Wide Band Bipolar RF Amplifier

January 15th, 2014

By Ford-NØFP

Many RF projects require modest amplification in various spots. Picking the right active stages, the right circuit topology, the right collection of components, seems to be more like an art form than science. No perfect solution fits every circumstance. The correct “choice” must balance competing parameters. The design must determine a suitable, rather than perfect, compromise. Finding the right balance of gain, dynamic range, noise figure, power output, power consumption, voltage available in the circuit, cost, and complexity, is never perfect. For me, a rank amateur at the design game, the process is more like guesswork than art or science. I recently set out to do some basic measurements of a wideband bipolar amplifier just to see how it performs. To use sailing vernacular, it would be nice if my designs “looked smart” rather than “cobbled.”

The initial choice of a bipolar amp for my first design is borne of a desire to use readily available components instead of fancy modules, which perform to spec for a handsome price. The readily available cheap transistor makes this initial design a simple choice. So I shamelessly copied a two transistor preamp design from page 13 of the Guidebook of Electronic Circuits, © 1974 by John Marcus.

The description uses a pair of MPS-918 NPN transistors, a Motorola part. The author states:

100 MHz Preamp—Broadband general-purpose amplifier has gain of more than 40 dB below 5 MHz, dropping to 30 dB at 10 meters and 23 dB at 6 meters. Will improve noise figure and image rejection of any receiver. Operates with either 50-ohm or 75-ohm receiver input impedance. D.K. Bercher and A. Victor, A general purpose Solid-State Preamplifier, QST, Sept. 1971, p 32.

My goal was to design this circuit board, crank out a few copies on my CNC PCB engraver, and measure the design using various common available NPN transistors instead of the unavailable transistor used by the author. A quick check of my junk box shows several SOT-23 surface mount components I’d like to evaluate and compare side-by-side.


Vceo = Maximum Collector Emitter Voltage (V)
Ic=Maximum Continuous Collector Current in milliamps (mA)
Pd=Total Device Dissipation in Milliwatts (mW)
fT=Current Gain x Bandwidth Product in MHz

Using KiCAD to draw a schematic, build a circuit board using the on-line Java based autorouter, and generate the Gerber files, the design was uniform for comparison purposes. My homebrewed CNC router table requires GCode instead of Gerber so I converted the Gerber output of KiCAD to GCode using PCB2GCODE for Linux, which runs the EMC2 software driving the router table.

Click on drawing for a larger view and here for a PDF of the KiCAD drawing.

Engraver software EMC2 for Linux screen shot

Cutting isolation bands around pads using a 0.2mm router bit.

Shearing the board into groups using a small shear / finger brake.

Cutting the small boards apart.

Of the 9 boards, one was flawed due to variation in board thickness. Notice the connector trace at the bottom where the isolation band didn’t quite route out completely.

Plating the boards using “Liquid Tin.” The plating looks complete after only 15 seconds but the directions call for 5 minutes of soaking in the liquid brine. Washing in clean water when done. Soldering to the Liquid Tin is wonderful. It’s like soldering a pre-tinned board.

The completed board has power and ground pins fabricated from a wire wrap socket. The input on the left and output is on the right. Connectors are MCX style board edge connectors, which are WAY cheaper than SMA style connectors.

Initial testing showed a total gain at 12 volts of about 40 dB at 8 MHz with 0 dBm output before any spurious harmonics started to appear. Gain dropped to about 30 dB at 2 MHz and continued to drop as I went lower. At 20 MHz the gain started to drop to about 30 dB and at 50 MHz down to about 23 dB, which seems consistent with what the author indicated. More testing is needed after building some more units using alternative active components for comparison.

The design would easily facilitate a narrow band version using tuned circuits instead of resistors in the output portion of each half of the design. For what it is worth, a tap after the first stage indicated that about ½ the gain is distributed to each stage. This is anecdotal at this point since my testing was to simply probe that location with the analyzer and was hardly scientific.

Simple Bench Supply

November 11th, 2013

By Ford-N0FP

I am building a N2PK Vector Network Analyzer (VNA) http://n2pk.com/ using VE3IVM boards http://www.makarov.ca/vna.htm . Since this uses surface mounted parts, I want to test the device in sections while mounting parts. The VNA uses a series of power supplies feeding a series of subassemblies and I wanted to test each supply rail before mounting the rather expensive ICs, which become fuses if not done correctly. For the completed project I need 12v at about 50mA and 5v at about 310mA to run the VNA. For testing purpose, a simple 3-terminal 1.5A regulator seems like a natural choice. A simple device would have proved useful in the past. Now I have the perfect excuse to build this handy tool. While I can always use the high powered and metered bench supply for the same function, I wanted a work area supply that is small, noise-free, and easily adapted to whatever fixture was on the bench.

For the final VNA assembly I will be using LM2596 class pre-regulators to muscle an unregulated power cube 15v supply down to 12.0v and 5.0v needed to run the VNA. More on this for a later blog post. For now, searching the junque box for parts, I found all that I need to make this useful tool. An Altoids tin seems like a perfect enclosure for my newest workbench gismo.

AltoidsDualSupply

There are two LM340-T5 class 1.5A regulators. Each is the 5V version and use a basic heat sink. One is fixed at 5v and the other has a 2K pot in the ground return to make it somewhat variable. The raw power comes from a ten pack of AA cells. The 9V style connector on the pack makes a convenient source and is easily adaptable to a single 9V cell. The ten-pack with alkaline batteries installed has an unloaded terminal voltage of just over 15vdc. The regulators need 2v of headroom to work. I have not tested the terminal voltage under load but the 2K pot seems to give me decent regulation between 5.4v and 12.8v. The intended service is to bias intermediate and unloaded power supplies while the board is being built, which should be at a very low current.

The picture shows two TO-220 regulators. On the left 5V fixed running to red-black wires. On the right is the variable version adjustable from 5.4v to 12.8v using the purple-green wires. The mounting tab on the adjustable TO-220 is insulated from ground.

Hustler Resonator Adapted to a Pickup

July 17th, 2013

By Ford Peterson – NØFP

 

I found a set of 80-40-20-15-10 Hustler resonators at an estate sale.  The system had a tip-over mast but the lower element was missing (likely due to damage).  The system has many virtues that interested me.  Having the resonator mounted above center increases the area where maximum currents flow and radiate.  Photons radiate because of current.  Above the resonator is the high voltage area.  Moving the loading higher in the element makes the hustler a more efficient radiator than a vertical with base loading.  The tip-over hinge allows me to store the element quickly so I can park in the garage without too much hassle.

 

 

The upper vertical element simply presses into the lower section.  The element(s) were constructed from ½” aluminum tubing.  I happened to have some ½” copper plumbing tubing, which is actually closer to 5/8” diameter.  Using several inches of 5/8” diameter solid rod, I hollowed out a pocket in one end to hold the upper-hinged element.  I also had to turn down the element to match the diameter of the original element.  The end-plug was turned down to fit tightly inside the copper tubing.

 

 

This particular Toyota Tundra has a ladder rack that can slide into rails of the cover.  I exploited the availability of the rails to fabricate ¼” plexi-glass brackets to support the long element.  ¼” bolts are fixed to a ¼” lock-nut that slides into the rail.

 

 

About a year ago I fabricated a bracket to hold a 3/8” x 24 tpi threaded element.  Although I had primed and painted the channel stock, it is quite rusted but continues to provide good continuity to the truck’s cargo box/fender.

 

 

When folded, the elements are secured to the truck to prevent taking out the mirror or windshield of a passing vehicle.  Two keepers were needed to facilitate quickly attaching the resonators, whether they are the longer 80M resonator, or the shorty 10M resonator.

 

 

The two keepers are identical.  A slot along their 3.5” length has a slot notched with a hand saw to allow me to gently bend the element and slide it into the pocket.

 

 

 

80M (all-band) NVIS (N.ear V.ertical I.ncidence S.kywave) Antenna

November 3rd, 2012

November 3, 2012  By Ford-NØFP

 

4 tower, 28′ high, 284′ long loop

4 supports, 72’ apart, 28’ high, 284’ of wire.  Looking west.

Design Concept:

More than one antenna is an operating feature.  When the two antennas have different performance, noticeable differences can be observed when doing an A-B switch.  I currently have an all band vertical.  It’s an elevated feed (8’ off the ground) aluminum tower about 68’ high, making the radiator about 60’ vertical.  An SG-230 tuner matches the system on any frequency from 1.8MHz to 30MHz.  I have many friends located 60 to 100 miles from my QTH.  HF propagation is tough, especially during the day.  The vertical has a very low launch angle and my friends fall inside the skip zone.

Horizontally polarized antennas are, in many ways, superior performers on HF.  While the vertical omni-directional antenna may provide superior performance in all random directions, and occupy a small footprint, they fail to take advantage of the additive effects of the reflection of the wave front in the area near the radiator.  The so-called “ground effect” of the Fresnel zone reflection can account for 6dB of additional gain!

Horizontal Dipole (1/2 wL high) vs Vertical Performance

Here we see a horizontal dipole at 66’ (just less than ½ wL) in black and compared to a ground mounted vertical in blue.

The physics behind the phenomena is simple.  The wave front emitted directly by a horizontal dipole combines with the wave front reflected off ‘ground’ beneath and in front of the dipole.  A vertical has no parallel reflective surface so it can only launch a direct wave front.  A horizontal antenna effectively combines the apparent radiation from two points, thus providing about 6dB of apparent gain.

So far, this discussion has only considered the peak radiation at approximately 30° above the horizon.  The launch angle is important.  Skywave propagation reflects signals off the ionized upper atmosphere to extend signals well beyond the visible horizon.  Multiple skips permit Trans global communications.  Suppose you don’t want to talk to the next continent, but the next county?  Or two counties over?  Ground wave communications, which is polarization sensitive, can be limited to a few dozen miles.  Skywave propagation becomes impossible when the target is inside or just beyond the “skip” zone. 

The classic propagation pattern of a vertical or ½ wavelength high dipole is perfect for long distance skywave.  By building a horizontal antenna low, you can introduce skywave propagation that virtually eliminates the problem of skipping over your target.  Compare the 28’ high 284’ square loop to a full sized 80M flat top dipole at 130’.

Low Loop (black) compared to 1/2 wL high dipole (blue)

Here we see the 28’ high square loop (black plot) to a full sized 130’ high dipole (blue plot)

Clearly the high dipole does a better job of launching the wavefront to the horizon.  The NVIS antenna shows a 15dB improvement at zenith (straight up).  Even at 30° it is roughly equivalent to a omni-directional vertical.  And it is a true omni by nature.

A 130’ dipole mounted horizontal at 130’ in the air is nothing short of darn near impossible.  What if the dipole were also mounted at 28’?  How would it compare to the loop?  Very similarly:

Low Loop (Black) versus Low Dipole (Blue)

The loop (black) plays only a couple dB better at zenith than the dipole (blue).

At these low altitudes, they are both omni directional, similar launch response, but the loop is far more complex.  The loop takes 284’ of wire, and 4 supports.  The dipole requires only 135’ of wire and two supports.

Multi-band feature

The dipole is clearly easier to build.  When mounted low, it becomes difficult to match, narrow bandwidth, and virtually impossible to match on the upper ham bands.  Here we see an easy match at 3.5, 11, 18, 25.5, and only 3.5 is within a hamband.

 

Dipole Frequency Response 3.3 MHz to 30 MHz

135’ long dipole mounted at 28’.

What happens with the loop?  Using a 50 ohm line, the match is fairly easy on most ham bands.

 

Loop Frequency Response 3.3 MHz to 30 MHz

When matching to 50 ohms, the natural dips fall within ham bands.  80M and 40M are fairly easy match, although quite narrow.  What happens if you use a 4:1 transformer and match it to 200 ohms instead of 50 ohms?  Additionally, suppose I add a 255pF capacitor and tune out some of the inductive reactance of the loop—effectively raising its resonant frequency just a bit.

Loop Frequency Response with 4:1 transformer

Here we see an easy match on 80M and 40M with a very easy match on all the upper bands, including WARC bands.

Here is a graph of the 80, 40, and Upper Bands.  As can be seen, this is an easy match for any auto-tuner in most radios.

Measured SWR Curve w/4:1 (200 ohm) feed

Performance on Upper Bands:

40M Performance

 

20M Performance

 

15M Performance

  

10M Performance

 

Building the NVIS:

Each tower has two 10’ sections of 2” PVC (schedule 40) coupled with a pipe coupler.  The bottom section is 8’ section of 2’ PVC but schedule 80.  A special coupler was fabricated using the 2” schedule 80, ripped in two on a table saw, and drilled to accept ½” ready-rod.  A ¾” strap was laminated between the bolts to increase rigidity and strength.

 

Schedule 80 to Schedule 40 Connection

The “dead man” anchors are 16” pieces of the 2” schedule 80 pipe.  A 36” hank of stainless cable threaded through a 2” piece of ¼” copper tubing, flattened, and bent in a loop.  A cable clamp secures the cable.  The cable and tube is buried at right angles inside a trench approximately 12” deep.  5 gallons of water was used to cement the clay around the tube.

 

“Dead Man” Anchor

Here we see an anchor before it was buried.  You can barely see the outline of where the sod was lifted to expose the loam and clay beneath.  A 24” trench, approximately 12” deep, secures the anchor.

The top guy is at the 27’ level.  The pulley is just short of 28’ high.  The plexi-glass guy ring was fabricated from ¼” stock.  A table router was used to round down the edges to prevent it from cutting the ropes.

Feed Point Pulley

RG213 was attached to a SO-239 chassis mount.  A 4:1 balun was fabricated from a 1.5” diameter type 61 ferrite toroid.  8 turns of RG58 was used to make the transformer.  The entire system was weather sealed using a hot glue gun and coating the entire connection point in hot glue.  The high impedance end of the transformer was connected to two 510 pF Silver Mica (500v) capacitors in series to form a 255 pF 1000v capacitor.  This brings the resonant point of the loop up about 125 KHz on 80M, about 60 KHz on 40M, and is virtually invisible on the upper bands.

Sky Hook Pulley and Insulator

The other 3 towers have a simple insulator fabricated from 0.4” this HDPE stock.

The rope was 130 lb test baling twine purchased from the farm supply store.  This orange rope is about 1/8” diameter and does not stretch.  It is also UV protected for durable operation.  They make two varieties of rope:  Biodegradable and non-biodegradable.  I don’t want anything ‘degrading’ so I’m using the rugged variety.  9,000 feet of twine is $26, which is $0.003/ft (1/3 of a penny!)

Guy Ropes and Tensioner

The rope tensioner was made of ¾” strips of oak trim with a 1/8” diameter hole drilled in each end.

Cable Connection

The guys are secured through a chain repair link looped through the cable anchor.

Lifting Rope Cleat

A rope cleat is used at about the 5’ level.  A 2” strap clamp secures it to the tower.  The cleat is handy to secure the lifting ropes.  Winding the feed point tower a few turns secures the RG213 to the tower and prevents it from blowing in the wind.

Next summer, the system will be lowered and move to the wall of my shed.  The dead man anchors will not be in the way of the mowing operations that happen every week.  If pushed in an emergency, this system could be transported on very short notice and reassembled using fence stakes as anchors.  The performance is excellent out several hundred miles during the day.  The military uses this style antenna for its large area operations.  You often saw a multi-turn version of this mounted to the Hum-Vees of military troops entering Iraq a few years ago.  They used it because it works!

 

Field Day Solar Panel

July 17th, 2012

By Ford-NØFP

I completed the first of two solar panels this weekend.  The goal is to solar charge deep cycle batteries for the ARRL’s Field Day activities.  Although I have only completed one panel, the plan is for two in series to make 19V Open Circuit and 15V at load.  Building this has been a bunch of artsy-craftsy nonsense.  It is HEAVY!  I am glad I designed it as two panels as each one weighs just over 20# and measures almost 30” square.  A 60” x 30” panel would have been twice the weight, bulky to transport, and mechanically more complex.  I decided not to attempt a weatherproof enclosure.  It is too complicated mechanically for building in your garage.  Commercial panels for all weather use are a series of specialized materials laminated together.  This design was to be 100% portable and intended for day light use in good weather.  I plan to put them away to protect them if it’s raining and no sun.  These are weather resistant but not weather proof by any means.

 

Completed Solar Panel

I used the Monocrystalline Photo Voltaic type cell, which is supposed to put out slightly better power density than the Polycrystalline PV materials.  I bought 36 cells, including tinned wire, on eBay for $74.  I received 42 with some broken and cracked pieces.  I need 32 cells to make the 19V panels.  During email exchanges with the seller, I discovered that the seller had acquired them from a defunct assembler of panels.  He was liquidating the assembly line inventory on eBay.

Blank 6" x 6" monocrystalline solar cells

These are amazingly fragile.  I was startled when I got my hands on one, which immediately broke!  No wonder the company went broke (pun intended)!  They measure only 0.010” thick.  Pick up a corner and it will break off from the stress.  Bump them with a tool or your finger and it will fracture or break completely!  Contact with a hard object and they will explode into a thousand shards of glass!  Handling is simply awful!  Now I am understand why the commercial all-weather panels are so dang expensive!  ($4/watt is typical)  Manufacturing perfect units must be painfully fraught with rejected materials due to handling mistakes.

Notice the 0.1” tinned wire soldered to some?  I broke these cells during the solder process.  The blue face needs to have two tinned wires soldered to the blank on each side.  I started soldering the gray back first and it turns out that was a big mistake.  I fractured 50% of the first six until I discovered that if you soldered the blue side first, the rigid wire/solder seems to handle the heat stress better than soldering the back in three places.  The added tinned conductors are necessary to handle the load.  5A in a few thousands thick conductor is insufficient.  Each line you see on the cell is a conductor.  The skinny lines carry electrons back to the two thick conductors.  On the blue side, the traces are all connected.  So if you fracture one after soldering, no worries.  It still puts out full voltage and power.  Break or fracture it in such a way as to prevent a DC path back to the tinned wire, and that part of the cell has failed.  I have several cracked pieces in this panel but the soldered wires are holding them together nicely.  If this were a commercial unit, it would be a rejected panel!  This is for Field Day, not space travel!  I say “Good enough!”  Notice the completed panel picture above.  You can see a fracture on the top and bottom second column from left.  The bottom left also has a corner missing completely.

My attempts at soldering the tinned wire was not working.  Using fine 0.010” solder there was not enough flux to clean the surface of the cell.  Then I discovered that Kester makes a pen to do exactly what I needed.  I ordered some flux pens before starting the solder process.  Some eBay sellers provide a pen and the tinned wire with the DIY kits.

http://www.all-spec.com/products/Kester/Soldering_and_Rework%7CSolder_and_Soldering_Chemicals%7CSOL-08/?utm_source=msn&utm_medium=cpc&utm_term=kester%20%23959%20flux%20pen&utm_content=367069932&utm_campaign=4053587006

I found that the #951 worked better when using old style flux solder.  The #959 worked better using solder with a more modern flux.  I also found the Kester “Flux 245” solder that was 0.010” diameter worked better than some old solder (0.010”) I had laying around.  Are these optimum materials?  I have no idea.  I had them and made them work.  The pen is like a felt tipped pen the width of that trace.  Take care to prevent breaking the cell when you push down the pen to release flux.  Painting on the flux to clean the trace made a dramatic improvement in soldering.  I used my Weller WCC-100 pencil and high heat to solder the length of the cell and left a 1.5” pigtail on the end to lash it to the next cell.  I also found that a solid sheet of aluminum was the best surface to do my soldering.  The first cell I soldered on a wooden bench and the varnish melted and glued the cell to the bench, which broke on removal.  Each cell took 20-30 minutes to complete the soldering.  If careful, they were not broken or fractured.  I immediately slipped them into the pockets (see below) for safekeeping.

My initial thoughts were to attach these to hardboard for field day use.  Wow!  Was that a messed up design!  These things are excessively fragile.  Protection using a rigid piece of glass seems like the only viable approach. 

It is important to understand that glass does not transmit 100% of the light.  There is a nasty reflection off the front and back surface of the glass.  This reflected light is GONE before it can work its magic in the cell junction.  There are special coatings to minimize this (the losses can approach 40% with the wrong glass) but the plating requires noble metals like platinum and gold.  This would be very expensive.  A lot of glass sold today is “Low E” or energy efficient.  This is clearly BAD for PV panels as the Low E rated glass reflects as much light as possible to minimize heat transfer into a building.  I need all the light to transfer through to the cell.

I used a 27” square piece of 3/16” thick glass (untreated) and ½” Styrofoam (high density pink stuff) insulation material.  Larger panels for commercial use ¼” thick glass, which is both heavy and expensive.  As it was, these two pieces of glass cost $56.  Then I used the router to build 16 pockets 0.1” deep just over 6-3/8” square to hold the 6-1/4” solar cells.  Using a steel ruler and carefully spacing it as a guide, I could clamp the ruler and use it as an edge for the router.    This formed the edges of the pocket.  For the middle part of the pocket, I used a 12” x 6” piece of Plexiglass and attached it to the router bottom so it would float on the 0.1” foam between the cell pockets.  This gave me a uniform depth for the pocket.

Insulation with 0.1" deep routered pockets

The back is hardboard painted with white oil based paint in hopes that it would be somewhat resistant to water splashing.  I used an oil based Rustoleum to paint the flat portion of the 1/8” hardboard.  It makes the hardboard somewhat resistant to moisture and provides some protection of the wiring.

Back view of corner details. Note the 4" shelf bracket wedged between wood and frame.

The frame is ¾” aluminum angle iron on top with ½” aluminum angle iron on the back—all 1/16” thick.  I then took some ¾” spacers and milled them down to 5/8” to make a perfect fit with 6-32 screws holding it all together.  Then the corners I used a 2”x4” ripped in half and made corners for the aluminum frames.  I chopped the aluminum using my carbide tipped radial arm saw.  Picture frame 45 degree cuts with a 4” x ¾” shelf angle to make some rigidity.

The front frame is ¾” stock, screwed to a shelf angle and wooden block.  The back frame is ½” stock and is held in place with 6-32 hardware to the hardboard on the back, and slides inside the ¾” front frame.  Again, this thing will shed water but is not watertight.  The cells need to be dry because the junctions are exposed.

Some thoughts on what it means to be “weatherproof.”  It is not possible to weather proof any box that has air passing inside to outside and back.  Have you ever heard of the “Dew Point” of air?  Humidity condenses into moisture (water vapor into water) at that temperature.  A cold can of beer will condense moisture because the beer temperature is below the dew point.  Have you ever seen a damp garage floor?  In the spring, the cement is cold.  Warm air rushes in and you have standing water in minutes.  The same thing happens inside a tube, a box, anything you would like sealed completely.  If outside air passes, it will condense when the weather is right.  Pressure changes will force an exchange of air through a toggle switch, a wire grommet, an unsealed surface to surface, or an enclosure screw.  Weatherproof is nearly impossible.  The box makes its own weather inside.  If you try to make it weather tight, you will seal that moisture inside with no way to escape.  Deal with this by making a path for water to exit using gravity as our friend. 

The solar cells are blanks.  The blue side is (-) and the gray back is (+).  Each one comes in at 0.59V with no load and about 0.55V with a small load and slightly less than 0.5V at full power.  At optimum load they deliver about 5.5A each!  16 in series (in a 4 x 4 pattern) makes about 9.5V in full sunlight.  My goal is to charge a deep cycle battery to 15V using a charge controller.  The eBay seller provided 0.1” tinned strapping material.  Two runs of tinned wire on each side were soldered to the cells.  I must have broken or fractured at least 6 in the process of soldering.  I found after the 6-8 cell that the blue side needs to be soldered first as the reverse side is soldered in three places.  The stress of the heat on the tinned wire and the dissimilar coefficient of expansion of the materials was enough to bend the cell and fracture it!  The soldering process took about 20-30 minutes each!  The first 16 cells I damaged about 7!  Ouch!  I can make 32 but only three spares left!

Final assembly I used some blue painter’s tape to hold the insulation to the glass.  Soldering the cells in series included a wire to short the cell left-to-right.  E.g., an external tinned wire soldered between the two halves of each cell.

1/2" frame attached to back sits inside 3/4" frame holding in the glass.

This was a pile of work.  I cannot tell you how many hours I have into it.  Include at least four trips to Menards!  I only have one completed!  I will post an update to this when I get the second one done and the solar charger circuit in place.  I can provide some specs on what to expect out of this in terms of performance.  I am expecting about 4A to 5A at charging voltages when in full sun.  We will see how close I can get to this expectation.

Budget so far:

Surplus Solar Cells and wire          $74

Glass (Glass Shop)                             56

Aluminum (Menards)                        37

Foam (Menards)                                 11

Hardboard (Menards)                       12

Rustoleum Paint (Menards)             11

4” angle brackets (Menards)             8

Brass wood screws (Menards)         11

Flux Pens (Kester)                            15

Misc Hardware                                    6

Total out-of-pocket cost              $241

The panel is about 85 Watts of power.  That works out to just under $3/watt and I’m counting my time at $ZERO!  This is an embarrassingly expensive education in solar power using Photo Voltaics!  But I’m having fun!  Some people fish for $300/lb fish.  I build!  Go figure!

Short Dipole for Portable Use

September 5th, 2011

By Ford-NØFP

I have had a short flat top dipole for portable use on my bucket list for some time.  Verticals are ubiquitous on Field Day.  A truly horizontal polarized antenna is very handy for rejecting near proximity transmitters operating with vertically polarized antennas on the same band.  Using 80 and 75 meters on Field Day is very nice.  Same with operating both ends of 20 and 40 with one station on CW and the other on SSB and even a third on Digital modes.  Having a pair of antennas crossed polarized provides upwards of 30dB of attenuation between the stations.  Experience has shown this to be effective in eliminating desense in even simple HF equipment.

A dipole is horizontally polarized when it is horizontal and parallel with the ground.  This provides some so-called “ground gain” when the direct ray combines with the wave reflected off the ground.  Modeling suggests as much as 6dB gain in some elevations.

Elevation Plot

This is the plot of a 20M dipole at ½ wave length above average ground.  Imagine looking down the end of the wire.  Imagine this plot with a direct ray shooting off the dipole at about 30 degrees.  Imagine another ray shooting off towards the left, reflecting off ground, and combining with the direct ray to provide over 6dB of gain.  There are also direct rays shooting off in all elevations, including up and down.  The down wave reflects off the ground and actually cancels the wave front going directly straight up.  This is how you end up with the classic dipole looking elevation pattern with a null at zenith.  Below  1/2wL in height you get a big bubble pointing straight up (a cloud burner).  Heights above 1/2wL you will see multiple elevation lobes developing as different ground reflected rays combine with direct rays at different elevations.  But I’m digressing from the topic.

A general rule for dipoles states: ‘higher is better.’  An inverted V does not provide the ground gain you see in a flat top as it contains both horizontal and vertical polarization.  The V shape does not allow the wave front to uniformly reflect off the ground (also true of verticals) to recombine and provide gain.  A height of 130 feet up is considered ½ wL on 80M.  Making a true flat top at that altitude is well beyond the capability of most Field Day sites.  Frankly, achieving 30 feet can be a challenge, as a full sized (130’) wire dipole will sag almost 10’ to 20’ in the middle.  This will produce a pattern roughly equivalent to a 6M beam mounted on a footstool only 11 inches off the ground!  But they do radiate, forming a near vertical incident radiator.  A cloud burner for certain.  On 80M this is still good for many hundreds of miles. 

Short dipoles can be useful radiators if made to be efficient.  Below about 8% of a wavelength, a short dipole becomes very difficult to load.  Making them take power is the name of the game.  The current in the wire (not the coil) is what radiates.  For this reason, top loaded antennas are considered the best radiator, center loading next, with base loading being the worse.  The total difference, as measured on a field strength meter, may be 2:1 in favor of Top verses base loading (e.g. 3dB).

Design goals:

  • Adaptable to any 80M thru 6 M band
  • Portable form factor
  • Flat top design
  • Auto-Tuner at the feed point

I found some surplus fiberglass poles originally used for pruning tools.  The pole quickly collapses down to about 7’.  I adapted the center using a 6” round plastic rod shaved down to the exact diameter of the fiberglass and attached it to a mounting plate from an old 2M beam that long since went to recycling.

Test Setup

The fiberglass poles were fitted with an end cap and 3/8 – 24 tapped hole to accommodate a 48” stainless steel whip recycled from a pair of CB antennas.

Element End

The LDG AT-100Pro tuner was mounted in a Tupperware-like container obtained from the local big-box outlet store.  The unit was inserted into a plastic bag first to hopefully repel water that may sneak into the loose fitting lid.  It also allows me to push buttons on the tuner’s front panel.

http://www.ldgelectronics.com/c/252/products/1/1/1

The whole thing was screwed to a small piece of plywood.

Water Resistant Bag

I wound a simple current balun using a pair of type 43 ferrite beads and RG58 class coax.  I’m not sure how effective (or even necessary) it is but 8 turns made it through those cores.  This sits between the coax back to the radio and the DC Bias-T (more on that later).

Input Balun

The output balun was wound on 4 type 43 beads using #12 stranded electrical house wire.  About 8 pairs of turns made it through those cores.  This sits between the tuner’s ANT1 port and the dipole wires, which are also #12 stranded wires.

Tuner Details

I wanted to run the 13.8V DC to the tuner using the coax back to the shack.  This requires the use of a so-called “Bias-T” arrangement.  It’s basically a RF choke with a large value high voltage capacitor as a DC block to keep DC out of the tuner.  The capacitor may be superfluous but without a schematic of the tuner it is difficult to know.  The cap is a safety precaution. 

The AT100Pro is rated at 13.8v @ 500mA DC.  I used a pair of 3300uH ferrite wound chokes rated at 500mA in parallel to form approximately 1A of current carrying capability.  I used a 470,000pF (0.47uF) 650v capacitor as a DC block to the tuner (Xc=0.1 ohm at 3.5MHz).  The chokes in parallel are 1650uH, which at 3.5MHz is about XL=36K of inductive reactance.  A 0.1uF and 1000pF are installed on the DC side to provide some bypass of RF.  It seems to work.

Tuner Enclosure

The coils were fabricated inside a 4” sewer pipe.  The optimum coil size turns out to be 3.5” diameter, 42 turns at 9 tpi for a total of 85.4uH.  Wrapping the coil with bubble wrap made a perfect snug fit inside the pipe.

Coil Details

The trimming of the coil was tedious in that I made it quite a bit bigger than needed.  As it turns out, my first attempt (see my other blog posting on the subject) came in at 106uH, which resonated the system at 3.3MHz.  Removing 1 or 2 turns at a time brought the resonant point up about 50KHz per turn.

Trimming the Coil

 The coils are identical.  The coil assembly attaches to a small piece of plastic with ¾” EMT clamps for quick release.  Stainless wing nuts eliminate the need for tools when swapping the 80M coil for the 40M version.

Coil Mount

Hoisting the system up and down using my handy motorized skyhook was a snap.  Any time you can put a motor on tedious motion is a good thing.  The boat winch on a tripod is handy for getting good antenna measurements.

Sky Hook

Through modeling techniques and www.nec2go.com I found that a 1000pF capacitor would resonate the antenna without the tuner at 50 ohms and a perfect SWR at 3.74 MHz.

Tuning Cap

Unfortunately, this decreased the usable bandwidth on 80M by a substantial amount.  The natural 2:1 SWR point was only +/- 9KHz for a total bandwidth of about 18KHz!  That’s a system Q of 207!  With the tuner and the added capacitor I could only extend the usable bandwidth to about 120KHz of the band.  Without the capacitor, the tuner was able to achieve 350KHz of bandwidth automatically.  I can now effectively QSY from 3.550 to about 3.900 and just let the autotuner make a match.  Nice!

With the 80M coils in-place I cannot obtain a match on 40M.  Modeling this antenna allowed me to anticipate the problem on 40M.  With the 80M coils mounted at this +/-   9.2’ point allows me to tune 350KHz of 80M-75M, all of 30M, 20M, 17M, 12M, 10M and 6M.  Removing the coil would make it a 32’ dipole with an autotuner.  This would also be an easy match for the tuner on all bands except 80M and 10M.  80M because it just too low an impedance, and 10M because it is exactly a full wavelength long.  Removing one of the whips (off-center fed Windom) or both (a 12’ dipole) is also an option.

The assembly collapses down into roughly 7 pieces.  The two coils, the two elements, two whips, and the tuner assembly.  I’m going to fabricate 40M coils too, just because I can!  The system can be hoisted from a single rope and remain perfectly horizontal.  So far my 75M signal reports are quite favorable.  So I’m pleased with my shorty dipole.

Winding Air Coils-Version 2

August 12th, 2011

By Ford-NØFP

As I indicated in my version 1 posting, my quest to make an air coil mandrel will likely take three attempts.  This is version 2.  It made me a good coil so I might run out of ambition to do a third and final version of the mandrel.  We shall see.

In this version, I wanted a much longer mandrel.  An available hank of 3” schedule 80-vinyl pipe (3.5” O.D.) was 13” long and my last available sample in my junk box.  This is also approaching the limits of my 1948 Atlas lathe bed.  15” or 16” would be about the maximum length.  Schedule 80 pipe is heavy.  The typical Menard’s PVC is schedule 40.  You would have to try cutting grooves but I think a length of schedule 40 this long would be difficult without a mandrel inside the mandrel. 

Version 2 Assembly

The basic assembly is a pipe with about 120 degrees removed from one side, and two small 1/8” aluminum stock to be used as spacers.  8 slots of 0.040” deep by 3/8” wide were cut along the length every 45 degrees.  12” Teflon sheet stock (1/32” thick) was cut into 0.375” strips—8 of them.  The strips placed into the slots before wrapping the wire allows me to place copious amounts of hot glue along the length without permanently gluing the wire to the mandrel.  The Teflon sheet was available at MSC and was not that expensive.  In my case, the junk box Teflon was a leftover from a previous high voltage capacitor for a small loop.

Version 2 Cutting

Turning the 9 turns per inch slot ACME thread was easy enough with the power feed of the lathe.  I took about 0.010” on each pass with the lathe turning very slowly.  Groove depth is about 0.040”.

Here is another picture like above but with the flash this time to freeze the frame.

 Version 2 Cutting w/flash to freeze frame

And then another picture cutting up close.

Version 2 cutting close-up

The end cap from Version 1 becomes an end support for the end post.  It also holds the mandrel assembly together on that end.  Notice the small hole drilled diagonally through the end cap to allow me to start the wire and wrap it around the end post support.

Version 2 cap end

The chuck end of the pipe was fabricated in Version 1, along with a suitable stainless band clamp to hold the assembly together.  3” diameter is the maximum size for this old lathe. 

Version 2 chuck end

Winding the wire was simple enough.  I used an office chair with the spool of #20 enamel wire draped over the arms of the chair.  I used the tool post, the auto feed to rest my hand, pinching the wire, and holding it tight as the lathe pulled the wire off the spool on onto the mandrel.  It took a few minutes at very slow speed to pull (over) 100 feet of wire on the mandrel.

Here are left and right pictures of the wire filled mandrel.  I inserted the Teflon strips and then wound the coil.  The strips can also be inserted after the wire as the slot is 0.375” and the strips were 0.350″ with plenty of clearance to the wire.  The slots were 0.040” deep and the strips 0.032”.  In hindsight, making the grooves 0.050” deep would have allowed more hot glue to flow behind the wire.  Using 0.040” also worked just fine.

Lathe setup left

Looking at the above from the other side.

Lathe setup right

Installing the hot glue was a bit of a chore.  My hand was sore after squeezing that trigger for 20 minutes.  Squirting copious amounts of glue into the wire took several sticks of glue.  I then let it cure for about an hour to return to room temperature.

As a first step in disassembly, I pulled the chuck fixture off the end by releasing the band clamp.  This provides a good picture of the spacers too…

Version 2 end removed

Pulling the spacers released the section of mandrel.  I had to work at it with my fingers from both ends to get it loose.  If there were 3/16” gap instead of 1/8,” it would have worked better.  Of course this would have required a 3/16” spacer, which I don’t have on-hand.

Spacers removed

Placing the end of the mandrel in a bench vice was the trick to getting the system apart.  By compressing the mandrel with the vice, the wires were free from the ACME screw guides and I could just lift off the assembly.  Working the Teflon strips with a long-nose pliers helped release the wire

Version 2 complete

Once released from the mandrel, I put a second thick layer of hot glue on all 8 seams.  This time I could let the goo squish through the wires without being concerned about excess amounts on the inside.  I took a picture down the end, which appears distorted but is actually an optical illusion.

View down center

The coil is somewhat fragile.  This absolutely needs to be much heavier wire to survive without an enclosure.  With the 4” sewer pipe enclosure, it is robust enough to survive bumps and rough handling. 

Version 2 came in with 101 turns, 11.25” long, 3.5” diameter #20 wire calculated using NEC2GO.EXE (www.nec2go.com)  was predicted to be 249uH.  Using my trusty AADE.COM L/C meter (www.aade.com) measures 246uH.  I am still quite surprised at how closely the NEC2GO Coil Designer when compared to actually building the coil.  Predicted Q is 520 @ 3.75MHz.  My goal is to cut this into two halves and trim until I get to 103uH.  Two of these become a center load on a shortened dipole for 80M-75M. 

 

Winding Air Coils–Version 1

August 9th, 2011

By Ford-NØFP

I’m working on a portable antenna design and am going to be center loading a short dipole for the low bands.  More on that later…  The long-and-the-short of it is that I intend to use a LDG AT100Pro Autotuner at the center of the dipole to find the match.  The specs indicate that it will make any reasonable match, defined as 6 to 1000 ohm loads. 

The adjustable short dipole (24-32 feet) will require about 103uH of loading at the center of each half element to satisfy the requirement throughout the 80M band.  About 23uH will do the trick on 40M.  No load will be needed on 30M through 6M as the dipole will be within range of the autotuner.

103uH coil is no small coil.  Doing it without introducing a low self-resonant frequency (SRF) will require some finesse.  High Q is always a feature but the tradeoff is a low SRF.  A quick look around for some wire wound coils leaves me in a funk when I look at the prices on new coil stock.  Center loading requires the use of two loads!  Finding suitable coil stock in surplus is a trick.  Finding TWO identical units of the right size is priceless.  The odds of finding surplus are about the same as being struck by lightning—TWICE!

Various Junque Box Coils

Pictured is the www.AADE.com L/C Meter IIB.  A 71uH roller inductor, hand wound 4.5uH, a 3.3uH and 5.8uH wound on fiberglass rods, and a 11.7uH wound on ceramic.  My target goal of 103uH is going to be HUGE by comparison to these.

The need to be able to make reasonable air wound coils has always been of interest and this weekend I decided to tackle the chore of making a mandrel for winding coils to order.

Design goals included: 1) a large diameter coil; 2) suitable for 20 awg and larger; 3) reusable; 4) fit inside 4” plastic sewer pipe; and 5) reliable and reproducible.

Winding on a Mandrel:

My first attempt at forming a large diameter coil was a limited success.  True to most projects, you have to build the first one to figure out how you should have done it.  By the time you get to version 3, it’s pretty much the way you like it.

Plastic sewer pipe has a 4” ID.  Using a length of 3” vinyl the schedule 80 variety has a 3.5” OD.  Perfect!  Cutting ACME style screw threads into the heavy pipe required an end cap with a centered hole to rest on the tool post of the lathe.  My small lathe had a chuck capable of an inside clamp (expand the chuck instead of clamping on the stock).  This allowed me to clamp the stock well enough to cut a 10 turns per inch groove.  Using the power feed on the lathe; I was able to take several passes and formed a 0.035” groove.

Cutting ACME threads

The above was a picture from version 2 but it gives you an idea of what I mean by ACME screw slots.  I used a cutoff tool on the lathe, which was 0.035″ wide.  Set the feed for 10 TPI and let her cut at about 0.010″ each pass.

Winding the wire was simple enough.  I could use the power feed on the lathe to pace my fingers as I feed #20 wire into the slowly turning mandrel.  I was able to make 37 turns on the length pulled from the junque box.  Although I know I will need a bigger coil, this version 1 will nudge me up the learning curve to understand how to make a more suitable mandrel.

The turns are to be held together using hot glue.  Knowing that the hot glue will also stick to the mandrel I needed a way to release the coil after hot gluing.  I tried a few layers of cellophane plastic wrap, which worked but became messy, as the hot glue would bind with the wrap.

Version 1 still on Mandrel

I cut the tube end-to-end using a handsaw in two vectors making the mandrel 4 pieces.  This allows me to collapse the mandrel after winding and hot gluing the wires together.  Four hacksaw blades became spacers to restore the circumference to original. 

Removing the end cap and pulling spacers

After picking away at the excess cellophane wrap and putting on a second layer of hot glue, the coil was complete.

Version 1 complete

The whole assembly fits nicely into some 4” sewer pipe.  End caps will protect the rather fragile wire assembly.

Version 1 coil sample fits nicely in 4" sewer pipe

Version 2 will include some improvements.  Slots along the length will allow for 3/8” wide strips of Teflon sheet material beneath the wire to act as a non-stick backer for the hot glue.  It will be longer to hold up to 110 turns of wire (roughly 250uH).  The end-to-end cuts made with a 1/8” cutter instead of a handsaw will provide more precision.  This will allow me to use 1/8” flat aluminum stock as spacers for the void, making it a perfect 3.5” diameter mandrel.

Version 1 came in with 37 turns, 4.08” long, 3.5” diameter #20 wire calculated using NEC2GO.EXE (http://www.nec2go.com/)  was predicted to be 74.1uH.  Using my trusty AADE.COM L/C meter (http://www.aade.com/) measures 74.6uH.  Pretty darn close in my book.

My 40M loading coil for the shortened dipole requires 23uH coils.  I’ll use this one by cutting it into two coils.  I’d hate to waste 35′ of perfectly good enamel wire!

 

 

 

 

Wavetek 2510A Problems

July 16th, 2011

By Ford-N0FP

The Wavetek 2510A is a 0.1MHz to 1100MHz signal generator with a calibrated output.  I’ve tested it against the Boonton Power Meter and found it to be pretty accurate.

Wavetek 2510A

The inspiration to tune up an old “American 5 Tube” radio resulted in damage to my Wavetek 2510A signal generator.  Silly me.  I was probing around inside the radio to inject some signal to see what the problem might be.  I forgot that the power supply on these old radios is a Kevorkian widow maker with one side of the chassis hot.  Probing on the loop antenna I inadvertently placed 120VAC on the signal generator output and fried the programmable attenuator.

Ouch!

The attenuator has a label with “0.25dB insertion loss at 1100MHz” penciled on the sticker.  Measuring the output stage prior to the device shows a maximum output of 13.25dBm according to the Boonton Power meter.  The level control on the generator can vary the output about 30dB through the front panel control.  This analog adjustment in combination with the 10dB steps provided by the DUT allows for precise handling of the output levels.

 

Final Amplifier Stage

The next logical step is to see if the Attenuator is being properly ‘keyed’ by the digital controls.  A quick check using the O Scope verifies that the controls are working as expected.

 

Attenuator section provides up to 130dB isolation in 10dB steps.

 

The attenuator is a step attenuator set to steps of 10, 20, 20, 20, 30, and 30 for a total of 130dB of insertion loss.

The technical reference manual shows the attenuator as a black box.  Disassembly is impossible due to the unit being soldered together as an assembly.  I guess I need to find a parts unit to harvest that assembly.  Off to eBay!

Balloons anyone?

May 29th, 2011

Everybody has a bucket list.  Mine is long.  On that list is a project to launch a weather balloon to some ridiculous height, take pictures, and return to earth with a payload of pictures.  A ham radio beacon will deliver GPS coordinates and other data, like temp, altitude, etc.  It’s a complex project.  The FAA gets nervous when you play in their airspace, so there are rules–loads of rules.

It turns out that Kelley-WØRK has a similar item on his bucket list.  Kelley contacted me last week to discuss the prospects of doing a balloon.  I was eager to discuss more.  We’ve been meeting on 3620 LSB to discuss the possibilities and the problems involved.  We now have a plan.  And we’d welcome the involvement of other like-minded hams in the neighborhood (and beyond for that matter) who want to get involved.  “Involved” may be just cheering from the peanut gallery.  Or dig in and get more than just your feet wet by assisting in the development of the balloon and/or the payload.  Certainly the opportunity to pitch in more than sweat when it comes to paying for Helium or latex balloons.

We are in the very beginning stages of planning.  We’d welcome your participation by becoming a member of the MinnDX@yahoogroups.com list server.  To participate, you need to send an email to: MinnDX-subscribe@yahoogroups.com

Once subscribed, you can participate in the chatter and keep informed as to the progress of the project.  If you have any interest, by all means, send that email to MinnDX-subscribe@yahoogroups.com and we’ll get you included in the fun.

73

Ford-NØFP