Im in the design stage of a new PPP amp (4 off KT88) using Van der Venne Toroidal Output Tannies.
DC Balance of the output stage is vital. Rather than design a bias servo to adjust fixed bias I'm considering using active Current Sources in each cathode bypassed by Blackgate electrolytics.
Does anyone have opinion / experience with the pros and cons of this arrangement.
Since it is effectively cathode biased, does this mean I could use the higher Cathode Bias Rg values safely, and hence reduce requirements on the driver stage?
Cheers,
Ian
DC Balance of the output stage is vital. Rather than design a bias servo to adjust fixed bias I'm considering using active Current Sources in each cathode bypassed by Blackgate electrolytics.
Does anyone have opinion / experience with the pros and cons of this arrangement.
Since it is effectively cathode biased, does this mean I could use the higher Cathode Bias Rg values safely, and hence reduce requirements on the driver stage?
Cheers,
Ian
SY are you sure?
SY,
I don't think the Class A ONLY comment is correct. Are you absolutely sure?
I've got a little EL84 PP Ultralinear Amp - basically the design out of Morgan Jones' "Valve Amplifiers". Its running with a ring of two (BC557 and MJE340) 40mA current source in each cathode (i.e. as a direct replacement for the usual 270R resistor) with 470uF/50V Blackgate Standard Electrolytics across them. I haven't noticed anything objectionable with the sound - in fact this little 10W per channel amp is quite stunning. It must surely be transistioning into Class B (rail is +300V, bias voltage developed across the current sources ar idle is approx 10.5V). The reason for the post was to see if anyone had experience of scaling this up to Parallel Push Pull KT88s or similar.
Can you please respond with a "Class A ONLY is definitely correct" OR "On second thought ...."
Thanks,
Ian
SY,
I don't think the Class A ONLY comment is correct. Are you absolutely sure?
I've got a little EL84 PP Ultralinear Amp - basically the design out of Morgan Jones' "Valve Amplifiers". Its running with a ring of two (BC557 and MJE340) 40mA current source in each cathode (i.e. as a direct replacement for the usual 270R resistor) with 470uF/50V Blackgate Standard Electrolytics across them. I haven't noticed anything objectionable with the sound - in fact this little 10W per channel amp is quite stunning. It must surely be transistioning into Class B (rail is +300V, bias voltage developed across the current sources ar idle is approx 10.5V). The reason for the post was to see if anyone had experience of scaling this up to Parallel Push Pull KT88s or similar.
Can you please respond with a "Class A ONLY is definitely correct" OR "On second thought ...."
Thanks,
Ian
Thoughts!!!
Well NO the current is NOT constant in Class AB BUT thats what the bypass capacitor is for.
I was mainly trying to get my head around things like what its recovery from overload would be like etc.
I guess I need to re-proiritize my hobby work and get my modelling program up and running - mostly conversion of various PSpice models to Spice 3F5 is what is required. That is to create the tube libraries.
I also spent some time on the weekend doodling on some scrap paper to see if I could adapt your SYclotron circuit to an output stage (using MOSFETS as the Source Followers). I think this can work - requires AC Coupling to the cross coupled grids unless full bias is developed across the MOSFET Source resistors (to the tube cathodes) BUT allows shunt feedback from the anodes to the MOSFET gates. Another circuit where my pencil and paper analysis is not sufficiently advanced to check what happens as one side cuts off, DC balance etc.
So - I guess its put away the soldering iron and fire up the PC for the next week or so.
Cheers,
Ian
Well NO the current is NOT constant in Class AB BUT thats what the bypass capacitor is for.
I was mainly trying to get my head around things like what its recovery from overload would be like etc.
I guess I need to re-proiritize my hobby work and get my modelling program up and running - mostly conversion of various PSpice models to Spice 3F5 is what is required. That is to create the tube libraries.
I also spent some time on the weekend doodling on some scrap paper to see if I could adapt your SYclotron circuit to an output stage (using MOSFETS as the Source Followers). I think this can work - requires AC Coupling to the cross coupled grids unless full bias is developed across the MOSFET Source resistors (to the tube cathodes) BUT allows shunt feedback from the anodes to the MOSFET gates. Another circuit where my pencil and paper analysis is not sufficiently advanced to check what happens as one side cuts off, DC balance etc.
So - I guess its put away the soldering iron and fire up the PC for the next week or so.
Cheers,
Ian
For several years I have a push-pull setting, cathode coupled, with EL34's in triode wiring. This means that the cathodes are floating. For biassing I have used a constant current source based on the LM7805. It has a 51 ohm load resistance. The IN-side has a set of capacitors (10 mu foil 63V, ERO 1613; 100mu electrolytic). The current is balanced to within 0,5 mA both sides. This means that the toroid will never saturate.
The transformer is a Métalymphy 110/110/13V toroid, a French firm, and the toroid were also used in some experiments described in the Audiphile magazine..
The amplifier is straight down to under 5 Hz. No feedback over the output. Driving is an ECL82. The high-end roll-off is about 100 kHz.
My speakers are QUAD ESL 57. Notoriously difficult to drive, with this setting the output impedance is calculated to about 1 ohm, without feedback. Not bad and in fact just what the ESL likes. Its has all the dynamics you like. The effect: the most gorgeous depth in the sound (well, one has to have something to cuddle
The transformer is a Métalymphy 110/110/13V toroid, a French firm, and the toroid were also used in some experiments described in the Audiphile magazine..
The amplifier is straight down to under 5 Hz. No feedback over the output. Driving is an ECL82. The high-end roll-off is about 100 kHz.
My speakers are QUAD ESL 57. Notoriously difficult to drive, with this setting the output impedance is calculated to about 1 ohm, without feedback. Not bad and in fact just what the ESL likes. Its has all the dynamics you like. The effect: the most gorgeous depth in the sound (well, one has to have something to cuddle
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.