This page is a summary of a few things I made over the years, in roughly chronological order.
Back in the 1950's, most kids started out building a basic crystal set.
I was fortunate in that my parents' house had a large antenna & a good earth, put in
when the house was built in the 1920's, & radios were far less effective than they
later became. I paid for the parts for the set, & my father matched me by providing a pair
of proper high-impedance earphones. I had ordered the parts from my local radio shop, & was
quite disappointed when they supplied a ready-wound tuning coil, rather than the spool of
wire I had expected (I intended to wind the coil myself.) Nevertheless, it worked very well.
It was built in an old wooden cigar box.
If you want to try this yourself, 2 tips:
This was a logical progression from the crystal set. The parts were bought as a kit;
I simply had to assemble it. I have used an available pentode symbol: the 3S4 tube is
directly heated, so there is no separate heater. The schematic is from memory, &
very much E&OE.
The learned will at once recognize the (in)famous super-regenerative design. A station is tuned in using L1/C1, then a degree of positive feedback is applied by adjusting C2 (coils L1 & L2 are magnetically coupled.) The danger with this circuit is that the most sensitive spot has it just on the edge of outright oscillation: it readily radiates broadband interference. Back then, people didn't worry so much about this, but today it would be asking for trouble. If you want to experiment with a super-regen, put a separate RF amplifier between it & the antenna.
As with the crystal set, a high-impedance earphone is required.
This design was published in the UK magazine Practical Wireless, February 1961 & following. My build followed the magazine quite closely, the major difference being the use of a Collaro Studio tape-deck, rather than the Motek which was specified. However it worked very well. This project taught me the importance of careful layout with high-gain tube amplifiers: they can readily pick up from their own outputs, & burst into oscillation.
The 8-bit computer I built back in 1976; it has its own page.
A Z-80 powered CP/M board, sized to fit on a 3-inch disk drive. It also has its own page. Although I no longer sell this, it still has a following.
This is a vacuum-tube audio amplifier, built around 2007. An OTL design using ex-Soviet 6C33 output tubes directly connected to the speakers, it delivers 25+25W continuous power.
Why was it named Igor? The first breadboard version used a collection of miscellaneous transformers to generate the numerous voltage rails, & looked like something from a 1930's Frankenstein movie. “See Igor, Russian tubes! Now ve shall ze vorld astound!” (Sorry🙂)
The first version was purely experimental, & went through several changes in topology. Version 2 settled on a conventional totem-pole output stage, with a deliberately asymmetrical phase-splitter to compensate for the different gains in the upper & lower limbs of the totem-pole output stage.
This sequence shows how it was built. The mechanical design began with a 3D model using Solid Edge, from which the sheetmetal drawings were taken off. The parts were professionally fabricated, & the external parts powder coated.
The schematic of one channel is shown. The accompanying “PCB layout” is a dummy, which was drawn up to assist in planning the component layout in the chassis.
V1 implements a cascode first stage, which provides most of the gain.
The phase-splitter, V2, has a very small cathode resistor (R7), this causes the stage to run unbalanced, with more gain in V2A than V2B. This compensates for the extra gain provided by V5, the lower limb of the output block. The upper limb, V4, is a cathode follower, & so contributes no voltage gain. The cathode supply (via R9) is switched in the control module, enabling the controller to provide a “soft-start” function, to avoid thumps on turn-on.
The cathode follower V3 provides the ability to drive the output tubes into grid current, so greatly increasing the available output. R3 & R17 provide some 12dB of negative feedback.
The power supplies are shown here. The HT & LT transformers were custom built. Each is controlled by a solid-state relay, which switches instantly, not on zero-voltage. The control module uses this ability to soft-start the supplies, gradually increasing the duty cycle. A slow start is provided for the tube heaters, & a rather faster one for the HT rails. The edge connector at centre-right is the control module.
Control is implemented by software in an Atmel AtMega-8 microcontroller.
A basic remote-controlled garage door opener. Yes, these are quite inexpensive to buy, but I took it as a challenge to build one (both the electronic and mechanical parts) using stuff in my junk box.
The motor came from an ancient “washing machine” disk drive, feeding a 30:1 worm drive from the junk box. This is followed by an intermediate shaft (top in the picture), with a dog-clutch operated by the large black knob, providing for manual operation.
A commercial 433MHz receiver feeds the control PCB. The software is available.
The motor has 2 identical windings spaced 90° apart. To start the motor, one winding is powered directly, & the other received power phase-shifted by a series capacitor. Once started, the phase-shifted current is stopped, & the motor runs on one winding. Interchanging the 2 windings runs the motor in reverse.
The door motion is sensed by industrial reed switches at the top & bottom of the track.
Another simple project, the point of this one was to include backup via a supercap, to enable it to survive the too-frequent power outages in the Perth Hills. 3 copies were built, & used around the house.
The specification provides a basic functionality, using 3 rear-mounted buttons. The software is available.
A cheap benchtop oven is re-purposed to solder SMT circuit boards. Two soldering profiles are stored in EEPROM, & can be changed by reloading the software.
The original controls were removed, & replaced by a custom design. A panel was made from the fireproof material used for electrical switchboxes, & fitted to the side of the oven chamber. Control components are fitted to this panel.
Temperature is sensed by a K-type thermocouple, visible at the right, just under the transformer. The control PCB uses the same Atmel CPU as my previous projects. The software is available. Inputs are the thermocouple & panel switches. There are 3 isolated triac outputs: the 2 heating elements & a fan, salvaged from an old microwave oven.
Several fine-pitch SMD boards have been successfully soldered in this oven.