Embedded Files
The End mk.4a .... late night test before the engineer's verdict for thursday's hifi!! 2018
The first test of a combination between mk II input stage and mk 3.1 voltage amplifier and output amplifier
This composition to avoid SMD components and to achieve the same stability of standard TO92 transistors everything works fine!
But first back to 1989 where it all started
This first “The End” mk.1 diagram suddenly set new limits for what ordinary self-taught electronics people could achieve in terms of good sound.
By precisely measuring transistors, zeners and resistors, anything seemed possible with this circuit
"The End" 1989-2018
A thread about "Self-build or DIY amplifiers" where the amplifier principle "The End" without feedback from 1989 was a hot topic at the time.
Around 1989 there were also other interesting DIY amplifiers such as Leach amp, Goldmund or Mitronic's mosfet PA100.
That's why I contacted Lars Clausen years ago about the rights to the circuit and about being allowed to show The End diagram on a website like this. Lars Clausen referred me to the owner, who is "L C Audio Technology" and Gert Frederiksen. Gert Frederiksen bought the company "L.C. Audio Technology" from Lars Clausen and thus also took over the rights to the principle diagram "The End".
The development in the amplifier area has changed in many ways from the "The End" mk 1 diagram, and I have been given permission in writing to show the old diagrams, together with changes I have made to the later mk3.1 diagram.
Already in the shown The End mk.II version, the input impedance was raised to approximately 20k, an input impedance that several owners of tube preamplifiers were very happy with.
In one of the first "The End" mk. II versions I introduced some DC safety relays with delay on right and left channel output. A DC amplifier that is thermally balanced and without fuses in the plus and minus power supply will draw all the amps out of what is available from the mains transformer, and the result will be melted transistors and melted voice coils.
The End mk.II can easily deliver 30A with a 500V/A and 4 x MJ15003/15004
The turn-on delaytime is about 3 seconds.....by setting C20 to 15uf the delay time can be increased. R31 0g R32 are DC-sensors, and the threshold is about 4VDC...in practice lower...D17 and D18 react if the power supply tilts to either plus or minus...in my versions I use 2 double 24VDC relays which I connect in series to 48VDC. This gives me 4 contact points for each channel and I can't hear the difference, even though the relays are bypassed.
Relay/delay CadInt layout
Two Erni 24VDC relay's ~ 48VDC as output safety
After trying many enormously large mono-built power supplies, I eventually found the necessary capacity that, together with the double bridge, gave the amplifier the desired "punch and attack". ...the mk.II printed version is distinguished by a compressed but correct printed layout, where the small signal paths do not exceed the necessary width, and where the printed circuit is at the same time free of "open" areas....the whole challenge lies in the cable routing. The version shown runs with +-39Vac and delivers ample power for quite intense sound pressures.
In both Norwegian, Swedish and Danish forums where the topic of "The End" mk.I has been in the threads, the words self-oscillation and burnt transistors have been repeated, which was also the case with version mk.II?
But all problems are due to insufficient knowledge about the length of the power supply wires and other parameters.
The first "The End" mk.I made in 1989 with the affordable TIP35 and TIP36 from Texas still works in 2018 without a hint of self-oscillation 36 years later. 2 years ago I changed the old TIP35/36 to MJL1302/MJL3281 as an experiment without any problems. The new Motorola/Onsemi outputs may have given a little more air around the instruments, but at least it gave the owner peace inside when the DC-safety relays were installed at the same time.
I have never understood the arguments about "The End" as an amplifier that goes into self-oscillation and burns out output transistors? ... I have I have constructed quite a few "The End" amplifiers with MJ15003/4 and 2SA1943/5200 for people with speakers where the sensitivity was below 86dB/1W and very difficult impedance curves, and the amplifiers have been driven hard......all the amplifiers are alive in 2018.
I think people underestimate thermal cooling and transistor/component match when a dc amplifier has to balance without "lifelines" like DC servo and the like .... ....
It is absolutely essential to make the power supply wires short.... that all transistors and resistors should be sized when this dc amplifier has to balance thermally. In addition, sufficient cooling fins are required such as Fischer Elektronik Heatsink, 0.35K/W, 150 x 300 x 40mm.
If you perform the mechanical construction of cooling and measurement of transistors well, "The End" principle delivers hifi sound beyond the ordinary standard.
"The End mk.II" approx. year 2009... 2SA1943 / 2SC5200, approx. +- 54Vdc, 47.000uF/80V Hitachi, single power supply, 2 x 35A diode bridge, 750W toroid, double 10A Omron relays for each channel, 5 sec. delay, DC and AC detector.
Due to the measured transistors, both outputs balance within < +-10mV ...the small change in quiescent voltage is due to thermal variation in the transistors and the two 4.7V zeners.
The two measured zeners are also light-sensitive and therefore adjustment of the offset should preferably be done without direct light...provided that the amplifier is adjusted with the "lid" on !
Above the 750Watt toroidal transformer and +-54VDC on the rails
Concerning speakers together with The End mk.II the AR9 with double bass, and the AR10pi with crossover filter incl. transformer is two of the most difficult speakers to drive in terms of impedance curves and sensitivity.
A well-adjusted "The End mk.II" can draw up to 30A at certain attacks in the music when the speaker approaches 1.5 ~ 2.0 Ohm. 30A is not a problem if the heatsink is dimensioned according to the speaker that the amplifier will draw.
In other words, a passive heatsink cannot be too large....if the heatsink is oversized, 110mA in idle current will barely make the heatsink lukewarm.
After many experiments with different transistors in the mk.II version, I think that Toshiba's 25mHz 2SA1302/2SC3281
appealed to me the most after running in. When the matched transistors have found their resting position, you can start comparing the sound. 2SA1943 /2SC5200 is close with many fine details.
MJ802/MJ4502 2mHz , 30A , 90V, 200W mono-built sounds "warm" at 33Vac = approx 44Vdc and can handle almost anything...easily drives 15" Coral with thundering bottom.
MJ15003/MJ15004 2mHz , 20A, 140V, 250W sounds good and effortless on +-56Vdc
Just to name a few...
Awaiting the sound from mk.4a with Sanken 2SA1303 / 2SC3284 150V approx 50mHz ring emitter
Do not use faster drivers in the mk.II version than TIP42C/41C as it creates instability.....however it should preferably be Motorola 42/41C
Only in mk 3.1 and mk 4a are faster drivers is recommended such as 2SA1370/ 2SC3467 from Sanyo, 150mHz, and 2SB1186 / 2SD1763.
The End mk. 3 DC-servo module is an option
Around 1994 Lars Clausen came up with "The End" version 3.0, and in 1995 with version 3.1
I wasn't entirely sure that the many emitter resistors for the LAPT 2SA1216 and 2SC2922 outputs would have a positive effect on the sound reproduction....but it would have to be tested.
After testing the first four versions of The End, the sound image from the mk.3.0 with one set of Sanken transistors could be characterized as smooth with many details in the music, and with good "punch" in the bass. It was not possible to distinguish the mk.II and mk.3.0 from each other at moderate sound pressures in a blind test.
The difference between the two versions occurred at higher sound pressures where the mk.II version with the extra amps from several transistors in the output proved superior to the mk.3.0 version. There were simply more amps to take. We considered an extra set of Sanken in mk.3.0, but that would require a whole new outlay.
The End mk. 3.1 with servo sounds good in my opinion, but perhaps with a little less dynamics which is probably due to greater control of the entire amplifier stage. Mk.3.1 will never be my favorite in this company.
After many listening tests, my opinion is that The End mk.II or the mk. 3 without DC-servo version takes first places, even though mk.3.1 measures well compared to mk.II.
The End mk. II and mk. 3 is a "loose" amplifier .....It is the amplifier where the sound depends on measurement....measurement of all components.
....but also the amplifier where "the best sound experience" is produced...
in my opinion..............
The End mk. 3.1
After the jpg’s of “The End” mk. 3.1 pcb’s incl. add-on servo, follows a construction manual of 19 pages explaining the “The End” principle as used for “The End” mk. 3.1 , “The End” XP and “The End” Millennium.
I would like to point out that L.C. Audio Technology stopped supplying the complete amplifier and parts for “The End” mk. 3.1 , “The End” XP and “The End” Millennium more than 20 years ago.
I would also like to inform you that the following 19 page manual is shown with written permission from L.C. Audio Technology.
Copying and distribution of this material is permitted, but commercial use may not take place without written permission from
L.C. Audio Technology.
After the 19-page XP / Mk.3.1 manual shown, follows a modified version of “The End” that I have composed of the input stage from Mk. II and the output stage from
Mk. 3.1 / XP.
This hybrid version of the “The End” principle without global feedback I have called “The End Mk. 4a” where SMD components or servo circuits are NOT used.
The electronic fuse circuit has also been removed from the Mk. 3.1 circuit and replaced with a double relay delay, exactly as for the Mk.II version.
The explanation is to avoid the electronic circuit in favor of a “mechanical” fuse circuit, because I think that the amplifier plays more freely when “The End Mk. 4a” is not limited by both DC servo and electronics that automatically protects the output transistors.
"The End Mk. 4a" thermally balances without the help of auxiliary circuits and returns to the adjusted idle current without any problems as all transistors, zeners and resistors are sized and matched.
Dual power supply boards for mono power stages 17,800uF/80V Nippon Chemicon for + and - rail including Wima film capacitors
After "The End mk.II" came version 3.1 where the input transistors were delivered as SMD components. All transistors in both the voltage amplifier, output stage and the current generators themselves should be measured for amplification. It is the emitter resistors themselves of 0.22R that determine the damping factor. The amplifier should have suitably large heat sinks as the amplifier balances thermally.
To avoid SMD components and be able to build "The End mk.3.1"... I changed the input stage to version mk.II with 2240/970, at the same time the output transistors were changed to Sanken ring emitter types 2SC3263/2SA1294. Since it was no longer version 3.1 the "new" amplifier was given the number "4a". "The End mk. 4a" is easy to adjust and shows no hint of instability. The complete design shown above.
100kHz without the slightest hint of instability or spikes and "free air" wired...must be better on a pcb....
Page updated
Google Sites
Report abuse