Pass JFET Power Amplifier

Status
Not open for further replies.
hitsware said:
With Jfets aren't the source and drain interchangeable?

This tends to be true in the small signal "resistive" region, but
I wouldn't bet on it holding up a higher voltages and currents.

Interestingly, the same is also true of Bipolar devices. I once
had matched diff pairs that worked fine but would break
occasionally. I found out that the metal can's tab did not
denote emitter - it was the collector, and the break down was
around 8 volts.

😎
 
Nelson Pass said:


This tends to be true in the small signal "resistive" region, but
I wouldn't bet on it holding up a higher voltages and currents.

Interestingly, the same is also true of Bipolar devices. I once
had matched diff pairs that worked fine but would break
occasionally. I found out that the metal can's tab did not
denote emitter - it was the collector, and the break down was
around 8 volts.

😎

I tried your preamp remote with various devices including JFET's and BJT's. As it turned out, only BJT's produced anything near symmetric results. Thus it can be said that BJT's are even better than JFET's when used in both directions at the same time.

Petter
 
Other option with powerJFET

Nelson and the rest of the threaders,

Impressed with your contributions, I find this the time and place to throw in another idea.
For a long time I have been thinking about a topology to use a power JFET in a (almost) rail to rail output on an amp.
With normal bootstrap one output device can be driven to its rail, the other by far not. Using the power JFET also the top transistor in the schematic can be driven close to the rail.
This will improve the poweroutput for a given rail voltage and thus the efficiency.
When the Lovoltech parts will be available I want to try this idea. Any comments yet?
Rick Ruijsink
 

Attachments

  • bootstrap-mosfet-jfet.gif
    bootstrap-mosfet-jfet.gif
    4.2 KB · Views: 1,153
I believe that will work, but I notice that you pair the JFET
with a P channel Mosfet and use bootstrapping to drive it
to the negative rail.

If you want to use complementary Mosfet followers rail to rail,
you can bootstrap on both sides of the bias source, and drive
it from the center.

😎
 
Nelson Pass said:
I believe that will work, but I notice that you pair the JFET
with a P channel Mosfet and use bootstrapping to drive it
to the negative rail.
😎

Yes that's the idea.

Nelson Pass said:

If you want to use complementary Mosfet followers rail to rail,
you can bootstrap on both sides of the bias source, and drive
it from the center.
😎
In that case the driver has to go rail to rail still, shifting the problem.
My idea a single ended driver can go from almost top rail to below bottom rail.
Still don't know if and how an NJFET and PMOSFET will team up.

Rick
 
Just thought I'd throw this in:
There is now a second source for the 2SK170. Linear Systems, aka Linear Integrated Systems, manufactures the LSK170. Supposedly, it is directly equivalent to the 2SK170.
The datasheet for the LSK170 mentions a dual version they call the LSK389, but I have not been able to determine that they have actually released the device yet.
No word on P-ch devices.
Neither Digikey nor Mouser carry the LSK170. I haven't had time to pursue the retail aspect further. Perhaps someone with more time on their hands can run down a retail source for these things. Given the difficulty in sourcing Toshiba parts, it would be nice to have them available.
I, for one, could use some VI parts right about now.
www.linearsystems.com

Grey
 
Hi, Mr.Pass,

If you want to use complementary Mosfet followers rail to rail, you can bootstrap on both sides of the bias source, and drive it from the center.

I've been thinking about this. It is about your NP-PMA. This is an error correction circuit.

Does this NP-PMA always needs CCS for feeding its upper and lower? Or this NP-PMA can use the Double bootstrapped like the quote above? NP-PMA error correction always needs high impedance CCS or it can work with just double bootstrapped?

I know that CCS and bootstrapped differs in output impedance, but each has own pro's and con's in different areas.
 
Hi, Mr. Pass,

Hi and thanks Dimitri,

In 7 post above, you mentioned about using double bootstrapped for powering the VBE multiplier and inject the signal from the middle.

Since NP-PMA is not ordinary VBE multiplier, and it is Error Correction, can it be powered by double bootstrapped or it MUST be powered by double CCS?

This is the example of double bootstrapped for driving bias generator. Can this method used for driving Error Correction?
 

Attachments

  • double.jpg
    double.jpg
    54.3 KB · Views: 1,248
Status
Not open for further replies.