EL156 SE – The Wind

  • After tasting the smoothness of the tube sound with my EL84 se project called the small one, I just wanted more power! To have more, I started the quest for a more powerful output vacuum tube. First thought was to pick-up a triode for the task. But then I realize that ‘exciting’ the grid of a 300B triode is not easy, as needs a 70 volts swing for full power, if you want low distortions and simple zen topology. Russian 6C33C was next, but here some other problems rose: despite this beautiful valve has a very low internal equivalent Rp and needs a quite low anode voltage, it need a lot of current to drive the grid for maximum power, not to mention the difficulty to build a high current /high voltage PSU (180 mA – 200 Volts) cheap and without humm infection ; also, due to the spread of the each unit parameters you need a true pair of these babies, which is hard to get, plus the mention that a high current output transformer is also hard to get at decent money. The 813, 211 and 845 triode options were using too high B+ for a rookie like me at the time L.  The chance is smiling to the dumb one from the unexpected corner: I just received as a gift a pair of  Telefunken EL156 pentodes.Almost fifty years ago, Telefunken introduced the EL156 as a replacement for KT88  It was used in LP mastering labs to drive cutting lathes. At the time, the tube had a ten-pin base socket,Y10-A , and wasn’t used outside recording studios. As the LP faded, so did the EL156. In 2005, when I started this project there was not much documentation on EL156 used in single ended triode mode, except Sakuma San and Jogis,  who use it in their projects. I started the project by calculating the final stage parameters, to be able to send a proper query for the output transformer at AE Europe, as they need some time to process customer’s orders. A good reference on how to handle output stage maths could be found at Valve Wizard, but I was using some old communist radio book formulas. Tube Cad or SPICE simulation methods should be the way for the modern tube diy builders! Still, I was using the graphic method for all the calculations; I scanned the original graphs from the tube data sheet book, and I was using analytic formulas found in book called “Radio Receiver’s Systems”, vol I, pages 301-312,  by Viniciu Nicolescu and Andrei Vladescu, Technical Publishing House, Bucharest 1958 ( inspired by ‘Radio receiver’s design’, vol I, Chapman & Hall, London, 1947). There is a gallery here, with hard-copies of all the pages on output stage calculations that I have found after 5 years. The working point calculated for the EL156 strapped in triode mode is:   Ia = 100mA; Ua = 480 V, Ug = -27 V.  I decided to order a 3k5:8 Ohm output transformer, demanded survive to a current of 125mA. A hard task, but for a fair amount of money the Dutch winding master Will Blaaw from AE Europe create two OT wanders! I decided then to use the fixed BIAS method, which I never used before. It was working from the first time flawlessly, and I will recommend it to anyone who wants to build a serious amp. Half of the diagram was solved, but how about the entry stage: the driver? Ohhh, while waiting for the output transformers I was digging … Having stability problems with a MU follower before, I was thinking to use a pentode too for the driver. 417a or EC 8010 were dreams, being hard to find. The help was coming from Russia with love: 6S45P E high mu triode. I was using it in common cathode configuration, self bias. Despite what everybody was saying on the forums, this 6S45PI triode does not produce a good sound at higher currents, according to my ears, sorry. The sound was too much compressed, with narrow soundstage and dry clinic presentation! I started then to play  with 6S45P’s bias and then listen to the amplifier sound for different plate currents; when I  reached the point I liked the sound, I stopped: 168 Volts for the plate at a 10 mA was the answer! I played with this valve a week, but whatever I tried in my amp configuration, that was the magic spot. The signal was feed to the EL156’s grid via a PIO 0.68uF, Russian made. Auricap’s value was too little when testing it (0.33uF), and there was a lack of lower register during the test, so I’ve drop it and preserved the sturdy military Russian one, which gave a well balanced sound   Some words about the original Telefunken EL156 longevity: I have found data that one of these survived like 60.000h in a laboratory PSU, at an university. I did not believe it at the time. All I can say is that I am still using the same EL156 pair since 2005, after replacing the 6S45PI drivers twice, and the VdH DDT II mc cartridge once!!! It’s true, I am using the tubes at just 75 mA instead of 100, with a slight quality loss into the sound compared to the 11. The power supply  Despite I tried to make a sturdy passive PSU, I failed: the toroidal transformer needs to be bigger than my expectations. At the full throttle, you could see the current dropping trough the final tube when cranking up the volume. The passive psu uses silicon rectifiers BYW96E, two in parallel for each position, a CLC for the high voltage B+ needed for the final tube. Capacitors are sturdy Russian PIO rated 1000 Volt, and L is a toroidal choke, an experiment made with Romtoroid. From 675 Volts the needed for the final stage, I dropped the B+ to b+, the voltage needed to feed the driver, with a bunch of resistors. Not very elegant, but these days that will be changed with an active stabilizer. For the b+ the filtration was done again in CLC, this time for each channel. For power up I used two rocker switches, one for mains and the other one to give first (when closed) half of B+ and after snapping it, gives the full tension. The mains transformer has four separate windings: B+, final tube filament, driver tube filament and one more for the fixed bias circuit. For EL156’s filament alternative tension is used, as for the driver a regulated one do the job. 10.000uF were used for filtering the fixed bias tension, and high quality expensive variable resistors were mounted in position. You don’t want to smoke your cigar from the final valves because you put a cheap potentiometer to adjust the bias!!!  The Wind amplifier was hardwired, using 0.6mm silver wire for signal and GND buses. The grounding was a mix between star ground and rail; I had no hum problems, except those from the PSU. All the connections wires were kept very short, with equal length for the both channels. Except the wires from the input connectors to the volume control, which are balanced shielded cables VDH Patch Bay, all the wires are no longer than few centimetres. As for the volume control, an ALPS black beauty did the job. The chassis was made by massive wood, covered by a 4mm thickness brushed aluminium top plate. The underneath of the chassis is covered by aluminium 2mm thick mesh, in order to ensure proper ventilation. The sockets were purchased from Jan Phillip Wuesten. For the EL156 tubes, socket covers machined from massive aluminium were mounted over the top plate, in order to hide the attaching screws.
  • Here is a gallery with various pictures of EL156 SE “The Wind” all tube, no global feedback amplifier:

© attitube 2011

find me:
grid (at ) attitube (dot com)
wordpress counter