Hmm, no that didn't work. I need to get different types of MOSFETs and figure out what is going on.

So you are saying the with a resistor from gate-to-source and no other gate bias (gate open) the transistor won't turn off?
That makes no sense.

 

Hmm, where do you see that in the datasheet?
View attachment 92842

The line right below the one you circled. Rdson is specified at Vgs = 10 V.

ak

 

So you are saying the with a resistor from gate-to-source and no other gate bias (gate open) the transistor won't turn off?
That makes no sense.

I tested this setup that you suggested and the switch didn't work right. I don't remember now how it failed but the switch didn't turn off when 0V was applied to the gate or it didn't turn on when 5V was applied to the gate. I need to do more testing.

 

The line right below the one you circled. Rdson is specified at Vgs = 10 V.

ak

They made Vgs = 10V only for testing purposes. Here is the definition of Vgs from Wikipedia,

"The threshold voltage, commonly abbreviated as Vth or VGS (th), of a field-effect transistor (FET) is the minimum gate-to-source voltage differential that is needed to create a conducting path between the source and drain terminals."

 

... minimum gate-to-source voltage differential that is needed to create a conducting path between the source and drain terminals."

Yes which amounts to 1mA according to the datasheet, so it is barely conducting.

 

Yes which amounts to 1mA according to the datasheet, so it is barely conducting.

Hmm, good point!

 

Have you tried testing the MOSFETs from the second circuit individually to make sure they're good? If you remove the gate bias voltage, do both transistors turn off?

I have to try that.

I suspect that the problem has something to do with the input/output capacitance of the MOSFETs at a high frequency (1MHz). The appnote that you mentioned above is been very helpful. It talks about the Drain-Source Capacitance, Cds, of the transistor.

"The series-pass capacitance, CDS, not only creates a zero in the response in the ON-state, it degrades the feedthrough performance of the switch during its OFF state. When the switch is off, CDS couples the input signal to the output load as shown in Figure 10.
Large values of CDS will produce large values of feedthrough, proportional to the input frequency. Figure 11 illustrates the drop in OFF-isolation as a function of frequency. The simplest way to maximize the OFF-isolation is to choose a switch that has as small a CDS as possible."

The switch works fine with the TSM2N7000K transistor. This transistor has very low input/output capacitance. However, the switch won't turn off with the BS170 transistor. This transistor has a high input/output capacitance. It all makes sense. What do you think?

 

They made Vgs = 10V only for testing purposes. Here is the definition of Vgs from Wikipedia,

"The threshold voltage, commonly abbreviated as Vth or VGS (th), of a field-effect transistor (FET) is the minimum gate-to-source voltage differential that is needed to create a conducting path between the source and drain terminals."

No. The sentence you quoted is correct, but it is not what you want for your application. You are confusing Vgs and Vgsth. These are not the same thing. The threshold voltage is a specific parameter for what basically is the worst conduction the FET can have and still be considered a tiny bit "on". Vgs is not a specific condition. It is the abbreviation for the gate-source voltage for whatever *other* condition is being discussed. In this case it is the condition for the absolute minimum Rds, called Rdson, the best conduction the FET can achieve. This is the exact opposite of the threshold voltage, and I don't know how to state that more clearly. The data sheet says that the only way to guarantee that the device is operating at its minimum Rds is for Vgs to be 10 V or greater.

ak

 

No. The sentence you quoted is correct, but it is not what you want for your application. You are confusing Vgs and Vgsth. These are not the same thing. The threshold voltage is a specific parameter for what basically is the worst conduction the FET can have and still be considered a tiny bit "on". Vgs is not a specific condition. It is the abbreviation for the gate-source voltage for whatever *other* condition is being discussed. In this case it is the condition for the absolute minimum Rds, called Rdson, the best conduction the FET can achieve. This is the exact opposite of the threshold voltage, and I don't know how to state that more clearly. The data sheet says that the only way to guarantee that the device is operating at its minimum Rds is for Vgs to be 10 V or greater.

ak

Okay, I see your point now.

 

The switch works fine with the TSM2N7000K transistor. This transistor has very low input/output capacitance. However, the switch won't turn off with the BS170 transistor. This transistor has a high input/output capacitance. It all makes sense. What do you think?

The datasheet for the 2N7000 doesn't make sense.

Coss = Cds + Crss
Ciss = Cgs + Crss, Cds shorted

From the 2N7000 datasheet: Cds = Coss - Crss = 3.42pF - 7.63pF = -4.21pF
For the BS170: Cds = 17pF - 7pF = 10pF

Crss is also called Cgd or Miller capacitance.

Below is the cap data for Fairchild 2N7000:

Cds = 11pF - 4pF = 7pF (typical, 20pF worst case; same as the BS170)

 

Two additional things to note:

To turn off the MOSFETs, the gate voltage must be no greater than the most negative voltage. With a 3Vrms input that means the gate should be no more positive than -4.2V (say -5V).

If feedthrough capacitance is a problem, then use a lower impedance load.
The oscilloscope presents a very high impedance load.
What is the real load you will be driving?

 

The datasheet for the 2N7000 doesn't make sense.

Coss = Cds + Crss
Ciss = Cgs + Crss, Cds shorted

From the 2N7000 datasheet: Cds = Coss - Crss = 3.42pF - 7.63pF = -4.21pF
For the BS170: Cds = 17pF - 7pF = 10pF

Crss is also called Cgd or Miller capacitance.

Below is the cap data for Fairchild 2N7000:
View attachment 92947
Cds = 11pF - 4pF = 7pF (typical, 20pF worst case; same as the BS170)

No, I'm not using the Fairchild 2N7000. I'm using the TSM2N7000K from Taiwan Semiconductors. They work just fine.

 

No, I'm not using the Fairchild 2N7000. I'm using the TSM2N7000K from Taiwan Semiconductors. They work just fine.

My point was that the TSM datasheet appears to have a mistake and that the capacitances of the two MOSFETs are closer than the TSM datasheet would make them appear.

In an ideal world, all companies second sourcing a standard part would make the significant parameters match.