Transistor Bypass

[Gobotak]

I am trying to use transistors for switching audio but am having issues with them not fully turning off.  I am basically using the boss method of bypass but without the flip flop section of the circuit.  The transistors I am using are 2SK3019s.  I have tested using a single 9V supply and biased the drain and source to 4.5V with 1M resistors. I have a 4148 to the gate and a 47nF capacitor to ground and 1M resistor to either 9V or ground to turn the transistor on and off. The transistors work when turned on to pass audio through but still let some signal through when turned off.  I have tried multiple different transistors (all 2SK3019) with the same results. 

Does anyone have any ideas of what could be causing this or ideas to remedy the issue?

[antonis]

A draft schematic would help..
(and you should save two rows of typing..)

[Gobotak]

Apologies for not including a schematic on the first post.  Here is how I currently have it:  https://i.imgur.com/O0jEhfN.jpeg

[bluebunny]

Don't be shy!  Let's get that image inline for everyone to see:

[antonis]

I should place a Gate to Source pull-down resistor (in the region of some kilos) to make sure for V_GS = 0V

[anotherjim]

Could the mosfet body diode be an issue?

[merlinb]

the mosfet body diode

+1   Use a JFET, not a MOSFET

[Gobotak]

So using the 2sk3019 was a poor choice?  I thought I read that Boss used 2sk30s but maybe this specific variety is a no go. This is the datasheet for the one I am using.
https://www.mouser.com/datasheet/2/258/2SK3019_SOT_523_-2510059.pdf   Any suggestions for a substitute transistor?  Or what to look for when choosing one?

[GibsonGM]

What they're saying is that the 2SK3019 is  MOSFET.  The 2SK30 is a JFET.   There is a difference! 

Suitable JFETs you could try are 2N5292, 2N5485, 2SK118Y - according to a search on here I did of this topic.  There is probably more info on this 'archived' here, if you look.  FWIW, I've tried this and had the same 'not quite off' results...I chaulk that up to using a JFET not up to the task and not wanting to order any of the above-listed just for an experiment. 

[anotherjim]

You will find plenty of examples using discrete MOSFETs as switches, but not for small signal work. We want a bi-directional AC signal control with negligible distortion. The one place MOSFET technology works for this are in CMOS ICs where the body diode is turned into a substrate connected to a supply pin and doesn't pass over the drain and source of the FET as it does in the discrete devices. Examples are the cd4016, 4066, 4053 and also the 4007. To use, you carry on with the resistors holding the input and output at vref. However, the control input pins are logic level inputs so any RC slow-down on those pins may only delay the command and not introduce any fade in/out.
You may still get some slight bleed through while off but with the 4053, which contains SPDT switches, you can arrange for it to not only disconnect the input but also connect the output directly to vref so any bleed is effectively shorted out.

[amptramp]

I started a thread about silent switching some time ago:

https://www.diystompboxes.com/smfforum/index.php?topic=120006.msg1122277#msg1122277

Due to the problem with postimages, the actual schematics are in the last post on this thread.  The 4007 is available pretty much everywhere whereas specialty FET's are not always that easy to get or guarantee continuous production and there is a method as shown to get fast switches to provide silent switching.  There are more options than the Boss switching circuit.  In most circuits, the series resistance of several hundred ohms for a CMOS switch is not important and for double-pole switching, the output of the active pole usually swamps the leakage through the "off' pole.