JFET Voltage Controlled gain in opamp circuit?

[ashcat_lt]

I've never really played around with transistors.  Was always kind of mystified by how they work and generally turned off by what seems to be a very wide tolerance range which seems to mean you have to fiddle around and modify the circuit to get it to behave based on the specific part that you pull out of the box.  But I'm starting to at least get the theory, and seeing that some circuits can be designed to sort of deal with variations automatically.  But, with this thing I'm dreaming up I'm a bit over my head.

I'm looking at this:

and, to me, that first opamp stage looks a whole lot like a tubescreamer with voltage controlled gain.  Great.  Problem I have is that the control voltage must be more negative than ground.  I would like to be able to use positive control voltage here.  I suppose the easy answer is to use a p-channel jFET, but all I've got here are 2N5458.  Is there a way to get these to work with positive control voltages?  

I don't really care right now which way the gain follows the voltage:  more voltage = more gain or vice versa at this point doesn't matter.  Just having trouble figuring out how to make it work at all.

[midwayfair]

I think you can use another, more standard envelope circuit to create a positive reference voltage [edit: but I'm not entirely certain about that]. I think the one from, say, the Nurse Quacky will do it. You probably have to flip some polarities and play around with stuff on the breadboard.

Could be interesting.

Make sure you read Mark Hammer's and PRR's comments on the circuit here:
http://www.diystompboxes.com/smfforum/index.php?topic=99561.0

But if you can't get it to work, Tayda sells 2N5457s.

[Ronan]

Never mind, what Jon said...

[ashcat_lt]

That thread is what got me thinking on this.

To be clear I'm not worried right now about deriving the control voltage right now.  I'm actually looking to be able to add a CV input in order to get AM distortion effects similar to a ring modulator.  I figure I probably will include the comressor side chain as well, but that's a step or two down the line. 

The first step is to make that transistor do its thing when the control voltage is always more positive than ground.  I'm already convinced that a p-junction would be the right answer.  Dont need anything else from mouser, but I'm sure i can find something to make the shipping worthwhile. 

At this point it's a learning excercize.   

[Ronan]

Try a 2N5088/2N5089 or similar. I have used them with success in other circuits doing a similar function.

[midwayfair]

The first step is to make that transistor do its thing when the control voltage is always more positive than ground.

Well, there's the FET's internal resistance to keep it above 0R -- about 280-320 Ohms in the 245X series from what I've measured. Does that help?

Excited to hear what you come up with, sounds like an interesting project.

[ashcat_lt]

Okay, can anybody help me understand how the JFETs do their thing in a Phase45?

Edit - Could it be as simple as the Lovesqueeze thing with the Source connected to 9V instead of ground.

[R O Tiree]

> Try a 2N5088/2N5089 or similar.

2N5088/5089 are NPN BJTs, not FETs

@ ashcat - How about connecting the Source of the FET to Vb?  The opamp doesn't care much and the Gate of the FET will be able to get comfortably below that virtual GND.

D3 is the wrong way round, as is C10.  D2 you can lose.  If you connect the Source of the FET to Vb, you can also lose C11.

C12 needs to be 47 - 100pF?

C1 value not shown?  10nF would be fine with that huge 1M pull-down resistor at R2.

Lots of 1M resistors floating about in here - lots of thermal noise.

Sorry for being short and sweet - bed time and busy day tomorrow

[R O Tiree]

Phase45 - Have a look at R.G.'s article at Geofex about the technology of phasers.

Essentially, for each stage, one of the opamp inputs is connected to a resistor, the other to a cap.  The signal at the cap is bled away (or not) by the FET.  If the signal is bled away, then the signal through the resistor dominates and you get just about the original signal at the output of that opamp stage.  If the signal is not bled away, then the 2 inputs are out of phase and the output will also be out of phase with the original input signal.  Then we get to mixing "dry" and "wet" back together... if dry and wet are exactly out of phase, then the mixed output is zero.  Any other condition is somewhere in between full and zero signal.

Because the voltage output from the LFO is applied to the Gates of those FETs, they change their resistances.  When the voltage is low, their resistances go up, so signal is not bled away.  When the voltage is higher, their resistance goes down and signal is bled away from the caps.

It's kind of counter-intuitive and there are several logic stages to go through to get the final answer.  Simple and elegant when you get your head around it, though, much like an awful lot of these wonderful, fascinating circuits 

[Gurner]

Apparently, while I was typing ...this happened....  

@ ashcat - How about connecting the Source of the FET to Vb?  The opamp doesn't care much and the Gate of the FET will be able to get comfortably below that virtual GND.

...I agree that's a viable solution...i.e. source at VB, but then of course the gate's preceding peak detect circuit would need to be referenced to VB too, else the FET will be cutoff in the absence of a signal.

[ashcat_lt]

R O - I guess I was asking about the voltages at the S and D on the Phase 45.  Having trouble sorting that out.

Then I thought and edited my post while you were typing.  I thought I could just connect to 9V, but I guess 4.5 V will work too.  It just gives us less available swing before we hit ground, and I think I'd have to build another Vb at less than 4.5V for the fade where the raw VC swings around ground, if I want it to follow both sides of the waveform and avoid forward biasing the gate.  Not a huge deal, I guess.  Vgsoff in a 2N5458 is kinda tiny anyway, so the VC will need to be pretty small anyway. {Edit - Just looked at the datasheet again and it says that Vgsoff can be anywhere from 1 to 7 V?!?  This is why I've stayed away from transistors!}

Thanks guys.  Starting to clear up a bit for me.

[boogietone]

I've never really played around with transistors.  Was always kind of mystified by how they work and generally turned off by what seems to be a very wide tolerance range which seems to mean you have to fiddle around and modify the circuit to get it to behave based on the specific part that you pull out of the box.  

This last bit is exactly right. Remember that when all of this electronics stuff started with tubes and Ge transistors, the tolerances were even more all over the place. Everything needed to be "trimmed." JFETs are the modern day equivalent.