A few variations of the mu-stage © 2001 Alan Kimmel

Alan Kimmel

From the orginal TubeLabs site

Copyright © 2001 Alan Kimmel. All Rights Reserved 

A few variations of the mu-stage 

Over the years since my first article on this subject ("The Mu Stage", Glass Audio 2/93, p.12), I've experimented with ways to further improve it. I have no doubt that the best device for the upper half of the mu stage is either a pentode cathode follower (CF) or a MOSFET source follower. By the way, it is the use of these devices in the upper half of my mu stages that is the main distinction between my mu stages and other mu-mode stages. (And when better IGBTs come out I will probably add them to the list of devices suitable for the upper half of my mu stage.)


My basic mu stage establishes a fixed voltage across a fixed resistor thus setting up a constant current. This is accomplished by bootstrapping the fixed resistor as explained in the above article. This fixed resistor is the plate load resistor of the bottom tube in the mu stage.


Soon after I produced my basic mu stage I began to wonder what would happen if the fixed plate load resistor of the bottom tube were to be replaced with a device that is already a constant current source (CCS) in and of itself. I tried this with a JFET by using one resistor to program the JFET to give the desired current, which the JFET keeps constant. This is the familiar two-terminal JFET CCS. It worked very nicely in the mu stage.


A variation on a variation: I replaced the JFET with an opto-isolator "Photo-FET". These are JFETs you can program "optically" to provide the desired constant current. (Note: some generic Photo-Fets are slow and not ideal for high fidelity audio. Digi-Key has some fast ones.) It is easy to adapt this to provide isolated servo control of the quiescent point of the stage to prevent drift. A servo op-amp, which would drive the LED of the opto-chip, can be powered by the regulated DC filament supply of the tube. This arrangement works well, BUT the servo is completely unnecessary when a fixed bias voltage source is used for the upper device as in the Fig. 5 hybrid mu stage of my GA 2/93 article.


Another possible mu stage modification is to use a floating screen supply as I reported in Glass Audio 3/96, p. 30, "A Direct Coupled Mu Stage". It was born out of my assumption that the extra load imposed by the screen dropping resistor of a pentode CF was detrimental to the CF. But it turns out that some pentode CFs actually like the extra load. Nevertheless, fitting the mu stage with a floating screen supply provides other advantages such as some increase in an already very good voltage swing ability plus making possible direct coupling from the lower tube's plate to the upper tube's grid.


Another variation of the mu stage is to use a power MOSFET as the upper device as in Fig. 5 of the GA 2/93 article, but with the addition of a low-mu type power triode such as a 6BL7, 6BX7, etc. to the MOSFET as follows: Instead of connecting the Drain lead of the MOSFET to B+, connect the Drain to the 6BL7 cathode. Connect the 6BL7 grid (through a grid stopper resistor such as 150 ohm) to the Source lead of the MOSFET. Lastly, the 6BL7 plate goes to B+. The main advantage of this is that it in effect increases the voltage rating of the MOSFET in the hybrid mu stage because you're dropping the voltage across the 6BL7 instead of the MOSFET. With this technique a 60V MOSFET is adequate even though B+ is several hundred volts.


Another variant mu stage is a high voltage version based around a special kind of triode such as the 6HV5 or 6HS5 tube.


This type of tube is unique in that these are "beam triodes" and they have unusually large amplification factors AND unusually large gm for a triode. The caveat is that these tubes require considerably higher 3+ voltage than most tubes. This type of tube is used in my ESL Direct Drive Amplifier (DDA). This amplifier uses a split-load cathodyne phase inverter to drive a push-pull pair of high voltage mu stages operating from a B+ of 4KV. The high voltage mu stage develops all the voltage gain and it is direct coupled to the output stage which is a PP pair of HV White CFs.


All the above work well. I've experimented with other variations of the mu stage which also work well. Many more mu stage variations are possible.


Just remember that, as explained in the GA 2/93 article, only a triode should be used in the bottom half of any mu-mode stage and the device used in the top half can be anything; but again, the best devices for the top half are either a pentode CF or a MOSFET source follower. This causes the top and bottom halves of the mu stage to best compliment each other and bring out the very best of each device, especially the triode, as explained in the Introduction of this website.