And yet ANOTHER high power Class D amplifier!  I just built  David’s (G0MRF) 300 W PA and the details are at the end of TX, just above Homebrew Highlights…a really nice bang-for-the-buck project.

Information on the ULTIMATE 3S  (QRP LABS) micro transmitter added, just above Homebrew Highlights, at the bottom of this page.

A new section, HOMEBREW HIGHLIGHTS, has been added at the bottom of this page.  Send your pix and details so others can marvel!

It’s hard to go out and buy a transmitter that’s kHz capable.  JUMA makes a kit for a full bore transmitter – lots of surface mount and $$$ for 60W out.  The current Kenwood TS-590 has provision for low level (0 dBm, 1mW) – but you’ll need some kind of outboard amp to get on the band.  The KX3 same deal, very low output, plus $$$.


The JUMA TX500 60W Ttansmitter

I was fortunate and acquired the JUMA TX pictured above … the original buyer had no experience in surface mount soldering.  I would not recommend this rig to someone not very experienced in soldering fine pitch components…nice rig, however, and worked perfectly upon initial powerup…this a CW only transmitter, including a RX upconverter.

So what to do?  The easiest way to make some power on our band is via a transmit converter, driven by your existing HF rig.  This device is essentially a mixer – for example, assuming you want a signal at 474.2 kHz, USB – the MF WSPR dial frequency.  Your existing HF rig is set to 4.4742 MHz USB, and a one or two watt signal is passed to the converter.  Again, assume that this converter has a mixer employing a crystal oscillator producing an LO frequency of 4.000 MHz.  The output of the mixer is of course a sum and difference set of frequencies, in our case 8.4742 MHz (sum) and 474.2 kHz (difference).  A simple low pass filter strips the unwanted sum frequency, leaving our desired kHz range product, ready for amplification.  The beauty of this scheme is that ALL the functionality of your HF rig is used to produce whatever modulation is desired at the MF.

Such a converter can be made with a handful of cheap, common parts.  Go here:


Rog, G3XBM, across the pond, developed just such a rig.  His site has a wealth of ideas on the subject and is a must visit.  His transverter develops about 10W out at the MF.  John, WG2XIQ pointed me there when I started all this MF thing in 2012.  Here is my incarnation of that transverter:


Above is that original transverter under test.  In the foreground, left is an attenuator that reduces 2 watts from the HF rig to about +7 dBm.  Above, L, is the crystal controlled oscillator at 4.000 MHz.  Rear, center, is a MiniCircuits SBL-1 mixer device.  Dead center is the LPF, which drives a simple Class D amplifier to 10W out at the MF.  The output LPF is far right, rear.

This can be built for WAY under $50 using all new parts.  This rig is NON LINEAR, and is suited for the so-called digital modes like WSPR, JT-9, and CW.  We’ll discuss a linear rig later…  This concept is easily realized using different parts, mixers, FETs, etc.  Pretty much anything works at these frequencies…no surface mount, totally non-critical layout, etc.  When homebrewing, I highly suggest using powdered-iron toroids with #2 mix to create the inductors – more efficient and much smaller.

HPIM1697 Scope shot of tune-up session — tweaking the slug tuned coils results in a perfect sine wave.  Homebrew MF SWR bridge sits above the scope.

HPIM1700 Transverter driving a 100W out PA, above the Astron.


The schematic for Rogers rig is on his page, refer to the link.

I made many variations to this basic rig, using different mixers, air inductors (no tweaking needed!) as well as toroid inductors using #2 mix, such as a T-114-2 or T-96-2 (iron).  The inductance values are non-critical, +- 20%, it just works.

Being never satisfied, I developed my own version of this transverter, using a SA-612 VHF mixer chip with built-in oscillator, modified intermediate LPF, and a robust 5-pole LPF with second harmonic suck-out trap to meet current FCC emission specs. (Rogers version is barely legal as designed in the USA, it’s not as strict over there.)  I developed a custom PCB for the circuit – and amazingly enough due to layout and toroids, it makes 25W out – a 3 dB boost over the original!  Power supply is 13.6V @ about 3A.  Additionally, I incorporated provisions to drive the mixer with an external reference such as a GPS stabalized 10 MHz source.  A PTC thermistor stabalizes the crystal.  A tap is provided to drive an external power amp with a low level signal.

I should mention the term ‘transverter’ implies receive capability.  These rigs simply pass the antenna RX signal straight back to the HF rig without upconversion, thus requiring ‘split’ operation.  See ‘RX’ section of these pages…

IMG_0070 The WA3ETD converter, set up for external 10 MHz GPS reference LO, thus the unpopulated areas…

IMG_0071 A closer look – the unneeded LO crystal would be located just to the right of the mixer (the 8 pin DIP).  The 10 MHz reference enters at the lower, R BNC connector.

IMG_0066 One of the highly organized ETD bench, testing a converter – the IF driving rig is the FT-817ND on the left. (Set for CW output)


Best Intentions, Part 1

Back Then’ (2012) we all thought the USA would get 630m as a ham band…I figured the best way of getting guys past the TX issue would be to offer up a transverter kit, as seen above.  And so it came to pass…as of this writing about 20 of these rigs are on the air worldwide.  I’m on the 2nd run of 20 PCBs now.  Ya want one?  See the MF Solutions section…cheap, foolproof.  All documentation and schematic is available for download.


A 10 or 20 watt power level will get you on the air – surprisingly well when employing the digital modes such as WSPR and JT9.  However, when considering the brutal reality of the highly inefficient antenna systems we employ, having ‘QRO’ capability definitely helps.  Much more regarding antenna efficiency in a bit.

Fortunately, making real power on the band is very easily realized.  Recalling that most of the activity (currently) on the band involves digital modes, a suitable non-linear power amp is easily created … 100W is almost trivial, 500-1000W easy.


The amplifiers are non-linear in nature – meaning that an AM, SSB, or PSK-31 signal when applied will not be faithfully reproduced at higher level.  However, modulation such as MSK (minimum shift keying) or CW will.  Our ‘digital’ amplifiers are essentally switches – a bipolar transistor, or better choice FET, is switched fully on and fully off by a square wave.  These pulses at the operating frequency are applied to a LPF (low pass filter), which very simply stated, produces a sine wave at the output of the filter.  The gain block(s) do not operate in their linear region – the bonus we reap is that the amplifier efficiency is very very high – efficiencies in the mid 90%s are easily obtainable … meaning little power lost to heat.  For example, my 1 KW Class D amplifier, design by Jay W1VD, operates at 95% efficiency.  The smaller, single-ended amps to be described here achieve about 85%.

The sine wave at operating frequency from the converter mixer, post intermediate LPF, is applied to a ‘squarer’ circuit, which is nothing more than an overbiased bipolar transistor.  The sine wave overdrives the device, resulting in a clipped, essentially square wave  as output.  This low voltage drive, about 12-15 v P-P, is applied to the gate of a FET operating at anywhere from 20-100 volts of drain voltage.  If designed correctly, the chosen FET will be fully turned on and off by the drive, resulting in a high voltage square wave that is presented to the output LPF.  Depending on complexity, an intermediate FET driver chip or circuit may be required to overcome the very high gate capacitance of the FET which tends to slow the rise time of the driving signal, resulting in heating.


Once again the British invasion continues!  Many operaters use an amplifier designed by Roj, GW3UEP, known as the QTX series (Quick TX)  His amp uses an el cheapo IRF-540 FET, about 90 cents in quantity.

Again, view his extensive site for all details and schematics … at http://www.gw3uep.ukfsn.org/100W_QTX/100W_QTX_Text_090711.htm .  (Also in ‘HOT LINKS’)

This amp can be built with a VFO for a self contained CW transmitter, or easily driven by the converters previously described.  I have built three of these PA, one on a 3″ x 5″ PCB card – the mini, pictured below.  John, WG2XIQ, uses two of these amps combined to produce over 200W to the feedline.  A single amp can make upwards of 120W if the PA capacitors have sufficient voltage ratings…don’t push it!  The amplifier requires ~25 VDC at ~4.5A.  All parts available from Mouser…

HPIM1604 GW3UEP VFO under test

HPIM1612 Original completed 100W CW transmitter.  VFO at bottom, amp in middle, built ‘Manhatten’ style on a double sided PCB, LPF at the top.  I used this when I first got on as a CW beacon at 472.5 kHz.  Inductors are RTV’d right to the chassis.  This unit is now my emergency backup rig.

HPIM1718 WG2XKA original WSPR station.  Amp driven by low level signal from original converter prototype…25 V power supply is two Astrons in series…

TOTAL COST for the above station is under 100 bucks and some sweat equity!

IMG_0110 The Mini QTX on a 2.5″ x 5″ PCB, testing at lower power into a dummy load.

Trolling the web for ‘500 kHz amplifier’ will uncover a multitude of designs using scrounged parts.

This amp can be built using toroid inductors – #2 iron mix, like a T-114-2…I achieved 90% efficiency with the toroid design – don’t hesitate to experiment.

IMPORTANT NOTE!  The capacitors specified by Roj in the post-FET and LPF circuits can’t be pushed past 100W safely due to voltage ratings.  The WIMA caps he specifies can be obtained in higher AC voltage ratings at the same value…see the WIMA section in Mouser.  ALSO, ‘any old cap’ ceramics, etc. CANNOT be used in the PA and LPF…these caps MUST be pulse rated.  Don’t be tempted to skimp on a 1 dollar capacitor as the wrong cap just flat won’t work !!!  Otherwise, this amp just works.

UPDATE:  John WG2XIQ / KB5NJD is offering a ready-built/tested beefed up version of the 100W amplifier, with custom enclosure and all connectors…see the NJDTechnologies link…


100W not enough?  Building a large Class D or Class E amplifier will solve that issue.

Jay Rusgrove, W1VD has designed several large amplifiers and accessories for the band.  Again, these are non-linear and suited only for CW and the digital modes (PSK modes excepted).  Jay’s site is :  http://www.w1vd.com   .  His ‘prototype’ amps are production quality and a joy to look at.  I first duplicated his 500W power block, seen here:

IMG_0165 ETD’s version of the W1VD 500W Class D amp.  The heatsink was in my scrap metal box, and by sheer luck exactly fit the Bud box used.  At full output the heatsink is hardly warm, running at 94% efficiency.  This is my main MF station amp; typically I run about 250W out at night – power output is controlled by a Variac in front of the power supply – more on that subject below.  This amp employs two $1 FET devices.

Since I enjoy building stuff here, I also built his 1KW amp that uses four devices, seen below:

IMG_0166 A W1VD 1KW Class D amplifier.  Four devices drive the large toroid on the right.  To the left is the driver board with FET gate driver and frequency divider – this design requires a driving signal 2X of the desired output frequency.  The toroid is critical and substitutions result in low efficiency – see Jay’s detailed discussion on his site.

These amps both require 50 VDC at full output, at 10 and 20 A .  The LPF is external to these amps.

So, how to get that kind of DC power…I lucked out, and was given a large filament transformer N1DYL had scrounged from a scrapped RCA FM broadcast transmitter.  It has two, 17V secondary windings, rated at 50A continuous duty.  I placed those windings in series to a bridge rectifier…filtering is brute force, using ~300,000 uF worth of computer grade electrolytic caps in a series/parallel configuration – these found new from ALL Electronics.  Metering V and I is via an el-cheapo LED block with current sense resistor found on the auction site.  A Variac controls primary AC voltages, thus allowing continuous power output changes from the amplifier.

Xformer2 The transformer – those stud rectifiers were not used.  Moving that beast will not be discussed!  It, and the capacitor bank/metering/control stuff sits under the operating bench.


As previously noted, the above rigs are suitable for most ‘digital’ modes, but not for anything requiring linear output.  I desired to experiment with PSK-31 as my Part 5 grant allows such operation.  I created a linear system, described here:

IMG_0169 Driver for the 50W linear amp.  The original transmit converter was stripped of its non-linear PA and LPF.  Now the output of the mixer drives an intermediate stage ( the blue PCB, R), which makes about 1W if fully cranked.  The pot in the middle controls the drive to that stage.  The small green pot on the blue PCB controls the bias for the intermediate PA (FET w/heatsink).

IMG_0168 The 50W linear PA. Two IRF-540 FET devices drive the output toroid on the left.  Two (blue) bias boards receive regulated 15V and individually control the gate bias for the FETs.  The driver resides on a subassembly, far R.  The secondary of the toroid transformer is tapped to allow impedance matching.  The entire amp sits on a finned heatsink, which is mounted to the stripped chassis of a computer power supply.  The fan is most definitely required as this amp, at best, is less than 50% efficient, but it is linear!

So much for amplifiers. As you can determine, one must want to homebrew gear for the band.  I have derived tremendous enjoyment from building this gear – hopefully you have some ideas of your own.  Again, the web has a wealth of projects undertaken by others and is well worth researching…this is supposed to be fun, right?

FLASH…The link below describes the new KB5NJD/WG2XIQ 100W power amplifier for 630m!  Check it out.

NJDTechnologies 630m Amp





The U3 series of micro transmitters have proven extremely popular with the QRP gang.  These inexpensive rigs (base unit, around $35 USD) consist of a microprocessor controlled synthesizer driving a low power (1W max) PA and swappable LPF modules for the band of choice.  The kit is a snap to build (no surface mount components) and features an extensive menu driven operating system.  Of interest to us is that it supports all popular digital modes including WSPR on 2200 and 630m.  Options, such as a GPS receiver and custom enclosure, support for six switched filters and the like are offered.  See the link above for all the details and documentation downloads.  A really informative and detailed website, don’t miss it.  This rig is ultra fun and the kit is of the highest quality.

I have a bare bones original U3 that I use as a driver for my 80m I-F transverters.  I’ve also used it on the air on 630m WSPR (no computer required!!) – lots of fun!  Of note, the GPS patch antenna is factory tuned to the board dimensions – the PCB is the groundplane.  This results in very high GPS performance indoors, unless your house has a metal roof and spun aluminum window screens!  Placing the outboard GPS option near a window is all it takes to timelock the U3S for digital modes.

Recently I built a full blown, all option U3S for another ham.  The workmanship, fit and finish, and PCBs are all first rate.  Pictures of that build follow:


U3S Start up Splash Screen


A U3S on 630m WSPR


Full Bore U3S in custom enclosure with 6 band LPF switch board


U3S Optional GPS Receiver – a current very high performance GPS RX – I added the SMA connector to use an outside antenna – fast satellite lockup and battery backup.  (Distorted shot angle – the SMA really is square to the receiver chip!)


GPS unit topside, featuring a large high performance patch antenna and LED status indicators.


An extremely complete full transverter is available from Monitor Sensors, please see this link:    http://www.monitorsensors.com/ham-radio/630m-transverter  .  The MS rig makes 50W out, is fully linear and incorporates extensive self-protection systems.  The IF used by the MS box is in the 160m band.

THE G0MRF 300W Class D Amplifier

As I’m always looking for fall/winter projects, I decided to build the G0MRF PA.  This is a really nice rig and can be built by anyone with average homebrewing skills and a bit of care.  This amp is loaded with features such as built-in SWR bridge, overcurrent and SWR protection, and complete TX/RX control.  It will make 300W out, requiring a power supply capable of 35V at 10A.  Power is continuously variable by using a Variac with your chosen power supply.

All components (except the cores) are available from Mouser and DigiKey, and the toroid cores are common items – I got them on Ebay with “free shipping” (right!) .  David will provide the PCB, output transformer and reflectometer cores as well as a complete resistor kit if desired.  His webpage g0mrf.com provides very detailed construction and setup tips.

A few notes:  I wound the LPF cores with fairly heavy #16 magnet wire which was a bit tedious…one needs to insure the inner winds are tight against the toroid – it took 78″ per core.  Take extra care with the output transformer.  David now recommends 16 instead of 15 turns for the secondary, wound first.  I used #14 ZIP cord for the 4 turn primary.  It is absolutely mandatory that the turns thru the toroid be drawn down tight over the secondary layer to insure efficiency!!  Also, the pigtails from the primary wind, when splitting the ZIP cord (if used) must be exactly the same length…

Below,  the unpackaged PCB and test leads associated with the first power up.  I used an available chunk of heatsink for the power devices and snubber resistors, drilled and tapped for 4-40 hardware.  Excuse flash highlight.  Now to package this project …….. .


G0MRF 300W Class D PA under initial test


A look at 630m projects.

First off is from Phil, VE3CIQ, an obvious breadboard/dead bug specialist.  His 60W amplifier is based upon a MRF454 bipolar transistor, running Class C and fan cooled.  According to Phil “3db input attenuator is the box on the left, LPF is on the right. Parts were mostly pulls from a couple of TV chassis, had to parallel pieces to get the values.”  The amp is driven by a MS Solutions converter kit, throttled back.  A classic!


VE3CIQ Class C 60W PA


VE3CIQ 60W amp exciter, throttled back from 25W, using XKA downconverter


One thought on “TX

  1. Pingback: Quiet terrestrial and geomagnetic conditions return but was there enough spark for strong signals?; DJ0ABR -> UA0SNV; WD2XSH/17 -> G8HUH; WH2XCR -> VK6XT and ZF1EJ/1; WG2XKA completes G0MRF amp build; Server relocation in the coming days

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s