Build an Audio Amp from discrete components

Introduction

Warning – this project requires use of 120 V AC mains voltage and also uses +- 40 V DC for an amplifier power supply. If you are uncomfortable working with these voltages, do not attempt this project.

Ever since I started dabbling in Electronics, it has always been my ambition to build an audio amplifier. Somehow there is a certain mystique associated with been able to mount components on a printed circuit board, hook up a source and speakers and listen to music resulting from a collection of electronics

So I decided to take the plunge and build an audio amp from discrete components. I soon discovered that old school component amplifiers have been superseded by Chipamps (or Gainclone) and Class D amplifiers. But never fear, discrete component amps have some advantages:

• There is greater control over individual component values allowing some “tuning” of amplifier parameters
• It is more accessible for a DIY project with a greater chance for success
• It costs less than other techniques

To start I bought one of the venerable books on the subject: “High-Power Audio Amplifier Construction Manual” by G Randy Slone. It boasts an inset on the cover which says “Includes complete amplifier projects you can build”. The book was published in 1999 and is currently out of print but can be purchased second hand on Amazon and eBay for about $25.00

Everything you would ever want to know about discrete component audio amps from the 90’s is covered in this book – enjoy the read!

The back cover claims that the book contains 12 complete ready to build amplifier designs. In reality all you get is a circuit diagram and a single black and white photograph – so you have to do plenty of “fill in the blanks”

I decided to build Design 3 (Figure 11.4). To quote from the book: “It incorporates all of the best design attributes for low-distortion performance in all three stages, bringing class B performance to levels often considered impossible” (Wow).

Before getting into the build, a little background theory. All audio power amplifiers follow a three block design, simplified as follows:

The input stage is a differential amplifier with global negative feedback from the output. The voltage amplifier is responsible for amplifying the signal from the input stage to a full voltage. The output stage primarily increases current to fully drive the speakers.

A more detailed block diagram shows all of the blocks that make up a class B amplifier. This topology is essentially the same as the amplifier in this project and is used exclusively in all discrete component audio amplifiers.

Here is the circuit diagram of the Randy Slone Design 3. The output stage uses a Complementary – Feedback topology. This has some advantages with regard to thermal management of the output transistors. This schematic only represents one channel. Stereo requires two of these circuits.

If you need all the theory behind the workings of this circuit, it is covered extensively in the book.

The gain of this particular circuit is given by R10/R12 + 1. Substituting the values of the circuit gives a nominal gain of 31.

Supplies

Download the BOM for the project. This gives all the components that appear on the schematic.

I sourced all the components from my stock, eBay and Amazon. None of them appear to be very scare or difficult to get. Just one word of warning – pay attention to the voltage rating of the capacitors. Some of the capacitors require a rating of 50V or better.

Of course, these components are not the most expensive in the build. If you attempt this project, the most expensive items are likely to be the mains transformer, reservoir capacitors (required for the power supply – more on this later) and the heat sink.

To get these components, I purchased an old non-functioning amplifier on eBay. You will find lots of these amplifiers – typically they have the caption “for parts or repair”. The particular model I bought was a SONY STR-DE345 Audio/Visual Control Center at a cost of about $40.00 (including shipping)

With a bit of effort, I managed to salvage the following components:

• Mains transformer
• Reservoir Capacitors
• 0.22 Ohm 5 W resistors
• Heat sink (I hack sawed it in half for this project)
• Speaker connection post
• Bridge rectifier
• Screws to hold transistors to heat sink
• Output Darlington Power Transistors (not used in this project, maybe later)

To retrieve these components, you will need to do some desoldering. If you have never tried this before, I suggest you search for “desoldering” on YouTube and watch a few videos for helpful hints.

Some additional supplies needed, if you do not have them on hand

• TO-3P TO-247 Transistor Silicone Insulator Pads (eBay)
• 18 AWG Solid Copper Wire (eBay) – insulated
• 18 AWG Solid Cooper Wire – enameled (this is required for the 1uH inductor used on the amplifier output)
• Miscellaneous screws, posts etc
• Connection wire

Power Supply

Power supplies such as constructed here should be treated with healthy respect. Not only are there high DC voltages, but the reservoir capacitors hold a substantial amount of energy and can provide fireworks if discharged improperly or accidentally shorted.

Of course the book does not give a design for the power supply required to drive the amplifier. The schematic merely specifies +- 40 V DC. The first challenge become building a suitable power supply that could provide the required voltages. I used a simple design utilizing the scavenged components.

Circuit diagram below:

BOM also included and can be downloaded.

A few notes on the Power Supply construction:

• I used a piece of perfboard to hold all the components.
• The +- rails, common and capacitor connections were all wired using 18 AWG wire.
• Fuse holders are type BLX-A PCB Mount Fuse Holder for 5mm x 20mm fuses
• To quote from the Randy Slone book “The correct way to to interconnect the reservoir capacitors is to run a wire directly from the rectifier to the capacitor terminal. Then connect the rail supply wire directly to this same point, right at the capacitor terminal.”
• The resistors R1 and R2 are so called “bleed resistors”. They serve to safely discharge the reservoir capacitors once power is removed. D3 and D2 provide visual indication of the state of charge.
• Secondary fuses are a must – don’t leave these out

Another recommendation from the Randy Slone book is the requirement for a “High Quality Ground Point”. All the major grounds should be bought to a single point. This will become apparent during the construction.

Once the build is complete, you can switch on the power and check the DC voltage on the rails. This should be around 40V for the circuit shown.

Once again, be very careful of the reservoir capacitors, they can retain dangerous voltages even after power is removed. The LED’s in the circuit should be out before the capacitors are safe.

KiCAD Design

The book offers only a black and white picture of a PCB layout which is designed for home fabrication. I don’t have that capability, so I needed to get the PCB made by an outside firm. For that you need the so called Gerber files, an output format that is common across the industry and that can be fed into an automated machine.

To get the Gerber files (and the drill file for the holes), I recreated the complete project in KiCAD. This is a free ECAD design software (https://www.kicad.org/) which takes a design from schematic to printed circuit board. The software is very powerful and requires a learning curve to use; if you want to master the process check out the tutorials available on their website or watch many YouTube videos on the subject. An alternative to KiCAD is Fritzing.

Included here are the KiCAD files for the project

Also included here are the Gerber files for those who would want to have the PCB manufactured.

My original PCB design was incorrect because the collector and emitter leads on the 2N5551 and 2N5401 transistors were swapped. The KiCAD schematic and PCB included with this project have had this error corrected.

Fabricating the PCB

At this point, I had the PCB made by an outside firm.

There are multiple companies that will make a PCB for hobbyists. A good site to start with is https://pcbshopper.com/ which allows you to compare prices from various manufacturers. There is normally a minimum quantity that you have to order. In my case it was 5 boards.

Here are the specs of the boards ordered:

2-Layer FR4

1.6mm thick

1oz HASL Tin-lead Green Soldermask

White Silkscreen

The boards took about two weeks to reach me and were of very good quality.

Pictures of the final board included. Remember that one PCB = 1 channel. If you want a stereo system, you will need to build two channels on two separate PCB’s.

After collecting all the parts, it was time to start building the amplifier.

Again, let me stress that the photos show my original PCB which is incorrect – the emitter and collectors of the 2N5551 and 2N5401 are swapped. The Gerber files incorporated with this project have been corrected. But if you attempt this project, check carefully before soldering any transistors.

For this you need basic soldering skills and a good soldering iron. If this is something you have not done before, practice on some perfboard first. Also, there are many articles on the internet and videos on YouTube that can help.

The suggested way to populate a PCB with parts is to start with the lowest profile components (usually resistors) and then work your way towards the tallest components. I used a hybrid approach of working from left to right on the schematic with moving from low to high.

To make the 1 uH output inductor, use the 18 AWG enameled wire wound around a ¼ inch shaft. A total of 12 to 13 turns in the first layer and 10 to 12 turns in the second layer. Once it is conplete, slide off the shaft and arrange the ends to suit your requirement. Use a small piece of sandpaper to remove the enamel off the ends that need to be soldered. A better explanation of this process at http://buildaudioamps.com/make-an-output-inductor-pics/

There are 5 transistors that need to be mounted on the heat sink: Q12 & Q14 (main output transistors), Q11 and Q13 (driver transistors in output stage) and Q7 (VBE bias transistors). The reason for this is that transistor characteristics drift with temperature. By having these transistors all on the same heat sink, they all are at the same temperature (more or less). This causes them to drift equally and reduces potential sources of distortion. If you need a full explanation of this effect, consult the book.

You will need to drill holes to mount the transistors. I salvaged the screws that hold the transistors to the heat sink from the old amplifier. Make sure you insert a piece of silpad between the transistor and the heat sink. This provides electrical isolation (transistors have their collectors connected to the metal pad built into the package) but allows for heat conduction.

See the photographs for details

Otherwise, populating the PCB with the components is fairly straight forward. I added some small wires to various points on the PCB to act as test points. For the specific points, see the testing section later in this project.

Final Assembly

At this point, you should have all everything completed. Connect the amplifier to the power supply. Speaker ground should be connected to the High Quality Ground Point.

Don’t forget to add the fuses.

See the photographs for the final assembly

Testing

Before switching on power, thoroughly check all your connections and components. I vaporized two sets of output transistors before successfully getting the circuit to work. Fuses will not protect transistors, so check and double check.

My initial test was to have nothing connected to the input or the output. In this condition, the quiescent voltages in the circuit are shown in the attached diagram. These were the tet points that

If all is well, there should be no smoke or strange popping sounds and you can proceed to the next stage of testing.

Next test needs a function generator that can produce a sine wave. Attach a ½ watt resistor to the output with a value between 500 – 700 Ohm. I used a 680 Ohm resistor. Before switching anything on, make sure the function generator is set to a very low output, say 700 mV peak to peak (about 250 mv RMS) with a frequency of 1 kHz and then connect to the amplifier. Use an oscilloscope to check that the input is present on R4 (10K input resistor) and that it is symmetrical about 0V.

Switch on. Another oscilloscope probe (set to 10x input attenuation) connected to the output should give an in-phase sine wave with a peak to peak voltage of about 22V. The gain can then be calculated as 22/0.7 = 31.4. Which is pretty close to the design gain of 31. See oscilloscope trace captured

If this all works, you can proceed to the sound test. I used an old iPod as the music signal source and a Yamaha speaker. Connect the iPod to the input via a 3.5 mm audio jack. Switch on the amplifier and start the music. If all goes according to plan, you should be playing music over your DIY audio amplifier.

I don’t have the sophisticated equipment necessary to test the amplifier for THD and frequency response, so just listen and enjoy.

Final Words

This project proved to be more challenging than I initially anticipated. However, I learned a lot of useful skills. I am sure that the final result would be panned by the audiophiles out there to which I say – enjoy the sound.

My next step is to build the other channel for a stereo system and get the whole assembly into a reasonable enclosure. Good luck!