ad: L-HROutlet
ad: l-rl
ad: l-gcopper
ad: K3QNTad-1
ad: Subscribe
ad: L-rfparts

Login is required for additional detail.


Email: Login required to view

Subscriber Lookups: 60091

Licensed in 1972 as WN9JCG while I was a junior high school student in Joliet, Illinois. Then obtained WB9JCG in 1974 and N9AZ in 1977. During this time, I was very active on the H.F. bands together with several friends from high school until I entered college at Northern Illinois University in DeKalb.  My OM was W9EAC (SK), who became active as a high school student in Cudahy, Wisconsin during the late '40s.  My cousin Thor is KL7JJN.

QRT during the entire '80s decade. In the early '90s, I began rebuilding a station at my present location in Florida. I still prefer the low HF bands and mostly operate CW, with some PSK31, JT modes, and SSB.  Desktop transceivers include a SunSDR2-Pro, ANAN-200D, and Elecraft K3 for remote internet operation. Amplifiers are SPE-1K-FA and Alpha models 77Dx, 87A, 70V, and 70A.  AC Power for the station runs from dedicated, isolated-ground circuits and further isolated through a pair of 1 KVA balanced iso-transformers.

The Alpha 70V uses steam from a one-pint distilled water reservoir as the sole mechanism for cooling. A column of purified water separates the 4KV HV supply from the chassis. I maintain a resource list for current Alpha 70 owners. If you own one of these amps, contact me with your amp's serial number and I'll supply you with a copy of the resource list.

The reference oscillator for the ANAN and Sun transceivers comes from a customized Trimble Thunderbolt GPS-locked frequency standard.  A Frequency Electronics FE-5650A rubidium standard is used to lock the frequency of a HP 5385A counter and several pieces of test equipment.  Bench test equipment includes two spectrum analyzers: Agilent N1996A with tracking generator and HP 8560E with phase noise testing module, Sencore LC102 L/C analyzer, GenRad/IET 5000 LCR Meter, HP 8640B+8657A RF signal generators, Rigol DG1062Z arbitrary waveform generator, Array VNA2180 and N2PK vector network analyzers, Agilent 34461A and Fluke 8060A DMMs. Scopes include: Agilent MSO-X-3024A (digital), and Tektronix 2465B (analog).  On the station desktop, a Tektronix TM-500 series test set incorporates a dual-trace scope, function generator, and two ultra-low-distortion audio oscillators with distortion analyzer.

I am an attorney working in the telecommunications industry and business immigration. In addition to a law degree, I have earned MBA, BSEET and BSCS academic degrees in Illinois and Florida, and spent a portion of my legal studies in the U.K. at Trinity College, University of Cambridge.


N4CC and I have teamed-up to construct a remote internet station near Hilliard, Florida - approximately 30 miles NW of Jacksonville.   The towers are solid-leg, Pirod self-supporting with a linear taper from top to bottom.  Tower #1 is 140 ft. and supports: (1) a full-size M2 4L 40m OWA and; (2) a 9-element M2 6m Yagi at 155 ft. on a 48 ft. boom.  Tower #2 is 100 ft. and supports a SteppIR 4E for 20m-10m and a M2 30m 30M3 monobander at 110 ft.   Each tower is rated for 30 sq. ft. of wind loading. 

A surplus Sprint/Nextel telecom equipment shelter is placed adjacent to the towers.  The shelter was manufactured by Rohn for the cellular industry.  Standard equipment includes: dual HVAC units for cooling redundancy, coaxial bulkhead, 200-Amp AC panel, TVSS surge suppression, gen-set input, lighting, and security sensors.  From start to completion, the construction phase took ten weeks.  Equipment provisioning was complete two weeks later.

Prop-pitch rotators, reworked by K7NV turn both arrays and are controlled by M2 RC2800PX units.  As decribed below, the controllers were modified by us to better withstand the very lightning-intense environment of north Florida.  

The site stays connected 24/7 no matter how bad the weather.  Grounding and bonding conforms to Motorola R56 Standards and Guidelines for Communication Sites.  Ground rings are formed around each tower and the communications shelter.  In total, spproximately 50 ground rods are used: 15 rods are used in a hub/spoke arrangement at the base of each tower.  Four rods are 24 ft. in length and are placed at the tower piers, mid-way between the towers, and at the External Ground Bus (EGB) "common point."  The EGB brings all grounding rings together at the building bulkhead cable entrance and also bonds the utility power ground.  At points where Cadwelding wasn't convenient for bonding, we used silver-solder bars and an oxy-acetylene welding torch.  After measuring ground resistance with an AEMC 3711 earth resistance tester, we achieved numbers well below 10-ohms.  Getting anything close to that number is very difficult in the sandy soil of North Florida.  In fact, increasing the number of 8-ft rods did little to drive down earth resistance.  We achived nearly all the low-resistance effect with the 24 ft ground rods.  

Although the terrain in this part of Florida is mostly flat, the station is located to the east and south of the St. Mary's River, on the southeast edge of the Okefenokee Swamp.  The river valley creates a downward slope in an arc from ZL to central Asia.  Toward Oceana, JA, and deep Asia, this produces a low take-off angle just a few degrees above the horizon.  The computed HFTA analysis of the 40m bearing into JA is shown below.  


Predicted 40m performance at 140 ft. height, short-path into Japan on the 325 deg. bearing shown in green.  For comparison, the aqua-colored line plot shows predicted peformance at the same height over flat terrainThe lavender colored bars show arrival angle above the horizon as a percentage of time, with 3 degs. being the highest repetitive arrival angle from Japan into north Florida.

In the HFTA graph above, 40m morning long-path (LP) prediction into India.  The red, green, and blue line traces show predicted performance along a 10 degree LP arc between 205-215 degs.   The aqua-colored trace shows a hypothetical 4L Yagi at 140 ft. but over flat terrian.  Notice the statistical mode from Florida into India is only 1 degree above the horizon on 40m.  Height and stacking alone can't achieve super low angles on 40m; sloping terrain into the St. Mary's river is responsible for the low-angle field strength at 1-3 degrees above the horizon.  This makes the difference between marginal QSOs and half-hour S9 ragchews with VU stations on 40m during the early morning hours of the winter months.  These same S9 VU stations are inaudible on my backyard dipole.

Rohn telecom Shelter, surplus from Sprint/Nextel Cellular       200-amp AC panel, GenSet Input, TVSS suppression, Dual HVAC units


140 ft. Pirod tower ready for the crane. Base pier to the right.   100 ft. Pirod tower.  John Deere tractor with fork lift used during assembly.


     Tennadyne T14 30m-10m LPDA assembled.                                           Looking 42 ft. down the dual boom.


140 ft. (43 mH) tower set with M2 4L 40m and 6m monoband Yagis.    100 ft. (30 mH) tower with crane bringing up 30m-10m LPDA.  N4CC at the top.


LPDA set on 100 ft tower.                                                                N4CC removing hoist harness on 140 ft. tower.


       Tennadyne T14-HD truss detail showing custom boom-mast plate and Philystran cabling

Prop Pitch rotators on both towers.                                                    Completion: both towers shown from a distance


Inside the comm shelter: Alpha 9500, Elecraft K3 with RemoteRig, Alpha 8406 and M2 rotator controllers



Shortly after installtion, we installed a vacuum relay at the M2 controller output to isolate the switching MOSFET from lightning exposure.  But, we later realized that a Polyphaser MOV short to ground on the controller's return lead could result in the rotator turning out of control.  Prop-pitch rotators don't normally use limit switches unless the user customizes his own protection system.  Recently, we added a second vacuum relay to isolate the +48V supply lead.  Now, two SPST vacuum relays form a DPST "knife switch."  One pole isolates the supply, the other isolates the MOSFET drain lead.  You can see both relays in the photos below.   Due to cabinet space constraints, the relays are not installed adjacent to each other.

In the first photo you can see the controlling MOSFET, the device known to ground-short and also cause a prop-pitch rotator to turn out of control.   I have seen many references to MOSFET failure in these controllers.  Because the +48V supply side is always active to the prop-pitch coils, it's imperative to protect the rotator in the event of a source-to-drain short.  The MOSFET gate is normally pulsed with a PWM circuit.  The pulse duty-cycle is what determines rotator speed.  After modification, if the MOSFET shorts, the relays will still activate/deactivate -- and certainly hot-switch, but the rotator will still stop at its target bearing.  The MOSFET is now mostly vulnerable only during a turn in a thunderstorm.  But even with the inclusion of the vacuum relays, "anything goes" with a direct strike.  For nearby induced surges, the relays should adequately protect the controller, at least better than the stock design that offers almost no protection at all.  With the remote site being 30 miles to the NW of my control point, the lightningmaps.org website has been indispensible for letting me know when strikes are too close to turn the arrays.

In the controller's stock configuration, a thin Bergquist Sil-PadÒ insulator is used between the MOSFET and chassis.  You can see the Sil-Pad is now replaced with an AAVID Thermaloy aluminum oxide ceramic insulator that offers greater breakdown potential between the MOSFET and chassis.  In the first photo, a new terminal strip can be seen with a 2N7000 switching FET.  The small FET drives the relay coils and was only going to be used for brief testing.  However, according to the Fairchild data sheet, it will handle 4X more switching current than the present 100 mA draw.  So, it stays for now.

Inside the M2 RC2800PX Rotator Controllers (the disorganized wiring belongs to the manufacturer).



In Progress:  Homebrew 8877 amp using dual S-QSK boards, built on an Alpha 77D chassis.

Key features:

- 100% microprocessor controlled, programmed in C++.  Fault control/status of Ip, Ig, HV supply, and QSK timing;

- External HV power supply;

- Back-lighted 250 degree Western-Electric style panel meters with PEP/AVG for the wattmeter using my PEP amp board shown further down this page;

- Vacuum variable caps on both plate tune and load controls;

- Mil-Spec panel components from Raytheon, NKK, Dialight, Electroswitch, Hoyt;

- Complete enclosure anodized and engraved;

- More detail to come...



Control switch printed circuit board shown above that connects to the primary S-QSK microcontroller board.  The yellow-colored pot at the circuit board's edge adjusts hue between red and green to create a yellow switch button from an RGB LED. 


The photos below show a custom DDS VFO under construction for use with older vacuum tube transmitters.   It covers 160m through 10m on non-WARC bands.  The electrical design is based on the excellent N4YG DDS Board with modifications written by W9BHI.  I designed an output amplifier board in DipTrace and it provides both Z matching and 1:4 level amplification from a single transistor stage with toroidal output transformer.  RF in/out connectors are SMA type.   Once completed, I'll measure phase noise with a HP 8560E spectrum analyzer and post results here.

The extruded aluminum cabinet was purchased from Buckeye Shapeform.  The end panels were designed in Front Panel Express software and CNC-milled.  A sub-chassis was designed to slide into the bottom retaining fins of the enclosure.  The software allows the designer to include automated installation of PEM nuts, threaded spacers, and threaded mounting studs.  With adaquate planning, I was able to create a cabinet design that required no hand machining; all mill work was done in the computerized CNC machining process.   The VFO knob is weighted on an optical encoder and was purchased from 73cnc.com.

The DDS VFO sub-chassis slides into the Buckeye enclosure.



My latest projects are focused on the design of several station switching utilities, all controlled by an Arduino Nano or PIC microcontroller. Through a choice of microcontroller programs, one Sequential QSK (S-QSK) Board can manage many different types of switching applications. Presently, three programs have been written for the S-QSK Board and are described below.

In the photo above, the Nano and Uno boards are compared. They are functionally equivalent and both are programmed in the C++ language. A powerful attribute of the Arduino boards is that they come pre-burned with a bootloader, allowing the user to upload new code to it without the use of an external hardware programmer that is commonly needed for PIC and EEPROM chips. Just grab a mini USB cable and you're ready to program. The user can also bypass the bootloader and program the microcontroller through the ICSP (In-Circuit Serial Programming) header.

I use the larger Uno for quick code testing; the tiny Nano microcontroller board is used in the S-QSK Board. The microcontroller and peripheral circuits are powered by the USB port during testing, and an external +7 to +12V supply during normal operation. After writing code to perform several different station switching operations, it became apparent to me that one board could be used for many switching and timing projects, using the Nano to perform all of the complex switching and timing operations.

Sample code, line-by-line documentation, and structured flowcharts are provided to assist the novice programmer with code customization. Structured logic is used for easy code modification. Here's a complete parts list for the Main and Remote S-QSK boards. More information on the S-QSK Project is available in the 2014 ARRL Handbook.




S-QSK Main Board. To the left, a BNC connector is used for RF sensing.  To the right, a CAT5/RJ-45 cable is connected to an optional remote PhotoMOS relay board


S-QSK Remote RF Sensor shown above

The S-QSK Board contains the following feature set:

- A common hardware platform used to control many R.F. switching applications where precise timing is required between switching events. The hardware remains the same; only the source code changes for different applications;
- Nano or PIC microcontroller plugs into the S-QSK motherboard;
- Programming a Nano requires no special hardware to burn the code into the microcontroller. Just grab a USB cable and upload like a photo from your camera;
- For advanced programmers, a 16F88 PIC chip can be used in place of the Nano;
- Uses screw-down Euroblock I/O connectors. Each I/O connector can be unplugged from the motherboard for easy disassembly and servicing;
- Four digital input channels; eight digital output channels. Header for access to optional analog channels;
- Optically isolated I/O for maximum RFI immunity. Photo-Darlington transistors on the input and solid-state PhotoMOS relays on the output;
- Each input channel can be selected for dry contact closure, a solid-state open collector, or any other solid-state switching device;
- Each output channel can be selected to float or reference circuit ground. Each output can be jumper-selected to function as a current sink or current source;
- RF sample with choice of BNC connector or 2-pin header. RF is converted to a DC level and conditioned into a photo-coupler. Sensor activates with less than 100mW of RF power;
- Remote RF Sensor board can be used to sample RF at a distant location or where a "T'eed" RF connection presents an unacceptable line mismatch;
- On-board LED diagnostic status indicators to show RF presence, and logic state of all output lines;
- Uses a +12V muRata DC-DC converter, bootstrapped to the +12V supply to provide +24V for optional vacuum relays;
- Two optional relay coil accelerators for driving external frame relays;
- RJ-45 connector routes four digital output channels to an optional remote photoMOS relay board;
- Board can be populated with only the circuits of interest, thereby saving on construction cost and assembly time;
- Precise control and delay of all sequencing steps in 1 ms. timing increments;
- Significant fault protection built into the code (e.g., hot-switch protection). Before anything can switch in between steps, RF is first sampled and judged with the state of the input key line;
- PIN diode T/R daughterboard under development;
- Small board size. S-QSK measures 5.3" x 3.5". The Remote RF Sensor is 1.7" x 1.6".


For the QSK Control Box project, the S-QSK Board is used as the foundation of a fast and silent QSK T/R system for quickly switching classic transmitter and receiver "separates." Initial tests look promising with QSK switching speeds that approach near full-duplex operation. A switching limitation of modern transceivers is that a common oscillator is used between Rx and Tx modes. For several reasons, stabilization of the oscillator is not instantaneous with the switching transition. Since separate VFOs are normally used with a split receiver-transmitter combination, it's possible to let the VFOs free-run without Tx offset interruption and perform the T/R switching function external to the transmitter and receiver.

Hot-switch protection is accomplished by sampling the presence of the complete RF envelope. Detection of lead and tail timing errors inhibit T/R switching and activate fault LED indicators.

Of interest to CW operators is an auto keyed element length retention algorithm that adds back the amount of lead-in delay to the RF envelope. Generally, it's the lead-in delay that otherwise results in so called "dit shortening" in modern transceivers.

With high isolation PIN diode RF switching, there's no need for a CW sidetone line; the receiver is tuned to the frequency of the transmitter. For testing, I am using a modified Ameritron QSK-5 PIN diode RF switch with the S-QSK Board. The microcontroller accepts an input from any electronic keyer, straight key or semi-automatic bug. Through software, lead-in and tail times are independently adjustable in 1 ms increments. The only optional modification required to the transmitter (for SSB mode) is a simple control of the bias line. At a later date, I'll supply intormation on a matching PIN diode T/R board presently under development.

The C++ code can be found here: QSK Control Box along with the system timing chart and the structured  logic flowchart. Simply copy and paste into the Arduino software's edit window, then upload to the Nano board with a mini USB cable.


Need a way to time your homebrew amplifier's RF in/out relays and bias control?

I have developed another Arduino program to precisely sequence all critical timing elements between an RF power amplifier and transmitter/transceiver. The S-QSK Board will work with QSK and non-QSK amplifiers. A sequenced electronic bias switching (EBS) system is created and supports both two-state and three-state bias. Switching between types only requires a simple change to the microcontroller code.

The delay time between events is independently adjustable to accommodate various transmitter and amplifier timing characteristics. The input key and RF sensing lines are polled in a loop. Depending on the line states and the state of a flag bit, the S-QSK Board's output key line, solid-state relays, and bias control are switched with time sequencing to avoid "hot-switch" effects. The board offers hot-switch RF protection by sampling the presence of the complete RF envelope. If RF excitation is present at the input to the S-QSK Board before the input key line is active, a switch from Rx to Tx is inhibited. Likewise, if the RF envelope has not decayed to zero after transmission, the S-QSK Board will not switch back to Rx. If either type of timing fault occurs, one of two LEDs will illuminate, showing a fault. The LEDs remain lighted until the timing fault clears. In the event of brief timing faults, each LED is pulsed to remain on for 0.5 second. Upon detection of a post-switch fault, bias lines are deactivated, providing further amplifier protection. The RF input sample is capacitively coupled from the transceiver's RF output to a photo-transistor. The microcontroller's digital sampling input is optically-isolated from RF.

For 3-stage EBS systems, "hang" bias is supported and is adjustable by the user from 0 ms to 255 ms in 1 ms increments.

The C++ code can be accessed here as a text file together with a structured logic flowchart of the timing sequence:

In addition to establishing precise timing between amplifier switching elements, the S-QSK board can accelerate the relay activation time of frame-type output and input relays. The board contains two optional relay coil accelerators, each powered by a choice of +12V or +24V power buses. The accelerator circuits are engaged with header jumpers. For mechanical T/R switching, Jennings RJ-1, Kilovac HC-1, and Gigavac GH-1 are all good choices for the RF output relay when they are powered at their rated supply voltage. For the input relay, consider the Aromat/Matsushita RSD-12V.

The RF input sample is capacitively coupled from the Input RF line to a photo-transistor. RF sampling is optically isolated from the microcontroller digital inputs. A remote RF sampling board (also shown above) can also be used in instances where a "T'eed" RF connector presents an unacceptable line mismatch or where the sample point is at a long distance from the main S-QSK Board.The sample line will sense at less than 100 mW of power.


Here's the C++ code to emulate KD9SV's "Front End Saver," a device first described in the February, 1997 issue of CQ Magazine. The microcontroller code provides better sequenced timing between switching events as well as hot-switch protection. Again, the S-QSK board is used. The Front End Saver minimizes the possibility of damaging your receiver's front end when using an auxiliary receive antenna (e.g., Beverage, flag, or loop). The circuit is designed to: (1) disconnect the receive antenna during transmit; (2) deactivate a receive antenna preamp; (3) ground the receiver input port when transmitting; and (4) key an RF amplifier -- all with sequenced timing and RF hot-switch fault protection. The C++ microcontroller code can be found here. At a later date, I will post a logic flow chart.


The ANAN 200D and 7000DLE transceivers use a clever adaptive pre-distortion algorithm when controlled by either PowerSDR mRx or Thetis desktop software.  The algorithm is named "PureSignal" and was developed by Warren Pratt, by NR0V.  PureSignal minimizes transmitted IMD by applying an opposite amplitude and phase non-linearity to the RF waveform.  In the ANAN 200D, three relays are used during the T/R transition (K5, K19, K36) and they are quite loud when sending CW.  K19 and K36 route RF for PureSignal linearity correction.   Since a CW transmission doesn't require optimized RF linearity, there's no reason why K19 and K36 need to engage in CW mode.  Unfortunately, K19 is hard-wired in parallel with K5, the T/R relay.  As such, there's no easy means to separate the two relay lines on the microscopic multi-layered PC board.  Adding to the dilemma, my S-QSK board discussed above is too large for the ANAN cabinet.  A new controller board was designed for this application and fits within cabinet space constraints.  

The circuit shown below monitors input key line activity from the ANAN's CW paddle "dit" and "dah" lines.  When one or both lines are activated during a CW transmission, an Arduino Nano microcontroller prevents the ANAN transceiver's PureSignal relays from activating.  Two solid-state photoMOS relays on the new circuit board (K1.1 and K1.2) are used in series with the ANAN PureSignal relay coils.  The photoMOS relays are pulsed to remain active during CW operation at any CW speed -- then deactivate after a period of key line inactivity.  The result is greatly diminished relay noise as only the T/R relay (K5) remains active in CW mode.  K5 was changed to an Aromat/Matsushita RSD-12V RF reed relay and is noiseless in operation.  The C++ microcontroller code is here.

Schematic of the ANAN 200D PureSignal Relay Controller

Above: DipTrace CAD design image and the completed circuit board (in middle) installed in the ANAN 200D transceiver. 




I had been looking for a Narda 0-99 dB RF decade attenuator to use when making RF measurements, but these units are scarce on the used market.  It looked like one of the more common programmable step attenuators could be used together with a binary-switched interface.  I settled on a Weinschel 3200 model with 0-127 dB of attenuation in 1.0 dB steps.  The price was reasonable and has specs very close to the Narda model although it's limited to 3 GHz.  That's well within my measurement needs. 

Shown below is a photo of the completed board.  Next to it is a PCB design image of a motherboard I created in DipTrace.  The Weinschel attenuator mounts on the board with four #4 machine screws.  A header socket forms the connection between the attenuator and the toggle switch matrix. 

The arrangement is not as elegant and easy to use as a decade attenuator since some mental math is needed to keep the binary switch values straight.  It also requires a +12V power source.  I actually started on a decade interface design that uses a 2-1/2 digit thumbwheel switch with 4-bit Gray code and an Arduino microcontroller.   In the end, I elected to stick with simplicity but may revisit this option at a later date.   





The Shure SM-7B is a dynamic microhone that's been popular in the broadcast and recording industries since the early 1970s.  While the mic has some nice audio attributes, its output level is very low -- too low to drive most common transceivers.  To overcome the level disparity, I designed a high performance mic prepamp based on two key components: (1) A Jensen JT-115K-E transformer with a 1:10 turns ratio; and (2) an ultra low-noise, low current Analog Devices ADA4841-1 op-amp.  When the transformer secondary is terminated into its design load impedance of 150K-ohm at the op-amp non-inverting input, the JT-115K-E transformer is capbable of adding 20 dB of voltage gain with only a 1.5 dB increase in the noise figure.  The ADA4148-1 can provide additional gain and has an input noise voltage of only 2.1 nV/√Hz.  The quiescent zero-signal current is just over 1mA. 

Since the preamp will primarily be used during portable remote operation, powering comes from two 9V "transistor radio" batteries.  As the op-amp's maximum input voltage rating is +/- 6V, a 78L05/79L05 regulator pair limits the supply rails to +/- 5V.  Photos and schematic are shown below.




I recently designed a PC board that includes footprints for three types of I.F. filters: (1) the FA style Collins mechanical type as used in the S/Line; (2) the PS tortional mechanical filter currently being manufactured by Rockwell/Collins (UPDATE: Rockwell ceased production and sales of their mechanical filter product line on January 1, 2016); and (3) for those who want to experiment with inexpensive ceramic filters, the board accommodates either the muRata CFR455H or the identical NTK LF-D6. The purpose in designing this board was to let users add sharp skirt selectivity to various communications receivers that utilize a 455 kHz I.F. stage.

One board accomodates all three filter types and includes pads for all necessary input/output Z trimming. The builder simply populates the components on the board for the particular filter type. The board size is approximately 3.6" x 1.0" and will fit nicely into many VT and solid-state receivers and transceivers. As can be seen in one of the photos, gold-plated sockets allow the user to easily change between filters if different bandwidths are desired. Also seen at the lower-left edges of the photos are a pair of abandoned double-tuned I.F. transformers that are now replaced by the Collins filter board.

I have completed design of a dual filter board that uses miniature Teledyne RF relays to perform the filter switching function. The relays, in association with attention paid to board layout, offer excellent skirt selelectivity and high Input/Output isolation.


Dual filter version using miniature Teledyne 411 RF switching relays



After acquiring Drake SW-4 and 4A receivers, I wanted to experiment with a low-distortion AM detector, but not go through the hassle of using an outboard synchronous detector. In reviewing on-line literature, K2CU has developed an interesting low-distortion, full-wave circuit that seems ideal for use in older receivers where the I.F. is 455 kHz and lower.
I took K2CU's schematic and converted it into a PCB as shown below. There are a few changes to K2CU's design, mostly in power rail conditioning, power supply decoupling, polarity protection, grounding options, and the addition of an optional LDO voltage regulator. In the image below, the board looks much larger than its actual size of just 2.5" x 2.0". An SMD layout could have been used to make the board even smaller, but that would preclude performance experimentation with different op-amp types.
In some receivers where the AGC sense is post detector, it may be better to keep the existing square-law diode detector while routing audio from the new detector to the AF stage. I'll have more info here on my QRZ.com page when the boards have been fully tested and ready.

For the last two years, I've been using the excellent SM0VPO audio amplifier design as a replacement in some Drake receivers. It's class AB biased using a string of diodes. In a class AB amplifier, the transistors are biased in such a way so as to never fully turn off. A small amount of precise conduction occurs by biasing three 1N4148 diodes that allow a small amount of collector current to flow even when no signal is present. This means that the two output transistors will be on for more than a half-cycle of the waveform but much less than a full cycle, giving a conduction angle between 180 and 360-degrees - or 50 to 100% of the input signal depending upon the amount of additional biasing used. The amount of diode biasing voltage present at the base terminal of the driver transistors can be increased in multiples by adding additional diodes in series. Most often, two series diodes are used in the AB transistor designs. However, at quiescent levels, the push-pull output transistors still run cold to the touch when using three diodes and the elevated bias accounts for complete elimination of crossover distortion and its super low THD performance. The performance is extraordinary for a +15V, single supply amplifier.

A few upgrades to the original design:

- An optional on-board, low-noise 2SK170 or LSK170 FET source follower can be selected from a panel switch. This is useful when tapping audio from a Hi-Z source, like a 500K volume control potentiometer;

- Isolated bridge rectifier accepts external power from either AC or DC sources;

- A Zobel network was added to increase stability when driving reactive loads -- and when headphones are (un)plugged between speaker and headphones;

- On-board LDO voltage regulator with additional power supply filtering and decoupling;

- Optional Jensen input transformer with jumper header option for non-polarized coupling from audio sources with a DC clamp or offset;

- Audio input selection is made from a 3-positon rotary switch on the front panel. Position 1 selects an unbalanced input; Position 2 selects the same source but through the 2SK170 Hi-Z FET buffer; and Position 3 selects a balanced input through a Jensen line input transformer. Connection from a balanced audio source is on a 3-pin Euroblock connector;

- "No wires" design. All components that are normally hard-wired on a panel (e.g., switches, potentiometer, jacks) are mounted on the PCB. There's no hand wiring of any controls;

- Noiseless Bourns conductive plastic volume control -- the same type used in many professional audio mixers;

- Circuit board slides into the retaining fins of a Hammond 1455NF extruded aluminum enclosure.

- No "Pin 1" issues as a single point ground connection is made only at the audio input. By using an AC-output wall transformer, DC is generated on-board and no ground current flows back to the power source. This results in zero noise current flow on the signal ground plane. A dedicated ground plane on the back side of the board is used for the DC current return of the output transistors and speaker/headphones. This keeps all heavy DC current away from the high-gain stages that could otherwise modulate the signal ground buss. It's difficult to attain compliant grounding when using on-board connectors (rather than chassis), together with the use of common extruded enclosures. In this case: (1) the speaker/phone connector and power input share no common ground reference; and (2) only the audio input is connected to a grounded source. Both factors contribute to a low-noise design;

Go ahead, plug in your most sensitive audiophile headphones. There's nothing there but an ultra-low distortion output that's completely free from hiss and hum/buzz. Measured THD at 1 kHz/1-watt is less than 0.005%. The unusually low noise level is attributed to careful ground path design and the use of Toshiba BC550/BC560 low-noise-figure transistors in the high gain front-end stage. Although the output is not DC-coupled (a limitation of a single power supply), it has a whopping 4500uF/50V Nichicon MUSE audio output coupling capacitor that will easily drive speaker loads down to just a couple ohms owing to the circuit's high damping factor. While the amplifier satisfies the most demanding headphone requirements, its 4-watts of clean power will blow you away when listening to a moving coil speaker. This circuit is an excellent alternative to the LM386, LM383, and TDA2002/3 series amplifiers that are prone to high levels of residual hiss. When listening to music at the workbench, I use this circuit as a power amp for my Apple iPod and iPhone.




In this design, a zero-crossing, solid-state relay (SSR) is used to take a high current load off an AC power switch. Designed primarily for older vacuum tube receivers and transmitters, it's added right after the line fuse and will switch up to 2A of current at 120VAC. The board is approximately a 1.7" square and after installation, the power switch passes only a few mA of current. Installation in the equipment is easily reversed if desired at a later time. Unswitched Neutral connectivity is preserved and it is not ground-referenced so it's possible to use in U.S. and international 240V systems with an additional line fuse. The two AC power connectors are 3-pin Molex types, but damage cannot occur if accidentally reversed. In the photo below, a Solid-State AC Switch board is installed in a Hallicrafters receiver. In this installation, I made direct connections to the board although the board supports use of Molex connectors rated for the switching voltage.



The circuit board shown below is a drop-in PEP amplifier for the Drake W-4 wattmeter that can be switched to display average or peak power. The existing meter wires are moved to the barrier terminal strip and the board mounts directly to the meter's 0.25" terminal bolts. The circuit is powered from either a +9V battery or +12V from a standard 2.1mm coaxial DC plug with on-board +5V LDO regulator. The DC power jack is deliberately mounted upside-down to allow easy connector access when the directional coupler is mounted inside the case. The PEP/AVG function can be controlled from the on-board switch -- or a switch may be brought out to the front panel. A low DC offset dual op-amp is used that does not require an offset trim adjustment for meter zero. The input op-amp stage isolates the directional coupler and provides a low-Z source into a diode peak detector. The second stage is a unity-gain source-follower and provides an accurate source Z to drive the moving-coil meter movement. Ample RFI filtering is used on the power and input meter leads. Full-scale trim and peak decay rate are adjustable with 10-turn Cermet potentiometers.

PEP Amplifier Board Attributes:

- Uses a LM358 single-supply op-amp as a peak voltage detector. It has exceptionally low DC offset characteristics in this application -- No DC offset trim needed;
- Easily reversed at a later date with no "modification residue" left behind;
- Excellent scale tracking at all power levels;
- Uses a standard 2.1mm coaxial DC power jack for connection to a +12V source;
- On-board 78L05 5V low drop-out regulator. Circuit draws about 5mA of current;
- Can be powered from a 9V battery;
- Board fits nicely in the W-4 case with easy access to all controls and terminal screws -- even with the directional coupler installed;
- "No wires" design; there's no point-to-point wiring. All parts mount on the board and the build can be completed in less than 30 minutes;
- Board mounts on the back of the Drake meter's 1/4" terminal bolts. No other mounting screws are used. The existing meter wires are simply moved from the meter lugs to a two-screw terminal barrier strip on the new board;
- Adjustable full-scale meter trim with 10-turn precision potentiometer. Trim tracking is very smooth;
- Adjustable peak decay rate with 10-turn precision potentiometer;
- Switch for PEP/Average mounts on the board -- or a small DPDT switch can be added to the case. With the W-4 having an open back, it's easy to just feel for the switch lever as it can be blindly identified with ease. However, once you try the PEP mode, you'll want to keep it there;
- RFI mitigation on the DC supply and input metering;
- Responds faster than a Bird model 43P PEP meter amplifier.



Similar to the Drake PEP board, this circuit is designed for use with the Bird Thruline series wattmeters and will operate with sensitive meter movements below 30 uA for full-scale deflection. In addition to the features of the board shown above, it uses a temperature-compensated, ultra-low DC offset op-amp at U2 but without the need for a servo-controlled stage to null a DC offset.



Here's a link to an article I wrote for QST titled "An SWR Null Meter," QST, Feb, 2010, pp. 30-32.


Shown below are two versions of a "no wires" keying interface that accept simultaneous inputs from either logic-high or low sources. For example, a computer LPT or serial port can connect to J4, while a key, keyer, or transistor switch can be connected at J5. A choice of RCA connectors or headers is provided for I/O lines. The logic-high input is double-isolated: an opto-transistor and photovoltaic isolator buffers this input from the external power supply and swiched load.

Typical applications include: keyer to transmitter interface; computer to transmitter interface; transmitter to amplifier interface; anti-shock key-to-rig interface for older vacuum tube gear; and as a power relay for station accessories. Device powering is provided by a standard 2.1mm coaxial DC jack, or a two-pin header, although the board will support a direct input at the logic-high input without the need for external power.

Output switching is ground-independent and rated at >500V AC or DC @ 2.5 A of current. The boards are sized to slide into the retaining fins of a Hammond model 1455C801BK extruded aluminum enclosure. Or, they can be integrated into existing equipment by using the header connectors and mounting holes.  The enclosure end panels were designed using Front Panel Express software.  They are blue-anodized, engraved and filled.

Shown on the left, an optically-coupled CPC1978 solid-state relay (SSR) from IXYS is used to perform output switching. On the right, a pair of IRF-820 power MOSFETs perform the same switching function.




Schematic and circuit board image shown below of an over-voltage protection circuit compatible with most Astron linear-series power supplies:


After measuring the SPE amp's T/R time, it is not QSK compatible, although SPE appears to have addressed the problem with newer production amplifiers. A modified Ameritron QSK-5 PIN diode switching unit now controls the amp's T/R and MOSFET bias functions. Control logic between the SPE and QSK-5 is optically isolated. Switch time was reduced from 16ms to under 4ms and hot-switching is eliminated. A schematic of the QSK-5 and SPE bias modification is shown here:





Three versions of the Astatic D104 FET Transformer Circuit appear above (with schematic below) and all are designed to work with a 9V battery -- or any transceiver where +8V is available at the mic connector. The circuit board to the left measures 1.75" x 0.70" and is slightly smaller than the original 2-transistor board. The two extended mounting tabs make it a direct drop-in replacement into the D104 microphone. The board in the middle is a SMD/SMT version that also uses Astatic mounting tabs. The board to the right uses surface-mount (SMD/SMT) components and is about the size of a postage stamp. This is a universal FET Transformer Circuit that can be dropped into any crystal microphone base including those made by Astatic, Turner, and Shure. In all three board versions, a ground plane pour is used to assist with RFI mitigation.

Like all crystal mic elements, the D104 performs best when the load impedance is significantly higher than its source impedance - especailly when the load impedance is mostly resistive. A typical crystal cartridge's source impedance is highly dependant on its internal capacitive reactance with equivalent series capacitance values being in the range of 1000pF - 2400pF. By lightly loading the crystal element with the extremely high input impedance of the JFET's gate, the D104 can attain reasonably good low-frequency response to approximately 80 Hz -- approaching the quality of a dynamic broadcast microphone.

If the stock D104 base contains a 2-transistor "power mic" circuit, it should be completely removed with substitution of the FET transformer circuit. In computing the input impedance of the old "power mic" pre-amp, circuit analysis shows approximately 470K -- still too low for the attributes of the crystal cartridge if one desires a low-end response down to 100 Hz.

The SMD boards were designed after trying to source J201 and MPF-102 FETs in the traditional thru-hole style case. The major semiconductor manufacturers who once mass-produced these transistors in TO-92 style cases have now "obsoleted" these two components. In fact, even the SMD versions are now in danger of going obsolete. Presently, only InterFET Corporation and Linear Systems appear to be manufacturing both through-hole and SMD versions of high-performance JFETs.

Be careful when sourcing any semiconductor. The components stated here are widely available on the Asian market but when making a purchase through many on-line retailers, it's not possible to trace the supply chain of the component to its origin. For example, some Toshiba low-noise bipolar transistors are in fact re-labled 2N3904 devices. Unless the purchaser has access to a curve tracer and a means to perform critical noise and frequency response testing, one is left to trust the seller for a clean "chain of custody" between the time of manufacture and point of sale. Because of this, I only purchase semiconductors direct from the OEM or through well-established distributors including Newark, Mouser, Digi-Key, Arrow, AvNet, and Allied. For an interesting perspective concerning the severity of the counterfeit semiconductor market, see the following report issued by the Semiconductor Industry Association (SIA).

The FET Transformer is configured as a classic source-follower (common drain amplifier) where gain is less than unity. It uses either a J310, MPF-102, or equivalent n-channel JFET transistor. Note that the JFET as a source-follower is self-biased and does not require a "leak" resistor (R3) between the gate and circuit ground.   Due to the extremely high input Z of the JFET, there's ample leakage current though the cartridge to stabilize bias.   However, notice that R3 is shown as an option in the diagram below. R3 can be set to 10 meg in low humidity areas to prevent JFET gate damage from static electricity. Moreover, R3 can be reduced in vaue to deliberately form a simple 6 dB/octave high-pass filter. An R3 value of 100K will result in a -3dB turnover point close to 250 Hz.

Self-bias stabilizes the quiesecent operating point against changes in JFET parameters (e.g.,Idss, gfs, etc.). Here's the idea: Suppose we substitute a JFET with a forward transconductance (gm or gfs) value that's twice as large. Then, the drain current will try to double. But since the drain current flows through R(s), the gate-source voltage V(gs) becomes more negative and reduces the original increase in drain current. A gate voltage equal to 1/4 of V(gs-off) results in drain current equal to approximately (Idss/2). To set-up the mid-point bias "Q point," R(s) is computed simply by taking the reciprocal of the JFET's forward transconductance value. This value creates a drain current equal to (Idss/2).

With some difficulty, transconductance can be measured. In the alternative, simply research the range of gm or gfs values from the manufacturer's data sheet and compute the geometric mean. For the Motorola MPF-102, the computed R(s) value is approximately 270 ohms. This becomes the driving source impedance to the transceiver. And so, the input impedance of the FET transformer circuit is greater than 10-megohm while the output impedance is less than 300 ohms. Within limits, R1 can change in value based on the JFET's gfs value. As the value is lowered, one benefit is the source Z decreases, but at the expense of greater current and a shift of the Q point along the FET Transformer's bias curve. Keep this in mind when powering the circuit from a 9V battery.

Output coupling capacitor C4 is 4.7 uF and is adaquately large enough to pass a low-end response below 100 Hz as the typical mic input impedance of most modern transceivers is > 1K-ohm. Click here for a display of the LTSPICE simulator output.

The FET Transformer Circuit is an excellent choice when using a crystal mic element on a long mic cable run into the grid of an audio triode, for example. Many owners of older vacuum tube-based transmitters incorrectly assume that the extremely high input impedance of the control grid is enough to offset the crystal element's Hi-Z source impedance. However, a recipe for disaster is set up any time a high parallel capacitance exists across a Hi-Z source impedance. Since the source Z of the crystal element is mostly a capacitive reactance, the cable capacitance creates a uniform voltage divider across all audio frequencies. To mitigate the effect of distributed cable capacitance from the D104's Hi-Z mic element, the FET Transformer Circuit will convert the Hi-Z mic output to a Lo-Z source at a point where it matters most: the input end of the mic cable. In doing so, the low output impedance from the FET Transformer is set up by R(s) and swamps out the detrimental attenuation caused by a combination of cable capacitance in parallel with a series capacitive source. A low driving source Z also minimizes noise pick-up on the mic line.

The D104's crystal element and mechanics have evolved little since its first introduction in the 1930s. However, during its 70+ year manufacturing history, the "grip-to-key" arm was added in the 1950s, and the 2-transistor amplifier in the TUG8/TUG9 base style was added in the late '60s. Finally, the Silver Eagle version included an additional PTT arm in the mic base. Below, I show several D104 wiring diagrams for use with the FET Transformer circuit. This is not an all-inclusive set of diagrams. For example, I do not yet have a wiring diagram of the non-amplified D104 with the grip-to-key arm. Also, very early D104s have no PTT switching whatsoever: the mic cable is connected to the mic element, runs through the stem and out the base to the transmitter. So, some brain power will be required when using the FET Transfomer in these D104 variations.

When considering cost, circuit simplicity and performance, it's difficult to beat the FET transformer circuit. However, for those interested in obtaining gain from the circuit, see the MOSFET and Op-Amp variations of the FET transformer circuit shown further down the page. In each of these configurations, the existing Astatic 5K pot may be used for level control.


Input impedance:  Greater than 10 megohm. 
Frequency response:  +/- 0.5 dB, 50 Hz - 10 kHz (incl. crystal element equivalent source Z of 1000pF in series with 10K resistance);
Audio filtering:  On-board provision for optional 6dB/octave HP filter by adding two components;
Output impedance:  Less than 300 ohms and will drive both Lo-Z and Hi-Z terminations;
THD, ref. 1 kHz at -10 dBu input level:  0.1%;
IMD (4:1 SMPTE RP120-1991 test standard): 0.15% at -10 dBu input level;
SNR, no weighting, > 80 dB ref. -10 dBu input level; 
Gain:  Less than unity as a source-follower.  Generally, no amplification required as the high-output-voltage crystal element is completely unloaded;
Output level:  Adjustable with external control as shown on our schematic and circuit description.  The existing Astatic 5K potentiometer may be used to control output level;
Powered from +5V to +15V and only 2 mA of idle current with the J310 and MPF-102; approx. 1 mA with the J201.  


D104 FET Transformer circuit board installed in the base of an Astatic D104 microphone; powered from the transceiver.




The circuit shown below is a 2-stage MOSFET preamp for the Astatic D104 and makes use of the existing 5K base-mounted level control - or an on-board Bourns sub-minitaure potentiometer. The first stage sets circuit gain and utilizes a BS170 MOSFET (Q1) in a common-source configuration, followed by a J310 JFET (Q2) as a source-follower for driving Lo-Z terminations. As the component count is increased over the simpler, one-stage source-follower, this board uses SMD/SMT surface-mount components.

By using fixed bias on the gate of the BS170 MOSFET, the circuit maintains a constant Q point on the FET's load line, regardless of R10's gain setting. Moreover, the noise figure of the amp is not fixed; it will vary only to the extent of needed gain. Although a voltage source is used on the MOSFET gate rather than being self-biased, the input Z is still 10-megohm. Zeners D2 and D3 are used to protect the MOSFET gate from static discharges when using the D104 in low humidity conditions. These diodes may be omitted when using a MOSFET with internal ESD protection -- like the 2N7002K.

The circuit output Z is set by R12's low value of 270 ohms. As in my previous D104 circuit designs, the power bus is completely decoupled and seperately filtered at the bias voltage divider as well as the drain lead of each stage.

This circuit design taught me a valuable lesson concerning the use of multilayer chip capacitors (MLCC). Note the use of several high-value MLCC caps in both the audio and power supply filtering paths. When using Class II & III (X7R, X5R & Y5V) MLCC types, the voltage coefficient de-rates the effective capacitance value and this can have a disastrous effect on the operation of a circuit. For example, what you think is a 100 uF cap, may be derated down to 10 uF or less as the voltage nears the MLCC's maximum voltage rating. NP0/C0G MLCC caps do not exhibit this behavior but beyond about 0.5 uF, this class of MLCC capacitors don't exist. When a high value capacitor is needed, use an electrolytic type or use it in parallel with the MLCC.  Miniature electrolytic SMD versions are plentiful and can be seen in the improved op-amp design of the equivalent circuit shown further down this page. For a detailed look at the effect of capacitor voltage coefficent, see this document from Vishay Intertechnology.

UPDATE (06/16/2016): Vishay and AVX now produce high-C tantalum capacitors that are almost unaffected by voltage coefficient.  While audiophiles grimace at the thought of using tantalum coupling caps, the reality is that distortions are likley masked by other circuit factors.  According to this AVX white paper, typical THD is still less than 0.2% across the audio spectrum when using older tantalum caps.  Notice the near identical distortion curves of the new tantalum type when compared to low-ESR electrolytics.  With older tantalum capacitors, one can mitigate distortions by applying bias (e.g., back-to-back tantalum caps in series, with + pull-up in between),   As a point of reference, Denon, Marantz, and Thorens all used used tantalum coupling caps in their legacy audio product lines during the 1960s and '70s.  




At the suggestion of Jack Smith, K8ZOA (SK), I designed a D104 transformer circuit built around an ultra-low bias current op-amp to perform the gain and Z transformation function. It wasn't long ago that when a designer wanted to optimize a circuit based on a given op-amp performance parameter that the performance of other op-amp attributes was sacrificed. Today, there's little sacrifice in audio applications and several manufacturers offer op-amps with superlative noise, distortion, transient, input bias current, and offset voltage characteristics.

In the circuit shown below, a high performance LTC6240HV op-amp is used. Both the top and bottom PCB layers are evenly populated with SMD components. Because this op-amp has an exceptionally low input bias current of 0.2 pA, a high-value bias injection resistor (R5=10 megohm) can be inserted onto the non-inverting input to bias and stabilize the op-amp at 1/2 of Vcc. The resulting input Z is commensurate with that of the self-biased JFET circuit. Perhaps a decade ago, it would have been a difficult challenge to create a biased Hi-Z op-amp input while trying to preserve some semblance of good low-frequency audio response from the D104's highly reactive crystal element.

The upper frequency response is mostly limited by the 390 pF feedback capacitor (C2), but only under conditions of gain since the non-inverting input is used. No matter the gain control of a non-inverting input, the output level is not less than unity since the gain equation is Vout = Vin[1+(Rf/Rg)]; where Rf represents the feedback resistance and Rg is the inverting input resistance.  A caveat concerning the feedback network: notice the feedback resistior (R2) is formed after a 100-ohm build-out resistor but the frequency/noise-limiting feedback capacitor is connected directly to the op-amp output.  In a prior design, I had both R2 and C2 connected after the build-out resistor.  In that version, Chris Farley, KC9IEQ discovered a high frequency oscillation at ~ 500 kHz.  The two feedback components placed after the build-out resistor (R8) created excessive phase shift beyond the specified phase margin of the op-amp.  The current configuration solves the problem. 

Click here for a display of the LTSPICE simulator output. The graph shows worst-case low-end frequency response when the gain pot is set to zero ohms, resulting in maximum gain. The circuit board photos below show a virtual reality image from the PCB software as well as the top and bottom layers of the completed board.

Beginning with the op-amp version of the D104 transformer/amplifier, I am now using extensive "via stitching" between top and bottom copper ground plane pours of my new PCB designs.  Via stiching generally becomes effective with an increase in frequency relative to board size.  Multilayer PCBs, especially when designed for RF, may have many layers of ground; even the signal layers typically have ground flooded on them. These grounds need to be all tied together to keep them from acting like stubs or transmission line segments themselves. Or put another way, they need to be a low impedance at all the operating frequencies of the PCB, otherwise they don't look like ground anymore - they start looking reactive to the signals on them.  Rule of thumb: If you space ground vias at 1/8 of a wavelength or less, your ground plane will look and behave like a solid ground.



Those interested in powering their Astatic D-104 mics from the Elecraft K3 transceiver may be interested in the following common-source buffer circuit. It uses any common n-channel JFET transistor (e.g., MPF-102, J310, J201 or equiv.). The K3 supplies power to the interface through +5V bias injected onto the mic lead. The input impedance of the circuit is 5 Megohm. The output impedance is approximately 5K-ohm and is equal to the JFET's drain resistance. The drain resistor is internal to the K3 as shown. An alternate diagram is shown for interfacing the D-104 microphone into a PC sound card where the card supplies a mic bias voltage.


I've been thinking about ways to best manage AC power distribution for my desktop amps. Until recently, I've been manually switching power plugs under the operating table. One branch circuit is used, and consisted of a 30A OCPD (breaker) and a run of #10/4 electrical cable to a 30A Twist-LockÒ wall-mounted receptacle. 4-wire cabling is used since the older Alphas make use of a neutral for the blower.

To simultaneously feed AC power into three amps, I had several choices: (1) add more branch circuits; (2) add a J-box in the wall and split one branch circuit to feed more than one receptacle; or (3) construct a relocatable power tap (i.e., a 240/120V version of a multiple outlet strip). I decided against options 1 & 2 since the house will one day go to a new owner and I didn't want to add even more wall trauma. Further, UL and NEC limits option #2 to 20A circuits.  However, my amps are all legal-limit and 20A service is adequate as the maximum current demand is less than 15A. The #10 cabling to my shack helps to reduce voltage drop losses over an 80 ft run.

Option #3 became the plan but what does UL have to say about 240/120V RPTs? UL 1363 sets the RPT product standard although a similar UL provision addresses portable power distribution of the type generally found at construction sites. Even though I will not mass-produce these RPTs , I wanted to know how close I could come to making one that is 100% UL 1363 compliant. I got pretty close. In fact, it complies but for the fact that the PVC enclosure has permanent mounting tabs. UL 1363 states that any physical mounting must be effectuated without the use of tools. Typically, we see home RPTs with mounting blocks designed to slip over screw heads so that the RPT is easily attached and removed without the use of hand tools.

UL 1363 has a significant limitation: use of an RPT is limited to 20A service and ALL device connectors must be of the same type and current rating. This had a "trickle up" effect. Since the RPT is limited to 20A, then so are the connectors. But now we move from UL to the NEC since the NEC is a premise wiring standard and not a products wiring standard. As the plug is limited in size to 20A, then the mating premise receptacle must be 20A. Finally, NEC 210.21(B) in this application states that the serving OCPD must match the receptacle rating. In essence, use of the 20A RPT forced a re-design of the branch circuit from 30A to 20A components although wiring is allowed to stay at #10/4.  UL 1363 Section 10.1.4 limits the length of the supply-side cord to no more than 25 ft., and not less than 1.5 ft.  My cord length is 6 ft.

130A silver-plated copper bus bars are used to distribute service into three outputs. You'll see this in the photos below. It was actually possible to forgo the bus bars and use jumbo #10 winged wire nuts to join four #10 conductors together. Kinda' scary, but they're used all the time by electrical contractors. However, most wire nuts are not rated for fine-stranded portable power cable. I really didn't consider this a viable option.

A larger J-Box could have been used to install panel receptacles rather than pig-tails but the box size gets pretty large when you consider space needed for panel receptacles. Short pig-tails also allow the use of different styles of 20A plugs/receptacles, provided they remain the same type.

If secondary-type surge protection devices (SPD) are considered for use inside the RPT, installation and type must comply with UL 1449.  In such case, place only between Line 1 and Line 2, and one each from the line to neutral - with no SPDs connected to the grounding conductor.  Even then, I would not use secondary SPD protection inside the RPT unless "whole-house" Type 1 or Type 2 protection is first used at the service entrance as the primary SPD. This is another reason why adding a neutral may make sense even if your present amp only requires 3-wire service and no neutral.  Adding the neutral also ensures that you can properly use older Alpha and Henry amps in the future.

Some features of the RPT:

- 1 x 3 device, designed for either 3 or 4-wire powering (240/120V);
- #10 wiring and hardware used;
- Heyco compression connectors;
- Nickel-plated copper bus bars (with protective covers) rated at 300VAC and 130 amps;
- Dual fuseholders with 20A ABC ceramic fuses;
- Dual neon light AC power indicators; and
- 20A Hubbell Twist-Lock connectors used on all cable ends.

Images below (shown before the neon pilot lamps were wired):


RPT shown above with exposed neutral bus.



Each 20" LCD monitor mounts on a 60-inch steel arc. Movable vertical compartment boards float between the desktop and the riser, allowing for fast equipment changes. The boards are 2/3 the depth of the riser so that frequent cable changes between compartments are easy. Tektronix TM test set, Elecraft K3 transceivers, and W8ZR StationPro II shown on the desk tier.


Homebrew CMOS CW Keyer, SDR-IQ, LP-100A, Alpha 4510A, custom SWR Null Meter, and SunSDR2 shown on the desk riser, just below the monitors.



WWV accuracy in a box. It's a precision GPS reference oscillator for locking transceivers and test equipment to a common 10 MHz frequency standard. A Trimble Thunderbolt GPS-DO module is housed in a customized enclosure. Power for the Trimble unit comes from a PowerOne triple-output linear power supply. A series 20 encosure was purchased from Par-Metal (www.par-metal.com). I designed the front panel using Front Panel Express (FPE) software (www.frontpanelexpress.com) . The panel is CNC precision engraved, anodized, then lettering completed with white filler. With FPE software, the designer creates any type of panel that comes to mind, uploads the design file to FPE, then the panel shows up at your door about a week later. Notice the power supply test jacks and LEDs. The LEDs are series Zener regulated such that if any one of the supply rails drops by 0.5V, the LED turns completely off. Otherwise, a failed supply could still result in a dimly lighted LED. The test jacks are short-circuit isolated through 1K metal film resistors allowing for accurate power supply measurements with a standard Hi-Z input DMM. A pull type locking toggle switch is used to prevent accidental hits to the power switch. The AC line uses a standard CORCOM IEC power entry module with RFI/EMI protection.


Left pod shown with the control unit of an AG6K style motorized balanced antenna tuner. A 100-foot, 25-pin cable connects the desktop control unit to a nitrogen-filled, NEMA Type 4 enclosure where the tuning components are housed. Components from a commercial Bird Thru-line wattmeter panel were salvaged and installed into a custom anodized panel from The Panel Authority in Lockport, IL. http://www.panelauthority.com Preston Wakeland's panel work is exceptional.

Also shown at the top of the rack is an Ethernet power controller and Gentner XAP400 processor unit that features: 8x8 matrix crosspoint router, internal phantom power supply for condenser mics, AGC, compressor, expander, 16-band parametric EQ, and auto-nulling digital phone patch.


Touch-Tone control pad for the Gentner digital phone patch.  An audio mix-minus bus is created inside the Gentner unit.   To activate the phone patch, the "call" button is depressed.  Immediately, a 100 ms. pink-noise burst is sent down the telephone line to null the trans hybrid to create a full-duplex connection on a single-pair telco line.   The touch-tone pad becomes active and a call can be placed either on or off the air.


Detail of the balanced tuner control panel. In addition to dedicated metering of forward and reflected power, the two smaller meters are used to index the position of the motorized vacuum variable tuning capacitor and Collins ribbon inductors.  Close inspection shows a pair of pulse-width-modulation (PWM) motor controls to allow for precise tuning speed of the vacuum cap and ribbon inductors.  The motor speed control pots can be seen below the key switches.  As expected, the PWD circuits generate RFI, but they are powered only during the tuning procedure.  LEDs illuminate to show out of range (Max-Min) conditions for both the tuning capacitor and ribbon inductors.


The balanced tuner's outdoor ATU unit in a sealed NEMA Type 4 enclosure. Starting at the upper left, two Jennings 75-amp vacuum relays form a "knife switch" to automatically connect/disconnect the balanced transmission line when the control panel power switch is activated. ITT/Cannon Military control cable connector to the right. Schrader valve for the nitrogen feed is near the control cable connector. Immediately below the vacuum relays, a motorized Jennings vacuum tuning capacitor. At the far bottom, a pair of Collins motorized silver tape inductors. A cogged drive belt synchronizes the two inductors and can be seen at the lower cabinet edge. Above the tape inductors, the solenoids for a pair of Eimac glass vacuum relays can be seen. These are used to switch the vacuum cap either in front of, or in back of, the tape inductors to allow for Hi-Z and Lo-Z terminations.



My first Internet remote station was completed in 2007 and was built around the Kenwood TS-480 transceiver and controlled by W4MQ IRT software. A special USB-to-RS232 adapter was designed that allows plugging-in the TS-480 control head into a netbook PC USB port while the base unit remains at home.

The control head and adapter were powered through the USB connection. Audio streamed over the Internet through IP-Sound software with a choice of a dozen different CODECs. CW mode was controlled by two synchronized K1EL WinkeyerUSB devices - one functioning as the host, the other as the client. The WinkeyerUSB kept CW keying perfectly timed over the Internet, despite network latencies. The client keyer was modified with a low-distortion "Twin-T" sinewave side-tone generator and voltage-controlled-amplifier (VCA) to cleanly ramp the keyed waveform into the netbook PC without keying artifacts.

Internet station AC power was controlled remotely with a Digital-Loggers Ethernet Power Controller. Through an Ethernet port and router, any one of sixteen home outlets could be turned on/off through the Internet. Antenna switching and environmental status was provided by an N8LP Remote Board. Transceiver and amplifier switching was accomplished through a W8ZR StationPro II device with a CAI Networks control board and embedded web server.

Finally, two time-out-timers (serial data detection) monitor the presence of a continuous and valid Internet stream to keep the station FCC-compliant. If the stream was lost for more than ten seconds, the TS-480 automatically powered down.




The circuit shown below is a small modification made to the Harbach Engineering SK-220 Soft Key interface, designed for the Heathkit SB-220. If Q1 fails with an open base, the modification ensures that +120V from the amplifier does not appear back at the tranceiver's amp key line. Notice the inclusion of the +2.5V Zener diode inside the blue-colored oval. Without it, the only mechanism isolating +120V back to the transceiver is the fragile emitter-base junction of Q1 where the junction, together with D2, holds the base at +1.4 V with respect to circuit ground. The added Zener is current-limited through R1 and will clamp the transceiver's key line to no more than +2.5V.



The adapter board shown below was created in DipTrace PCB design software.  The board mounts on the key's two wiring posts and a 3.5 mm (1/8") stereo mini jack is used (tip and GND) to connect the key to a transmitter.  I have several Begali iambic paddles that all use a 3.5 mm TRS jack.  Now, the bugs are interchangeable with the paddles by moving the mini plug from one key to another.  Two adapter board versions were designed since the wiring post polarity is not consistent across all Vibroplex models.  Even though a key is a simple switch, the wrong polarity exposes the operator to a transmitter's key line voltage when touching the base.


Various Link-Coupled Tuners















HRO E76 from the second production run in 1935. 

Attributes of the HRO D and E series: white mother-of-pearl S-meter switch; off-white coil charts; cloisonné "NC" dial pointer; German silver PW dial; round IF transformers; and S-meter reads to "S/5" as the RST reporting system was introduced late in 1934 by Arthur W. Braaten, W2BSR.  RST was adopted roughly at the time the S-meter plates were stamped.  

I modified this HRO with High Frequency Oscillator (HFO) temperature compensation using Collins uncased silver mica caps with a negative coefficient, and also added Zener voltage regulation to the HFO plate and screen circuits.  E76 performs very well and maximum drift on 10m is 1 dial division after a 10-minute warm-up.

HRO L73 (early 1936). 

Attributes of the L series HRO Seniors: silver "NC" dial pointer; black lacquered dial; panel lamp power indicator; black coil charts; "S/9" meter; and "pull" type meter switch.  L73 is a rack version with an alligator finish.  Like HRO E76, I modified L73 with HFO temperature compenation using Collins uncased silver mica caps with a negative coefficient, and also added Zener voltage regulation to the HFO plate and screen circuits. 



This NC-101X took a bit more work to restore than the HROs.  Using 6.3V tubes, it sometimes exhibits HFO "hum modulation" on the upper bands.  A chapter could be written about hum modulation in these receivers.  The cause, effect, and cure is an interesting subject.  Apart from that, audio quality is superb since it's a push-pull design with a pair of 6F6 power tubes. 

Drift is more pronounced in this reciever as the power supply and PP audio section is contained within the same cabinet.  The creation of cabinet heat is substantially more than that of the HRO design and affects HFO drift to a greater degree.   In fairness to National, the HFO sections in these receivers are not heterodyne/mixed like later receivers.  For example, receivers designed twenty years later typically run with a VFO that operates after the first mixer and crystal LO.  In those receivers, the VFO is running at only a couple MHz and high stability is easily attained.  However, in these early National receivers, the HFO operates on the operating frequency all the way up through 10m.  For National, it was very difficult to design an HFO with reasonable stability -  and one reason why HRO production was delayed in 1934.  

Like HRO E76 and L73, the inside of this NC-101X almost looks like it left the factory.  The entire NPW tuning system was painfully removed and given an hour-long dishwasher treatment.  After washing, the parts look new.  It's especially important to keep the area under the NPW drive system in very clean condition since the catacomb coil pins are switched just below the tuning capacitor. Any debris that lands in this area affects receiver performance.       


8382286 Last modified: 2017-10-11 12:07:34, 106188 bytes

Login Required

Login is required for additional detail.

Apply for a new Vanity callsign...

You must be logged in to file a report on this page

Please login now...

ad: giga-db
Copyright © 2017 by QRZ.COM
Fri Oct 20 23:29:27 2017 UTC
CPU: 0.101 sec 166179 bytes mp