Softrock Lite 6.2
Adventures in Electronics and Radio
Elecraft K2 and K3 Transceivers
26 July 2009
I've added more data for BG Incorporated's KS transformer,
this time with low impedance (zero ohms) op-amp drive. The details are towards
the bottom of the page
Non-Linear Transformer Behavior or follow the links at the top of that page.
26 July 2009
It's necessary to make certain performance tests on
amplifiers such as the Z10040B in a shielded enclosure, and after fumbling
innumerable times with short 4-40 screws when a new amplifier is built and
tested, I made a set of four thumbscrews.
The thumbscrew consists of a short length of 0.5 inch
diameter acetyl (Delrin) rod partially drilled through. I then pressed in a
one-inch long 4-40 male/female standoff. Since acetyl is soft, I used a bench
vise to press the standoff into the cap.
25 July 2009 PM
It's now 1900 hours and repair is finally underway for my
CATV and Internet connection. Total outage is approximately 80 hours.
We've wound up with a temporary connection, coaxial cable
laying on the ground, with a subsequent visit required to trench it into the
25 July 2009
Joop, PE1CQP, sent several BG Industries model "KS" transformers for measurement
and I've added the data to my
Non-Linear Transformer Behavior page. These are quite decent transformers,
with good harmonic distortion performance below 1000 Hz compared with other
inexpensive parts. At some signal levels, however, the intermodulation
performance plateaus above 1000 Hz and is not as good as some similar
transformers I've measured.
25 July 2009 AM
As I suspected, yesterday did not see my CATV and Internet
service restored, despite spending most of the day on the telephone with
various Cox Cable support people.
I was repeatedly assured that someone would be out on the
24th, between 9 AM and 9 PM. Didn't happen. A couple of times the telephone
support people said that they had asked the assigned technician to phone me with
an estimated schedule. Didn't happen either. The work order is "carried over" to
today I am now told.
24 July 2009 AM
As of 0830, the CATV cable is still on the ground and my
Internet service remains out. I was promised yesterday that a bucket truck and
crew would show up before 2000 hours that evening and restore service. That
didn't happen, and this morning I was told that a crew would be out "sometime
today." No specific time, except before 9 PM.
23 July 2009
I had planned on updating this page yesterday, but my
Internet service is out. In fact, at the time this entry is drafted (1530 EDT,
23 July 2009) it's still out.
My Internet service is from our local cable television
provider, Cox Cable, and it's generally been reliable and fast. Yesterday,
Verizon Telephone had a crew with two or three bucket trucks working on the telephone lines
running along the (shared) electrical power and CATV poles near our house.
Shortly after the work started, our CATV
video and Internet service went out.
Walking out to the utility right-of-way quickly showed
what happened. For some reason the Verizon work crew cut my CATV service drop at
the tap point (about 200 feet from our property line) and pulled the twin coax
cable through the lashings and left it on the ground.
My CATV drop after Verison cut it and pulled it down.
This was not a small undertaking, as not only was the cable cut, it was pulled
out of the lashings.
Clean cut of the two coax cables and the attached
The first Cox technician arrived this afternoon but realized
stringing 200 feet of aerial cable requires a bucket truck, not his van
and a ladder. I've been promised a repair crew will arrive before 2000 hours
22 July 2009
My initial order of active antenna printed circuit boards arrived a couple
days ago and I built one today. It looks very good from my initial tests, but it
will take some off-the-air listening to verify the test gear data.
The PCB is 2 inches (50 mm) x 3.4 inches (85 mm) and fits nicely inside a
Hammond weatherproof die cast enclosure. I've decided to not install a fixed
coaxial connector, but rather use a waterproof cable gland with a short RG-58
pigtail cable. The cable will be equipped with the purchaser's option of BNC,
UHF or F connectors. Likewise the antenna rod input is via a second cable gland
that enters at the top of the enclosure. The side hole is left over from an
abandoned idea and won't be in the production units.
I'm still at least a month away from being ready to offer a complete kit, as
I'm waiting on a price quotation for the mounting bracket (see 07 July entry
below) and then need to revise the DC power coupler PCB to fit a somewhat larger
enclosure than the current 2 x 2 inch (50 mm x 50 mm) die cast box.
Production PCB installed in a test enclosure.
The image below is the frequency response of the production active antenna,
plotted from 300 KHz to 300 MHz. The 3 dB roll off point in the high frequency
end is 139 MHz. I've intentionally added high frequency degeneration via ferrite
beads as a parasitic suppression measure (the output transistors have an fT
of 5 GHz, so parasitic oscillations are always something to be concerned over
when laying out the board. Still, the antenna looks to be quite usable up
through 150 MHz, although it's not what one normally uses for VHF reception.
I've also taken some design measures to roll off the low
frequency response below 15 KHz with the thought that it will aid in keeping 60
Hz power line and harmonics from overloading the amplifiers. I did not plot the
response below 300 KHz but my earlier data shows a useful response down to 20
KHz or so. The limiting factors are associated with the DC power feed and the
shunting effect of the isolation inductors.
The mid band gain is -1.6 dB, which is reasonable for this
sort of antenna. Gain is not the object of an active antenna, but rather
transferring a signal from a high impedance short rod antenna to a 50 ohm
receiver is the job it must perform. A short rod antenna will pick up nearly
(within a couple dB) as much signal at 1 MHz as a full size vertical. The
problem is efficiently coupling that signal into a 50 ohm receiver. A full size
antenna has an impedance nearly the same as the receiver's 50 ohm input, so
signals it intercepts are efficiently coupled to the receiver. A short rod
antenna, on the other hand, has a very high impedance, and if coupled to a low
impedance amplifier or receiver, most of the received signal is lost due to
voltage divider action.
A 1 meter long rod antenna near ground looks roughly
like a perfect low impedance voltage source with a 10 pF output capacitance. At
1 MHz, 10 pF has an impedance of 16K ohms, so into a 50 ohm receiver, the
voltage divider loss is around 50 dB.
An active antenna has a high input impedance (several
hundred thousand ohms) and a low output impedance and hence obviates the voltage
divider loss. 1 microvolt induced into the rod antenna yields nearly 1 microvolt
at the active antenna's output. Hence, the very short antenna delivers nearly
the same signal to a receiver as does a full size antenna.
As good as this sounds, the world is not overrun with
active antennas because there are some remaining problems. One is it's difficult
to filter the active antennas input, and hence it is exposed to potentially very
strong medium wave AM broadcast signals, or in Europe even stronger long wave
signals. Another is that the active antenna is not noiseless and hence amplifier
noise can offset part of the potential gain. And finally, being a high input
impedance device, the active antenna can couple to electrical noise from nearby
sources, such as pickup on the coaxial cable feeding the active antenna.
And, because the active antenna is small, there's a
temptation to use it inside the radio room. This is usually a recipe for seeing
how much computer hash is generated.
I'll have more to say on these subjects over the next
weeks and months.
14 July 2009
I've modified my Si570
kit review page to correct my mistaken understanding that the controller
firmware did not allow correcting for frequency error. I've added a section with
step-by-step instructions on correcting for frequency error and also added a
longer term frequency stability run.
13 July 2009
I've updated my Si570
kit review page to add an output power versus frequency plot.
12 July 2009
I've added a new page with a detailed review of the Si570
controller kit developed by John, K5JHF, and Kees, K5BCQ.
Si570 Kit from K5BCQ.
I've also revised the canned oscillator phase noise page
to add data for two new "synthesizers in a can" as well as the Si570.
Canned Osc Phase Noise.
These pages, by the way, represent about a week's worth of
measurements, analysis and writing.
10 July 2009
Jeff, AC0C sent a couple of newer design inexpensive
synthesized crystal oscillators for me to look at, with the thought that they
might be usable in a Softrock receiver as a substitute for a custom quartz
crystal or a less expensive single frequency alternative to the Si570.
That request lead me to purchase the Si570 oscillator and
controller kit sold by Kees, K5BCQ,
I received the Si570 kit earlier this week, assembled
it and started looking at the synthesized oscillators. My plan is to add a new
page reviewing the Si570 kit and also add the new synthesized oscillator
measurements to my Canned Osc Phase Noise page.
With some luck, I'll have it all finished by Sunday, July 12th.
The image below provides a sample of the data. The AP3S
synthesized oscillator has a great deal of noise compared with a high quality
signal source, the HP 8640B signal generator. The Si570, from my preliminary
measurements, appears to have excellent phase noise, similar to the HP 8640B.
HP 8640B signal generator (green) and Abracon AP3S
synthesized oscillator (yellow). 100 KHz total span.
07 July 2009
While waiting for the multi-coupler board to arrive, I'm
working on the Z1501C active antenna and the associated Z1202 DC power injector.
I completed the revised and, I hope, final PCB layout for
the Z1501C active antenna today and ordered a small quantity of printed circuit
boards. Delivery is probably 10-12 days away, but that does not mean the kits
will be ready that soon. I'm still working through the mechanical design, but
that is far enough along to permit finishing the PCB layout. The main change to
the PCB from the last version is increasing the board size to add mounting
holes. (The last prototype used the BNC connector for mechanical support.) I've
also removed the PCB-mounted connector and changed from transformer coupled
output to shunt fed DC duplex power as implemented in the Z10040B.
One way I work out the mechanical issues is with paper and
cardboard mockups of the hardware. The photo below shows a cardboard version of
the bracket I designed to attach the electronics module to a support pipe and
also support the antenna rod. I'm waiting on the U-bolts to arrive so that I may
complete the bracket layout and prepare drawings for my sheet metal supplier.
With some luck, I'll be able to order the brackets next week. That places
delivery into late August or early September.
I also plan to finish the enclosure layout for the Z1202
DC power coupler. It may be used to power either the Z1501C active antenna or a
remote Z10040B Norton amplifier.
My target for kit deliveries is early September, with the
limiting factor being the sheet metal work.
Cardboard mockup of mounting bracket.
02 July 2009
I've completed a major revision of my antenna
multi-coupler prototype and have a set of new printed circuit boards ordered. I
expect to receive the boards in 10 days or so. Assuming the new design works as
well as the individual pieces spread across my workbench, I'll then revise the
enclosure layout and wrap up a couple of associated tasks. If all goes according
to plan, the multi-coupler kit (or assembled) will be available around the first
I've also resumed the mechanical work on my active antenna
kit. These tasks involve revising the enclosure and antenna rod elements. When I
have a final enclosure, I will revise the PCB layout to fit in the enclosure.
There are a couple of related elements to the active antenna that I've
also been working one. One is a common mode choke and the second is a revised DC
power injector. These fall into the same category as the active antenna - the
electronic part is finished, but I have more work to do on the packaging.
I've also been working on several small kits. One if a 3
dB hybrid combiner or splitter. The difference between a combiner and a splitter
is the direction of signal flow. A splitter divides an input signal into two
equal outputs. A combiner sums two signals into a common output, with isolation
between the two inputs. Same device, the difference is which ports are used for
inputs and which for outputs.
A prototype version of the coupler is pictured below. The
extra holes are from an earlier project housed in the same enclosure.
Version two of the prototype 3 dB hybrid coupler.
My objective is performance, measured in terms of splitting
loss, input return loss, and isolation, to at least equal Mini-Circuit's
ZFSC-2-6+ hybrid, at a lower price, when assembled by the user, over the range 1
MHz to 50 MHz. So far, my prototype is performing quite well, superior to the
The image below show port-to-port
isolation of the ZFSC-2-6+ and my two prototypes over the range 300 KHz to 100
MHz. The lower the curve, the greater the isolation. The more isolation, the
better. At 1 MHz and below, all three hybrids are about the same. Above 5 MHz,
both of my prototypes have significantly better isolation than the ZFSC-2-6+.
The plots below show the hybrids used as a splitter. The upper plot is the
return loss (similar to SWR) of the input port when used as a splitter and the
two output ports terminated with 50 ohm loads. The horizontal reference line
(marked with the > symbol) is -20 dBm return loss, corresponding to an SWR of
1.22:1. The greater the return loss, the lower the SWR. Greater return loss
shows as a lower trace on the plot.
The lower plot
is coupling loss, i.e., the power out of one of the output ports
compared with the input power. Since the input power splits two ways, a coupling
loss of 3.01 dB represents a perfect hybrid. Anything over 3 dB represents loss
within the coupler. In this plot, the higher the trace, the lower the 'excess'
loss and the better the coupler performs.
In both return loss and splitting loss, my prototype
designs are performing at least as well as the ZFSC-2-6+, and in most
frequencies, my prototypes perform better.
02 July 2009
As usual, I've moved June 2009 Updates to an archive page.
Click here to view it, or use the archive navigation
table at the top of this page.