First attempt 10 years ago
As a DIY hifi guy with an interest in amplifier technology, it happens that you are guided into buying ready-made amplifier kits where measured input transistors and other components are pre-assembled, usually at higher prices than you had imagined.
This scenario unfolded in connection with the release of a DIY amplifier kit that was sold under the name “The End mk.3.1”. Both output boards were delivered with mounted measured SMD transistors for a fixed price that I don’t remember, but was annoyed enough not to buy the delights.
It was my irritation at being forced to buy some half-finished boards with SMD components that got me started experimenting with my own version of this no-feedback amplifier. But first to my attempt to reuse some of the surplus material that was in previous amplifier designs.
Photos of the beginning of a box for my experimental amplifier
First: I have recieved permission to show the circuit in writing from the owner L.C. Audio Technology and Gert Frederiksen.
Next: My idea
...and of course everything is better in my idea than the professional calculations....no it is not! ....but to my surprise the circuit worked from the start.
”The End mk. 4a”
A final variant of the ”No-Feedback” amplifier ”The End” assembled in August 2025 with the following transistor layout.
2SA970/2SC2240 Toshiba
2SA1145/2SC2705 Toshiba substitute ~ 2SA1370/ 2SC3467 Sanyo 150mHz 1W 150V
2SD1763A/ 2SB1186A Toshiba
2SA1294/2SC3263 Sanken ring-emitter
Simple PSU ...17.800uF/63V... one pcb for each channel
My own assembled ”The End” amplifier with vintage 1987 input stage from ”The End mk.II” and driver – as well as output stage taken from ”The End mk.3.1” ...plus small circuit changes.
In the attempt to get the “last micro details” out through the “The End” amplifier design, all “unnecessary” circuits have been removed....no servo....no electronic fuse circuits that can interfere with the output. (A mechanical fuse circuit comes as a suggestion or accessory for those who take special care of the bass voice coils....).
All transistors measured. For example, the four outputs on amplifier no. 1. are measured to: 2SC3263 (15A) ~ 2 pieces with 126 times amplification / 2SA1294 (15A) ~ 2 pieces with 119 times amplification.
The choice was 4 Sanken LAPT 2SA1294/2SC3263 (15A) output transistors per channel instead of one set of 2SC2922/ 2SA1216 (17A).
Next, the 11 pieces of 1 Ohm metal flim emitter resistors (0.09 R) were removed and 4 pieces of 0.22R 5% Vitrohm resistors were reinstalled, one for each of the four Sanken output transistors.
Since “The End mk.4a” is thermally balanced, all transistors from input to output have physical contact with each other, and the drivers 2SD1763A/ 2SB1186A and the four outputs 2SA1294/2SC3263, together with T9 (2SC2240), have physical contact with the heat sink.
In order to maintain a stable uniform temperature, T9 was moved to the heat sink between the driver transistors to stabilize the idle current as much as possible. Despite all the considerations that were made regarding measurement and temperature stability, the amplifier output will fluctuate approximately plus – minus 25mV around 0 volts DC out. . On the other hand, the output transistors quickly return to the resting position 0 Volt DC (+-25mV) after having been activated.
Note that the thin black guide-lines is the "jumpers" drawn by the software!
The two “home-etched” printed circuit boards were produced as prototypes with a single copper side. Therefore, you can see in the drawing several “jumpers” that were necessary, due to the desire to place all transistors in physical close contact. If these “jumpers” are to be avoided, a new layout and a double-sided printed circuit board will be necessary.
I can immediately reveal that the single-sided printed circuit boards shown including “jumpers” did not give rise to “parasites” or other forms of “spikes” and self-oscillations on the oscilloscope.
With reference to previous optimistic distortion measurements regarding non-feedback-coupled amplifiers with the same type of emitter output as on this “The End mk.4a”, I can only state that when you add the resistance in the last three base-emitter sections in the mk.4a together it will be almost impossible to get below 0.121% as HP’s distortion analyzer showed in distortion on the output of this “The End” emitter configuration.
I must frankly admit that the expectation was distortion data below 0.1% after a lot of work measuring all the transistors, zeners and resistors.
Based on the explanation (and experience) I got while looking at the oscilloscope, the reason why a transistor in this amplifier distorts is mostly because a resistance can be measured between the base and the emitter inside the transistor. The resistance found internally in the transistor plays a role together with the external emitter resistance that is included in the calculation, and on this basis there are three base-emitter stretches in the “mk.4a” (and all the other The End variants) that must be added. One for T13 (2705), one for T15 (1763) and one for T16 (3263), respectively T14 (3467), T18 (1186), T19 (1294), which contributes to the shown distortion figures.
If the distortion on the HP distortion analyzer is to be reduced below 0.1%, the amplifier topology at the output must be changed...and thus a completely different amplifier. Having said that, the professor from DTU believed that the tested amplifier was very stable and that the amplifier actually maintains a high hi-fi standard. As he pointed out, this "The End" 4a version is a non-feedback amplifier that can easily fall back to thermal balance after testing.
We have only measured the 4a amplifier up to this point. We have no knowledge of the amplifier's sound qualities, but if the "mk.4a" maintains the same convincing sound quality as previous versions of the "The End" principle, the self-build construction can be recommended.
1 kHz 10 kHz
45 kHz 100 kHz
Measured on the HP Distortion Analyzer:
At 1,000Hz 1W/8 Ohm HP: 0.121% distortion
At 12.5W/8 Ohm HP: 0.11% distortion
At 25W/8 Ohm HP: 0.15% distortion
Linear over >300kHz ~ Gain 43
Input impedance approx. 20K
The lowest distortion achieved at a supply voltage +- 40 VDC, 200mA idle current. However, only marginal increase at lower idle current around 100mA to 150mA.
Seen in the infinitely bright light of the rearview mirror, a single set of Sanken 2SC2922 /2SA1216 with their 17A and approx. 90Watt RMS on the output of this “The End mk.4a” version would have more than done the job as a linear uncolored power stage. An output stage that would have easily revealed sine wave errors or distortion from any transistor or tube preamplifier.
But when you own a pair of Edgar Villchur's AR3a or AR11 you know, that the amplifier must be able to deliver amps (enough), so the choice was a double set of transistors per channel.
Then the soldering og the two single-side copper PCB's
Wooden cabinet combined with 3mm aluminum base plate
Then all the pieces must be put together.
Still missing a 3D printed top - and front plate.
The last power amplifier of the “The End” type is built from components left over from the stock because the later versions of the “No Feedback” amplifier were realized with SMD components.
In version 3.1, the output circuit board was delivered from L.C. Audiotechnology with measured SMD transistors (I think it was 2240/970), which made it difficult for the average “do-it-yourself” man to “be there from the start”. It’s just a hassle to solder SMD transistors with soldering irons so that it looks professional!
I admit that it “nagged” my “do-it-yourself” soul that you had to buy a “half-finished” amplifier circuit board with SMD input transistors soldered on to achieve the best result.
Therefore, my starting point was to put together a stable output stage with the “No Feedback” idea, and with components that anyone interested could buy in hardware stores. That goal seems to have been achieved.
At the same time, I wanted to have a good test amplifier with an input impedance of approx. 20K for all my tube preamplifiers, and “The End mk.4a” seems to fulfill that wish.
The power supply will come later as an add-on to “The End mk.4a”, which will hopefully also include a listening test for the circuit shown.
But up to this point the mk.4a version fulfills all the wishes I had for a super good power amplifier, where the ring emitter transistors from Sanken have really surprised with a fantastic reproduction of details, and then on top of that with the power I have been missing for my AR11 I am extremely satisfied with the whole process which has given me lots of experience. I can only recommend this feedback-free power amplifier
”The End” mk.4a
K.J.Hansen
Mechanical relay/delay fuse of “The End mk.4a” comes later
to be continued.....