Not EC factory work — my own modifications and designs from the NRF era (1979 onward), after leaving Electrocompaniet.
After I left Electrocompaniet in 1979, I tried to start up a new audio company in Asker together with two good friends, Paal Rasmussen and Espen Evensberget. We called the company “Norsk Radiofabrikk” (in English, the Norwegian Radio Manufacturing Company). We wanted it to sound old-fashioned, and really had a lot of good fun with this. However, our own original designs never made it into reality. In 1980, or thereabouts, we started to modify EC amplifiers. We modified between 50 and 100 amplifiers, and a somewhat lesser number of preamplifiers. The sound improvement was very good — not too surprising, since I knew these amplifiers in and out. Later we did some 50 amplifiers from the ground up, called the Special Version — the modification taken to its logical end. The components for these amplifiers came partially from Electrocompaniet. I don’t think Per liked it, but I also don’t think he really minded — if he had, he wouldn’t have sold us components, like cabinets and so on.
I have had some trouble locating all the schematics and component lists — don’t really know where I have (mis)placed them. However, I have redrawn the schematic for the modified amplifier below. Please be aware that the schematic is not to be used for commercial purposes without my permission. All private use is encouraged.
Components with red values are changed, components with blue values are new, otherwise equal to original values. Please get in touch regarding proper frequency compensation of this amplifier.
I’ve also included some pictures of the modified circuits below, so that you can see for yourself how we did the practical part of the modifications. Note that none of the traces were cut on the circuit board — all new components were wire-mounted (“bird’s nests”). Not the prettiest thing, but it sure works.
The modification significantly reduces the distortion of the amplifier, bringing much more clarity to the midrange, a tighter bass, and a smoother, less harsh top.
(A later note also mentioned a further schematic variant that popped up — apparently a later version that also never made it into reality, with component values and general setup equal to the modified ones described here, but with an extra double pair of output transistors and emitter followers added before the third stage. That particular schematic image itself hasn’t survived.)
Later we did some 50 amplifiers from the ground up, called the Special Version — the NRF modification taken to its logical end.
Note the heavily reduced values of the output-stage emitter resistors — down to 0.33 ohm from 1 ohm! This reduces the AB nonlinearity, as described in my 1982 AES paper.
Note also the changed input stage, with a JFET differential pair as a source follower. It was found that the bipolar input stage added distortion caused by its nonlinear input base current acting on the input resistors. (Which again shows: when we reduced the input resistors from 6k8 to 2k2, we heard a sound improvement and believed it was caused by improved frequency-response behavior above 20kHz — it may have been the 10dB distortion reduction that we heard!) To reduce this distortion further, the JFET pair was added. Further, a cascode pair was added to eliminate the Miller effect and the nonlinear voltage modulation of the critical first stage.
The gain distribution was also changed — the 3rd stage gain was increased by raising the load resistors from 2k2 to 3k3. More gain was needed to increase the overall feedback to 40dB. It was found that increasing this particular point (3rd stage) was the less critical of the options. However, we got slightly more relative modulation effect on the 3rd stage than we wished for.
All in all, approximately 100 amplifiers were modified, and an additional 50 Special Versions were made.
By the end of the ’80s and the beginning of the ’90s, I tried a few more changes to the amplifier design, which never made it beyond the lab bench. Only one amp exists with these modifications. They included emitter followers added before the 3rd stage, to reduce the modulation effect mentioned above.
(Captions translated from the original Norwegian file names.)
This is the phono stage, handling both MM (Moving Magnet) and MC (Moving Coil) pickups. The design uses active RIAA compensation, which was untraditional for the EC design — up to this time it used passive high-pass and active low-boost. The gain stages are designed to handle the active feedback, and note (!!!!!), the active RIAA compensation used shunt feedback. A tough job to calculate, but when that was done, it worked wonders!
The line-stage section is more traditional, with a normal series-feedback line stage. Note the extra buffer on the output.
For both block schematics above, note the roman numbers inside the amplifiers, denoting the particular gain block used.
This drawing shows gain blocks 3 and 4. Note the completely symmetrical design. The other gain stages also used an ingenious method of utilizing all the current from the last differential stage. Also note the current generators — ever seen something like that? Well, it works!
Details on compensation and other design choices will be described further in a later write-up. Don’t try to build the gain stage as shown above without that context — smoke will arise. Be patient.
The following image is the first of an analysis done on the 25W amplifier. The date, 4 July 1980, indicates that this was done after I left EC (which was in ‘79, at least on a full-time basis — after a short break in ‘80 or so, I returned on a consultancy basis to do more work for them until sometime in 1982).
I will upload the whole analysis. For those technically inclined, it should be interesting — lots of similar analysis was also done earlier. I got the feeling that many people believed we were only some kind of non-serious hippies. Well, in a way we were outside the establishment, but we did our mathematics! Also note the point made at the end of the note, saying that feedback below 20dB is no good idea when it comes to distortion.
