New gain stage .. Opinions???

[nitrate]

Hi all,

I've been working on this new gain stage for my amplifier. I wanted to run it past you all before i hack up my old design and change the gain stage to this one. The stage comprises of Electro Harmonix's 6922EH. The B+ is rarther low and constricting at 120V. This is due to utilising a bootstrap B+ supply from the main amp supply of approx 60V. My main goals are to go from line level input to a +-30V ( 60v Pk-Pk )drive signal at a distortion of less than 0.1%. The simulator says that this circuit does the job but i don't trust simulators when simulating valves at low voltages. The valve model i used is here:-

ECC88 PSpice model
from Raytheon 6922 datasheet

.subckt ecc88 P G K
Gp P K VALUE={(0.06987627466m)*limit(V(P,K)*log(1.0+(-0.09315593712)+exp((5.807723862)+(5.807723862)*((2 9.18392242)+(-191.2885442m)*V(G,K))*V(G,K)/sqrt((51.21619923)**2+(V(P,K)-(25.47300401))**2)))/(5.807723862),0.0,1.0e16)**(1.367133435)}
Cgp G P 1.4pF
Cgk G K 3.3pF
Cpk P K 1.8pF
.ends ecc88


This stage is intended to drive a mosfet power output stage. The global feedback being derived from the output of the mosfet stage and not the output of the 6922 cathode follower as shown in the schematic.

What i'm particularly interested in is how the circuit will react to ageing. I don't know what properties of the valves will drift and how much they drift over time. I don't wish to end up with a biasing meltdown over time LOL


So does anybody have any opinions good or bad?? Do you think nested feedback loops are a good idea? Will it work in the real world LOL

Regards
Leigh

[45]

You are right in not trusting the simulator!

Basically with 60V/0.3mA you are running the valve at the cut-off: mutual conductance close to zero and very high internal resistance.

I don't think in real world you will get anything decent, especially the sound.

If you have limitations on the supply voltage then go for the ECC86 at the input.

You could go for a single gain stage: JFET/ECC88 cascode .
In that case you don't need huge load resistor nor low idle current.

Cheers,
45

[nitrate]

I have some ECC88 lying around, i might give it a go although i cant use a single stage as i need to apply feedback to reduce the final stage output impedance.

I knew the simulator results were too good to be true LOL!!


Leigh

[45]

A JFET/ECC88 cascode will have a gain like an EF86, depending on how you do it.

[45]

You could get a gain between 100 and 150 with 120V supply.
Say 120, i.e. 41.5 db.
Then with 14-15 db overall feedback and considering a 0.9 gain for the cathode follower, you should get 21-22V RMS with 0.7V RMS input voltage.

I would not be worried too much about the damping factor. It is just a static parameter.
Poor bass perfomances of most valve amps are not due to low damping factor in my experience.

Many speakers actually behave better if driven by costant power or costant current sources. That means damping factor approx. equal to 1 (valve amp!) or much less than 1, respectively.


45