new active antenna with bootstrapping of the SFET to minimize the input capacitance

Am 27.12.20 um 18:43 schrieb snipped-for-privacy@highlandsniptechnology.com:

Not really. Doubling the number of FETs doubles Cin and halves the voltage from the probe.

It also doubles gm, but the noise of the FET is given by root Cin/gm ratio, so quadrupling the # of FETs only halves the noise voltage. Does not look like a win when you are Cin-challenged.

Doubling the drain current also increases gm by a factor of 1.414. The effect on noise is also only 4th root.

BTW the circuit has a negative real part of the input impedance. Yellow trace. That means that more energy comes out of the input than there goes in.

Even the generator voltage rises visibly above 200 MHz. (green trace) With the right input impedance it will oscillate.

< >

The noise voltage density is 2.8 nV/rtHz in the flat part. Below 100 KHz it goes through the roof. The corner frequency depends strongly on the 18 pF cap. The rise is much steeper than 1/f.

< >

Cheers, Gerhard

Indeed.

In an electrically short (

Mo lived in Germany for a year or so. She was impressed by how rule-driven the Germans were. They wouldn't walk across a street against a red light even if there were no cars for a mile in either direction. If Mo sat on the grass in a park, Germans would walk up and tell her that's not allowed.

That attitude seems to apply to electronics too. Follow rules, even if they are applied wrong.

--

John Larkin      Highland Technology, Inc 

The best designs are necessarily accidental.

It's not necessary. A simple 1 pF phemt follower wouldn't load 18 pF much, but could have much better noise behavior. Like maybe 10x better.

18 pF does sound like a lot for a small rig like that. A beer can is about 8.

--

John Larkin      Highland Technology, Inc 

The best designs are necessarily accidental.

Not if the end is terminated. Source termination just wastes half the signal.

Suppose I built a grounded copper wall that has three bulkhead BNC connectors, and you are on one side with all the test equipment you want. On the invisible far side, one has a 50 ohm resistor. One has a few meters of good hardline coax terminated with a 50 ohm resistor. One has a light-year of coax with no termination.

What measurements would you do to tell which is which?

You might change the subject, to what produces the best signal from the receiver.

--

John Larkin      Highland Technology, Inc 

The best designs are necessarily accidental.

The group benefits from occasional comic relief.

But seriously, only a fraction of the people who fool around with electronics actually understand the basics.

--

John Larkin      Highland Technology, Inc 

The best designs are necessarily accidental.

Not if each phemt looks like half a pF. The source is said to be 18.

The best s/n happens with maximum power transfer into an amplifier. We can't conjugate match a wideband amp, but we can capacitive match it. Going from 0.5 to 1 pF is a win if we can cut the noise by sqrt(2) and increase the voltage into the coax too. The Gm of a phemt is a lot higher than a jfet, so follower output impedance is low, so a little Mini-Circuits stepup transformer is noiseless voltage gain.

In this case, we're not.

Good point. More noise, too.

Peesumably there is a shortwave receiver at the end of the coax, not a

50 ohm resistor. It may have a lot of equivalent input noise; they don't need to be super good, because they generally assume a lot of atmospheric noise. So some voltage gain in the active antenna might really help. The cited 0.9 uV sensitivity could be 10x or so better.

Is there anything interesting to hear on the HF bands nowadays? I used to listen to BBC and Radio Moscow, but that sort of thing is on the internet now.

--

John Larkin      Highland Technology, Inc 

The best designs are necessarily accidental.

A standard beer can is about 9.

Two beer cans in an e-field dipole could be interesting.

--

John Larkin      Highland Technology, Inc 

The best designs are necessarily accidental.

We discussed a nice solution here a few weeks ago. Use an OPA858.

It's annoying when you design some really cool circuit then some smarty-pants IC designer puts it on a cheap chip that anybody can order from Digikey.

--

John Larkin      Highland Technology, Inc 

The best designs are necessarily accidental.

Joe Gwinn wrote in news: snipped-for-privacy@4ax.com:

Here is a link to a QRZ article by Rhode about short active antennas.https://forums.qrz.com/index.php?threads/dr-ulrich-rohde-electrically-short-antennas.740851/

--
This email has been checked for viruses by Avast antivirus software. 
https://www.avast.com/antivirus

It's a nice summary of the issues.

Joe Gwinn

On LF, MF and HF (below 30 MHz), it is far more important that the amplifier is extremely linear and tolerant to large signals than absolute best sensitivity.

On a careless amplififer, all the crud on the band creates far more crud from intermodulations.

--

-TV

Nobody wants to play? How dull.

--

John Larkin      Highland Technology, Inc 

The best designs are necessarily accidental.

It seems like a rhetorical question. did you seriously expect an answer?

so you attach a thermometer, and an ammeter to to the three ports and inject a DC voltage, the one that heats up first has the resistor.

The one that shows a current decline in about three years has the open coax.

--
  Jasen.

========================

** Errrrrr - wot is an SFET ??

Wiki tells me it's a "Spin FET", an experimental device long theorised but never in production.

The device in you schem however is a regular JFET, one optimised for the AM band - ie car radios.

BTW:

I see from you other posts that you are not playing with a full deck.

..... Phil

Well, you could probably measure the gravitational torque from the long one. ;)

Cheers

Phil Hobbs

The eN*C noise goes up linearly with the number of FETs, whereas their voltage noise goes down like sqrt(N).

So for brute-force parallelling, the optimal noise occurs when the total FET capacitance is half the source capacitance.

Cheers

Phil Hobbs

If they are unsupported, there could be mechanical resonances. A sonar ping might work too.

If they are ideal parts, no electrical measurement can tell them apart. In real life, TDR would.

An ohmmeter would spot the long one.

--

John Larkin      Highland Technology, Inc 

The best designs are necessarily accidental.

0.141 tinned copper semirigid weighs 44 kg/km, so a light year's worth would be 4.16E14 kg, a good-sized asteroid.

It would look like a strange lossy capacitor.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC / Hobbs ElectroOptics 
Optics, Electro-optics, Photonics, Analog Electronics 
Briarcliff Manor NY 10510 

http://electrooptical.net 
http://hobbs-eo.com