be built using around 3000-4000 transistors, although if you simplified it as much as possible, it might be possible to make a 4-bit CPU with as few as 900 transistors. An 8-bit CPU like the 6502 could be made with around 5000-8000 transistors, or possibly 2000-3000 if it were stripped back to basics. Clock speeds for these devices would be on the order of a few kilohertz and they would be hard to debug. Debugging such a machine built from discrete or homemade parts would be extremely difficult due to wiring complexity, variability in component performance, noise and the sheer number of connections. By way of comparison, early 4-bit computers often used 800-1500 transistors, depending on the instruction set and register design. 8-bit designs (eg, something like a homebrew 8080 or 68000 clone) easily reach 2,00010,000+ transistors when built with individual gates. ICs require fewer transistors than discrete designs for the same logic siliconchip.com.au Fig.61: an improvised light globe using pencil graphite and CO2 as the inert gas. Source: www. instructables.com/HomemadeLightbulb/?linkId=75697704 Fig.62: the AYTABTU, a computer made of discrete components. Source: https://github.com/mengstr/aytabtu Fig.63: Ben Eater’s 8-bit computer. Source: https:// youtu.be/ HyznrdDSSGM Australia's electronics magazine October 2026  21 Fig.64 (left): the MOnSter 6502, an Mostek 6502 microprocessor clone built from discrete components. Fig.65: a homemade 1024-pixel image sensor. Source: www.instructables.com/DIYImage-Sensor-and-Digital-Camera because they can use tricks to provide multiple functions from a single transistor. They also don’t need to drive (relatively) long wires. Some examples of computers or other digital devices built from discrete parts include: The AYTABTU (All Your Transistors Are Belong To Us) project was an attempt to create “a usable computer built entirely out of discrete components like transistors, diodes, resistors and capacitors just as they did it back in the [1960s]” (Fig.62). See siliconchip.au/link/acbj and https://github.com/mengstr/aytabtu The Ben Eater (Fig.63, https:// eater.net/8bit) is an 8-bit computer built using 74xx logic chips on breadboards. The creator of this computer is offering it as a kit; see https://eater. net/8bit/kits The FACOM 128B is a Japanese-­ designed computer from the 1950s that was made using relays (siliconchip.au/ link/acc6). Since then, many people have designed their own relay-only computers, with memory being a big constraint. The MOnSter 6502 (Fig.64, https://monster6502.com) reproduces the Mostek 6502 microprocessor as used in early consumer computers such as the Apple ][, Commodore PET, Atari 400 & 800 and the BBC Micro (the Commodore 64, Atari 2600 and original Nintendo Entertainment System used close derivatives, like the 6520). There are 4769 components on the ▪ board; 3218 transistors and 1019 resistors comprise the ‘functional’ part of the 6502. Additional components include LEDs to visualise its operation and a variety of other miscellaneous components related to supplying power, driving LEDs, supply filtering etc. Digital cameras Sean Hodgins made a 32 × 32 pixel image sensor and digital camera from parts. Each of the 1024 image sensor pixels was hand-soldered; see Figs.6567 and the YouTube video titled “I Made My Own Image Sensor! (And Digital Camera)” at https://youtu.be/ PaXweP73NT4 If you want to make one, you can download the required files from https://github.com/idlehandsdev/ diycamera Rather than soldering over 1000 parts, you could consider a DIY pickand-place machine to automatically place surface-mount components onto a PCB. One example is OpenPnP (https://openpnp.org) – see Fig.68. Display adaptors Ben Eater made a video graphics array (VGA) display card from ▪ ▪ ▪ 22 Silicon Chip Fig.68: an example of an OpenPnP home-made pick-and-place machine. Source: https://github.com/openpnp/openpnp-openbuilds/blob/develop/Images/ overview_top.jpg Australia's electronics magazine siliconchip.com.au Silicon Chip PDFs on USB Fig.66: the completed DIY digital camera using the sensor in Fig.65. Fig.67: a sample image from the camera shown in Fig.66. standard ICs and prototyping boards – see Fig.69 and the video at https:// youtu.be/l7rce6IQDWs and even basic active elements from scrap. The fundamentals of electronics remain accessible even today. Rediscovering the ‘lost’ knowledge of previous generations can be thrilling, such as adjusting a cat’s-whisker contact to find the sweet spot, building an LC tank that commences self-­ sustained oscillation, or modulating a bias voltage to send a weak signal across a room. Experiments with improvised electronics show that amplification, oscillation and rectification don’t require sophisticated factory-made components but can emerge from your own garage. So grab some galvanised sheet, scavenged wires and a multimeter can allow you to start tinkering and push the boundaries. The next breakthrough in scrap-built semiconductors might even come from your own bench. Conclusion We have provided a window into the world of improvised electronics, where everyday scrap, some ingenuity and a willingness to experiment can yield active devices like homemade negative resistance diodes, liquid rheostats, carbon-rod electrodes, transistors, diodes and oscillators that function without a single off-theshelf part. The home experimenter can still do incredible things, from coaxing quantum tunnelling out of oxidised zinc to building amplifiers, transmitters ¯ A treasure trove of Silicon Chip magazines on a 32GB custom-made USB. ¯ Each USB is filled with a set of issues as PDFs – fully searchable and with a separate index – you just need a PDF viewer. ¯ Ordering the USB also provides you with download access for the relevant PDFs, once your order has been processed ¯ 10% off your order (not including postage cost) if you are currently subscribed to the magazine. ¯ Receive an extra discount If you already own digital copies of the magazine (in the block you are ordering). Further reading and viewing Fig.69: a VGA display adaptor made from standard ICs on a breadboard. Source: https://youtu.be/l7rce6IQDWs siliconchip.com.au • “How to rebuild civilisation. The Ultimate Guide to Rebuilding a Civilization: Dynamic Practices and Core Principles for Building a Sustainable and Ethically Grounded Future”, paperback by Jackson Ridge (2024) • Video: “How long would it take to build a TV in ancient Rome?” (https:// youtu.be/rJAFUu8RcvM) • A 275-page document of Nyle Steiner’s improvised electronics: siliconchip.au/link/acbc • Ben Krasnow’s Applied Science YouTube channel, including a homemade electron microscope, see: www. youtube.com/<at>AppliedScience • Peter Parker’s channel and website with lots of amateur radio DIY content: www.youtube.com/<at>vk3ye SC and https://vk3ye.com Australia's electronics magazine EACH BLOCK OF ISSUES COSTS $100 NOVEMBER 1987 – DECEMBER 1994 JANUARY 1995 – DECEMBER 1999 JANUARY 2000 – DECEMBER 2004 JANUARY 2005 – DECEMBER 2009 JANUARY 2010 – DECEMBER 2014 JANUARY 2015 – DECEMBER 2019 OUR NEWEST BLOCK COSTS $150 JANUARY 2020 – DECEMBER 2024 OR PAY $650 FOR THEM ALL (+ POST) WWW.SILICONCHIP.COM. AU/SHOP/DIGITAL_PDFS October 2026  23