100w Hi-End Hybrid Amplifier
started on March  12 ^st , 2010 

INTRODUCTION

This amplifier born to create a reference in the hybrids using the best design and no compromise components.
Any solid state amplifier is composed by two sections:  voltage and current amplifier.
All the hybrid amplifier use as voltage amplifier vacuum tubes and transistor or mosfet in the current amplifier.
The vacuum tubes are not good to drive low impedance without transformers but have a good linearity also for large voltage swing.

VOLTAGE AMPLIFIER

To get a good output distortion decay in any condition I have decided to use a single tube in this stage.
I have searched a tube with a very good linearity for large swing and low anodic resistance to drive the output stage without problems.
This  tube should have an amplification factor greater than 40 to connect this amplifier directly to the CD player or phono pre-ampl.
To get 100w on 8ohm we need:
     P=V * I  and  I = V / R  =>   P = V * V / R  =>   V = sqrt (P * R ) = 28Vrms
so if we want a sensibility of 0.5Vrms we need
    Amplfier factor tube = 28 / 0.5 = 56

In this amplifier it is possible to use the Russian 6C45 (instead of D3a) without many changes (disconnect only 2 pin on the socket) but in this case will be necessary 1Vrms in input (instead of 0.5Vrms) to drive this amplifier.
To use the 6C45 it is necessary limit the power supply voltage of the driver reducing the relative secondary of the transformer because this tube have a max anode voltage of 150V.
There are 4 types of single ended configuration

	Anode/Plate resistor This is the more simple and cheap configuration but it give a very good sound with a little high distortion value. The max output with a good distortion spectrum decay is 22.8Vrms   1.78% thd - 1KHz FFT - frequency response
	Anode/Plate choke In this case all the power supply voltage is on the anode and the choke duplicate the voltage swing. It give a very good sound with a very low distortion value but it need a good and expensive choke specific for this use with low parasite capacitors. The max output with a good distortion spectrum decay is 44Vrms at 0.81% and 22.3Vrms with only  0.24% thd. As visible in the frequency response plot it is necessary a large cathode capacitor to reduce the low frequency resonance generated by the choke.   - 22.3Vrms 1KHz FFT - 44Vrms 1KHz FFT - frequency response with Ck 440, 1000 and 1440uF
	Output transformer As the plate choke configuration also in this case all the power supply voltage is on the anode and the inductance of the transformer duplicate the voltage swing. It give a very good sound with a very low distortion value but it is necessary a good and expensive inter-stage transformer. This transformer is responsible of the sound and it should have low parasite capacitors to get a good high frequency band and large primary inductance to get good bass frequency. Using an inter-stage with a primary inductance of 45H, like my new Headphone Amp.  we will get 20Hz at -3dB. The transformer should have also a low primary resistance, near to 1/10 of internal resistance of tube to lose low voltage.
	Constant current generator This configuration is completely different from the hybrid mu-follower tested and not used in my 2A3 PSE cheap. Here the solid state current generator is fixed and it is open for the signal so will not influence the sound. To get the max output swing the anode is at half of power supply voltage like for the anodic resistor configuration. The cost of this solution is near to zero and the result are good. The max output voltage with a good distortion spectrum decay is 44Vrms at 0.5% and 23Vrms with only  0.26% thd like the choke. Here no resonance are present and the frequency is always flat. - 23Vrms 1KHz FFT - 44Vrms 1KHz FFT - frequency response

This driver stage use a cathode bias so it is necessary a very good audio capacitor in parallel to the cathode resistor 100ohm.
The value of this capacitor is the result of many test because its influences the low frequency response (lower value can create a positive peak in the low frequency area for a resonance).
I have used 3 x 220uF 6.3v Sanyo OS-Con in parallel to have the 660uF value but you can insert a single 680uF.
A valid alternative to the OS-Con are the old Black-Gate N/NX/STD or new ELNA Silmic II.

Improved constant current generator As the previous there is a fixed solid state current generator. I have use the Kand Audio modules to implement this current generator (12$ each). - 35Vrms 1KHz FFT - frequency response

CURRENT AMPLIFIER

The most simple design to create a current amplifier is the darlington with 3 stages.
This solution have good performances and it give few problems in fact this is used in many amplifiers.
If we want increase the slew-rate performances and reduce the open loop distortion is possible to use a design like the Audio Research D100 or Audio Research D-400.
In the D-400 to prevent the limitation of this design, as explain very well in the article Building better buffer from Electronics World Nov. 1992, the Audio Research have put on emitter of first stage a dynamic current generators with high feedback error correction.
Using this design the slew rate is very high and the distortion very low, as visible in the following plot obtain with a Spice simulation using Winspice.

If the design is not extreme like the D-400 it is possible avoid the dynamic current generators in fact others amplifier like the Audio Research D100 use static current generators.
The Pioneer A-09 use a simpler method to implement a dynamic current generators but this circuit can lost stability.
My choice for this project is a current amplifier inspired to the Audio Research D100 output stage.
In the current amplifiers using the darlington is necessary use a vbe multipliers as thermal effects compensation but in this design is not necessary if all the transistors are fixed on the same heatsink.
The LM308 op-amp. is used as an integrator to eliminate the dc offset on output and there is no manual regulation to do.

• Eagle files

Follows some data obtain with the Winspice simulations where is visible the best combination in red with the output stage on perfect class AB bias.

• spice list

The simulation output at 40Vp on 8ohm load has been saved in the following images (red current combination).

Follows the first measurements of the output stage on 8ohm load with only one pair of output transistors and bias current set to 0.110V / 0.47ohm on each.

Follows the first measurements of the output stage on 8ohm load with all the pair of output transistors and bias current set to 0.070V / 0.47ohm on each.

POWER SUPPLY

To get the max sonic performances the my choice is a slow turn on power supply for the filaments using a common LM317 and a virtual battery operation power supply for the anodic with a power mosfet.

• Eagle schematic file
• Eagle board file

	The power supply for the solid stage output stage is composed by a custom toroidal  transformer  created on my specification by Italtras. electromagnetic flux 80% than normal 33-0-33   6A 9-0          2A 200-0     0.3A
	FWB1 rectifier module with Schottky Full Wave Bridge Rectifier got by Avondale Audio
	4 x 22000uF 63v Jensen electrolytic mod. 022670 (dimension 50x65 mm)

PROTECTION

MEASUREMENTS

Both stages at 28Vrms (100W) at 1KHz and 100Hz.

Follows the measurements at 12Vrms (18W) on 8ohm

Follows the frequency response.

PHOTOS

PROTECTION + SOFT START + PUSH BUTTON (It will not be used!)

Any serious solid state amplifier need a protection circuit because a fault on output transistors can destroy the loudspeaker.

I have integrated in the same circuit a soft-start to eliminate shock on electric environment during switch on, a push button to get a beautiful look and the protection on dc output.

I have found 2 little problems in the pcb so there are some corrections (in red).

• Eagle files