Best available transistor for jfet audio switch?

[Tubis]

What would the best jfet in regards to getting maximum headroom when "on" and minimum leakage when "off" be? One that is available at for example Mouser Electronics - Distributör av elektroniska komponenter

The type of switch im talking about is this. Control voltage 9V, bias voltage 4.5V

[Mick Bailey]

Most Jfets will work, but the optimum performance and current capacity comes with a dedicated switch device such as the 2N4391. This is used frequently in sample and hold circuits where leakage is very important to minimize voltage 'droop'. On resistance is 30 Ohms.

[Tubis]

I have two fets now at home that i tried. The 2sk880 doesnt work, it opens slightly at every peak signal, is that beacuse the cutoff voltage isnt negative enought? The 2sk545 works but seems to start to clip slightly on the positive peak just above line level. Maybe thats the maximum i can get from any transistor?

[Dave H]

The 2sk545 works but seems to start to clip slightly on the positive peak just above line level. Maybe thats the maximum i can get from any transistor?

Perhaps try a 10M resistor from gate to source to remove the stored charge?

[Tubis]

Most Jfets will work, but the optimum performance and current capacity comes with a dedicated switch device such as the 2N4391. This is used frequently in sample and hold circuits where leakage is very important to minimize voltage 'droop'. On resistance is 30 Ohms.

That transistor wont close with the voltages i have.

[Mick Bailey]

Jfets vary a lot, that's why many parameters are given as a range that is sometimes quite wide. Because you've got AC superimposed on the 1/2Vcc, you may need to select a device with more suitable Vgs characteristics.

[Tubis]

Perhaps try a 10M resistor from gate to source to remove the stored charge?

That doesnt seem to work in spice simulation. The transistor wont close properly in off mode.

[Dave H]

The transistor wont close properly in off mode.

Do you mean on mode?

Edit:
Wait, I see it now. You mean close off the signal path and I'm thinking close the switch (FET).

[Mick Bailey]

If youlre simulating in Spice make sure you simulate at min and max Vgs for the particular fet. Actual fets vary so much that you have to pick from a bunch to get an ideal component for a given application, unless there's a good deal of tolerance in the design. Maybe a J113 would work better in this particular application. I've certainly had success with the 2N4391 in 9v circuits, but I've had the luxury of plenty to pick from.

Your problem is that the AC peaks are taking the device out of range of Vgs and causing it to allow the signal to pass on peaks. What is your maximum peak-to-peak on the audio signal?

[Tubis]

Your problem is that the AC peaks are taking the device out of range of Vgs and causing it to allow the signal to pass on peaks. What is your maximum peak-to-peak on the audio signal?

Its basicly guitar pickup level. That would be fine if you only look at the ratings from the manufacturer, for example 400mV(dont know if that is peak or rms) . But when i look att the guitar signal in my scope there is plenty of content way over 2Vpeak on big chords. And with a realy hot pickup that would be even worse.

Im beginning to think that this is why boss and ibanez bypass chains seem to alter the tone. Boss and ibanez uses this even when they boosted the signal, like in an eq.

I could by a lot and pick out the best, but how would one go about testing an smt component? I dont wont to have to solder in and out, that could be a real hassle.

[Mick Bailey]

A more predictable way is to use a bilateral CMOS switch. I use plenty of these for full-range signal switching, usually a CD4066. They're easy to use, cheap, and gives you 4 switches in a single package and will switch +/- 7.5v.

[Tubis]

A more predictable way is to use a bilateral CMOS switch. I use plenty of these for full-range signal switching, usually a CD4066. They're easy to use, cheap, and gives you 4 switches in a single package and will switch +/- 7.5v.

Yes but it switches abruptly, se the other thread I just started. http://music-electronics-forum.com/t40908/

[Dave H]

If it's a circuit you are designing from scratch I think I'd be inclined to use a second battery and have +/-9V supply.

[Tubis]

If it's a circuit you are designing from scratch I think I'd be inclined to use a second battery and have +/-9V supply.

No, i want it to be compatible with a DC9V adaptor. +/- supply would make a lot of things easier but this is going to be for guitar pedals and the 9V supply is a standard.

[Mick Bailey]

If you look at the specs for the 2SK30A-Y, in the context of the circuit you show it would suggest that the devices needed to be specifically selected for that particular application, as the Vgs would be too high at the upper end of the tolerance range. You'll probably face that with any other device.

Dave H's suggestion gets you out of the trap, but if you wanted to stick to a 9v input you could use a charge pump IC to give you 18v and keep the rest of the circuit just the same. Make sure the caps are rated for 25v, though. You'll get much more headroom with 18v.

EDIT: the other device I use for signal switching is the Silonex NSL-32SR3S. It has an off resistance of 25M Ohm and on resistance of 60 Ohms. Switches very smoothly.