Dont understand N Mosfet transmit on page 149 of ITMD

[brucester]

Hi,

Sorry for a beginner electronics question but i just dont get this.

On page 149 of inside the metal detecto there is an N mosfet coil pulse circuit.



I have made the circuit on a few simulators and the electricity just flows through the diode to groud making the mosfet pointless.

For example, here is it simmed on falstad:

http://www.falstad.com/circuit/#%24+...625E-5+5+-1%0A

Is this right?

It works if i move the diodes up in line with the coil and damper and dont connect them to ground.

Is this a mistake of the book or am i missing something?

Thanks

[Mechanic]

HI Bruester,

Looks like the mosfet needs to be put where the coil currently is and the coil to where the mosfet is, as well as the damping resistor. Also the value for the series input resistor looks a bit too low to me. I would expect to see at least 500r~1k for this configuration.

Check out the surf pi schematic > http://www.geotech1.com/forums/showt...55-Surf-PI-Pro

Cheers Mick

[brucester]

Hi Mechanic,

I think that is for a p type. N type has the load above the drain.

[brucester]

As for the 2 ohm resistor, that just simulates the batteries internal resistance as this sim outputs infinite amps :P

[Qiaozhi]

Hi,

Sorry for a beginner electronics question but i just dont get this.

On page 149 of inside the metal detecto there is an N mosfet coil pulse circuit.



I have made the circuit on a few simulators and the electricity just flows through the diode to groud making the mosfet pointless.

Is this right?

It works if i move the diodes up in line with the coil and damper and dont connect them to ground.

Is this a mistake of the book or am i missing something?

Thanks

Figure 11.11 on page 149 does indeed have a small mistake. The bottom end of the clamping diodes should be connected to 0V, and not to the negative battery terminal. The way it is drawn at the moment will cause the preamp output to be driven up to the positive supply rail. There are no component values shown in the diagram, as it is "supposed to be" an example of how to configure the detector to use an inverting opamp, together with the more widely used NMOS switch.

Well spotted!
I will add this to the errata section.

Simply disconnect the bottom end of the clamping diodes and reconnect to 0V, then it should start working. In your simulation, I would also advise you to increase R1 from 40R to 1k, and R2 from 40k to 1meg. Using a low resistance value for the feedforward resistor adds extra [unwanted] damping to the RX signal.

[brucester]

Great, thanks a lot Qiaozhi.

I realise they are just examples and ppl that know electronics can join the dots. I am from a software engineering background and have pretty much taken a 30 crash course in electronics, thats why the diagram stumped me :P

The sooner i get a simple analog front end sorted so i can start programming the better

Thanks again

[Ferric Toes]

Try this -


Scope ground also to battery + ground. Opamp is inverting so waveform is inverted. Make opamp non inverting if you wish.
Best to pulse gate rather than switch.

Eric.

[brucester]

Cool, thanks a lot Ferric

[brucester]

Another quick question about these pages from the book.

On page 148, it sets up the coil pulse circuit with a p mosfet. It then uses a non inverting amp and talks about the need for an inverting charge pump. Why not just use an inverting amp and output the negative flyback as positve between 0 and 9v?

http://www.falstad.com/circuit/#%24+...625E-5+2+-1%0A

One small thing, that may be a tiny error or just my ignorance, on page 149 it says Rs can be eliminated and R1 can be the clamping resistor. Should that not be, R1 can be elim, Rs used as clamp?

Now, onto my question.

My circuit thus far is setup using the n mosfet.

This makes 9 volt ground.

I have been banging my head against a wall trying to amplify the clipped signal using 9v as ground.

I have a signal that is +/- 0.8v about 9v.

If i feed that signal into an amp of only a gain of 2, 9 becomes 18.

I can't figure out how to amplify the clipped signal when it has a 9 volt ground

Its so simple with a p mosfet and inverting amp.

Impossible using N :P

Any help would be great.

Thanks

[brucester]

I noticed in PI5 in the book, the diodes are connected to ground like you said earlier.

When i ground this circuit, it goes crazy. Is this a fault in the sim? Or am i doing it wrong?

http://www.falstad.com/circuit/#%24+...625E-5+0+-1%0A

[Carl-NC]

In most PI circuits the coil is dead-shorted to the battery for TX. When the switch is opened one side of the coil is still connected either to +VB or GND. This becomes the settled value of the coil voltage.

In either case you can run this into an opamp that is powered from, say, +VB and GND but now the input signal is at the one or the other of the opamp's supply rail. Most opamps don't work with the input signal "railed out" so you need to power the opamp with additional headroom. Either +VB to -VB if the coil is GND-referenced, or +2VB to GND if the coil is +VB referenced. Either way keeps the coil voltage in the middle of the opamp range.

- Carl

Edit:
"Rs can be eliminated and R1 can serve as the clamping resistor."
or
"R1 can be eliminated and Rs can serve as the feedforward resistor."
Either one works.

[6666]

Thank you for the drawing Eric

[Qiaozhi]

I noticed in PI5 in the book, the diodes are connected to ground like you said earlier.

When i ground this circuit, it goes crazy. Is this a fault in the sim? Or am i doing it wrong?

http://www.falstad.com/circuit/#%24+...625E-5+0+-1%0A

PI5 is a different animal, as it has a DC blocking capacitor between the coil and the input of the preamp. Also, the ground is not tied to either the +ve or -ve battery terminals, but is a virtual ground created using a spare NE5532 opamp.
PI5 is offered as another way of skinning the proverbial cat.

[brucester]

I see, thanks a lot guys.