In trying to design my own version of the Leach Amp (I built, and still use, the Version 3 amp, and am getting ready to build the Version 4.5 design, but am taking the time to research some possible upgrades), I thought of a way to increase output power and possibly improve the sound. I was reading an article at http://www.aksaonline.com/discussion/discussion_papers_aksa.html and it seems that the Leach Amp meets all of the listed design criteria for building a SS amp with tube-like characteristics, with the possible exception of the power supply decoupling.
Currently in the Leach Amp, there is RC decoupling between the current drivers and the voltage amplifiers. My first thoughts were to break this connection on the PC board and feed the decoupling network directly from the capacitors with its own separate wiring. I think the ground for these stages would also have to be brought out to prevent a ground loop. A second thought was to bring each individual stage back to the main capacitors via separate wiring.
Then, I got to thinking about the excess voltage that the current drivers have to see in order to get the output voltage of the voltage amplifiers high enough to produce the desired output power. That is, if we need, say, +/- 50 volts at the collectors of the output transistors to provide the output power we need, we may need +/-58 volts, or more, back at the voltage amplifier stages to provide the needed overhead. Unfortunately, the power transformer has to be full power rated for that +/-58 volts, though only +/-50 volts of full power is actually needed. Seems to be a bit of a waste of iron!
The classical way to provide multiple voltages is to use a power transformer with either multiple or split windings. I believe some experimenters use a separate transformer to generate the low power voltages, and this may be an ideal method.
I thought of one other approach that we experimenters could use.
If one is using a toroid transformer, it seems like it should be relatively easy to wind an extra set of wires around the core to generate the needed voltage to power the voltage amplifiers. I don't know if the transformer manufacturer would provide the volts per turn number, but we could figure this out via experimentation. So, if the existing windings produced +/-58 volts, we could add a set of extra turns of, say, 28 gage wire (?) to generate, say, +/-68 volts. This voltage would drive the voltage amplifiers, and the existing +/-58 volt full power windings could drive the output stages. With the extra voltage overhead of the added winding, we could increase the average power output a little bit and the peak power a whole lot.
What do you all think? Seems like a cheap and relatively easy way to boost output power and possible improve the sonics as well. (Obviously, one has to make sure the voltage amplifiers can withstand the higher voltage.)
Currently in the Leach Amp, there is RC decoupling between the current drivers and the voltage amplifiers. My first thoughts were to break this connection on the PC board and feed the decoupling network directly from the capacitors with its own separate wiring. I think the ground for these stages would also have to be brought out to prevent a ground loop. A second thought was to bring each individual stage back to the main capacitors via separate wiring.
Then, I got to thinking about the excess voltage that the current drivers have to see in order to get the output voltage of the voltage amplifiers high enough to produce the desired output power. That is, if we need, say, +/- 50 volts at the collectors of the output transistors to provide the output power we need, we may need +/-58 volts, or more, back at the voltage amplifier stages to provide the needed overhead. Unfortunately, the power transformer has to be full power rated for that +/-58 volts, though only +/-50 volts of full power is actually needed. Seems to be a bit of a waste of iron!
The classical way to provide multiple voltages is to use a power transformer with either multiple or split windings. I believe some experimenters use a separate transformer to generate the low power voltages, and this may be an ideal method.
I thought of one other approach that we experimenters could use.
If one is using a toroid transformer, it seems like it should be relatively easy to wind an extra set of wires around the core to generate the needed voltage to power the voltage amplifiers. I don't know if the transformer manufacturer would provide the volts per turn number, but we could figure this out via experimentation. So, if the existing windings produced +/-58 volts, we could add a set of extra turns of, say, 28 gage wire (?) to generate, say, +/-68 volts. This voltage would drive the voltage amplifiers, and the existing +/-58 volt full power windings could drive the output stages. With the extra voltage overhead of the added winding, we could increase the average power output a little bit and the peak power a whole lot.
What do you all think? Seems like a cheap and relatively easy way to boost output power and possible improve the sonics as well. (Obviously, one has to make sure the voltage amplifiers can withstand the higher voltage.)
It will sound much better with a regulated high voltage tier for the front end. I would go 10~15V higher.
Leach designed this amplifier to clip in the voltage gain stages so as to avoid the high current devices 'sticking' .You must either add a NAD style soft clipping circuit to the input or a Baker clamp to the output stage.
Leach designed this amplifier to clip in the voltage gain stages so as to avoid the high current devices 'sticking' .You must either add a NAD style soft clipping circuit to the input or a Baker clamp to the output stage.
Thanks all for the comments.
The resistor could still be used in the front end to avoid the turn-on thump.
With the added windings, using a regulator would be a piece of cake and would not eat up into the overhead voltage of the main power winding. So, we could still power the mains off of +/-58V, but design the added secondary windings to produce the voltage needed to output a well regulated +/-70 V, say. Such a regulator would be easy to design due to the low current requirements. The design could also incorporate a soft start feature, so the series resistor can be eliminated. This will result in a power source with a very low output impedance which may very well help the sonics. With the higher voltages, some of the resistances and possibly the roll-off capacitor values will need to be revised, but this seems like the way to proceed.
The resistor could still be used in the front end to avoid the turn-on thump.
With the added windings, using a regulator would be a piece of cake and would not eat up into the overhead voltage of the main power winding. So, we could still power the mains off of +/-58V, but design the added secondary windings to produce the voltage needed to output a well regulated +/-70 V, say. Such a regulator would be easy to design due to the low current requirements. The design could also incorporate a soft start feature, so the series resistor can be eliminated. This will result in a power source with a very low output impedance which may very well help the sonics. With the higher voltages, some of the resistances and possibly the roll-off capacitor values will need to be revised, but this seems like the way to proceed.
I need an easy way to post simple schematics.
Anyone have any ideas?
"The most successful approach to limit storage time is to place a diode with low ON-voltage in parallel with the BC junction. This limits the forward voltage of that junction and prevents the transistor from going into saturation. For a discrete silicon transistor, this can be done with a Ge diode (which is then called a ``Baker Clamp''). In integrated circuits, the shunt is a Schottky diode. This is the basis of all high-performance..."
"fast recovery from clipping via Baker Clamp circuits"
"The first is a Baker Clamp diode configuration, and the second is the use of two additional supplies that are at a higher voltage than the main supplies."
http://164.195.100.11/netacgi/nph-P...,453,089.WKU.&OS=PN/4,453,089&RS=PN/4,453,089
4,453,089
Click on 'images', you need a TIFF viewer. All we need is CR6 and CR7 hooked to the pre-driver transistor (don't forget the mirror image), the anodes of these diodes would go to the collectors of the Vas stage.
Anyone have any ideas?
"The most successful approach to limit storage time is to place a diode with low ON-voltage in parallel with the BC junction. This limits the forward voltage of that junction and prevents the transistor from going into saturation. For a discrete silicon transistor, this can be done with a Ge diode (which is then called a ``Baker Clamp''). In integrated circuits, the shunt is a Schottky diode. This is the basis of all high-performance..."
"fast recovery from clipping via Baker Clamp circuits"
"The first is a Baker Clamp diode configuration, and the second is the use of two additional supplies that are at a higher voltage than the main supplies."
http://164.195.100.11/netacgi/nph-P...,453,089.WKU.&OS=PN/4,453,089&RS=PN/4,453,089
4,453,089
Click on 'images', you need a TIFF viewer. All we need is CR6 and CR7 hooked to the pre-driver transistor (don't forget the mirror image), the anodes of these diodes would go to the collectors of the Vas stage.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.