On 27 jul, 09:44, Ross Archer wrote:
On Jul 26, 6:28*am, Wimpie wrote:



On 26 jul, 03:23, Ross Archer wrote:

I'd like to try building an *untuned* loop amplifier covering HF and
wonder if anyone has seen links on this.

The thinking so far is using a pair of 2n2219a bipolar transistors
running a few hundred milliamps, with their collectors driving
opposite ends of an RF output transformer, and a reasonably large loop
(say 1 meter square or more).

I'd like the output to be "hot" enough to compensate for a losses from
a nice long run of cheap cable-TV type coax cable (through which power
will be fed up to the amp). *But not so "hot" as to create distortion.

Anyone here have experience with this sort of project? *If so, is the
2n2219a a stupid choice? *Would I be better with JFETs, and would a
single-ended amplifier work almost as well as a balanced amp? *(My
thinking is the balanced amp would potentially be quieter.)

It would be fun to try this, and my junk box is overflowing with
parts, so why not?

-- ross

Hello Ross,

I did some experiments, with same type of transistors. My loop was
made from chicken mesh to reduce the impedance in the high part of HF.
Low impedance reduces the contribution from the base current noise
from the transistors.

I had to do some tweaking to avoid parasitic oscillation. I also had
intermodulation distortion from VHF that was noticeable at around 5
MHz. The reason was the increased sensitivity of the loop around the
VHF broadcast band. *Adding RC section parallel to the loop did reduce
this significantly (below noise level at HF).

Other issue was intermodulation in my receiver (I checked this by
adding a tunable filter between the amplifier and the receiver).
Reducing the gain by adding some emitter resistance did help.

In the end I changed the loop into a loop with natural balun (same
principle as used in small loops for EMC purpose, the have a cut in
the middle [opposite to the feed point]). About 10 m low loss coaxial
cable (with common mode ferrites) *is between the loop and a simple LC
parallel circuit with selectable tabs to have some preselection and
matching.

I also built a 2 stage non-balanced amplifier with feedback via the
emitter of the 2n2219 input transistor (all inside the antenna loop
itself, fed over the coaxial cable). Using 2n2219 with 2N3906 resulted
in intermodulation.

Changing to BFR91A (input stage) and BF979 (buffer, is PNP, I used two
devices in parallel) gave a significant reduction of intermodulation
(because of higher open loop gain). *So if you want to use 2n2219A as
a single stage, a balanced amplifier will definitely better then an
unbalanced stage.

I live in an almost flat area (valid for most parts of the
Netherlands) with many signals (and noise) present. So I didn't need
additional amplification for HF (based on chicken mess loop with 1m
diameter). *When you live in a remote area you probably need some
gain.

Best regards,

Wim
PA3DJSwww.tetech.nl
please remove the obvious character combination in case of PM.

Very interesting, and thank you for the feedback.

Do I understand that the "natural balun" is the 10m length of coax
with ferrites to choke off any unbalanced currents? Is there something
important about the "break" in the loop opposite the feedpoint, do you
think?

I'm not set on any particular transistor, or even technology. *NPN,
PNP, JFET, MOSFET, IC, whatever works best.
My junkbox has the following in serious numbers:
BF998
2n4416a
J310
MPF102
2n3819
PN2222a
MPSH10
MPSA18
2n3904
2n3906
and a few:
NTE10
2n2219a
2n2218

I had a few 2n5109s, which might be even better, but can't find them.

The real reason I'm starting this is I was re-organizing and found my
parts stash, and figured it's time to build stuff again

-- ross

Hello Ross,

An example of a (small) loop with integral balun function is shown
he http://www.tetech.nl/divers/small_H_FieldLoop2.jpg

The coaxial cable is connected to the top of the loop (via the three
thin black wires). The coaxial cable goes down via one side of the
loop. Make sure the coaxial cable (bread) is connected (electrically)
to your loop at some points and at the point where it leaves the
loop.

Where it leaves the loop, you can see one of the ferrite cores that
provide additional common mode suppression. I use this loop for
measurement purposes.

The cut in the loop (where you connect the coaxial cable), must be
opposite to the point where the cable leaves the loop. My large loop
(not operative now, but same principle) is made of 2.5 m long 0.5m
wide piece of chicken mesh (so D = 0.8m). I checked the noise floor
of the complete active setup by approaching the loop with a large
aluminum sheet. When the aluminum (or other conducting material) is
about 30 cm separated from the loop, the inductance of the loop
doesn't reduce that much, but the outside noise reduces
significantly.

When you have good results with your unbalanced amplifier (with high
current), why not making a passive loop and keep the amplifier
indoors? You don't need the transformer and you have easy access for
various experiments. Don't worry about the 0.5*Vsupply issue, this is
only of importance when you want maximum voltage swing. Try to reduce
the current to the point where intermodulation isn't noticeable. It
may reduce the own noise production of the amplifier at the higher end
of HF (noise generated by the DC base current).

When you have more then sufficient gain, you may add a low value
emitter resistor. This reduces the gain, increases noise production a
bit, but enhances large signal performance.

Why I decommissioned the loop antenna? Up to some MHz, the loop
outperformed my small chicken mesh dipoles, but at higher HF, the
difference between the passive dipoles and both the active and passive
loop wasn't noticeable for me. I now only use the loop sometimes for
VLF measurement (RFID).

Best regards,

Wim
PA3DJS
www.tetech.nl
please remove first three characters of alphabet in case of PM