Millennium 2 Questions

Started by aziltz, January 28, 2009, 08:27:06 PM

Previous topic - Next topic

aziltz

I've been reading up on R.G's Millennium 2 and I'm looking to adapt it to use with a Jack as a switch.  This means I won't be using the output pull-down as part of the control circuit. 

What is a good value for a pull-down in general, and should I continue to use that to connect the control to ground?

I was thinking a 100K that would either float or go to ground to trigger the Millennium.



Secondly, can I use more than one Mosfet connected to the same Diode/Transistor/Resister Network to switch multiple things? 


Third and final question, does the Mosfet always have to connect something to ground? Can I use it as a simple open/closed switch? Similar to the way jFets are used in the Tube Screamer to Bypass the effect?

I want to use this circuit to connect/disconnect an Input from a Summing Amplifier.

aziltz

#1
Alternately, a simpler thing i can do is to just use FETs as switches.  How well do FET's pass and isolate well-buffered audio signals?  I'm wondering about a switch between the gain stage and output stage of a common overdrive...

Where can I find out more about FET Audio switches?

aziltz

Quote from: aziltz on January 28, 2009, 10:31:03 PM
Alternately, a simpler thing i can do is to just use FETs as switches.  How well do FET's pass and isolate well-buffered audio signals?  I'm wondering about a switch between the gain stage and output stage of a common overdrive...
Where can I find out more about FET Audio switches? And a schematic of what kind of resisters/capacitors accompany an FET to make a nice switch for audio.


edit: sorry i didnt mean to spam myself.   i'm not impatient, i've just been adding/editing my OP alot.

R.G.

What you want is the audio switching section of the National Semiconductor "Audio Handbook", printed in 1980. It's reprinted by Audio Xpress, about $15. There's a lot more in there that the effects hobbyist needs to know but doesn't. I guard my original copy with my life.

Let me just wander for a moment.

If you're switching DC, you want MOSFETs. They are clean, fast, and easy to drive. If you're switching audio, you want either JFETs or CMOS analog switches, depending on your application. MOSFETs (which are what's inside the CMOS) switch audio well, too, but discrete MOSFETs have an unavoidable substrate diode which limits their audio switching to about 25-50mV if you want low distortion and are unwilling to do complex driving circuits for them. They are, however, a great way to generated a logic signal for driving JFETs.

CMOS switches are quite good, and unless you have a specific, known application where they can't be used, they should be your first choice. There are a few tricks to using them, as I explain in the article on bypassing with the CD4053 at GEO.

If you must use JFETs, you must also understand JFETs. A JFET is a wide diode. The drain and source are on opposite ends of one side of the diode (which is called a channel) and the gate is connected to the other side. If you forward bias the gate to channel, current flows just like any diode. If you leave the gate open, current can flow between the drain and source, limited primarily by the resistance of the silicon between them - this is "Rdson", resistance between drain and source with the device on.

If you apply a reverse bias to the gate side, the electric field from the gate literally forces the charge carriers to flow on the remaining channel area away from the gate. It literally changes the channel resistance by what I'd call the "garden hose effect", squeezing down the area available. Eventually if you squeeze (reverse bias the gate) hard enough, current quits flowing at all.

N-channel devices must have their gates pulled negative with respect to the channel to switch off. P-channel devices must have their gates pulled positive with respect to the channel to turn them off. Everybody uses N-channel, which I personally don't understand, but then there's a lot I don't understand.

To switch audio with an N-channel JFET, you prop BOTH drain and source up at some voltage higher than you will pull the gate down to. You do this with high-value resistors to a bias supply to not load down the audio, and you couple the audio in and out of the JFET with capacitors to not change the DC bias conditions on the JFET. For audio going through (i.e. switch on), you merely open circuit the gate, and all the current that can flow does. To turn the switch off, you pull the gate negative with respect to the channel. The best way to do this is with a diode connected from the gate to a capacitor which goes to ground, and with a high value resistor to the junction of the diode and cap. The control signal is applied to the resistor's free end and must go equal to or higher than the bias voltage on the channel, and down at least to ground if not lower. The resistor/cap slows down the control signal applied to the gate and prevents popping. The diode prevents you from forward biasing the gate.

The subtlties are all in the drive signal. The gate must be driven negative by at least the JFET Vgsoff (i.e. the voltage needed to turn the thing off) PLUS the size of the signal. If you're switching 100mV signals with a 4V Vgsoff JFET, you must apply a signal to the gate of -4.1V or you will get distortion and bleed through because the signal is modulating the gate/channel resistance in the active region. If you have a 4V signal and a JFET with a 4V Vgsoff, you have to apply -8V to the gate to keep bleedthrough and distortion from happening.

Tell me what part's not clear and I'll elaborate. Or go find an Audio Handbook.
R.G.

In response to the questions in the forum - PCB Layout for Musical Effects is available from The Book Patch. Search "PCB Layout" and it ought to appear.

aziltz

I can't thank you enough R.G.

Sorry I didnt get my question streamlined before I posted.  You told me exactly what I needed to hear.  I'm just getting back into this stuff.  My undergrad electronics course was pretty useful, but the one I'm responsible for teaching now is even more in depth.  I will definitely get a copy of that book as it will probably be a lot of help in my research eventually.  (Grad School in Physics at W&M).

I'm trying to develop a clever parallel blend system that will mix two signals to a single output or split to separate outputs with individual level controls if a 2nd jack is inserted.  I'm just trying to avoid unused pots/controls, so I'm getting creative with the audio switching.

Thanks again.

Great Job with the Visual Sound stuff btw, I'm a big fan.