EF183 valves directly coupled to 300Ω headphones.
It is a transformerless stereo electron tube headphone power amplifier using 2 EF183 electron tubes operating in class A with just 12V for heaters and anodes. Class A operation means electrons are flowing all the time.
The headphones themselves play the role of the anode resistor. The 300Ω voice coils of the Sennheiser HD580 for example are directly connected to the anode. A small DC current passes through the headphones. So the class A principle is extended to the voice coils too, electrons flowing all the time.
The operation at only 12V anode voltage is made possible by a 1.6MΩ resistor from anode to grid, for freeing the electrons. Without it the electron tube would be biased negatively 2 times, one inside the tube by nature and one by the external components. Acknowledgement to Hliana Stalika for insisting that the electron tube should not be negatively biased.
Since this technique makes possible an anode current of the order of milliamperes at 12V anode potential (high current at low voltage), this means by R=V/I that the electron tube output impedance is reduced making possible the impedance match of various loads without the need of an output transformer.
The components used:
(See also part 2 including output transformer)
2 EF183 electron tubes triode connected (anode connected to screen grid and suppressor grid). One EF183 per channel.
2 1.6ΜΩ resistors from each anode to control grid for freeing the electrons.
2 1μF coupling capacitors connected to grid for not upsetting it DC wise by the music signal source.
1 stereo 47kΩ log RFT potentiometer.
1 12V battery.
Signal path, setup;
Various CDs including Ray Conniff - SONY portable CD player with line out - Pleiades Elektra III with 12V small lead acid battery - Sennheiser HD580 headphones
The sound is unbelievable.
Extending bass and treble, giving the impression of a frequency responce from 0Hz to MHz. It sounds more extended to Pleiades Elektra II. Why?
Amazing mid range.
Every instrument sounds real. And each instrument can be distinguished. It is much easier to find out what note each instrument is playing.
Transparent.
Immediately reveals the weak and strong points of a recording.
It sounds so musical one does not want to stop listening.
More technical information:
The heaters are in series and connected to 12V.
The cathodes are connected to ground or battery negative.
The headphones are connected directly to the anodes. The other side of the voice coils, ie the headphones common ground terminal is connected to +12V positive. Plastic ungrounded headphone connectors should be used.
The anodes are always connected to the battery when the headphones are plugged in. Then the heaters are powered and this is the smoothest possible turn on and off. Headphones should not be connected or disconnected while the amplifier is no. The clicking sound may damage hearing. And in general we never increase volume when listening with headphones therefore keeping intensity low.
The noise performance is amazing as the EF183 connected in a similar way can be used as a front end pre preamplifier for the faintest possible microphone, moving coil cartridge or even radio telescope signal. See Pleiades V series preamplifiers. No noise will be heard even if the output of this amplifier is connected to another amplifier.
Using 1.6MΩ from anode to grid makes the grid have a bias of -350mV.
The voltage across each headphone voice coil is 600mV.
A disadvantage is that when connecting the HD580 headphones for example, the frequency responce has a broad rise of 4-5dB peak at around 70Hz, and +1 dB close to 16-20KHz. This indicates the output impedance is higher than 300Ω, (3-5 KΩ?) and the change of the headphone impedance with frequency affects the voltage output of the amplifier. Subjectively it sounds a bit like room coloration, most rooms have an increase in then low mid range due to standing waves. It is not the same though as a room reverberates. This is at Ia=2mA, (600mV across the headphones voice coil). If the anode current is increased there should be improvement as the output impedance of the amplifier will be lower. The coupling capacitors of 1μF may be a cause and they can be reduced. Also experimentation can be made with an output transformer to reduce the impedance of the amplifier to 300Ω. And a tap can be made for low impedance headphones too.
Another way to reduce the output impedance without an output transformer is to connect 2 resistors in series as an anode load and take the signal from the low resistor. In this case if we take all precautions including protecting our ears while connecting, the headphones can be directly connected to the low resistor. Vb may need to be increased. If we do not want to use resistors many headphones can be connected in series so that the total load is comparable to the electron tune output impedance. This may serve many listeners.
Would a 6SN7 electron tube with a electron freeing resistor from anode to grid have a lower output impedance. It seems difficult to increase the anode current of the EF183 to more than 3mA at 12V. Much earlier experiments showed the 6SN7 aand 12SN7 electron tubes, triode connected to be quite promising. Currents can't be readily recalled but it is rememberedfor an anode old and take the that those tubes were giving kicks to the curre nt reading needle when the grid was connected through high resistors to anode for example.
. If we do not want to,use resistors
The amplifier starts gently overloading and sounding fuller at not a very loud intensity protecting the ears.
(Playing sine waves clearly shows that when the amplifier starts overloading 2nd harmonic distortion is produced which sounds nice. On the oscilloscope the waveform is unsymmetrical, the top becoming round and the bottom of the waveform thinner, confirming 2nd Harminic production. The amplifier stops showing an increas in VU at -10dBm. When playing music the amplifier sounds very nice at -20dbm average, leaving headroom for peaks, and being kind for ears as this is not a high intensity. The measurement was made by connecting a VTVM across the 300Ω heAdphone voice coil on one Channel.
Another strange thing about this amplifier is that although it sounds terrific, like a tape recorder running at 10x 30ips, the waveform of a sinewave at the scope looks completely unclear. It is still a sinewave but the line is 20x thicker. This indicates radio frequency oscillation. The line stops being thick when the CD player is stopped. It also happened connecting to the CD player's headphone out. It still happens if the anode to grid resistor is disconnected from the anode and connected to Vb. The amplifier sounds terrific as if a real orchestra is performing with the most high frequency yet smooth and untiring to the ear overtones. Could it be that a very high frequency exists and performs a similar function to the ultra sonic bias frequency that tape recorders use?
Another area that has to be looked at is why almost exactly the same configuration but with an output transformer connected to the anode and a full range loudspeaker, produced an anode current of the the order of 6mA when Rga was less than 1MΩ and Vb up to 18V. This is described on a previous euroelectron post.The EF183 used was a Philips instead of Telefunken, both NOS. Preliminary experiments to increase the anode current above 2mA where not as successful. And swapping tubes did not produce any immediate anode current change. The loudspeaker driving version may have to be re examined).
The sound seems 10 times as big.
It is smooth.
Reducing the 1.6mΩ carefully should increase power. 1.6mΩ was chosen by chance, see yesterday's experiments with 1MΩ.
When the heaters are disconnected the amplifier keeps playing for a long time, many seconds, indicating that the cathodes may be overheated for this application. Underheating may be tried.
Wishing everybody listening to the Pleiades Elektra III be as happy as I am.
References:
Operating features of the Audion - Edwin Armstrong
On preserving the transconductance of electron tubes at anode potential as low as 3V - euroelectron blogspot
Any developments of the Electra? Planning to build one.
ReplyDelete