开云体育

Thanks Steve,

I didn't have the article open at the time and had forgotten about
the pot that Dallas had. It amounts to sort of the same thing -
varying the DC bias on the gate - but keeps the impedance relatively
constant.

I'm aware of IP2 and IP3 although I've never measured them. One of
these days I'll set up the proper equipment to do that. Have you
ever tried either of these circuits in a balanced/differential
configuration?

Rob



--- In loopantennas@..., "Steve Ratzlaff"
<steveratz@...> wrote:

Hi Rob,
I adjusted the bias voltage to the jfet, with the corresponding

change in

the circuit total current.

The following are the IMD performance results I got using your 5 new fixed
resistor bias values. To be able to measure the bipolar current, I put a 1
ohm resistor between emitter and supply (with 0.1 uF bypass cap on the
emitter to ground); and jumpered out the resistor except when I was doing
the emitter current measurment. (I measured the voltage across the
resistor.)

Something here just isn't right. Regardless of whether you are applying
a bias voltage to the JFET gate or changing the resistors, the JFET drain
current remains virtually unchanged.

The values were determined by way of PSpice. I'll test this myself, but
I have to be out of town all day tomorrow so I won't have an opportunity to
look at it until Thursday.

Chris

,----------------------. High Performance Mixers and
/ What's all this &#92; Amplifiers for RF Communications
/ extinct stuff, anyhow? /
&#92; _______,--------------' Chris Trask / N7ZWY
_ |/ Principal Engineer
oo&#92; Sonoran Radio Research
(__)&#92; _ P.O. Box 25240
&#92; &#92; .' `. Tempe, Arizona 85285-5240
&#92; &#92; / &#92;
&#92; '" &#92; IEEE Senior Member #40274515
. ( ) &#92;
'-| )__| :. &#92; Email: christrask@...
| | | | &#92; '.
c__; c__; '-..'>.__

Graphics by Loek Frederiks

Hi Chris,
The following are the IMD performance results I got using your 5 new fixed resistor bias values. To be able to measure the bipolar current, I put a 1 ohm resistor between emitter and supply (with 0.1 uF bypass cap on the emitter to ground); and jumpered out the resistor except when I was doing the emitter current measurment. (I measured the voltage across the resistor.)
(I don't know how you calculated the resistor values, but my transistors don't give anywhere close to having equal current for jfet and bipolar, with the bipolar having much more current, which increases as the bias resistors change toward the "18 mA" values.)

1. "6 mA". +18 dBm OIP3, +29 dBm OIP2. 12.92 Vcc, 17.5 mA. Drain 12.32V/4.6 mA, Source 1.34V, Bipolar 12 mA.

2. "8 mA". +19.5 dBm OIP3, +29 dBm OIP2. 12.90 Vcc, 23.8 mA. Drain 12.34V/5.6 mA, Source 1.23V, Bipolar 16 mA.

3. "10 mA". +20.5 dBm OIP3, +29 OIP2. 12.88 Vcc, 32.5 mA. Drain 12.36V/6.3 mA; Source 1.15V. Bipolar 24 mA.

4. "12 mA". +21.5 dBm OIP3, +29 dBm OIP2. 12.84 Vcc, 50 mA. Drain 12.44V/5.9 mA, Source 1.21V. Bipolar 41 mA.

5. "18 mA". (I think you meant R1 to be 15 ohms, not 5.6 ohms. Using 5.6 ohms gave over 100mA current; using 15 ohms gave better results. I noticed the ratios of R2/R1 went from approx. 1.5, 2, 2.5, 3, so I guessed 3.5 should be the ratio, which gives 15 ohms for R1.) However the current was still very high, and the bipolar was very hot; the current was steadily increasing, but I made some measurements anyway. The circuit is unhappy with these bias values.
+19.5 dBm OIP3, +23 dBm OIP2. 12.79 Vcc, 82 mA. Drain 12.53V/5.5 mA; Source 1.22V; Bipolar 64 mA.

73,
Steve

I adjusted the bias voltage to the jfet, with the corresponding change in
the circuit total current. A two-tone IMD test set was used, looking at

the

2IMD level, and the bias was adjusted for minimum 2IMD, for the Lankford
circuit and for the best-IMD adjustment for the Trask circuit. (See
Lankford's www.kongsfjord.no article "Low Noise Active Antennas AC/DC

Power

Supplies" article for the active whip circuit under question--with the
arrangement of the jfet bias pot if you're not familiar with the

circuit--I

assumed everyone was referencing this article, in relation to this email
thread.) The object is to get the best IMD performance, which indicates

the

circuit is then performing at its most linear operation. In every circuit
that I'm aware of, best IMD performance comes from generous current
consumption; poorest performance comes from the lowest currents, as a
general rule.

I was just about to ask how the bias of my circuit was changed. If you
anjust the bias in my circuit by applying a DC voltage to the J309 gate, the
drain current of the J309 remains virtually constant and the collector
current of the 2N2709 varies proportionally with the applied gate voltage.
So, you are achieving a better IMD performance of the 2N2907 but not of the
J309.

Here's a short list of resistor values for my circuit that gives about
equal current for both devices [R1 = J309 source (2N2907 collector) to
ground, R2 = J309 drain (2N2907 base) to supply]:

Id/Ic R1 R2

6.0 82 130

8.0 51 100

10.0 33 82

12.0 22 68

18.0 5.6 47

Or you can increase the J309 gate voltage and decrease R2. In the
Lankford circuit, adjusting the J310 gate voltage increases the bias current
for both transistors without any alteration required.

Chris

,----------------------. High Performance Mixers and
/ What's all this &#92; Amplifiers for RF Communications
/ extinct stuff, anyhow? /
&#92; _______,--------------' Chris Trask / N7ZWY
_ |/ Principal Engineer
oo&#92; Sonoran Radio Research
(__)&#92; _ P.O. Box 25240
&#92; &#92; .' `. Tempe, Arizona 85285-5240
&#92; &#92; / &#92;
&#92; '" &#92; IEEE Senior Member #40274515
. ( ) &#92;
'-| )__| :. &#92; Email: christrask@...
| | | | &#92; '.
c__; c__; '-..'>.__

Graphics by Loek Frederiks

One additional comment on how Intercept Points are indicated (and measured). I measured and referenced the Intercept Points in the two circuits to the output of the device--thus "OIP3" means "Third Order Output Intercept", and of course the same for the second order--"OIP2" = Second Order Output Intercept.
With my test setup I'm looking at the output signals from the circuit, thus it's easiest to first determine the Output Intercept Point. And to convert to the Input Intercept Points (IIP3, IIP2), one simply subtracts the gain of the circuit from the Ouput Intercept Points. Thus, for a circuit with overall negative gain, which both these circuits had, the Input Intercept Points would be approximately 3 dB higher than the Output Intercept Points.
When comparing intercept points, it's important to know if the output or the input is being referred to; sometimes just an intercept point is given without stating which it is. In industry, normally Output Intercept Point is stated, or implied, but unless it's specifically stated as "Output" or "Input" intercept point, then you can't directly compare intercepts of other circuits.
Several ARRL publications give more information about distortion in amplifiers, intercept point, and such. The older (out of print) "Solid State Design for the Radio Amateur; any recent Handbook; the recently published "Experimental Methods in RF Design", and the one with the most extensive treatment, "Introduction to Radio Frequency Design" by Wes Hayward W7ZOI. The latter is aimed more at the engineer, with quite a bit of math, but still is readable for anyone with a basic electronics background.
73,
Steve

Hi Rob,
I adjusted the bias voltage to the jfet, with the corresponding change in the circuit total current. A two-tone IMD test set was used, looking at the 2IMD level, and the bias was adjusted for minimum 2IMD, for the Lankford circuit and for the best-IMD adjustment for the Trask circuit. (See Lankford's www.kongsfjord.no article "Low Noise Active Antennas AC/DC Power Supplies" article for the active whip circuit under question--with the arrangement of the jfet bias pot if you're not familiar with the circuit--I assumed everyone was referencing this article, in relation to this email thread.) The object is to get the best IMD performance, which indicates the circuit is then performing at its most linear operation. In every circuit that I'm aware of, best IMD performance comes from generous current consumption; poorest performance comes from the lowest currents, as a general rule.
Better IMD performance is shown by larger (positive) Intercept Points. Intercepts are measured for close-in, which is the Third Order (IP3), and for far-out, which is the Second Order (IP2) performance. Broadband circuits such as active whipamps and untuned loop preamps are better compared for their Second Order performance. Good Third Order performance is more desirable in narrowband, tuned circuits. Of course both Third and Second Order go together; you can't have one without the other.
Balanced, push-pull circuits naturally discriminate against second order products, thus a balanced circuit will have a higher Second Order intercept than an unbalanced circuit, and for a broadband amplifier will perform better, with respect to when distortion from strong signals is first noticed.
An example of a manufacturer with products with very good IMD performance is Wellbrook. Check out their IMD specs for their various active loop products. As far as I know, all their products use balanced amplifiers.
73,
Steve

Steve,

How were you adjusting the bias current in your tests? By changing
the value of the J309, J310 source resistor?

Rob

--- In loopantennas@..., "Steve Ratzlaff" <steveratz@...>
wrote:

I have just finished building and testing both the Lankford circuit

and the

Trask circuit on the bench, for gain/frequency response and IMD

performance.

The results are given below.
Note that Trask's reference to the Lankford circuit didn't include the
proper biassing of the original Lankford circuit. I used the original
Lankford circuit, and optimized the bias for best IMD performance. I

did the

same thing with the Trask circuit, using the same bias arrangement the
Lankford circuit used.

I have just finished building and testing both the Lankford circuit and the Trask circuit on the bench, for gain/frequency response and IMD performance. The results are given below.
Note that Trask's reference to the Lankford circuit didn't include the proper biassing of the original Lankford circuit. I used the original Lankford circuit, and optimized the bias for best IMD performance. I did the same thing with the Trask circuit, using the same bias arrangement the Lankford circuit used. However for the Trask circuit I tested it for the bias which gave best IMD performance, as well as at lower currents, down to the minimum bias level which Trask's circuit used (which was 13 mA for my particular devices picked at random from my semiconductor selection). I happened to have a J309 device on hand, which was listed ( on the Trask circuit, and used that same device for both circuits. (In other past experiments I have found no difference in performance between the J310 and J309; the U310 gives slightly better IMD performance.)(I first built and tested the Trask circuit, then tore it down and rebuilt for the Lankford circuit, so I could use the same device for both circuits.)

Trask circuit, using J309 and 2N2907 devices.
Vcc 12.8 volts (varied 0.05 volts with different bias/currents.)
1. Minimum bias, 13 mA for my devices, corresponding to the Trask circuit with the jfet 1Mohm gate resistor tied to ground.)
+14.5 dBm OIP3; +24 dBm OIP2

2. 20 mA bias. +21.5 dBm OIP3, +32 dBm OIP2

3. 30 mA bias. +26.5 dBm OIP3, +39 dBm OIP2

4. 40 mA bias. +29 dBm OIP3, +43 dBm OIP2

5. Bias for best 2IMD response, 62.7 mA. +29.5 dBm OIP3, +52 dBm OIP2. (The 2N2907 was so hot I burned my finger touching it, but it was still working fine, even with no heat sink.)

Gain/Frequency response, with 1 uF output capacitor: 20 kHz -3.3 dB, 100 kHz -2.1 dB, 1 MHz -2.1 dB, 10 MHz -2.7 dB, 20 MHz -3.2 dB, 40 MHZ -5.2 dB
------------------------------------------------------

Lankford circuit, using same J309 device, and 2N5109 with heatsink.
Vcc 12.8 volts

Biassed for best 2IMD response, 59.4 mA. +32 dBm OIP3, +75 dBm OIP2.
Gain/Frequency response, with 1 uF output capacitor: 20 kHz -4.4 dB, 100 kHz -3.7 dB, 1 MHz -3.3 dB, 10 MHz -3.6 dB, 20 MHz -3.4 dB, 40 MHz -3.4 dB
(I used a one-watt 120 ohm emitter resitor--it was only mildly warm; the 2N5109 with 3/4" diameter heatsink was also only mildly warm.)
------------------------------------------------------

Observations and comments.
The Trask circuit has very poor IMD performance when biassed for basic minimum current. Even when biassed for best IMD response, and the 2N2907 with no heat sink practically burning up, performance was only beginning to approach the Lankford circuit for 3IMD, and far from it for 2IMD. And the Lankford circuit could probably be optimized further by AC coupling the 2N5109 and rebiassing it for best IMD performance.
Gain and frequency response were basically the same for both circuits.

My test setup consisted of an HP3585A 40 MHz spectrum analyzer with builtin tracking generator used for the Gain/Frequency response test, and a pair of HP8640B generators along with a homebrew high-isolation 6 dB combiner used with the spectrum analyzer for the IMD tests. Both circuits were fed in all tests with an input circuit consisting of a 51 ohm to ground input resistor to the generator, and a series 10 pF capacitor to the gate of the jfet--simulating a one-meter whip antenna. This is the standard test setup for testing any high impedance active whipamp circuit; the series capacitor can be changed to simulate whatever actual whip antenna that will be used with the real circuit. A thin whip is characterized as approximately 10 pF/meter length.

73,
Steve

Signal gain can be achieved by adding an additional resistor
between the 2N2907 emitter and the J309 source. Doing this
will require a higher supply voltage so as to provide
sufficient overhead to prevent saturation.

Can you clarify this statement? It appears that this would be the
same as connecting a resistor from Vcc to the J309 source?

Ooops! It should have said between the 2N2907 COLLECTOR and
the J309 source. The output is then taken from the 2N2907 collector.

That's what I thought and makes more sense. I assume that the higher
the gain, the less feedback and less IP2/IP3?

As you increase the gain, there will be less feedback, however the
IP2/IP3 will degrade as less power is sent to the J309 source to correct for
distortion. There's sufficient open loop gain such that you shouldn't
notice this much for gains of less than, say, 10dB.

A 120-ohm resistor in the feedback path gives a gain of 3dB. Without
it, the circuit has a loss 0f 2dB, which is about 1dB better than the
earlier circuit.

BTW: The bandwidth can be improved by adding a 1.0uH inductor in series
with the resistor that goes from the J309 drain (and 2N2907 base) to supply.

Chris

,----------------------. High Performance Mixers and
/ What's all this &#92; Amplifiers for RF Communications
/ extinct stuff, anyhow? /
&#92; _______,--------------' Chris Trask / N7ZWY
_ |/ Principal Engineer
oo&#92; Sonoran Radio Research
(__)&#92; _ P.O. Box 25240
&#92; &#92; .' `. Tempe, Arizona 85285-5240
&#92; &#92; / &#92;
&#92; '" &#92; IEEE Senior Member #40274515
. ( ) &#92;
'-| )__| :. &#92; Email: christrask@...
| | | | &#92; '.
c__; c__; '-..'>.__

Graphics by Loek Frederiks

Hi Chris,

That's what I thought and makes more sense. I assume that the higher
the gain, the less feedback and less IP2/IP3?

Rob

--- In loopantennas@..., "Chris Trask" <christrask@...> wrote:

Signal gain can be achieved by adding an additional resistor
between the 2N2907 emitter and the J309 source. Doing this
will require a higher supply voltage so as to provide
sufficient overhead to prevent saturation.

Can you clarify this statement? It appears that this would be the
same as connecting a resistor from Vcc to the J309 source?

Ooops! It should have said between the 2N2907 COLLECTOR and the

J309

source. The output is then taken from the 2N2907 collector.

Chris

,----------------------. High Performance Mixers and
/ What's all this &#92; Amplifiers for RF Communications
/ extinct stuff, anyhow? /
&#92; _______,--------------' Chris Trask / N7ZWY
_ |/ Principal Engineer
oo&#92; Sonoran Radio Research
(__)&#92; _ P.O. Box 25240
&#92; &#92; .' `. Tempe, Arizona 85285-5240
&#92; &#92; / &#92;
&#92; '" &#92; IEEE Senior Member #40274515
. ( ) &#92;
'-| )__| :. &#92; Email: christrask@...
| | | | &#92; '.
c__; c__; '-..'>.__

Graphics by Loek Frederiks

Signal gain can be achieved by adding an additional resistor
between the 2N2907 emitter and the J309 source. Doing this
will require a higher supply voltage so as to provide
sufficient overhead to prevent saturation.

Can you clarify this statement? It appears that this would be the
same as connecting a resistor from Vcc to the J309 source?

Ooops! It should have said between the 2N2907 COLLECTOR and the J309
source. The output is then taken from the 2N2907 collector.

Chris

,----------------------. High Performance Mixers and
/ What's all this &#92; Amplifiers for RF Communications
/ extinct stuff, anyhow? /
&#92; _______,--------------' Chris Trask / N7ZWY
_ |/ Principal Engineer
oo&#92; Sonoran Radio Research
(__)&#92; _ P.O. Box 25240
&#92; &#92; .' `. Tempe, Arizona 85285-5240
&#92; &#92; / &#92;
&#92; '" &#92; IEEE Senior Member #40274515
. ( ) &#92;
'-| )__| :. &#92; Email: christrask@...
| | | | &#92; '.
c__; c__; '-..'>.__

Graphics by Loek Frederiks

Chris,

Can you clarify this statement? It appears that this would be the
same as connecting a resistor from Vcc to the J309 source?

Rob

--- In loopantennas@..., "Chris Trask" <christrask@...> wrote:

Signal gain can be achieved by adding an
additional resistor between the 2N2907 emitter and the J309 source.

Doing this will require a higher supply voltage so as to provide
sufficient overhead to prevent saturation.

Hmm I have a problem opening this URL.... Is there a possibility to
place it somewhere else, or email it directly to me? Or maybe the
files section?

If others have the same prob I could place it somewhere as well.

I just clicked on the URL in your posting and it downloaded properly.

Chris

,----------------------. High Performance Mixers and
/ What's all this &#92; Amplifiers for RF Communications
/ extinct stuff, anyhow? /
&#92; _______,--------------' Chris Trask / N7ZWY
_ |/ Principal Engineer
oo&#92; Sonoran Radio Research
(__)&#92; _ P.O. Box 25240
&#92; &#92; .' `. Tempe, Arizona 85285-5240
&#92; &#92; / &#92;
&#92; '" &#92; IEEE Senior Member #40274515
. ( ) &#92;
'-| )__| :. &#92; Email: christrask@...
| | | | &#92; '.
c__; c__; '-..'>.__

Graphics by Loek Frederiks

--- In loopantennas@..., "Jeroen Kloppenburg"
<3mensions@...> wrote:

Hmm I have a problem opening this URL....

Got it now! Thanks Cor & Paul :)

--- In loopantennas@..., "Chris Trask" <christrask@...> wrote:

Hmm I have a problem opening this URL.... Is there a possibility to
place it somewhere else, or email it directly to me? Or maybe the
files section?

If others have the same prob I could place it somewhere as well.

Most of the amplifiers used with active antennas have fairly common
flaws, and the amplifiers found in these two URLs are no exception. The
circuit from Electronic Design:



starts with a source follower and is then followed by a 703 monolithic
amplifier. Despite the fact that the FETs used for the source followers are
heavily biased, there is still going to be considerable IMD due to the Vgs
characteristics of the devices, and the 703 amplifier is only going to make
things worse. The circuit at least has some signal gain.

The circuit from Dallas Lankford:


%20Power%20Supplies%20II.pdf

further exacerbates the nonlinear characteristics of the JFET source
follower by adding a BJT emitter follower. Again, the circuit is heavily
biased and despite this it will not overcome the IMD problems due to the Vgs
characteristics of the J310 and the Vbe characteristics of the 2N5109.
Further, this circuit has no signal gain and only serves to couple the short
monopole to the coaxial cable.

I put together a very brief page showing the Lankford circuit and then a
far more linear and power efficient circuit that is sometimes known as a
current-boosted source follower:



This circuit has a very high open loop gain, and when the loop is closed as
shown in the schematic there is very little signal loss, a low level of IMD
distortion, and it operates on just 10mA of supply current. The supply
voltage can be as little as 5V. Signal gain can be achieved by adding an
additional resistor between the 2N2907 emitter and the J309 source. Doing
this will require a higher supply voltage so as to provide sufficient
overhead to prevent saturation.

Chris

,----------------------. High Performance Mixers and
/ What's all this &#92; Amplifiers for RF Communications
/ extinct stuff, anyhow? /
&#92; _______,--------------' Chris Trask / N7ZWY
_ |/ Principal Engineer
oo&#92; Sonoran Radio Research
(__)&#92; _ P.O. Box 25240
&#92; &#92; .' `. Tempe, Arizona 85285-5240
&#92; &#92; / &#92;
&#92; '" &#92; IEEE Senior Member #40274515
. ( ) &#92;
'-| )__| :. &#92; Email: christrask@...
| | | | &#92; '.
c__; c__; '-..'>.__

Graphics by Loek Frederiks

Thanks,

This will make it a serious candidate for experiments.

Regards,

Cor Beijersbergen van Henegouwen

Could this design be adapted to have the power supply through the coax?

Cor Beijersbergen van Henegouwen

Yes.

That is how r2000slwer built mine.

There is a power injector or diplexer at the receiver end
that consists of a choke to feed the DC and a small value
capacitor to pass RF to the receiver.

At the amplifier end a choke passes the DC to the V+ input
and a capacitor couples RF to the coax.

If you have a link relevant to the group, please post it in the Links section. If you just put it in a message, it will get lost in the shuffle after a few days. The coverage of the list is pretty broad.

All members can upload Links, Files, and Photos.

A few guidelines:

Try to place them in the proper folder, if it exists. If not, you can either just put it in the root folder and eventually I'll move it into the proper place, or create a folder.

When creating a folder or uploading a file or photo, please don't use spaces or punctuation. You can either use caps for the first letter of each word, or use underscore as a space.

Like this: "4x8Loop" or "4x8_loop" or "Steves_4x8_Loop"

Not like this: "Steve's 4 x 8 Loop." That goes for Folder names and File names.

A special note about the Photos section: Although it will allow arbitrarily large images, it will only show them to the moderator and whoever uploaded them. Everyone else can only see a resampled max of 300x400 pixels. So anything over that in Photos is a waste of storage. So crop in to the important parts and resample to about 300x400. For photos, JPG saves smaller and looks better, for line drawings GIF is smaller and looks better. If you need to post a schematic diagram that would be unreadable at 300x400, I strongly suggest the Files section as it has no such limitations.

Feel free to ask if you need help figuring out how to crop and resample. Irfanview is a fairly easy to use program that loads just about any file, you can crop and resample from there:

Or for more complex stuff, another free program is GIMP:

A last note: if you see something in any folder or in the messages that looks like spam or simply inappropriate, don't post on the list- email me directly and tell me about it and I'll take care of it with a minimum of fuss.
loopantennas-owner@...

Thanks,
Steve Greenfield

--- In loopantennas@..., "gfreport1" <gfreport@...> wrote:

I stopped by to have lucnh with Terry and this is what he has
to add.

Thanks for that info, much appreciated :)

I prefer the active dipoles designed by Dallas Lankford.

Gotcha =)

I'm still looking for the availability for some of the parts of the
wl1030 and for some I might need to get them per 10, so the wl1030
plan might simply be pushed of the table solely for that reason.

I will look into those dipoles as well. Maybe the cheap/easier to
built version first that is :) Still need to gain more building
experience anyway...

Thanks all!

--- In loopantennas@..., gandalfg8@... wrote:

In a message dated 31/05/2007 16:04:34 GMT Daylight Time,
rblum@... writes:

I think you are seeing the capacitors he added. He says he added 1 mf
ceramic capacitors, and that the diodes were not noisy, because they
"did not reduce the noise at the lower end of the NDB". He apparently
measured or at least noted the noise; added capacitors and inductors;
noted no significant noise reduction; concluded the capacitors and
inductors really weren't needed for that particular R/S supply.



---------------------------------
Hi Bob

The point Len was making is that you can see the added capacitors on

the

back of the board but that there are already capacitors across the

diodes,

mounted as normal through hole components on the circuit boards, and

that could

have been the reason there was no noise reduction in this example.
Whether or not the existing level of noise was considered acceptable

does

not seem very clear, altough I get the impression it might have

been, the

observation was only that it didn't change.

If you look at the lower photos in the reference given earlier

you'll see

that Len is right.
The existing caps look, to me anyway, to be fairly small ceramic

types but

they are definitely in circuit and part of the original construction.
You can see two of the original capacitors clearly, a bit of

another less

clearly, and can also see from the rear photo that they are across

the diodes,

in addition to the later ones mounted on the rear of the PCB.

regards

Nigel
GM8PZR

Thanks for the clarification, Nigel. This time I zoomed the thing up
to 300+% and I see the little fellas you are talking about.

Nice "talking" to you again.

Bob