Hi,

I'm trying to make a bidirectional switch using MOSFETs for a signal of 3V, 1 MHz, and around 100mA. When I use the TSM2N7000KCT MOSFETs the switch works fine. Here is the datasheet and schematics,

http://www.mouser.com/ds/2/395/TSM2N7000K_B12-522765.pdf



However, when I use another MOSFET like the BS170_D27Z the switch doesn't work. Here is the datasheet and schematics,

http://www.mouser.com/ds/2/149/BS170-39309.pdf



One MOSFET has a static protection and the other one doesn't. I don't see why the circuit works with one MOSFET, but it doesn't work with the other. Any help is appreciated.

Thank you,

Robert

 

What are you actually trying to do?

If this is some variation on level shifting perhaps you might look at http://www.nxp.com/documents/application_note/AN10441.pdf
for some guidance.

 

Driving the gate with only 5V may not be enough to turn the FET on hard enough to pass 100mA. Note that the two devices have different gate voltage sensitivities at their Vgs(th) values, so probably quite different Id/Vgs characteristics.

 

For the system to work for different transistors is necessary that the transistors are always included. You are applying to the gates of transistors 2.5 volt and signal at 3 volts (Vampl=3V*1.4142). it is not right. Remove the lower resistor and everything will work with your transistors (Vg=5V).
Transistors have a large variation in the threshold voltage! In addition, it is not clear what you mean about 3 volts. This acting value (RMS) or the peak-to-peak.

 

circuit works with one MOSFET, but it doesn't work with the other

Could you elaborate on the meaning of doesn't work?

Have you tried removing the lower divider resistor? 2.5V might not be sufficient for the second MOSFET.

 

For the two circuits above, the one with the TSM2N7000KCT MOSFETs works great. It works as a switch that can be controlled with the MOSFETs Gate voltage.

Now, the circuit with the BS170_D27Z MOSFETs is supposed to work as well, but it just doesn't work. The 'switch' stays in the closed position all the time. So, it doesn't matter what voltage you apply at the MOSFETs gate it is always conducting current. I was hopping that somebody spots a difference between the two transistors that might cause this.

The voltage of the signal is 3V amplitude or 6Vpp. If the 1K resistor that goes to ground is removed the transistor's gates won't discharge fast enough and the switch won't turn off fast enough.

 

The others have already told you the problem but you seem not to be listening.

The TSM2N7000KCT requires 5V Vgs to fully turn on.
The BS170_D27Z requires 10V Vgs to fully turn on.

There's your problem. You need to apply 10V Vgs to the BS170_D27Z, not 5V.
Also remove the top 1k resistor (not the bottom one) since that cuts the Vgs voltage in half.

 

The voltage of the signal is 3V amplitude or 6Vpp.

You still haven't stated what you're trying to do. Using N channel MOSETS to pass AC signals will introduce distortion as the channel resistance will vary with the signals being passed. Passing negative signals will be even worse. A better approach would be to use an N and P MOSFET with their bulk connected an appropriate supply voltage to give a more consistent "on" resistance.

If the 1K resistor that goes to ground is removed the transistor's gates won't discharge fast enough and the switch won't turn off fast enough.

It wasn't obvious from your schematic that the MOSFET was being switched. Change the ratio of the resistors so most of the supply voltage is applied to the gates when the devices are on and/or change the supply voltage to turn them on harder. I thought I read that the threshold voltage of the two devices was 3V max (but this is just starting to turn on).

 

Also remove the top 1k resistor (not the bottom one) since that cuts the Vgs voltage in half.

I assumed that MOSFETs were on constantly; so removing the bottom resistor would provide 5V to the gates through an unnecessary 1K resistor. If they're being switched, then the top resistor should be removed...

 

The others have already told you the problem but you seem not to be listening.

The TSM2N7000KCT requires 5V Vgs to fully turn on.
The BS170_D27Z requires 10V Vgs to fully turn on.

There's your problem. You need to apply 10V Vgs to the BS170_D27Z, not 5V.
Also remove the top 1k resistor (not the bottom one) since that cuts the Vgs voltage in half.

Hmm, where do you see that in the datasheet? The problem with the BS170_D27Z is that it is always in the on state. The Gate Threshold Voltage for the BS170_D27Z is from 0.8V to 3V.

 

You still haven't stated what you're trying to do. Using N channel MOSETS to pass AC signals will introduce distortion as the channel resistance will vary with the signals being passed. Passing negative signals will be even worse. A better approach would be to use an N and P MOSFET with their bulk connected an appropriate supply voltage to give a more consistent "on" resistance.

Hmm, what do you mean by their "bulk"? Do you have any schematics available for this circuit?

 

One interestin thing: the pinout for the BS170 is reversed from the TSM2N7000. When switching parts, you'll need to rotate the part you're placing by 180 degrees so that the pins connect correctly.

 

Hmm, what do you mean by their "bulk"? Do you have any schematics available for this circuit?

This is from an appnote on analog.com.

The bulk of the N device should be connected to a voltage more negative than the lowest signal voltage; P device should be connected to a more positive voltage.

 

I think Brownout found your problem.
If the MOSFETs won't turn off then it sounds like they are miswired.

 

DL324, the signal that I'm trying to switch on and off is the 3V, 1MHz signal. I'm using the gate voltage to switch that signal on/off. The 1MHz wave is a bidirectional one, so I don't see how I can implement your circuit.

CRUTSCHOW, no that wasn't the problem. I checked the pinout of the MOSFETs again and I wired it correctly. I can't figure out what the problem is.

 

the signal that I'm trying to switch on and off is the 3V, 1MHz signal. I'm using the gate voltage to switch that signal on/off. The 1MHz wave is a bidirectional one, so I don't see how I can implement your circuit.

Resistance of the MOSFET will change depending on the magnitude and polarity of the signal. If distortion matters, you need to use a CMOS switch. Below is a graph of relative switch resistance for NMOS, PMOS, and CMOS (from the same appnote I referenced earlier):

 

I think I see another problem.
Try connecting the 1 MΩ resistor between the sources and the gate connections instead of source to ground.
It might also be better to reduce its value, say to 100kΩ or less.

 

I think I see another problem.
Try connecting the 1 MΩ resistor between the sources and the gate connections instead of source to ground.
It might also be better to reduce its value, say to 100kΩ or less.

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

 

I need to get different types of MOSFETs and figure out what is going on.

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?

 

Why do you want to make an analog switch out of discrete mosfets? Unless you have mosfets with four terminals you need to use an analog switch. See dl324's posts #8 and #13, pay attention to where the middle bar from the mosfet symbol is connected in the switch, it is not the source like in a typical mosfet but rather Vcc or Gnd.