Fender Brownface Tremolo - With a SS Twist

[beasleybodyshop]

Hey Everyone! It's been a little while. I thought I would share a little idea i've been tinkering with for the last fortnight or two - A Brownface fender style circuit. I love the sound of the harmonic tremolo but think it could benefit from some Solid State help.

First of all, all credit goes to folks over at the Hoffman forum for the SS ideas in the trem.

So the idea is, the oscillator isn't in the signal path - so why not use a few LND150s to drive that?

Also, this circuit implement's Martin Manning's wonderful work around for the wonky tapped Treble pot in the Brownface circuit.

Lastly, I added a 3 way power switch setup that i've used in a few amp builds that works pretty well. Half power removes a pair of power tubes from the circuit, and the lowest power setting runs the remaining pair as cathode bias.

All this being said, I do have a question for the group:

The voltage amplifier stage right before the oscillator - Does it make sense to use an LND150 or SS device here as well? I was thinking that the way that a tube soft starts on and off might play a role in some of the mojo of the Trem - But it would make the amp design a hell of a lot easier to just get rid of that tube stage and remove a whole preamp tube from the equation. Let me know what you think!

[beasleybodyshop]

Correction - I said "Voltage Amplifer" but I meant to say "Phase Splitter"

[Tony Bones]

It wouldn't hurt my feelings. As far as I'm concerned, the inverter is part of the oscillator not the signal chain.

[martin manning]

A cathodyne is quite linear (as long as it doesn't clip), so I don't think there would any meaningful difference in substituting a FET. I have a Fuchs SS harmonic tremolo pedal, "Creme de la Trem," which is very good. Basically a SS version of the Fender circuit, but if I recall correctly it uses optocouplers to direct the signal to the HP and LP filters.

[beasleybodyshop]

Would substitution for the phase splitter be as straightforward as subbing in an LND150 in place of the triode? Or do I need to figure out some additional biasing components to make it work proper?

[R.G.]

For some other inspiration, see http://www.geofex.com/Article_Folders/s ... remolo.htm

Edit: Ach! Forgot http://www.geofex.com/FX_images/FakeFenderVib2.pdf

[martin manning]

Would substitution for the phase splitter be as straightforward as subbing in an LND150 in place of the triode? Or do I need to figure out some additional biasing components to make it work proper?

I think It will work with the same components as the 12AX7. LND150 curves are available here: http://ww1.microchip.com/downloads/en/D ... 041114.pdf
There are differences between this device and a 12AX7: it has pentode-like drain curves, and it is happy to run gate voltage above zero. Even so, you can just drop it in in place of the 12AX7. Check the operating point though, you might want to tweak the bias resistor. edit: see below...

[beasleybodyshop]

Thanks Martin and R.G.!

For biasing the FET, I would tap off the B+ line to Base with some high value resistor correct? Does that still work when using it as a phase splitter?

[martin manning]

Just use the same configuration as the cathodyne PI in your schematic.

[martin manning]

Looked into this some more, and to center the operating point I'd recommend a 500 ohm resistor for the PI bias, and about 1.2 mA idle current, 56k source and drain resistors.

These things are pretty interesting... A close approximation of a center-biased 12AX7 gain stage (100k Ra, 1k5 Rk, fully-bypassed) can be made from an LND150 using a 56k plate load and an un-bypassed 330 ohm cathode resistor.

[Phazor]

Where do you get a 10meg reverse audio pot as shown for
Intensity? I have som 3.3 RA for speed.

[sluckey]

Hoffman has a 10M-RA pot just for that circuit or the Revibe.

[R.G.]

For biasing the FET, I would tap off the B+ line to Base with some high value resistor correct? Does that still work when using it as a phase splitter?

MOSFETs can be classified in three categories by how you have to manipulate the gate-source voltage to get them to turn on (conduct current) and off. Modern MOSFETs are most commonly enhancement mode devices. Enhancement mode devices do not conduct all by themselves, and you must enhance the gate-source voltage to get them to let current through.

They also come as depletion mode devices: like vacuum tubes and JFETs, depletion mode devices conduct current if you do nothing to turn them off.

The third variety of devices are both enhancement AND depletion mode. The LND150 is one of these. An LND150 conducts some amount of current if its Vgs is zero. If you make Vgs negative, it turns progressively more off until it conducts no current. That's the depletion side. It also conducts more if you enhance the gate source voltage by making it more positive.

Actually, vacuum tubes do this same thing, a little. Normal vacuum tubes will conduct more current when Vgk goes positive. The reason this is not normally done is that the grid starts conducting and it sucks electrons out of the stream that would normally go to the plate, so it takes more and more grid current to keep on increasing the grid voltage for more current. It does this until you can't push any more current into the grid or something melts.

The zero-bias devices like the LND150 are very versatile - and potentially confusing, the dark side of versatile. I peronally like the enhancement mode devices, because I can bias them by setting a gate voltage, which sets the voltage on a source resistor, and bango, the device conducts the curerent set by the voltage across the source resistor. That, and they're widely available.

[beasleybodyshop]

RG,

Thanks for sharing this. I am trying to integrate more SS devices into designs to better understand them in audio context.

Here soon I'll firm up my schemo and try to breadboard some of these ideas and go from there.

[R.G.]

You're welcome.

In doing your design, remember that the transconductance - yfs, or what would be mu for tubes- is very much larger for MOSFETs. This parameter is the conversion rate for changes in grid/gate voltage to plate/drain current. Transconductance is in units of Siemens - amperes per volt of control change. Power MOSFETs are commonly about, roughly 1A/V. So a 1V increase in Vgs makes a 1A increase in drain/source current - if the external power supply and loads will let it, of course.

Tubes mostly have a much lower value of change in plate current per volt of change in grid-to-cathode voltage, as well as a much lower maximum current because there are limits in tubes to how many electrons can be boiled off a hot cathode, and how much your driver can push the grid into forward conduction.

The raw voltage gain of a tube or MOSFET is the transconductance times the load resistor if there is no feedback hanky-panky involved. So the gain of a tube is mu times Rp, the gain of a MOSFET is yfs times Rd. These vary, and especially in tubes, the bias varies, so it's common to add some resistance in the source/cathode circuit so that higher plate/drain voltage raises the source/cathode voltage, adding some negative feedback. This lowers the composite gain, but makes it much more predictable. Tubes in general are more predictable than MOSFETs in general, but you pay for that with lower gains, lower current, and lower maximum current/power ratings. But predictability is the name of the game, so most tube (and non-switching MOSFET) gain stages use cathode/source resistors to make things more controlled.

You can restore the open-loop, non-feedback, gain for AC by using a capacitor to "bypass" the cathode/source resistor above some frequency, while leaving the DC conditions stable - and low-gain.
That's what those cathode bypass capacitors do.