When I use a Jfet in LTspice as a capacitor multiplier, all the current goes through the gate.

None of it goes through the drain.
Is it spice, or me that's the problem?
I tried using different "named" models but they all draw current as well.

 

There is one massive thing wrong with this circuit - there is no bulk filter capacitor after the diode bridge and before the multiplier circuit. A capacitance multiplier circuit is not a replacement for the main filter capacitor. It is a method for dealing with the remaining ripple after the main filter, not the 100% ripple coming out of the diodes. The circuit you posted in#1 is not intended to be the entire power supply filter, and will not work in that function when the output of the system is fully loaded. Again, going back to post #14, calculate the component values based on the position you keep repeating and you will see that they will not work.

ak

 

I never intended to not have a filter cap before the multiplier. I'm going to use filter caps both before and after the multiplier.
My concern right now is not ripple but the fact that the Jfet gate is consuming current when it should not be.
Is this some thing that would happen in real life?

 

When I use a Jfet in LTspice as a capacitor multiplier, all the current goes through the gate.
View attachment 101794
None of it goes through the drain.
Is it spice, or me that's the problem?
I tried using different "named" models but they all draw current as well.

The gate has to be a few volts negative WRT the source to control the channel width (and hence conduction).

With no negative bias on the gate; the channel should conduct as much as its going to.

With regular silicon JFETs, the gate is destroyed by more than about 10mA conduction.

 

Yeah but the capacitor multiplier is effectively a source follower, shouldn't the source be higher than the gate naturally? How could I make the gate lower here?

 

Yeah but the capacitor multiplier is effectively a source follower, shouldn't the source be higher than the gate naturally? How could I make the gate lower here?

Not with a JFET - you have to reverse bias the gate PN junction to control the channel current. Its a depletion mode semiconductor.

AFAIK: you can also get depletion mode MOSFETs.

 

Hmmmmmmmmmmmm, there must be a way to make it work..

 

Yeah but the capacitor multiplier is effectively a source follower,

No, it isn't. It is an emitter follower. I don't know where you are getting these statements, but they are incorrect.

shouldn't the source be higher than the gate naturally?

No. You are equating the behavior of a depletion mode JFET to that of an enhancement mode MOSFET or an NPN bipolar transistor. None of these devices are equivalent. The circuit you started with has a bipolar transistor in it for a reason - because nothing else will work.

ak

 

No, it isn't. It is an emitter follower. I don't know where you are getting these statements, but they are incorrect.

ak

Not if the TS uses a FET it isn't.

 

No, it isn't. It is an emitter follower. I don't know where you are getting these statements, but they are incorrect.

There's no emitter in a Jfet

No. You are equating the behavior of a depletion mode JFET to that of an enhancement mode MOSFET or an NPN bipolar transisto

LTspice is showing the source of the Jfet at a higher voltage than the gate though, at least when I use it as a pure follower.

The circuit you started with has a bipolar transistor in it for a reason - because nothing else will work.

Lot's of people use mosfets it seems. For audio at least.

 

There's no emitter in a Jfet

LTspice is showing the source of the Jfet at a higher voltage than the gate though, at least when I use it as a pure follower.

.

Then LTspice is correct. Reverse biasing the gate PN junction increases the depletion region which pinches off the channel - that is the mechanism by which a JFET controls current.

 

How does a pure follower and the capacitor multiplier differ though? They both have a B+ voltage for the drain, they both have a lower-than-drain gate voltage, yet when I use a source follower in a circuit the Jfet becomes forward biased.

 

Lot's of people use mosfets it seems. For audio at least.

True, but you are taking an isolated fact way out of context. Power MOSFETs do very well in audio power amplifier output stages because the temperature coefficient of one operational parameter is the opposite of the equivalent parameter's TC for a bipolar transistor, enabling multiple MOSFET devices in parallel to "self-level". Also, because the device physics of some FETs resemble the physics of a vacuum tube more closely than do the physics of a bipolar transistor, there is the feeling that FET circuits have a "warmer" sound. This is partly supported by the nature of the harmonics generated at the onset of clipping, and partly debunked by rigidly controlled listening tests.

Separate from all of that is how devices perform in the schematic in post #1. With enough effort, any type of FET can be made to work as a capacitance multiplier. However, no FET device can approach the simplicity, efficiency, and performance of a BJT in that application.

ak

 

True, but you are taking an isolated fact way out of context

Actually I was referring specifically to capacitor multipliers. I was scanning some Diyaudio forums using search and lots of people claim that there is a significant "sound" improvement when using mosfets over transistors for the multiplier. I have no idea why that would be but I tend to trust it when many different sources claim the same thing. Especially on that forum since they laugh at "audiophoolery"

With enough effort, a FET can be made to work as a capacitance multiplier

Where would I begin? The reason I want to try to use a Jfet is because since people claim that mosfets sound better than BJTs for the multiplier, and it seems to be a common theme that people enjoy Jfets second only to tubes, a Jfet multiplier seems like a good experiment.

 

How does a pure follower and the capacitor multiplier differ though? They both have a B+ voltage for the drain, they both have a lower-than-drain gate voltage, yet when I use a source follower in a circuit the Jfet becomes forward biased.

You have to think in terms of where the gate is relative to the source.

With an enhancement device like most MOSFETS you have to take the gate positive to make it conduct.

With depletion devices like JFETs you have to make the gate negative to stop it conducting.

 

coinmaster: Yeah but the capacitor multiplier is effectively a source follower,

ak: No, it isn't. It is an emitter follower.

Not if the TS uses a FET it isn't.

Technically correct, but if he uses a FET it will not work. So, is it effectively a source follower if the source doesn't follow effectively?

ak

 

coinmaster: Yeah but the capacitor multiplier is effectively a source follower,

ak: No, it isn't. It is an emitter follower.



Technically correct, but if he uses a FET it will not work. So, is it effectively a source follower if the source doesn't follow effectively?

ak

Is an op-amp voltage follower still a follower if the internal circuitry is FET?

 

Technically correct, but if he uses a FET it will not work, so is it effectively a source follower if the source doesn't follow effectively?

Read the last few above posts. BJT is not the only thing that works.

You have to think in terms of where the gate is relative to the source.

With an enhancement device like most MOSFETS you have to take the gate positive to make it conduct.

With depletion devices like JFETs you have to make the gate negative to stop it conducting.

Yeah but I thought in terms of a source follower the source voltage will follow the gate voltage either way.
For example if the "depletion" aspect applied then if I had 200v B+ and I put 50v on the gate then I would have 150v on the output (linearly speaking at least) but in LTspice if I have 200v B+ and I apply 50v on the gate I get like 51v. It follows the gate but with a reverse bias.
A capacitor multiplier to me, looks no different. B+ voltage, gate bias, and source output.

 

Read the last few above posts. BJT is not the only thing that works.
Yeah but I thought in terms of a source follower the source voltage will follow the gate voltage either way.
For example if the "depletion" aspect applied then if I had 200v B+ and I put 50v on the gate then I would have 150v on the output (linearly speaking at least) but in LTspice if I have 200v B+ and I apply 50v on the gate I get like 51v. It follows the gate but with a reverse bias.
A capacitor multiplier to me, looks no different. B+ voltage, gate bias, and source output.

Did you tell LTspice it was a depletion mode part?

 

I used the NJF model.