Inverted cascode
One real advantage solid-state devices enjoy over the vacuum tube lies in having a "P" version of the FET, MOSFET, and transistor. Tube electronics would immeasurably benefit from a P version of the triode and pentode. In other words, a tube whose cathode received electrons from its plate. Such a tube is impossible, sadly. But by using P solid-state devices we can come up with topologies that would be impossible with just tubes.
The circuit below uses only one triode and is DC coupled. In function it is a cascode circuit. The triode is locked between the B+ connection and the MOSFET's source. As far as the triode is concerned its cathode and plate cannot move up or down. But its grid can. As the triode sees a varying grid voltage, the current conduction through the triode will vary in response. The varying of current through the 100 ohm resistor defines a varying voltage, which is in phase with the input, as the both the resistor and tube share the same current path. Roughly, the gain is equal the triode's Gm against the resistor's value.
The circuit shown above is also a cascode design. The bottom triode's cathode-to-plate voltage is locked by the MOSFET's source. (The plate resistor could be replaced with a top triode configured as a current source to decreased power supply noise making it way to the output.)
Any signal presented to the bottom triode's grid will provoke a current variation through the triode, which cannot find a exit path through the plate resistor, as the MOSFET holds a fixed voltage across the plate resistor, which in turn, fixes the current through the resistor. So where do the current variations go? Through the MOSFET and then into the 100 ohm resistor is the only path. What happens is that the MOSFET's source moves ever so slightly in response to these variations (the MOSFET has huge amount of transconductance compared to the triode) and this movement results in a varying current through the MOSFET. If the triode were pulled from its socket, the MOSFET's idle current would double, but the voltage across the 5k resistor would barely change. Locking the plate voltage is the main point of a cascode circuit. In fact, his circuit functions much like the previous one save for the phase inversion at the top of the 100 ohm resistor.
Adding both circuits together allows us to create a fully DC coupled hybrid amplifier. The amplifier shown below illustrates what is possible. The current variations through the input tube drives both output MOSFETs. So as not to confuse too many readers, no feedback loop is shown, but one could easily be added by bridging the bottom triode's grid to the amplifier's output. And a DC servo-loop can be added to the bottom MOSFET's gate, which would eliminate any DC offsets and the need for the potentiometer at the bottom triode's cathode.
The 5 volt zeners protect the MOSFETs from too great a current draw in the case of a shorted output. The 7370 is a 40 volt heater version of the 5687, which allows attaching the heater across one leg of the power supply. The BUZ900 might not be the best choice, as a greater gain could be realized by replacing the 100 ohm resistors with 400 ohm resistors. But the relatively low gate threshold voltage (usually an advantage) of this device prohibits using the higher value resistors.
한동안 속을 썩혔었던 회로입니다. 이 녀석을 흔히 구할 수 있는 6922(또는 12AU7)과 IRF610을 이용한 녀석으로 바꿔보려고 노력했지만 쉽지가 않더군요.
내용도 이해가 될듯 하면서 이해되지 않는 그런 녀석입니다. 아무래도 일반적인 형태에서 벗어나면 기초부족이 시리도록 느껴지는군요.
혹시 하스의 고수분들의 의견을 들을 수 있을까 해서 올려봅니다. 개인적으로 정말 특이한 회로라고 생각합니다만, 지식과 경험이 일천해서 그런 걸 수도 있겠군요.
원 링크는 http://www.tubecad.com/april_may2001/page6.html 입니다. 이것 뿐만 아니라 쭉 읽어보시면 하이브리드에 관한 재밋는 내용들이 많습니다. ^ㅅ^
One real advantage solid-state devices enjoy over the vacuum tube lies in having a "P" version of the FET, MOSFET, and transistor. Tube electronics would immeasurably benefit from a P version of the triode and pentode. In other words, a tube whose cathode received electrons from its plate. Such a tube is impossible, sadly. But by using P solid-state devices we can come up with topologies that would be impossible with just tubes.
The circuit below uses only one triode and is DC coupled. In function it is a cascode circuit. The triode is locked between the B+ connection and the MOSFET's source. As far as the triode is concerned its cathode and plate cannot move up or down. But its grid can. As the triode sees a varying grid voltage, the current conduction through the triode will vary in response. The varying of current through the 100 ohm resistor defines a varying voltage, which is in phase with the input, as the both the resistor and tube share the same current path. Roughly, the gain is equal the triode's Gm against the resistor's value.
The circuit shown above is also a cascode design. The bottom triode's cathode-to-plate voltage is locked by the MOSFET's source. (The plate resistor could be replaced with a top triode configured as a current source to decreased power supply noise making it way to the output.)
Any signal presented to the bottom triode's grid will provoke a current variation through the triode, which cannot find a exit path through the plate resistor, as the MOSFET holds a fixed voltage across the plate resistor, which in turn, fixes the current through the resistor. So where do the current variations go? Through the MOSFET and then into the 100 ohm resistor is the only path. What happens is that the MOSFET's source moves ever so slightly in response to these variations (the MOSFET has huge amount of transconductance compared to the triode) and this movement results in a varying current through the MOSFET. If the triode were pulled from its socket, the MOSFET's idle current would double, but the voltage across the 5k resistor would barely change. Locking the plate voltage is the main point of a cascode circuit. In fact, his circuit functions much like the previous one save for the phase inversion at the top of the 100 ohm resistor.
Adding both circuits together allows us to create a fully DC coupled hybrid amplifier. The amplifier shown below illustrates what is possible. The current variations through the input tube drives both output MOSFETs. So as not to confuse too many readers, no feedback loop is shown, but one could easily be added by bridging the bottom triode's grid to the amplifier's output. And a DC servo-loop can be added to the bottom MOSFET's gate, which would eliminate any DC offsets and the need for the potentiometer at the bottom triode's cathode.
The 5 volt zeners protect the MOSFETs from too great a current draw in the case of a shorted output. The 7370 is a 40 volt heater version of the 5687, which allows attaching the heater across one leg of the power supply. The BUZ900 might not be the best choice, as a greater gain could be realized by replacing the 100 ohm resistors with 400 ohm resistors. But the relatively low gate threshold voltage (usually an advantage) of this device prohibits using the higher value resistors.
한동안 속을 썩혔었던 회로입니다. 이 녀석을 흔히 구할 수 있는 6922(또는 12AU7)과 IRF610을 이용한 녀석으로 바꿔보려고 노력했지만 쉽지가 않더군요.
내용도 이해가 될듯 하면서 이해되지 않는 그런 녀석입니다. 아무래도 일반적인 형태에서 벗어나면 기초부족이 시리도록 느껴지는군요.
혹시 하스의 고수분들의 의견을 들을 수 있을까 해서 올려봅니다. 개인적으로 정말 특이한 회로라고 생각합니다만, 지식과 경험이 일천해서 그런 걸 수도 있겠군요.
원 링크는 http://www.tubecad.com/april_may2001/page6.html 입니다. 이것 뿐만 아니라 쭉 읽어보시면 하이브리드에 관한 재밋는 내용들이 많습니다. ^ㅅ^
