biasing a 2N7000 switched parallel to a MPF102.

[sean k]

As you can see I've got a switch to go from the MPF102 to the 2N7000 but the 2N7000 doesn't work.I thought at first it was just a bum transistor but tried another one and still no go so now I'm thinking you can't set up a mosfet the same way as a j-fet,or at least in this instant.With the biasing used in the multiface it works,I switch from j-fets to mosfets but in this case it doesn't work.
 Maybe the 1M resistor is the culprit and should go from gate to source?

 Any ideas to get the mosfet functioning?even if I have to take the mosfets source away from the j-fets source.

 Oh,and its the first stage of Joe G's decade and the 510k resistor at the input is actually 1Meg.

[davebungo]

The n-channel JFET requires a negative VGS whereas the n channel enhancement mode MOSFET requires a positive VGS to turn it on.  You need to form a bias network using a potential divider between 9V and 0V and feed this into the MOSFET gate via a 1Mohm resistor.  Connect your input directly to the gate so your input signal sees the 1M as the effective input impedance.

[sean k]

Dave,
        So what your saying is that I should lift the 1M,that goes from gate to drain,at the drain end and reattach that end to a voltage divider between the 9V rail and the 0V rail?,any particular voltage I should aim for,maybe 33k and 10k?

[davebungo]

Yes.  I would have though 2 or 3V VGS should do it but it would be better to have some adjustment.  Don't forget that because of the source resistor (the 25K) VS will be positive wrt 0V so VG will have to 2 or 3V higher than this.  To be honest I would maybe reduce or omit Rs for the MOSFET circuit because the reason it is there in the first place is to provide negative bias for the JFET transistor configuration.  This is not required for the MOSFET.  The only reason for having Rs in a MOSFET circuit like this would be to provide some negative feedback in order to stabilise the gain but as you have it bypassed with a capacitor it has little effect (except at low frequencies).

[davebungo]

I'm not so sure about this circuit because it appears to have negative feedback from the drain to the gate via the 1M resistor.  i.e. as the input signal tries to drive positive going, the drain will go down and mix some of this signal with the input hence negative feedback.  A similar effect could be achieved by putting some resistance in series with the source bypass cap.  i.e. the 25K would set the DC bias point (along with the trim pot in the drain) but this extra resistance would provide some negative feedback at audio frequencies also.  ie. if you remove the 1Mohm between the drain and the gate it may increase the resulting gain somewhat, but I'm not sure by how much without thinking about it a bit more.

[sean k]

Well it works now and this is what I ended up doing


 Though its kinda weird as when you switch from j-fet to mosfet it takes about a half second for the mosfet to come in.But switching the other way its pretty much instantaneous...maybe a touch of lag.Its almost as if the mosfet needs to charge.The drain on the mosfet is 6.1V and the gate is at 1.36V.
 Its the first stage of the decade and the only other changes are to the pot for the diodes which has been changed to 100k,which means my dual pot has 25k for the j-fet emitter resistance and 100k for the clipping diodes, and diodes switched between Ge and Si.

[davebungo]

A couple of other things I would probably do:
1) Move the 1Mohm to ground on the input to the input of the JFET stage so it only appears when the JFET is switched in.  This is because the input is already tied down via the other 1Mohm for the MOSFET stage.  
2) Put a decoupling cap across the wiper of the MOSFET bias pot to ground 10-100uF should do the trick.

The slow start-up of the MOSFET is because the bias voltage rises slowly as it charge the input capacitor.  You could avoid this by having separate circuits altogether and simply switching between the outputs so they are both always in a stable bias situation although this would mean that they are both presented to the input as a load albeit a fairly light one.

There are many ways to skin a cat.

[sean k]

Thanks Dave,
                   I'll do the two things you suggest and I think I might be able to get away with a much smaller cap on the input,I haven't got my roll off tables handy,but I would suppose a 100nano should be sufficient and will take far less time to charge up.It works nicely with the extra mosfet set at full gain as when you switch from j-fet to mosfet theres another lttle extra gain,crunch and graininess...maybe I should rename it ...dirt!

                       Thanks again,sean.