Noob question- How to choose transistors in modulation circuits

[Transistor-Transistor]

I've been looking at schematics to build a tremolo. I haven't worked with LFO's too much before, and in the schematics they always use some kind of transistor to control the modulation via the wave. What transistors should be used there? I've seen JFETS, MOSFETS, and even a standard BJT. Does it matter? What would make it matter? How do I know? What resources should I know about to learn more about this? Is there an official term for what I described?

[GibsonGM]

Don't forget opamps, too!    It will be interesting to see the replies to this, from a highly technical viewpoint. Why devices are chosen and how their characteristics are exploited or accounted for.

In my case...I usually copy am inspired by a complete LFO, and I don't wonder too much about the device characteristics...

[antonis]

A good BJT to deal with..

[Matthew Sanford]

First, pull up schematics on tremelos, phasers, etc., to study how they use different things for modulation with a control voltage. Some may use vactrols or VCAs for the same purpose. The thing to get into knowing is the difference on how they function; BJT uses base current to determine how much current goes from collector to emitter, FETs use gate voltage to control the same. They all can have different effects on what they are controlling, for instance a BJT's base current will make it out the emitter where as a FET should not really use any current on the gate. So on that end, how much current can the control voltage LFO provide?

Best thing to do is look at all the different ways you see in schematics and try to understand the current based on resistors that limit, etc. Try them out, see what has the sound you want. Make sure to change up your parts to accomodate whichever transistor you choose based on info from the schematic and it's datasheet - if it doesn't go as expected, post here with all the voltages and values of parts, which that kind of info people here can dig into your thinking and get you on to a better understanding. I'm pretty darn noob or trying to get past it, and really that is the kind of thing that brought me along - mostly because of the wonderful folk here giving me little tid bits that hit right to knock me back on the path.

[Rob Strand]

I've been looking at schematics to build a tremolo. I haven't worked with LFO's too much before, and in the schematics they always use some kind of transistor to control the modulation via the wave. What transistors should be used there? I've seen JFETS, MOSFETS, and even a standard BJT. Does it matter? What would make it matter? How do I know? What resources should I know about to learn more about this? Is there an official term for what I described?

From a general controlled resistance point of view:

JFETs work the best for a voltage controlled resistance.

MOSFETs will work sort of like JFETs but the MOSFET body diode can clip the signal.

BJTs are probably the worst since they only behave like resistors for very small signals.   The control voltage also leaks into the audio.  (The base current sets the resistance.)

You also have diodes.   Here's a good example of using diodes which avoids a lot of the down-falls of diode-based tremolos:

https://www.diystompboxes.com/smfforum/index.php?topic=119901.msg1124797#msg1124797

People have got all types to work.   The BJT units work but they are pushed closed to the line.

[amptramp]

Have you considered tubes?  My Hammond Solovox uses push-pull 6SK7 remote-cutoff pentodes to drive the push-pull 6K6 output tubes.  Volume control is provided by bias voltage on the 6SK7 grids.  This could be used to provide a tremolo function but the control voltage would have to be in the 0 to -10 volt range to be effective.

[duck_arse]

Is there an official term for what I described?

It will be interesting to see the replies to this

shunt.

[Transistor-Transistor]

Have you considered tubes?  My Hammond Solovox uses push-pull 6SK7 remote-cutoff pentodes to drive the push-pull 6K6 output tubes.  Volume control is provided by bias voltage on the 6SK7 grids.  This could be used to provide a tremolo function but the control voltage would have to be in the 0 to -10 volt range to be effective.

I would love to use tubes, but for the project I'm working on space is a problem. If you'd like to know I'm gonna try and make something like 2 tremolos in parallel going at different speeds, I thought that may sound nice.

[Transistor-Transistor]

Is there an official term for what I described?

It will be interesting to see the replies to this

shunt.

Thank you so much

[R.G.]

If I read correctly, you're asking about the element used to actually do the modulation. Rob's notes are good, as are the comments on diodes and vacuum tubes.

Since you asked about how to choose, I'll philosophize a bit. The effects world is crying for a good voltage controlled resistor (VCR) with:
- wide range of resistance variation
- narrow control voltage range - not too much voltage needed to change from max to min resistance
- big signal voltage handling range without a lot of distortion
- wide frequency response, not changing impedance or losing treble, etc.

As Rob notes, the JFET probably comes closest to a VCR. In fact, some semiconductor makers sold JFETs preselected for variable resistor use. Some early MOSFETs (the old 3N series) used a different semiconductor manufacturing process and were not prone to the body diode problems with modern MOSFETs, but these are about extinct now. If you are going to use a voltage LFO waveform, the JFET is probably where to start with a modulator.

In general, for bipolar devices, you will want to control them with a variable current through them, as they are more inherently viewed as current variable in this kind of application. This can be as simple as feeding an LFO voltage through a resistor to convert it into a varying current.

The BJT and diode VCRs both suffer from low signal handling, again as Rob noted. A "resistor" is anything that has a current linearly proportional to the voltage across it. As you apply smaller and smaller signals, every odd-looking curve of current versus voltage starts looking more like a straight line. At some point, you can call it a resistor with some contamination. For silicon junctions, the magic voltage of linearity is below about 25mV to 50mV. A signal in this range doesn't have too much audible distortion. This is the basis of the modulators that use BJTs and diodes.

As a side note, BJTs act like BJTs with emitter and collector reversed, but with different breakdown voltages. In the reversed format, a BJT generally has a better linearity as a modulator, so you will see designs using a BJT as a shunt-to-ground modulator in the inverted mode. In general, for a BJT modulator, an otherwise "good" BJT is best – high current gain, high linearity as an amplifier, etc.

Both BJTs and diodes are really current controlled modulators (CCR). They are usually used by forcing a control current through them. The signal to be modulated is impressed across the BJT/diode with capacitors to wipe off the DC levels in some way, and the current in the device is modulated. This is often done with an external resistor from the modulating LFO voltage to convert that LFO voltage to a current through the BJT/diode.

A great example of diodes used this way is the diode modulator bridge, as used in the Thomas Organ Vox amplifiers from the 1960s.  Four diodes are arranged with two diodes in series, and two of these strings in parallel. At the "top", two anodes connect, and two cathodes connect at the "bottom"/negative side of the bridge. Signal is fed into the middle of one diode string with a capacitor, and taken out of the middle of the other side. The top and bottom are fed mirror-image LFO voltages through large (100K-1M) resistors to control the currents through the diodes. This lets you use double the signal voltage, as it is divided across at least two diodes on its way across the bridge; diode distortion is first-order canceled by the bridge connection. It's handy, efficient, and low distortion, and suffers only from the small signal handling level of diodes.

The flaws of BJTs used as modulators are usually attacked by using the BJTs as a differential pair, and modulating the current through the diffamp by controlling the current in the shared emitters with a third BJT. This works well for either small voltage or current signals into the diffamp bases. This works so well that dearly departed Motorola made an IC of it – the MC1496 modulator. It is also the basis of all OTAs, as well as the commercial VCA designs. The down side is it uses a lot of parts and needs matching to be really good.

Two modulating schemes that didn't get mentioned are the Light Dependent Resistor (LDR) and Metal Oxide Varistors (MOV).  The scheme of modulating current through an LED which lights up an LDR has been the go-to modulator for pedals for a long time – at least 60 years, probably longer. MOVs are funny, and good ones are hard to find. MOVs are really voltage variable resistors. Like diodes, they have an incremental slope of current and voltage that makes them look like resistors for signal voltages that are small enough compared to the curve. Today's voltage-protector MOVs have been optimized to break over sharply, but early ones had a smooth curve of resistance versus voltage across them over a hundred volts or more. These old style MOVs were the basis of the Magnatone amp vibrato.

[amptramp]

You don't need to overthink this.  the Univox U65M amplifier does this for tremolo:



On the lower left, you will see the tremolo phase-shift oscillator and the transistor used to modulate the signal.  Even the word "transistor" says something about the device - it was originally called a "transfer resistor" before the shortened name came into effect.  The base current transfers its control to the collector current.

It is difficult for anyone to tell an ideal tremolo resistance profile from what this circuit achieves, so don't stress about getting an ideal solution.

[Transistor-Transistor]

I decided to experiment with a simple phase shift oscillator tremolo. The 2n2222 worked surprisingly well as a modulator. As you've all told me there was some noise/oscillation in the output so filtered it out and boosted the signal back to unity which worked perfect. Next I'm going to try the differential pair like RG suggested. Thank you all so much for your answers! I'm glad to be educated