The Story of the "Otala" Amplifier

A story of the legendary 25W “Otala” amplifier as seen by Terje Sandstrøm, one of the original designers.

Also see “The People Involved”, for other designers and people who contributed.

Introduction — the birth of the Otala amplifier

At an AES conference in 1973, Dr. Matti Otala presented a paper describing the design of a TIM-free audio amplifier. Present at this conference was Svein Erik Børja, a Norwegian record and broadcasting producer, and a great audio enthusiast. Svein Erik Børja was also one of the greatest Golden Ears of his time — he was able to hear even the slightest imperfection in an audio component. Having been dissatisfied with the sound of the transistorized audio amplifiers of that day, he saw an opportunity here for an audio amplifier of a new generation. Dr. Otala’s talk about TIM also explained the imperfections Svein Erik himself had noted in audio amplifiers.

He brought the paper to a friend of his who was running his own audio company in Norway, Per Abrahamsen of Electrocompaniet. Per decided to give the amplifier a try, to see what it might bring. They got the help of Nils Jørgen Kjærnet at Nera in Oslo, who did the circuit board design and also, to my knowledge, helped with the mechanical design of the amplifier.

They first made just a couple of amplifiers, but the sound was so good, fulfilling all their expectations, that they decided to manufacture a series of these amplifiers.

I started showing up at Electrocompaniet in the autumn of ‘74, and started to work for Per in the spring of ‘75. One of my first jobs was to assemble the first production series of the Otala amplifiers — a series of 10 amplifiers, based on the same PCBs as the original two prototypes. During the summer I started to look at the design, having been an electronics hobbyist since my early teens. At that time neither Per nor I knew too much about high-end audio design. That may well have been the factor behind the amplifier’s success, as it evolved over the next 5 years — we didn’t know how it should be done, so we worked it out from the framework given by the Otala design. We may well have been the first audio amplifier designers of the new school — the TIM-free designers, so to speak.

The first Otala amplifier in “production”

The first amplifier series used the same printed circuit boards as the first two prototypes. The PCB was a two-layer design — we called it a T-board, due to its shape. The power transistors were mounted onto the cooling fin from both sides of the stem of the “T”. One layer of the board was a ground plane (Nera being an RF company, that wasn’t so strange). The ground plane caused us a lot of mounting problems, because the component legs often got stripped when they were squeezed through the holes in the PCB, causing small metal pieces from the legs to curl up and make sweet, nearly invisible, shorts to the ground plane. We did see smoke …

In the autumn of ‘75 we started to get some attention, and a visit to Matti Otala at the VTT (Valtion Teknillinen Tutkimuskeskus — the Technical Research Centre of Finland), where he worked as a professor, became necessary. There were several reasons:

We had named the amplifier “Otala-Lohstroh” after its inventors, and Otala wanted to see the amplifier before he and Lohstroh could possibly allow us to use his name.

At the same time, we started to experience problems with the amplifier — it didn’t quite meet the specifications stated in the AES paper. We needed his help.

As it turned out, the amplifier was renamed to “The 2-channel audio power amplifier”, which was the name used for the rest of the amplifier’s lifetime. The relationship with Otala was established, and he visited the company on several occasions. There is still good contact between Otala and Electrocompaniet.

Period of confusion, start of revision

It was early 1976. We started to invest in measuring equipment. The company slowly turned into an audio amplifier company. It lost several of its old customers; the loudspeakers it sold became more neglected. We worked from early morning into the late nights. That would be the standard for the next 4 years. An amplifier brought to the US by Svein Erik blew up at the first listening test — embarrassing. We used the night to invent a new type of short-circuit protection. We didn’t want anything that could affect the sound, so we simply used a high-impedance circuit to sense the current through the output transistors, and then used a relay to switch off the voltage to the output transistors. It worked!

We now understood that we had to start changing the amplifier. Svein Erik’s golden ears became even more important than ever before. We were in uncharted territory; the old measurement methods could not be trusted. We had to rely on our own methods, our own interpretations of the measurement results. Now began an active period of theory, practice, and reading articles (Svein Erik provided us with articles en masse, on all aspects of audio design and electronics design in general).

We found that the compensation scheme used in the amplifier did not work correctly. As we measured the bandwidth to be far less than specified, we needed to redesign these filters. The amplifier used shunt feedback — one of its really strong points — with an input lag compensation; at each of the three amplifying stages, a lead-lag compensation was introduced. These did not match the actual poles we measured. One of the reasons for this, we assumed, was that the original design used “fresh,” newly developed Philips transistors. As both Otala and Lohstroh were working at the Philips labs in Eindhoven at the time of the design, we assumed they had access to better transistors. (I don’t remember if Otala ever confirmed this to be the case.)

In any case, a period followed where we changed the compensation back and forth with little or no effect. We did achieve something, but no significant improvement.

We had also now measured the distortion (having just bought our first distortion measurement set and spectrum analyzer). It wasn’t pretty — the distortion was far worse than specified, and performance was far below even our own expectations. Again we didn’t manage to do anything significant at first.

A major change came first when we started to change the quiescent current of the transistors. The design was in fact done as a mix between the old way and the new way. Older design books teach that you should use low current in the first stage due to noise. This was also done here (see “Noise Optimisation” for an explanation of why that’s not actually the case). We knew this theory was wrong, so we increased the quiescent currents, which decreased the resistance and again dramatically increased the bandwidth of the amplifier. The distortion also went down — not dramatically, but enough to make a significant change for the better in the sound (see “Transistors, Resistors, Current and Distortion” — not yet written).

The Audio Critic

At this time we started to hear rumours about a test in the US on our amplifier. One day our mail started to overflow, and a few days later we got hold of the test ourselves. It was in a magazine named The Audio Critic, and the test was fabulous. It started:

Audio freaks - Eat your hearts out: This is the worlds best sounding amplifier, and One: You can’t buy one in this country (The USA!), and Two: (not surprisingly - the only slightly negative in the whole test) It is too low powered to be counted.

Suddenly we had more requests for amplifiers than we could hope to handle. Sales boomed. It was mid-‘76.

We worked all summer, mornings, days, evenings, and into the night — Monday to Friday, Saturday and Sunday. Per’s wife, Anne, was not always too happy about this. Per had two small boys, and she had to take care of them a lot by herself. Anyway — her “sacrifice” made building these amplifiers possible, and the nice thing about it all is that both Per’s and Anne’s sons now work at Electrocompaniet! So in the end it turned out to be an investment in their own future as well!

The great change

Dr. Otala’s theorems on TIM very often came out as an attack on high-feedback amplifiers. Although it is correct that high-feedback amplifiers are more prone to TIM than low-feedback amplifiers, there is no magic here. The main goal of any amplifier is to reproduce the incoming music as perfectly as possible, neither adding nor subtracting anything. TIM is just one type of distortion; swapping one type for another doesn’t help. It is true that some types of distortion are harsher to the ear than others, but there still is a balance to be struck. If 0.1% of TIM equals 1% of THD in audibility, then an amplifier with 2% THD and 0.07% TIM will sound worse than an amplifier with 1% THD and 0.1% TIM.

So, we realised that the balance of distortions was the essential factor to consider. Not only do you have to balance THD against TIM, but also low-frequency distortion against high-frequency distortion, frequency and phase response against non-linear distortion in general, and so on.

This insight triggered the Great Change:

One night (it always happened at night!) we increased the feedback by 10 dB, for a total of 30 dB of feedback. The sound improvement was staggering — and contrary to common belief in our own community!

After this we only adjusted the amplifier slightly. It had found the form it should have for its remaining life at Electrocompaniet. And the amplifier was a success!

On marketing

We did brochures! The first one was rather technical. The second one was more professional. The rest of the marketing was done by word-of-mouth. The fact that the amplifier and the company were both hard to get hold of stimulated the market. Here was a mysterious company with a killer amplifier. Audio people love such stuff!

Technical things

See the schematics. Also see some of the calculations that were done.