Specifications

ManualsBrandsDyna ManualsAmplifierSCA-35

A New Look At an Old Friend

Improving the performance of Dynaco’s SCA-35 & ST-35 amplifiers

By David C. Gillespie Rev.2

In 1965 I was 13 years old and hopelessly hooked on music and vacuum tube amplifiers. I have

been, ever since I saw and heard my dad’s first hifi purchase, a used

Grommes Little Jewel amplifier (at right) he bought 8 years earlier. By

’65, anyone who knew our family knew of my addiction, so friends of

my parents asked me to “build” them a stereo system for their home.

For the heart of that system, I chose a Dynaco SCA-35 integrated

amplifier kit (above).

At the time, the SCA-35 was the finest amplifier I had ever built, and when I first heard the fin-

ished product, I was blown away. The clarity of sound and range to which it extended was

startling, and left an impression that remains with me to this day.

Recently, as part of an effort to build a Dynaco museum, I acquired a completely stock, very

pristine-looking SCA-35. Judging by the condition of the pc boards, the unit did not have a lot of

use, and as a bonus, sported real Mullard output and phono preamp tubes. The original power

supply and coupling caps were even still good. For a unit whose serial number shows it to be one

of the first few kits produced of this model, it was in remarkable shape to say the least.

Basis of Performance Upgrade Possibilities

As great as I thought this amplifier was, today I’ve always known that if I ever got my hands on

one again, it would be uniquely qualified for a modification I developed that could significantly

enhance it’s performance. The effort to find out all started with those wonderful output trans-

formers, the SCA-35’s specifications, and certain aspects of the original design.

Summary of content (18 pages)

PAGE 1
A New Look At an Old Friend Improving the performance of Dynaco’s SCA-35 & ST-35 amplifiers By David C. Gillespie Rev.2 In 1965 I was 13 years old and hopelessly hooked on music and vacuum tube amplifiers. I have been, ever since I saw and heard my dad’s first hifi purchase, a used Grommes Little Jewel amplifier (at right) he bought 8 years earlier. By ’65, anyone who knew our family knew of my addiction, so friends of my parents asked me to “build” them a stereo system for their home.
PAGE 2
In classic Dynaco fashion, the Z-565 transformers (at left) used in the SCA35 are excellent performers. The best examples are capable of delivering the full 20-20 kHz bandwidth at full rated power with very low distortion, although on the low end, the figure is often closer to 22 Hz. They also provide an optimized Ultra-Linear output stage configuration that gives the SCA-35 a clear advantage in damping and regulation of power delivered to a speaker load, versus that of an equivalent pentode output stage.
PAGE 3
As power output from the driven channel increases, it draws more current from the power supply. This drops the available B+ somewhat – to both channels of course. The idle channel reacts to this by reducing it’s quiescent current draw, making more current available for the driven channel to consume. In other words, the idle channel acts as a shunt regulator across the output of the power supply, acting to maintain a rather consistent B+ voltage for the driven channel to operate from.
PAGE 4
Earlier, it was shown that when a single channel was driven and measured, the idle channel actually helped contribute to the excellent performance produced by the driven channel, because it was actively working to help regulate the B+ supply. The common cathode connection has a huge part in the effectiveness of this action. Again, consider what happens on the test bench when a single channel is driven to full power.
PAGE 5
stereo operation, current draw conditions in both channels would rarely be identical, so given that understanding, the common cathode connection would generally work to hold the overall bias steady between both channels. In that light then, the connection is more than just an economic move. In this case, it really does improve output stage performance when the amplifier is used for its intended purpose.
PAGE 6
deviating from that approach really takes away from what the original designers intended it to be: a clean looking, simple, but high performing integrated amplifier of minimalist design. So what alternatives do we have? Is there a different approach that could achieve all the goals outlined, while still maintaining that wonderful KISS principle that Dynaco capitalized on so well? It turns out there is, but it’s probably not what you might be thinking.
PAGE 7
So existing efforts to regulate cathode bias are problematic at best or destructive at worst, because the problems they solve are done so at the expense of introducing others. The problems of the existing cathode bias system have already been established for the SCA-35, and conversion to traditional fixed bias operation has been ruled out on the basis of complexity. It appears that what is needed then, is a new type of bias regulator – one that apparently has not been developed yet.
PAGE 8
the SCA-35: With reduced B+, the EFB regulator reacts accordingly to maintain it’s set relationship, so the lowest distortion operating point is automatically maintained. And, because the regulator automatically reduces the voltage drop across itself in this condition, power output increases as well. Finally, from a distortion standpoint, there is now little need to actually regulate the B+ supply anymore.
PAGE 9
Figure-2. Complete schematic of the SCA-35 modified with EFB. Additions are shown in yellow. Cathode resistor to be removed is circled in red.
PAGE 10
Figure-3. The table readily shows the drop in performance of the original design when both channels are driven, versus performance when EFB is used. In most cases, EFB performance with both channels driven is better than single channel performance of the original design. Single Channel Driven Before Freq Max Power (W)1 Distortion (%) 2 20Hz 17.0 1.60 25 17.0 0.90 1K 18.0 0.17 20K 17.2 0.65 IM Distortion 3 0.80 After Max Power (W)1 18.9 18.9 18.9 18.5 Change (%) +11.2 +11.2 +5.0 +7.
PAGE 11
conditions in both channels. We have seen what that compromise produced: under-biased tubes at quiescent conditions which shortened their life, and over-biased tubes at full power conditions in both channels, leading to excessive distortion and power loss. This situation required the output tubes to draw about 35 mA each under quiescent conditions, and resulted in nearly 12 watts of dissipation for each tube. This also gave the SCA-35 a well deserved reputation for eating up output tubes.
PAGE 12
body, and then soldered stranded, colored, #22 gauge wire to each terminal. This allowed the terminals to then be easily extended and connected to their appropriate connection points, as seen in Figure-6. The bias control is mounted in the hole Dynaco originally provided for the user to install a possible headphone jack, or level control for a third channel speaker. Thus, no holes need be drilled for installation, and by all external appearances, there is no change to the appearance of the amplifier.
PAGE 13
The second point is obscure, but relates to a feature that Hafler always championed. The Dyna Biaset resistors in his amplifiers serve as a current measuring device similar to the 5.0 ohm resistors mentioned above. But they also act to significantly reduce IM distortion in fixed bias amplifiers. Hafler went to some length describing this fact with the introduction of his Mk III amplifier. It is a fact I have also confirmed for myself.
PAGE 14
4. The 100 mfd/25v cap on the output of the EFB cathode regulator circuit is actually the original cathode bypass cap used in the Dynaco design. Its presence is not needed to ensure a low impedance virtual ground for the output tubes to operate from, but it surely reinforces it, and aids in stability of the 337. 5. Preference for the manufacturer of various tubes is a personal choice.
PAGE 15
except at the very lowest frequencies, where compliance for THD was in line down to 25 Hz. Below about 22 Hz (for my particular transformers), core saturation ruled the day and distortion rose rapidly to the figure shown for 20 Hz performance. This could be improved somewhat by employing more NFB, but that topic is beyond the scope of this article.
PAGE 16
increased, but flattened as well. In short, it has fully succeeded in achieving the original goal of allowing both channels to perform at the same performance level as a single channel could of the original design, and has done so while lowering the dissipation level of the output tubes by 20%. The unit is now actually capable of developing 35 watts RMS on a continuous basis across almost the entire 20 – 20 kHz power bandwidth. 10.
PAGE 17
Since EFB provides not only fixed bias operation, but also the same distortion reduction abilities as does operation from regulated supplies, it is clearly the best choice of bias scheme for this tube/transformer combination, short of actually using regulated bias/B+ supplies. Biasing and power supply requirements, as well as overall performance results are all very dependent upon plate loading, screen loading, and power bandwidth qualities that a particular transformer offers.
PAGE 18
to each power amplifier module as well to implement this change. The original bias pot would then be used to bring the bias voltage in range, and then the individual controls would be used to tweak each tube or channel as needed for equal current draw. Once they were set, the original bias control could then be used to move the bias uniformly up or down as needed. 13. In the listening room, EFB makes itself known rather immediately in the SCA-35.