Cleanest Transistor for Audio ON/OFF Switching

[bool]

If you plan to stick with the bare-bone one-bjt muting circuit, you can increase the bjt's base isolation by simply adding a small-signal (fe 1N4148) diode in series with the existing base-drive (i.e. "mute") leg (at VOLTAGE/GATE-IN in above schem).

[bartimaeus]

Everyone has made it quite clear that the choice of transistor matters much less than the circuit designed around it. This makes sense to me, but I am a bit surprised that there isn't superstition about transistor use like there is about capacitors, etc. But maybe there is in other applications?

In fact, he has a whole page on (shunt) muting circuits at http://sound.whsites.net/articles/muting.html. Since this is slanted towards power on/off muting use in amps and the like, much of it will not apply to you, and there are some other details he leaves out too, like that there are CMOS analog switches (DG201/411 etc.) available that can operate at voltages up to 44V. If you want options, there are options.

Even though it's not totally directed towards my goal, I REALLY wish I had read this initially. Super helpful!

Transistor muting is done all over the place as Jonnyeye says above. Digital effects use it to mute while you change program and everything jumps to new settings which could cause all kinds of horrible noise. Hi-fi's do it to quieten un-selected sources.

...

In practice, x2 muting transistors, like this (note there's no coupling capacitors or pull-down as it assumes a ground referenced signal path)...

I tested out a shunt switch with a single BJT, but found that it did not sufficiently attenuate the signal. Looks like I needed to use a pair instead... If all of this extra circuitry is required to make a BJT work, I'd rather build a shunt switch with a JFET (assuming that I only need one JFET instead of multiple BJT's).

I'd say the switching JFETS like J201 and the like would be best easy to get single component,
they also tend to be low resistance.
BJT's used like AnotherJim mentioned are OK but probably not as clean as JFETs.

Why exactly are JFETS cleaner than BJT's? I assume that this only applies when they're used as series switches, not shunt switches.

The MC14007 data sheet from ON Semiconductor shows a 2-input analog multiplexer in Figure 1 that can be used as an audio switch.  The nice thing about it is the control lines are not buffered so you can slow them down to minimize switch popping.

https://www.onsemi.com/pub/Collateral/MC14007UB-D.PDF

If you run the device from the 9 volts and bias an inverting op amp at 4.5 volts, you can feed the op amp without using up the voltage excursion limits because the switch is held at 4.5 volts.  Of course, it may pay to have some resistance in series with the inverting input for stability so the capacitance to ground of the MC14007 does not cause a feedback lag that would promote oscillation.

I'm a bit confused as to the benefit of this device over a couple of BJT or JFETS, since it contains several transistors and seems like overkill.

A JFET + reverse diode solution (as used in BOSS/Ibanez/Maxon bypass circuits) looks simple and useful, but it looks like they're non-compatible with a gate signal that only goes positive.
Also, if I have DC-bias on the output when using a JFET, can't I just stick a capacitor on the output to filter out the bias?

Maybe. You're dealing with instrument level signals, on the order of 100mV peak. Any minor DC shifts that show up will be heard as clicks or thumps. A far better solution is to force the input side of the switch to a DC level with a capacitor and a resistor to a bias level, and do the same on the output. That guarantees that if some of the control signal happens to feed through the control pin to the switch itself, it is held in check by the DC level on the in and out pins. What would be best would be to make the "DC bias" be 0V by using a P-channel JFET, and use a 0V gate level for on, a higher voltage for turning it off. It is reverse logic, but then inverters are cheap.

That DC extra filtering sounds quite doable. But could you please explain further about inverting the JFET to accept 0V as on? I don't quite understand.

The price of a H11F1 is not ideal, though, running at about $3 compared to $0.50 for a lot transistors.

This is only an issue if you're making lots of them. The box it goes in will likely be more than that. And you're overpaying for transistors. BJTs are available for US$0.07 to 0.10.

BJT's go for pennies, but JFET's seem to go for around $0.50 even on Tayda. And for a circuit this simple, I could easily fit it in an old tin.

If you plan to stick with the bare-bone one-bjt muting circuit, you can increase the bjt's base isolation by simply adding a small-signal (fe 1N4148) diode in series with the existing base-drive (i.e. "mute") leg (at VOLTAGE/GATE-IN in above schem).

Thanks for the tip!


Also, does anyone have any recommendations for a through-hole alternative to the VO1400? None of Vishay's other SSR's seem to have similar on/off times: http://www.mouser.com/catalog/catalogusd/648/2006.pdf

[R.G.]

I am a bit surprised that there isn't superstition about transistor use like there is about capacitors, etc. But maybe there is in other applications?

Oh, yes, very much so. It is likely that the reason you aren't hearing magic-transistor blather on this topic is that shunt muting and JFET switching are not used in the old-time "magic" pedals. Otherwise we'd hear the same stuff about only transistors made in Sunnyvale in 1968 on thursdays by a team of Icelandic virgins being any good.

I tested out a shunt switch with a single BJT, but found that it did not sufficiently attenuate the signal. Looks like I needed to use a pair instead... If all of this extra circuitry is required to make a BJT work, I'd rather build a shunt switch with a JFET (assuming that I only need one JFET instead of multiple BJT's).

Be very wary about the results from any one circuit test. As you were noting the circuit matters. I was trying to point to this by the comments on how much attenuation you get from any series resistor and BJT pair. If you don't pick a low on resistance BJT, it will not attenuate enough. The series resistor and BJT you happened to use may not be well suited to the source and the load conditions, and so would and probably were not optimal. It takes some thought, calculations, and then getting the right parts for what those turn up to say whether the whole class of BJT attenuators is usable or not.
The series resistor and shunt switch make a "volume control" with the BJT to ground being the bottom portion of the volume control If you can't make that be zero ohms, you have the "volume control" still a little bit above fully off. There are no semiconductor switches that get down to metal-contact switch resistances, although big MOSFETs come close. I was trying to explain that the attenuation depends on both the BJT "on" resistance and the resistor, and that the resistor is constrained by other things than just the BJT, and that led to the possible need for multiple stages.

As for "all this extra circuitry", be careful what you ask for. You're comparing a fully developed two stage attenuator with your idea that you can do it with just one transistor. This circuit includes two transistors to feed the two bipolars with a clean on off signal of just the right amount and size. A single BJT **can** be made to work with just a resistor to feed it a signal on its gate. Exactly **how well** it works depends on (a) how good "good enough" is, and how much you have to add to the basic, trivial circuit to get to your "good enough". It also depends in this case on which BJT you pick. As I said, some BJTs are better than others. It's just not widely known what to select for. JFETs need a bias voltage on drain and collector, generally capacitor isolation on each end, usually a diode, resistor, and a cap for each FET, and then some kind of circuit to feed it a gate voltage; the first two bipolars in that circuit are this "circuit to feed it a gate voltage.

Why exactly are JFETS cleaner than BJT's? I assume that this only applies when they're used as series switches, not shunt switches.

As I described in my note, JFETs are inherently resistors made skinnier or fatter by the action of the voltage on the gate; they're variable resistors. BJTs are inherently diodes, and diodes are inherently non-linear. BJTs as a switch rely on their diode-ish-ness being hidden by their low relative resistance compared to other things in the circuit, and always retain a tiny bit of diode conduction distortion. This is least when they are off - infinite nonlinear resistors are almost as linear as infinite linear resistors - or fully on. It's between worse and horrible in the middle, the degree of worseness being proportional to how big the signal across it is.

I'm a bit confused as to the benefit of this device over a couple of BJT or JFETS, since it contains several transistors and seems like overkill.

(1) It is a bunch of matched MOSFETs in one package that can be wired up various ways.
(2) It is a bunch of matched MOSFETs that can be strapped to be a complementary MOSFET analog switch and use the inherent linearity of the MOSFET as a variable resistor without having the problem of the body diode intrude.
(3) It's cheap, around US$0.50 new, maybe much less surplus.
You seem to be confusing more transistors with too complex and overkill. You're hearing a lot of people tell you that using just one transistor may not do what you want. If it takes more stuff to do what you really want, you're stuck with more stuff not being overkill. Also, overkill isn't necessarily bad. If overkill makes what you want to get done be trivally easy for the solution to do well, then it's not necessarily overkill. In a manufacturing situation, it's reasonable to design right down to the fewest of everything. In making one or two, using more parts to do it easily and well isn't that bad a choice.

That DC extra filtering sounds quite doable. But could you please explain further about inverting the JFET to accept 0V as on? I don't quite understand.

There are two kinds of bipolars, NPN and PNP. They are the same in how they work, only the voltages have to be reversed. They're kind of mirror images. Likewise, there are two kinds of JFET, N-channel and P-channel which are mirror images in terms of how they operate. For N-channel devices, letting or making the gate be the same voltage as the source leaves the drain-source channel at its lowest resistance, and making the gate be some volts -negative- compared to the source makes it higher resistance up to effectively turned completly off. A P-channel FET is a low resistance when the gate is at the same voltage as the source, and turns off when the gate is pulled -positive- compared to the source.

So for N-channel, you have to attach the drain and source to some fixed voltage and turn it off by pulling the gate negative. For circuits with a positive power supply, you have to pull the drain and source up from zero by enough to let the JFET turn off if you pull its gate to ground. For a P-channel JFET, it is normally conducting if the gate is at the same voltage as the source, but turns off if you pull the gate positive. I've always thought that JFET switching in a positive voltage circuit would best be done with P-FETs. This allows you to make the "bias voltage" for the drain and source be 0V, or ground. The turn-off voltage then become yanking the gate up to the power supply. It is more expeditious, and doesn't force some restrictions on the bias voltage. The comment on "inverting" was just that the logic of the gate voltage is inverted, from off being zero volts and on being high for the N-FET to on being zero volts and off being high with the P-FET.

Frankly, I think that all the industrial switching with JFETs would be with PFETs if they weren't more expensive per device. This is another case of the right device being a tiny bit more expensive and not favored  for reasons that don't matter to us one-off guys.

BJT's go for pennies, but JFET's seem to go for around $0.50 even on Tayda. And for a circuit this simple, I could easily fit it in an old tin.

Yep. What you're telling us is that in fact, the absolute cost of the parts is a critical design decision to you in this matter. A whole lot of your comments here are asking for fewer parts and cheaper parts. I was there once, when I was a starving student, and another dollar for parts was a dollar I didn't have to feed myself and my wife. Dollars were worth about five times as much back then, but the concept is the same. I don't know if your situation is even comparable in any way, or if you're just being careful with your money for your own reasons. Over the years, I accumulated more money to buy parts, and also have a fairly large stock of leftover parts, so I am in a condition where how well a circuit works and looks is getting more important relative to the bare minimum parts number and costs.

For whichever reasons though, you are making the parts cost be a minimum be a design criteria. That's fine, but you need to realize that you're doing that.

Other people have other criteria. You may want to read "Effects Economics 101" at geofex.com. Preferring an old tin as an enclosure over a commercially made box is another way of saying that the cost of the box is more important to you than how the finished box looks and how durable it is. There are other consequences of making the design rule be "cheapest and fewest parts". Again, that's fine, but do consciously recognize you're doing it, and state it that way.

[digi2t]

I am a bit surprised that there isn't superstition about transistor use like there is about capacitors, etc. But maybe there is in other applications?

Oh, yes, very much so. It is likely that the reason you aren't hearing magic-transistor blather on this topic is that shunt muting and JFET switching are not used in the old-time "magic" pedals. Otherwise we'd hear the same stuff about only transistors made in Sunnyvale in 1968 on thursdays by a team of Icelandic virgins being any good.

Yeah.... like reverb tanks...

[Rob Strand]

Why exactly are JFETS cleaner than BJT's? I assume that this only applies when they're used as series switches, not shunt switches.

I'd say JFETs are cleaner than BJTs because with a BJT the DC level on the switched terminals changes causing a glitch.  There is no DC with JFET (with the reverse diode gate circuit like Boss/Ibanez use).

JFETs are clean because:
- The gate is largely isolated from the signal (drain/source).
- They have low capacitance from the gate to the drain/source.
- They have a variable resistance characteristic.   
   When you switch from one signal to another the two signals might not join-up and this
   discontinuity causes a glitch.   By slowly blending between the signals the join is smoother
   which makes it less audible.

JFETs do all the right things.  You can also get fairly low resistance.

[One caveat:  For the case of muting, where you want to kill the signal, the BJT can provide more attenuation than a JFET as the effective resistance is (/can be) lower.  This is outright suitability to the specific problem, nothing to do with glitches and cleanness.]

[R.G.]

It's common when using the BJT as a muting device to put a BFC* in series with the collector. This isolates the circuit being muted from the inevitable offset voltage at the collector. This itself is a problem, because the cap must be "large" to have a reactance that's compared to the BJT on resistance, which is quite low. This is commonly several hundred microfarads. It's another PITA with the BJT muting scheme that is often unappreciated until the last minute.

I'm surprised no one has mentioned the obvious: use two JFETs, one series and one shunt after it. You feed them the correct gate voltages and bingo, you have a really, really off "off-switch" and a much better/lower on resistance than with a shunt mute only.

This violates the OP's design rule to use only one transistor, but you get really good results. It generally requires opposed gate signals, as you want the series off when the shunt is on, and vice versa, but the application of an N-channel series device with a P-channel shunt device works well and only a single polarity of control signal.

*Big Freaking Cap

[Rob Strand]

It's common when using the BJT as a muting device to put a BFC* in series with the collector. This isolates the circuit being muted from the inevitable offset voltage at the collector. This itself is a problem, because the cap must be "large" to have a reactance that's compared to the BJT on resistance, which is quite low. This is commonly several hundred microfarads. It's another PITA with the BJT muting scheme that is often unappreciated until the last minute.

The BFC gets rid of the *steady-state* DC but it doesn't get rid of the glitch.  There is a DC *change* on the BJT before and after the muting; it is like injecting a step signal.  One component of DC is the (non-ohmic) saturation voltage, the other component is the ohmic drop due to the base-current.  At line levels, like on CD player outputs, the DC level is small compared to the signal but as the signal level drops it becomes more of an issue.

I'm surprised no one has mentioned the obvious: use two JFETs, one series and one shunt after it.

Yes, that will fix it!

You also see this configuration in modulators and synchronous rectifiers - often so the following stage sees 0V with low source impedance.

[bool]

Yes with BJT switching there can be a little (couple mV's) DC leak from the base to collector; but that can be minimized with some attention to the base "mute" signal current.

The BJT base has to saturate for efficient muting and the associated Vbe voltage would leak through via the B-C "diode"; but if the signal source impedance is sufficiently low; practically a high-enough beta BJT usually shouldn't cause too much DC leaks/troubles to get it working.

Getting the "mute" signal right would be one place too look at when designing and optimizing the simple circuit. But this usually boils down to selecting the right BJT ... and the right base resistor. And as I suggested, putting a small-signal diode there to isolate the BJT base when the "mute" logic signal is at "low" (iow, "play" position).

[anotherjim]

In some, like CE-1 chorus.

See Q12 for a J-FET shunt mute, But, note it is controlled by a negative envelope follower to ensure the gate of the N-channel can definitely pinch-off the FET.

[amptramp]

The Heathkit AJ-1510A FM tuner (the first digital tuner) used diodes wither forward or reverse biased to block or pass signal.  They used a set of diodes to ground and a set in series with the signal to get the degree of cutoff they needed.

In the MC14007 CMOS implementation I suggested earlier, you could connect on input to the signal and the other to the Vcc/2 ground so as well as the megohms of resistance in series with the signal, you could get ~100 ohms to ground.  This should be good for about 80 db of signal suppression in cutoff.

[R.G.]

Yes, the BJT muting scheme requires careful attention to both the transistor and the drive signal to it, as well as the source impedance, the series resistor that is the other half of the attenuator, and the load impedance it drives. The nature of the drive signal is important because of the leakage of the B-C junction; that junction has to be forward biased to get good saturation, as the base can't be any lower than 600-700mv, and the collector has to be significantly lower than that to be saturated to a low voltage, which is what you want for low "resistance". So you have to carefully tinker the base drive for both low saturation resistance and low feedthrough.

Then there is the issue of the transistor. In general, low saturation voltages are antithetical to high beta at equal current levels. Something to do with needing low resistivity silicon for low saturation voltages and that not being very good for high beta. That is old information and silicon has moved on to much better levels of refinement these days, so it may have been fixed. But I suspect that it's only been fixed for new-designed transistors special to the purpose. There's not a huge market for BJT switches these days, given how good FETs have become. Probably best is finding the types that were actually used in the Japanese hifi stuff back in the day. They had the incentive to design transistors to do that job.

Again, not that it can't work - but a pair of JFETs, well designed, will do better.

[bartimaeus]

Thanks again for all of the replies!

BJT's go for pennies, but JFET's seem to go for around $0.50 even on Tayda. And for a circuit this simple, I could easily fit it in an old tin.

Yep. What you're telling us is that in fact, the absolute cost of the parts is a critical design decision to you in this matter. A whole lot of your comments here are asking for fewer parts and cheaper parts. I was there once, when I was a starving student, and another dollar for parts was a dollar I didn't have to feed myself and my wife. Dollars were worth about five times as much back then, but the concept is the same. I don't know if your situation is even comparable in any way, or if you're just being careful with your money for your own reasons. Over the years, I accumulated more money to buy parts, and also have a fairly large stock of leftover parts, so I am in a condition where how well a circuit works and looks is getting more important relative to the bare minimum parts number and costs.

For whichever reasons though, you are making the parts cost be a minimum be a design criteria. That's fine, but you need to realize that you're doing that.

I apologize for being unclear about this. I'm lucky enough that I can currently afford the more expensive options listed above, especially since none of them are astronomically-priced vintage parts. I started this thread looking for a swap-in replacement for a circuit that was accomplishing it's purpose, if with coloration. I ended up trying to keep the circuit basic for simplicity-of-design reasons more than anything else – to see how much I could improve signal transparency with an equally simple schematic. As you surmised, I didn't quite realize these limitations which I was putting on the design.

That DC extra filtering sounds quite doable. But could you please explain further about inverting the JFET to accept 0V as on? I don't quite understand.

So for N-channel, you have to attach the drain and source to some fixed voltage and turn it off by pulling the gate negative. For circuits with a positive power supply, you have to pull the drain and source up from zero by enough to let the JFET turn off if you pull its gate to ground. For a P-channel JFET, it is normally conducting if the gate is at the same voltage as the source, but turns off if you pull the gate positive. I've always thought that JFET switching in a positive voltage circuit would best be done with P-FETs. This allows you to make the "bias voltage" for the drain and source be 0V, or ground. The turn-off voltage then become yanking the gate up to the power supply. It is more expeditious, and doesn't force some restrictions on the bias voltage. The comment on "inverting" was just that the logic of the gate voltage is inverted, from off being zero volts and on being high for the N-FET to on being zero volts and off being high with the P-FET.

Frankly, I think that all the industrial switching with JFETs would be with PFETs if they weren't more expensive per device. This is another case of the right device being a tiny bit more expensive and not favored  for reasons that don't matter to us one-off guys.

Thank you very much for explaining this!!! For my particular muting application, P-Channel JFET's seem like a good compromise between BJT's and SSR's. I ordered a couple P-Channel JFET's to test out.

I'm a bit confused as to the benefit of this device over a couple of BJT or JFETS, since it contains several transistors and seems like overkill.

(1) It is a bunch of matched MOSFETs in one package that can be wired up various ways.
(2) It is a bunch of matched MOSFETs that can be strapped to be a complementary MOSFET analog switch and use the inherent linearity of the MOSFET as a variable resistor without having the problem of the body diode intrude.
(3) It's cheap, around US$0.50 new, maybe much less surplus.
You seem to be confusing more transistors with too complex and overkill. You're hearing a lot of people tell you that using just one transistor may not do what you want. If it takes more stuff to do what you really want, you're stuck with more stuff not being overkill. Also, overkill isn't necessarily bad. If overkill makes what you want to get done be trivally easy for the solution to do well, then it's not necessarily overkill. In a manufacturing situation, it's reasonable to design right down to the fewest of everything. In making one or two, using more parts to do it easily and well isn't that bad a choice.

In the MC14007 CMOS implementation I suggested earlier, you could connect on input to the signal and the other to the Vcc/2 ground so as well as the megohms of resistance in series with the signal, you could get ~100 ohms to ground.  This should be good for about 80 db of signal suppression in cutoff.

It seemed like overkill to me because even some of the more complex solutions offered here don't utilize it's full capabilities. That said, I do see how it would simplify the transistor options, and it does seem like an economical solution.

Why exactly are JFETS cleaner than BJT's? I assume that this only applies when they're used as series switches, not shunt switches.

I'd say JFETs are cleaner than BJTs because with a BJT the DC level on the switched terminals changes causing a glitch.  There is no DC with JFET (with the reverse diode gate circuit like Boss/Ibanez use).

JFETs are clean because:
- The gate is largely isolated from the signal (drain/source).
- They have low capacitance from the gate to the drain/source.
- They have a variable resistance characteristic.   
   When you switch from one signal to another the two signals might not join-up and this
   discontinuity causes a glitch.   By slowly blending between the signals the join is smoother
   which makes it less audible.

JFETs do all the right things.  You can also get fairly low resistance.

[One caveat:  For the case of muting, where you want to kill the signal, the BJT can provide more attenuation than a JFET as the effective resistance is (/can be) lower.  This is outright suitability to the specific problem, nothing to do with glitches and cleanness.]

This is a very clear explanation, thank you for the help! So it seems like a JFET is much better as a series mute/switch, since the gate is isolated and you get no glitches, while a BJT is better for a shunt mute/switch, since it offers more attenuation and the audio signal doesn't flow through the transistor.

Do I need to worry about isolation of the gate/base in the case of shunt switches? It seems like  yes, but you can add extra circuitry to minimize it, which is why several BJT with supporting architecture is a solution used in many Japanese products?

I'm surprised no one has mentioned the obvious: use two JFETs, one series and one shunt after it. You feed them the correct gate voltages and bingo, you have a really, really off "off-switch" and a much better/lower on resistance than with a shunt mute only.

This violates the OP's design rule to use only one transistor, but you get really good results. It generally requires opposed gate signals, as you want the series off when the shunt is on, and vice versa, but the application of an N-channel series device with a P-channel shunt device works well and only a single polarity of control signal.

I find this solution very cool/interesting, since it mixes techniques, but if we're talking about dual-transistor solutions then I don't quite understand its benefit over using a pair of shunt or series switches, or even just a single shunt switch (which seems sufficient for many effects)?

In some, like CE-1 chorus.

See Q12 for a J-FET shunt mute, But, note it is controlled by a negative envelope follower to ensure the gate of the N-channel can definitely pinch-off the FET.

I'll try building something similar with a P-channel JFET once the ones I ordered arrive. They seem to get by with a single shunt switch. These old BOSS/Ibaneze/Maxon circuits all have slightly different support architecture for the switches, so I need to start figuring them out.

[bool]

Again, not that it can't work - but a pair of JFETs, well designed, will do better.

Sure will do much better but with fets you have to tinker with the gate drive logic vs tweaking two resistors for the bjt.

For fets, iirc D.Self had a series/shunt circuit block in one of his books.

[R.G.]

Sure will do much better but with fets you have to tinker with the gate drive logic vs tweaking two resistors for the bjt.

Yep. For better results you have to work a little harder, or smarter, or both. The trick is to work the least harder for the most better. That's always a personal tradeoff, or an  outgrowth of the "rules" you set yourself for the project.

Plus you have to take into account the rest of the circuit. You're at the edge of that point, but not quite on it. Generating the control signal nearly always means using whatever is most expedient in the rest of the circuit. So for the two BJT case, you have to use two resistors and two diodes, and generate a switched high voltage. Pretty easy.

For the series-shunt JFET version, you have the option of using a NFET and a PFET as the series and shunt elements, two resistors and two caps for voltage positioning of the NFET, and the >same< gate voltage on each one. The different polarities of the FETs do the inverted logic for you.

It takes one more cap than the BJT case, and one less cap than the BJTs if you do wind up not achieving acceptable feedthrough with carefully tinkered base drive.

Which is better? As always, it depends on your scoring criteria.The FETs will cost more. The BJTs will do a less total attenuation than the JFETs.  Either version may or may not have more thinking required for getting the control signal right, and whether you already have a simple way to generate the control signal in the rest of your circuit already is unknown until you have the rest of the circuit defined.

Like the answer to most interesting design questions, the answer to this issue is - it depends.

For fets, iirc D.Self had a series/shunt circuit block in one of his books.

Yes, he did. I believer it was in his small-signal audio book.

[bool]

The most hardcore and "audiophile" way to do it would be to use a small relay.

Higher cost than any of the semiconductor, but you'd get the bragging rights. (IF and only IF you choose one that has gold-plated contacts).

To make things manageable space-wise; you could use some small DIP8 relay. Also you can wire it in series/shunt config if the relay is a spdt/dpdt type.


Fets always made me a little un-easy; personally I wouldn't use them without some buffer amp after the attenuator. Bjts, otoh, I've seen them at the very end of the signal chain (output); with say "audiophile" computer audio cards, and I used them as a simple kludgy "bolt on" to an existing circuit as well - I feel they're a little sturdier.

[anotherjim]

I'm under the impression that the mute was needed for some kind of chopper/stutter effect?

JFET's (any FET) too have signal/gate interaction if you aren't careful.
1: If signal can wiggle the source voltage, it can change the gate-source voltage, turning it away from the intended switch state.
2: The signal can cause forward bias of the gate-source/drain diode. As with BJT base, this can have an extra diode in the gate for more headroom. Gate can be supplied via a reversed signal type diode. The diode reverse leakage is enough for the gate to see the control voltage - everything else sees 2 back to back diodes and the control path is effectively blocked from the signal.


No matter how perfect the muting action, it can make an audible click when signal is present if there's an abrupt switching action. As already mentioned, a little slewing of the control signal can give a short fade to smooth out a glitch. Up to 20ms will still be heard as a fast switch without causing a click.

With series/shunt, it can be a little tricky to get the right slope of the control. The series and shunt happening at the same time speeds up the switch between signal and silence. Also,  each FET may have overlapping thresholds. The result of that is a very short span of the control slope to get a fade in or out effect.
I tried this with 4007 MOSFETs. P for series and N for shunt with a common gate control should make this easy, well...
Once you find a working compromise between signal amplitude & the reference voltage to keep switch states unaffected by signal, getting a smooth, glitch free signal fade proved a complete PITA. It turned out to be easier to use only an N channel as shunt.

[bool]

When you factor all the pieces that you need for a simple shunt jfet version and a bjt version, and with a RC smoother for a click suppression; you get to approx 3x resistors, 1 diode (for gate/base isolation), one cap and one transistor.

In case of a fet; you will probably use a ceramic/poly cap in 100nf's range, with a bjt you will use a cap in 10uf's range. If smt; both circuits are going to be almost exact size (if you use a small tantalum in case of bjt; but there are also small ceramics in 10uf's range).

[bartimaeus]

R.G., did you see my post above addressing many of your points? I think you may have confused me with user bool – no worries if so!

I'm under the impression that the mute was needed for some kind of chopper/stutter effect?

Something in that ballpark, yep! Which is why a standard relay wouldn't work: I'm worried about its life-expectancy.

It seems I'd only need half of that circuit, since I'm not switching between inputs. Looks like this one would need 2x resistors, 2x caps, 1 diode, and 1 transistor. But does this have sufficient click suppression? I don't see any series resistors after the switch, although I know that those elec caps will help.

If using a combo of series and shunt introduces issues with synchronization, I'd really rather not bother. I also still don't quite understand why the combo would be necessary when muting a single signal? Does it just offer the most possible attenuation?

[mimmotronics]

If you are wondering how to slew the transition ON/OFF time know all that is needed is an RC circuit. I implemented this on a Decimator that was giving me popping issues past the 12:00 position. (See mimmotronics.com/er8)

[anotherjim]

If using a combo of series and shunt introduces issues with synchronization, I'd really rather not bother. I also still don't quite understand why the combo would be necessary when muting a single signal? Does it just offer the most possible attenuation?

JFET on-resistance isn't low enough for a total kill shunt unless used against a much higher series resistance from the input. Having a series FET cut off the input gives that higher input resistance. Think voltage divider.  It can be necessary to have 2 shunts in series to get closer to silence.
I only posted the above audio switch scheme as an example of the reversed gate isolation diode trick.