MooseFET is a low gain, power MOSFET based line stage that fills the gap between a passive preamp and a standard line stage.
The design goals were:
- Lower than usual linestage gain.
- Be simple.
- Be reliable.
- Be rugged.
- Be easily portable.
- Not sound "solid state".
- Not need a complex power supply, typically required of single ended MOSFET stages.
- Be inexpensive to build.
NEW - PCB's are available from my store, here!
Let's take a quick look at MooseFET's circuit:
The first thing you might say is, "Isn't that a BoZ?". The answer is not quite, but it was definately inspired by the magic of Nelson Pass (who contributed to this circuit with insights and suggestions).
The MOSFET chosen is the IRF510. It's a 100V device that can be easily obtained, inexpensive and has a low gate capacitance. You can use any of the IRF*10 MOSFET's without circuit modification.
The gain of the linestage is (Rd/Rs)*.8, a rule of thumb that's held true for every MOSFET we've dropped into the circuit. In this case, a measured gain of 5.3 is the result.
Should one change either of the source or drain resistors to change gain or Idd, the 14.8V drain bias voltage will no longer hold true. In that case you will need to use an oscilloscope and adjust the bias until the clipping of the amp at overload is completely symmetrical. It will then deliver similar performance as shown in the circuit above.
The 14.8V bias Vdd will hold true for any IRF510 (though I did not try other manufacturers than International Rectifier) and 14.95V for IRF710 (I used Siliconix here).
Another difference compared to other SE MOSFET preamps is the relatively low supply voltage. This does limit it's total dynamic range to the area of 7V RMS output. Remember, we're not going for typical linestage amounts of gain here, so you will be OK in most circumstances.
Some very critical components are the ferrite bead on the gate lead and the 470 ohm series output resistor. The ferrite bead is a must. Maybe add two. As with most high-bias MOSFET's, they like to act as VHF oscillators. In this design, position of the gate snubber and zener were irrelevant - it was a ferrite bead or instability! Most likely due to the tiny gate on these transitors. The 470 ohm series resistor on the output does the same if high capacitance and/or long patch cables are used, which could turn MooseFET into a "tuned line RF oscillator". If you look at the schematics of well designed pro gear, they all have the series resistor on the output.
I used metal oxide resistors in the drain and on the source not only to handle power dissipation, they are also low noise and non inductive.
The bias resistors were chosen for a centre-pot (or near to) when properly adjusted.
The 100K on the output side of the capacitor bleeds off any DC that might accumulate. Though, turn on MooseFET before your power amplifier, lest the output capacitor charging "thump" takes out your speakers. Low leakage, high quality caps like the Nichicon "MUSE" or Elna RFS series are reccommended here and are not expensive at all.
The power supply is part of this devices simplicity - you can't beat a wallwart and a LM317.
The LM317 is a wonderful regulator, low noise and stable. Once set to 24.00V, it will stay put within 100mV from cold to being in operation for days on end.
I had modified a 12V wallwart as a doubler, but since have found several 30-35V @ 500mA wallwarts at the local thrift store for $2. They work just fine and any noise they generate, the LM317 takes care of.
Being a simple circuit, it can be made quite compact. Here's the prototype and notes. The perfboard was slightly smaller than a buisness card.
Sonic impressions are pretty impressive. I brought MooseFET to the diyAudio Festival Victoria and had no bad comments. Matter of fact, when it was inline with the downstairs system (where the 2A3's, 45's and 300B's were being used), I was not permitted by the guests to move it out of the circuit upstairs :-)
It sounds like a cross between a triode and a SS device, yet both and neither.
Bah, rather than explain it, here's some FFT graphs of her output.
First is the spectrum of normal operation. The relation of fundemental and H2 remain constant, regardless on uncompressed volume level:
The H2 is the reason for her "musical" attributes and the triode-like sonics.
As we go up in volume to compression, the higher order harmonics appear, but still in an orderly, decaying fashion:
And finally, as MooseFET is starting to REALLY get annoyed:
MooseFET is one of those little linestages you can use when you need a little gain, but not a lot. Easily toss into your suitcase and bring to audio festivals or friends houses and demonstrate that improving sound to most systems doesn't mean a 45Kg piece of equipment, or a second mortgage.
So far, it has lived up to all the design goals and it filled another that wasn't even on the original list of targets:
- To be fun.