While studying for my GCSE exams - back in the late 1980s - I decided that a good project for the CDT (Craft, Design and Technology) course would be a new hi-fi amplifier. While this wasn't the first amplifier that I'd built, it was intended to be a significant upgrade on the one I had...
Looking back, it had some nice features, such as Record-out selection, and logic-controlled relays to switch the speakers and bypass the tone controls. After a number of tweaks and rebuilds it sounded pretty good, as well. Just before leaving home for university some 2 years later, I rashly decided to blow the money I'd earned in the summer on a replacement amplifier. I was able to compare it directly to a number of budget amplifiers - imagine my surprise at needing to stretch to a Musical Fidelity A1 to hear an improvement!
It's still working, and in daily use - my younger brother is using it along with the loudspeakers made at the same time for a different GCSE course. Recently, I got the chance to take a few pictures of it - it's not pretty inside, but I was only a kid when I built it!
The front panel
This is rather crude - it's just 3mm hardboard, hand-painted with Humbrol enamel paint, and labelled using rub-down transfers, which haven't held up all that well over the years.
Although the quality wasn't great, I was pleased with the control layout. The display window was from a Ferguson TV set - it had cut-outs for two rectangular LEDs on the right, and a 7-segment display would have displayed the channel number. Here, the green LED was a power LED, and the yellow LED was reserved for a future option - I just lit it up anyway! If I recall correctly, I think I was hoping to add a remote control volume control (using a motorised potentiometer), and was going to use that LED to indicate the reception of the remote signal (as a lot of kit from that era did). Behind the window, I simply had 3 red LEDs to indicate the state of the loudspeaker relays, plus one to show that something was plugged in to the headphone socket (why???!). Remember - LEDs were cool in the '80s. Especially rectangular ones!
The "direct" switch was a later addition. Originally, it was "CD direct" - which explains why it starts with a lower-case "d", and the text isn't centred on the control. The button itself - nicely back-lit - came from the remains of a car radio. At power-up, it defaulted to on, and the tone controls and other facilities were bypassed by a relay.
Inside
Although the original version was quite neatly done, by the time it had reached this state, it had become somewhat messy.
The pre-amp used to occupy the space on the left, but I built a different version and placed it directly behind all the potentiometers to shorten the wiring to the controls. Also, the power supply has been rebuilt on copper-clad board - it was originally built on matrix board. These three boards - the preamp, power amp and power supply - were on the same sized boards that were lined up nicely across the width of the chassis. All the wiring travelled in tidy bundles between the boards and to the front panel controls - these were all secured to the chassis by loops of solid-core wire (cheaper than tie-wraps). I simply drilled pairs of 2mm holes in the chassis about 5mm apart and threaded the wire through - you can see this most clearly at the top of the front panel. One downside of this layout is that the mains wiring to the power switch has to travel from one corner to the opposite one.
Needless to say, the power supply rebuild was a learning exercise in how to eliminate earth loops!
The chassis is made from thin steel which has signs of surface rust... I wanted to use aluminium, but that was far too expensive for the school! The top cover and side panels were made from 6mm MDF, glued together and covered with a sheet of self-adhesive plastic "veneer" in '80s black ash. There was a vent in the top. The overall dimensions were larger than normal hi-fi components, but were made equal to a record player that I had at the time (a Bush Arena unit fitted with a reasonable Garrard turntable). Sadly, the mains transformer induced hum into the pickup wiring, so I couldn't stack them.
The power amplifier
The power-amp was my first attempt at a DC-coupled design. The output devices are Darlingtons, and these are fed with a very simple Lin design, using a conventional VAS and long-tail-pair input stage. The VAS collector is bootstrapped - I hadn't really learned about current sources at this young age.
The short black wires that emerge from the middle of the power-amp board go to transistors on the heatsinks. These are the v-bias devices which I neglected to place on the heatsinks originally - I've no idea why, because I'd always done that previously. Either way, the inevitable thermal runaway happened, which cost me one output transistor and a bass driver!
The smaller board at the bottom is a simple relay-control circuit, and enables the 2 speaker outlets to be switched with tactile switches from the front panel. Despite the mentioned speaker-loss, I didn't take the opportunity to add DC-offset protection...
The board at the rear holds the loudspeaker relays and resistors to load the amplifier when all loudspeakers are disconnected. I'm sure they aren't really required - it would have seemed like a good idea at the time. From memory, I'm guessing they're something like 100Ω or thereabouts. Also, to the right of the relays are series resistors for the headphone socket, and similar resistors for an audio power meter that I had at the time. That was a Tandy/Realistic APM-500 - I probably still have it in the loft. I modified it - of course! - because the inbuilt mains transformer was rather noisy. Ditching that, all connections were made via a multi-core cable that terminated in a DIN plug (visible on the rear panel between the loudspeaker terminals and the mains inlet).
The pre-amp
The MM phono stage is visible in the image above - at the top-left - a simple 3-transistor design which worked well enough with the basic Shure cartridge that I had at the time. I'm sure I would have copied this from a text book or from an existing design.
The line pre-amp is shown here:
By this stage, I'd clearly sourced a bunch of Molex connectors - these would have come from ancient video recorders and similar products that I'd gutted for spares. The neatly pre-terminated coaxial cables were especially useful for this sort of thing.
Receiving my first CD player in 1989 is what prompted the pre-amplifier rebuild. Realising that the output level from a CD player was much higher than the rest of my sources, I tried feeding the CD player directly to the power amplifier via the volume control, and realised just how much the first pre-amplifier was colouring the sound. Simplification was needed!
The first stage was an emitter-follower to provide a low source impedance for the following tone control stage - it was during this project that I discovered that the active Baxandall circuit needs to fed from a decent voltage source. The tone control stage is built on the daughter-board, and is based around the TDA2310 - an SGS-ATES audio op-amp that was popular in the 1970s. I remember that this device seemed to be rather difficult to make work (it had to be externally compensated), but back then I didn't really have much experience with standard op-amps. The fact I'd made it a daugher-board tells you that I had spent quite a lot of time working on this...
A relay bypasses it all, controlled by the front panel 'Direct' function. This is logic-controlled, based on a simple 2-transistor "T" bistable circuit which is located between the front panel and the chassis. It's on the same board that holds the LEDs and switch, which you can see between the front panel and the chassis.
Immediately after the relay comes the volume control (you can see the Molex connector to the right of the relay), and then a single-tranistor gain stage. From memory, I've no idea how much gain this would have had, but it wouldn't have been much - no more than 10dB. With hindsight, it's fair to say that of the whole design, this was probably the weakest point - I would have been better off increasing the gain of the power amplifier. Still, back then, I wouldn't have known any better - my knowledge came from text books, electronics magazines, and from examining how commercial products did things. Of course, I didn't have access to schematics of contemporary hi-fi products - this was before the internet - so my references wouldn't have been state of the art...
Conclusion
Looking back, I'm amazed it worked! Bear in mind not only my age (14 when I first started the design work, 16 by the time of the final iteration), but also the almost complete lack of test equipment. I had an analogue multi-meter and a home-made bench power supply, and nothing else - not even an oscillator. I guess it was just a combination of patience, luck, and - it would be nice to imagine - intuition!
Importantly, the lessons learnt during the initial build and subsequent modifications were incredibly valuable, even if some of them were a bit painful at the time.
Shortly afterwards, I was given a broken oscilloscope and a few other bits of test equipment, and I built a couple more basic bench power supplies. Work started on its replacement, and I'd prototyped a much more sophisticated power stage that was much more powerful and undoubtedly would have had a much better objective performance. I never built that, because, as mentioned, I bought a Musical Fidelity A1 instead.
- Return to:
- Audio Projects »