This is my latest audio project: an amplifier for the legendary Sennheiser HD600 headphones made with the superb 4P1L directly heated tubes in triode configuration.
Some technical information:
-output power: 0.2W (maximum power for HD600 headphones) for 0.7V input signal, output impedances 300 or 32 ohms;
-frequency band (output power 0.1W) at -3dB: 18Hz-40KHz, at -1dB 40Hz-20KHz;
-no headphones noise, at maximum volume (the headphones have an SPL of 112dB, they don't forgive anything!)
-distortions, SNR and harmonic distribution according to the attached graphs;
-no global negative feedback loop.
I chose another approach for DHT heaters and establishing the static operating point: I used constant current sources on the positive and negative branches and a constant voltage source for the tube bias. Thus, I obtain a perfect isolation of the supply heaters from the audio path that closes through them. The noise attenuation on each branch of the heaters is approx. 40dB and the impedance of the voltage references in the cathodes is around tenths of ohms, making unnecessary frequency compensation with decoupling capacitors. All of these lead to maximum power delivered to the load, lower current consumption (compared to the option of using a cathode bias resistor), high signal/noise ratio, high frequency response.
The noise reduction of the power sources was achieved by using electrostatic shields between the primary and secondary of the power transformers, electromagnetic shields with GOS steel sheet, separate power transformers for the DHT tubes, low noise regulators for the anode voltages, efficient and low noise DC/DC converters for low voltage power supplies and separate power supply for each amplifier. A microcontroller sets the start-up sequence, which includes anode power transformer soft-start, heaters soft-start, timed anode and headphone protection.
I also included a three-band tone corrector, with triodes and CCSs. It's noise floor is somewhere around -110 -120dBu and the corrections are +/-12dB at low, +/-5dB at mid, +/-8dB at high frequency. It can be switched off during listening.
To have a reference of the input signal level, I also included a VU-meter with EM84 that can be switched off, too.
It sounds great! I'm delighted and I can't get enough of listening to it after almost a year of building it. Together with the HD600, they are a perfect pair.
Some technical information:
-output power: 0.2W (maximum power for HD600 headphones) for 0.7V input signal, output impedances 300 or 32 ohms;
-frequency band (output power 0.1W) at -3dB: 18Hz-40KHz, at -1dB 40Hz-20KHz;
-no headphones noise, at maximum volume (the headphones have an SPL of 112dB, they don't forgive anything!)
-distortions, SNR and harmonic distribution according to the attached graphs;
-no global negative feedback loop.
I chose another approach for DHT heaters and establishing the static operating point: I used constant current sources on the positive and negative branches and a constant voltage source for the tube bias. Thus, I obtain a perfect isolation of the supply heaters from the audio path that closes through them. The noise attenuation on each branch of the heaters is approx. 40dB and the impedance of the voltage references in the cathodes is around tenths of ohms, making unnecessary frequency compensation with decoupling capacitors. All of these lead to maximum power delivered to the load, lower current consumption (compared to the option of using a cathode bias resistor), high signal/noise ratio, high frequency response.
The noise reduction of the power sources was achieved by using electrostatic shields between the primary and secondary of the power transformers, electromagnetic shields with GOS steel sheet, separate power transformers for the DHT tubes, low noise regulators for the anode voltages, efficient and low noise DC/DC converters for low voltage power supplies and separate power supply for each amplifier. A microcontroller sets the start-up sequence, which includes anode power transformer soft-start, heaters soft-start, timed anode and headphone protection.
I also included a three-band tone corrector, with triodes and CCSs. It's noise floor is somewhere around -110 -120dBu and the corrections are +/-12dB at low, +/-5dB at mid, +/-8dB at high frequency. It can be switched off during listening.
To have a reference of the input signal level, I also included a VU-meter with EM84 that can be switched off, too.
It sounds great! I'm delighted and I can't get enough of listening to it after almost a year of building it. Together with the HD600, they are a perfect pair.
Attachments
Nice looking clean build, congratulations!
I also have 4P1L still waiting for final application. Maybe I will follow your direction 🙂
What is source of output transformer?
I also have 4P1L still waiting for final application. Maybe I will follow your direction 🙂
What is source of output transformer?
Thank you, @Triodos. Yes, go for it! I guarantee the results.
Initially I also used 6N1P instead of 6N5P, they have higher amplification. Then I realized, my DAC has a high output level, so for an optimal signal-to-noise ratio, I lowered the overall gain. The output transformers are made by a local master. If you need any other information, please don't be shy.
All the best!
Initially I also used 6N1P instead of 6N5P, they have higher amplification. Then I realized, my DAC has a high output level, so for an optimal signal-to-noise ratio, I lowered the overall gain. The output transformers are made by a local master. If you need any other information, please don't be shy.
All the best!