Tag Archives: Kostas (SV3ORA)

How to build a PC keyer and AM modulator for the EMTX emergency transmitter

Many thanks to SWLing Post contributor, Kostas (SV3ORA), for sharing the following guest post which originally appeared on his radio website. Note that this project builds on the EMTX emergency transmitter project:

PC keyer and AM modulator: A 15-components versatile keyer and powerful PSU modulator for the EMTX (Emergency Transmitter)

by Kostas (SV3ORA)

Schematic of the keyer and modulator (on the left) for the EMTX. The EMTX schematic is shown as well on the right, to determine the connections to the keyer/modulator.


My very successful emergency transmitter (EMTX) was only capable of CW or other slow speed ON/OFF keying modes. Then I thought, why not “give voice” to the design? CW is good, but it is half of the fun. If you could use your simple CW transmitter to send out your voice as well, this would be great. You could now chat comfortably on the nets or use any digital radio amateur mode and have much more fun. The simplest modulation you can apply to an existing CW transmitter, is the AM modulation. And whereas this is an old modulation, mostly abandoned by HAMs due to beeing inefficient, there are still AM nets on HF. But do not forget, AM can also be heard by SSB receivers by zero-beating the receiver to the AM carrier. So you could still use your simple AM transmitter to QSO with the SSB guys!

Along with the modulator, there is also a versatile keyer embedded to the circuit, so that the EMTX can be manually keyed with different ways or automatically keyed by audio tones from the PC. For more information on the keyer, keep reading.

The AM modulator

In the old days, the most common way to apply AM modulation was to modulate the high voltage to the plate of the tubes, using a transformer and a powerful audio amplifier. In low voltage solid state circuits, you can still do it using transformers, but you can also use series transistors instead of the transformer. All these things require many components and/or powerful AF amplifiers if one is to modulate higher power transmitters. This does not match the keep-it-simple design I am trying to achieve here.

So I thought of a simple trick with the use of the extremely common LM317 regulator, used as a modulated power supply. This modulator uses just a few common cheap components and it is able to achieve remarkably good modulation levels for it’s parts-count, just from line audio input. It juices every bit of the internal circuicity of the LM317, just look at where the base current of the 2N2222 comes from.

The AM modulator is a kind of novelty. Whereas there is nothing special in a modulated power supply, this circuit has some interesting properties. It is amazingly sensitive and it is able to provide lots of modulated current to any low power transmitter that it can feed. It can be easily driven by the line output of any laptop (around 20% volume) and provide a very good depth modulation to the transmitter. Charles Wenzel was kind enough to do a simulation on the circuit I developed, which is shown below.

His simulated circuit is a slight variation (for measurement purposes). The resistor to ground on the base stabilizes the bias and the ratio of R1 and R2 set the output voltage (0.6 volts across R2 gives about 8 volts across R1). He put in an emitter resistor just for good measure. Same for the series resistor from the source. Charles words, “I don’t know how believable these results are but it looks pretty darned good!”.

The circuit is being used as a current booster, the current being the supply to the transmitter and dependent on the voltage it produces. The LM317 always tries to keep 1.25V between it’s output pin and “adj” pin but where we benefit here is the current at the “adj” pin is very low, so it is easier to apply audio to it. Effectively, the error amplifier inside the voltage regulator is used as an additional amplifier stage. The output pin voltage varies according to the voltage on the “adj” pin so if we use it to bias the transistor we get negative feedback which improves the quality of the modulation. More output voltage = more bias current = lower output voltage. The result, is a very cheap, low components-count, very sensitive AM modulator that can supply lots of power to easily drive the transmitter and produce a clean and deep AM modulation!

The AM modulator bias is set with the 1M potentiometer. Depended on the bias level, the idle carrier on the EMTX can be set from about 0.5W all the way up to 8W. Needless to say that this modulator can modulate any similar power transmitter, not just the EMTX.

The keyer

If it is to modulate the EMTX from the PC, so as to use the different digital modes, there must be a way to key it also from the PC. This is why I decided to embed into the same circuit, a PC keyer which is triggered by the line audio of the PC, but also triggered manually (internal or external key). Keying by audio tones was decided, because modern PCs do not have LPT ports to trigger directly by DC. This keyer uses a reed relay to reliably, fastly and scilently key the EMTX, which is activated by a transistor. The base current for the transistor is derived from the audio signal after rectification. The incoming audio from the PC line passes through the mini audio transformer to increase its voltage, it is rectified and then charges the shunt capacitor to drive the base of the transistor. The keyer “speed” (decay) is determined by the shunt capacitor size. The circuit starts to trigger from about 50-60% of my sound card output signal level.

The relay used to key the EMTX, must be able to tolerate at least 1A of switching and carrying current. Note that the relay contacts switching current is not the same as the contacts carrying current. Reed relays are the best especially if you want long relay life, noiseless operation and very fast switching speeds, like the ones used in Hellshreiber. If you can’t find such a relay, you can use a reed switch capable of 1A of switching and carrying current and then place a suitable electromagnet close to it, so you can build the relay yourself. If you do so, find the best point where the reed switch responds to the electromagnet.

The keyer relay must be as close as possible to the emitter of the transistor used in the EMTX. The connectors at the back of the EMTX and the keyer/modulator have been physically placed so that when the two units are side by side, a very short link cable is required for this purpose. With the two devices placed close together, you can now use any length of cable for your manual external key, which is now connected to the “EXT” connector of the keyer/modulator.

The keyer does also have an internal mini straight key. I find this idea very nice, to avoid extra cables. It is not the most convenient key in the world, but it is there along with the transmitter every time you need it. By using a special panel switch from apem, I was able to triple this switch usage for the different modes of the keyer. The vinyl lever cap you see in the next picture, is the original part of the switch, to make it easier to key with your finger. But you may build such a part on your own, to fit on other switches types.

The switch is an ON-OFF-(ON momentary) switch type. In the default (middle) position, only the PC keying action is activated. In the top position (ON), the keyer is always active, which is useful for broadcasting audio (into a dummy load). The bottom (ON momentary) position, is the manual PTT action. This is used as a straight key on OOK operation, or as a PTT on AM voice operation. Simple and effective!

Initially, I used one channel of the PC sound card for triggering the keyer and also as an AF signal for the AM modulator, but this caused several problems of unreliable keying or distortion. So I decided to use a second separate AF input (KAF) to key the keyer. This second input, uses the other channel of the stereo sound card. With the addition of this input, there is no interaction between the keyer and the modulator. The AF levels that the keyer and the modulator require, can be set independently. Instead of adding more hardware for the purpose, I have chosen to set these levels by adjusting the volume and the balance of the sound card, which works great. Also, programs like Fldigi, have options for using one of the two channels of the stereo sound card as a keying interface (PTT channel), which makes the keying efen more reliable. When the program is in transmit mode, a continuous tone is heard on the PTT channel. This steady tone, is used by the keyer as a reliable keying signal, independent of the audio signal of the digital mode that modulates the modulator. This solution works very reliably for any mode. But if the program you are using does not have an option for a PTT channel, that is ok, as the keyer works reliably even without this feature. For voice communication or broadcasting music (into a dummy load) you just use the internal key switch as a PTT to handle these modes.


Prior to building the keyer and the modulator in the same device, I had tested the circuits independently quite a few times, to ensure the results can be reproduced. The modulation quality and depth out of the AM modulator have to be listenned to be believed. I have not made any linearity measurements, I just trust my ears on this one. It works great on music as well as on voice. Apart from that, this is the most sensitive AM modulator I have ever built, requiring only a small fraction of the line level output of the PC sound card.

When modulated by this modulator, the EMTX shows no audible signs of FM modulation. I switched my receiver to SSB and I could perfectly zero beat the AM modulated music signal which stayed on frequency and it’s tone did not change during loud audio signal music. Switching back and forth from SSB to AM modulation on the receiver, I did not notice any difference in the audio quality, apart of course from the narrower bandwidth on SSB modulation, due to the narrower IF filter inside the receiver on SSB.

The AM/OOK switch is used to select the modulation applied to the EMTX. When the keyer is set to be triggered by audio from the PC, at the OOK position, the EMTX is just switched on and off by the audio tones applied to the keyer, or by the manual key, internal or external (connected to the “EXT” connector). At AM position, the EMTX is switched on by the audio signal applied to the KAF connector and at the same time AM modulated by whatever audio signal is applied to the AF connector. On voice communications, the momentary position of the internal key is used as a PTT. On music broadcasting (into a dummy load) the non-momentary position of the internal key is used to keep the keyer always active.


Back connections to the EMTX.

Pictures of the finished keyer/modulator. You don’t have to build it that nice-looking if you don’t care.

Modulator prototype and EMTX built on a breadboard. Yes it worked just fine onto a piece of wood.

Thank you so much for sharing this brilliant and simple project with us, Kostas. Your handiwork is absolutely brilliant too!

Click here to check out Kostas’ website.

Spread the radio love

The EMTX: How to build an 8 component 40/30 meter QRP emergency transmitter

Many thanks to SWLing Post contributor, Kostas (SV3ORA), for sharing the following guest post which originally appeared on his radio website:

Emergency transmitter: An 8-component, high-power 40m/30m transmitter to get you quickly on the air

by Kostas (SV3ORA)


QRP is all about doing more with less. This is more than true, with the construction of this cheap, simplistic transmitter presented here. It is designed primarily as an emergency transmitter (EMTX) that can be built or serviced in the field or at any home. However, it can be used as a HAM radio transmitter as well. Do not judge by its low components count though. This transmitter is powerful, more powerful than anything the QRPers would dream of. It is just remarkable how 8 components can lead in so much output power, that lets you communicate with a big part of the world, when propagation conditions are right. It is very difficult for a circuit to match that kind of simplicity in balance with such performance.

Following my detailed instructions, the EMTX can be reproduced easily, within hours. The result is always success, this is one of the circuits that are not critical at all and a successfully working transmitter can be reproduced every time. I have built this transmitter several times, using similar components (even toroids) and it always worked. The transmitter meets the next expectations:

1. Output power (including harmonics): A few mW up to 15W (depended on transistor, crystals and voltage/current used) at 50 ohm.
2. It can drive any antenna directly, 50 ohm or higher impedance, without external tuners.
3. Bands of operation: Currently 40m, 30m
4. Mode: CW, Feld-Hell (with external switching circuit), TAP code and any other ON/OFF keying mode. AM modulation has been easily applied too.
5. Options like reverse polarity protection diode (useful in the field when testing different unknown polarities PSUs) and current meter (for easier tuning) are available.

The challenge

The purpose of this transmitter is to be used primarily as an emergency transmitter. This poses several challenges that influence the design of the transmitter:

1. It must be able to be built or serviced easily in the field or at any home, with components that could be salvaged from near by electronics sources or a small electronics junk box. This means that components count should be kept very low and they must not be rare to find but commonly available parts. As a side effect cost would also be kept small, if one is to buy any component. Also, the active components must be interchangable with many other devices without the need for the design or the rest of the circuit components to be changed.

2. It must be able to operate from a very wide range of DC voltage sources and at relatively low current, so that common house power supplies could be used to supply power to it. Such devices include linear or switched mode power supplies from laptop computers, routers, printers, cell phone chargers, Christmas lights or any other device one might have available.

3. It must be capable of transmitting a powerful signal, so that communication is ensured. An emergency transmitter that is capable of a few mW of output power, might be heard locally (still useful, but there are handheld devices for that already) but isn’t going to be of much usage if it can’t be heard really far away.

4. It must be capable of loading any antenna without external equipment required. In an emergency situation, you just don’t have the luxury of building nice antennas or carrying coaxial cables and tuners. There may be even extreme cases where you can’t even carry a wire antenna and you depend on salvaging wire from sources in the field to put out a quick and dirty random wire antenna.

5. Adjustments of the transmitter should be kept minimum without the help of any external equipment and there must be indication of the correct operation of the transmitter or the antenna in the field.

Components selection

The transistor:
This transmitter has been designed so that it can operate with any NPN BJT in place. This includes small signal RF and audio transistors and high power RF transistors like the ones used on HF amplifiers and CB radios. Despite 2sc2078 is shown in the schematic, just try any NPN BJT in place and adjust the variable capacitor accordingly. When you are in the field, you do not have the luxury of finding special types of transistors. The transmitter must operate with any transistor in hand, or salvaged from near-by equipment. Of course the power capability of the transistor (as well as the crystal current handling) will determine the maximum VCC and current that can be applied to it and hence the maximum output power of the transmitter. Some of the most powerful transistors I have used, come out of old CB radios, such as the 2sc2078, 2sc2166, 2sc1971, 2sc3133, 2sc1969 and 2sc2312. There are many others. As an example, the 2sc2078 with a 20v laptop PSU, gave 10-12W of maximum output power into a 50 ohms load.

Schematic of the 8 components EMTX for the 40m/30m bands. Components with gray color are optional.

The crystal:
This is the most uncommon part of the transmitter. You have to find the crystal for the frequency that you want to operate on. Crystals within the 40m or 30m CW segments are not that common. Further more if you operate the transmitter at high powers and currents, you will notice crystal heating and chirp on the frequency of the transmitter. The current handling capability of your crystal die inside the crystal case, will determine the chirp and the amount of crystal heating. You can still work stations with a chirpy transmitter provided that the chirp is not that high, so that it can pass through the CW filters of the receivers. However, if a small chirp annoys you or if this chirp is too much, then you have to use these vintage bigger size crystals (e.g. FT-243), that can handle more current through them. But these are even more uncommon today.

The approach I have used in my prototype, was to connect more than one HC-49U crystals of the same frequency in parallel, so that the current is shared among them. This reduced the chirp at almost unnoticeable levels, even at high output power, just if I was using a single FT-243 crystal, or even better in some cases. Again, this is optional, but if you want to minimize chirp (and crystal heating) without searching for rare vintage crystals, this is the way to go.

A bit of warning. If you notice a very high chirp when plugging in a crystal to the EMTX, you should consider this crystal as inappropriate for this transmitter, as it cannot handle the current required. If you continue to use this inappropriate crystal, you could easily crack it inside and set it useless. Don’t use these tiny HC-49S crystals, they won’t work.

The current meter:
A 1Amp (or even larger) current meter can be used to monitor the current drawn by the transmitter during key down. The recommended current operating point is anywhere between 450mA to 1A, depended on the output power (and harmonics) level you want to achieve. The current point is set by the variable capacitor. I would avoid setting the current to more than 1Amp, although it can be done. The use of the current meter is optional, but along with the incandescent bulb, will give you a nice indication of the correct tuning of the transmitter, so that you do not need to have an external RF power meter connected to the transmitter output. If you do have, then you can remove the current meter. If you don’t have a 1Amp analogue meter available, but a smaller one, you can parallel a low value power resistor across the meter. In my case, I only had a 100uA meter and I paralleled a 0.15 ohms 5W resistor across it to scale down 1Amp to 100uA, The resistor value depends on the internal meter resistance so you have to calculate this for your specific meter. When the 2sc2078 is used at 20V, 500mA in the current meter indicates around 5W of output power, 600mA indicates around 6W, 700mA 7W, 800mA 8W, 900mA 9W and 1A around 10W. So the current meter can be used as sort of power meter without the need to do any scaling on it.

The incandescent bulb:
A current meter alone, without the use of the incandescent bulb, will not give you the right indication of the operation of the transmitter. In some cases, the transmitter might be drawing current without actually generating much, or even any RF. When you are in the field you do not want to carry extra monitoring equipment with you. The incandescent bulb will light on when the transmitter oscillates. It monitors the actual RF signal, so it’s brightness changes according to the amount of RF power the transmitter produces. Along with the current meter reading, this is just what you need to know in order to set the variable capacitor properly. Note that the bulb will not lit at very low signal levels. The one used in the prototype starts to glow up from a bit less than 1W. Miniature incandescent bulbs may not be that easy to find nowadays. However, there is a good source of these, that almost anyone has in their houses. This source is the old Christmas lights. You do save old Christmas lights, don’t you? The incandescent bulb indicator as well as it’s single turn winding on the transformer, are optional components. If you have an RF power meter connected to the transmitter, you can remove these.

The diode:
The protection diode is an optional component to the circuit. If you are in the field, correct polarity of a power supply may not be obvious. Without a multimeter it might me difficult to determine the correct polarity of the PSU. A power diode (I used a 6A one) will protect the transistor from blowing up in the event that reverse polarity is connected to the circuit.

The Cx and Cy:
The Cx and especially the Cy capacitors need to be of good quality. The Cy will get hot on high output power if it isn’t. In the tests, I have used homemade gimmick capacitor and even double-sided PCB as a capacitor for Cy and they all got hot at high power. Silver mica capacitors run much cooler and they do make a small difference in the output power, so I suggest to this type. Cy must be able to handle quite a lot of voltage, so silver mica type is ideal.

The variable capacitor:
The variable capacitor can be air variable or ceramic, although I prefer air variables in tis application. In any case it must be able to handle a high voltage just as the Cy.

The key:
The key directly shorts the transistor emitter to the ground, therefore it is a part of the active circuit. For this reason, I suggest the key leads to be kept as short as possible. The key must be able to handle the voltage (20v) and current (up to 1A) on its contacts, which is usually not a big deal.

Transformer construction

The construction of the transformer is shown below step by step. Note that if you decide that you don’t need to drive higher impedance loads but just 50 ohm ones (eg. antenna tuners or 50 ohm matched antennas), you just need to wind 2t in the secondary and not 14t. You also don’t need any taps of course.

Step 1:

Take a piece of 32mm external diameter PVC pipe from a plumber’s shop. Alternatively, a suitable diameter pills box can be used, or any other suitable diameter plastic tube.

Step 2:

Cut a 4cm piece out of this tube. 4cm is the minimum length required.

Below a 4cm PVC tube has been cut in size.

Step 3:

Wind 16 turns of 1mm diameter enameled wire onto the PVC pipe and secure the winding in place as shown in the picture below. Notice the winding direction of the wire. This is the primary of the transformer, the one that is connected to the two capacitors. Notice that this winding is wound a bit offset to the right of the pipe.

Step 4:

Wrap the winding with 3 turns of PTFE tape. It can be bought at any plumber’s shop, just like the PVC pipe. The PTFE tape will help in keeping the second layer turns in place and it will provide extra insulation.

Step 5:

Wind 2 turns of 1mm diameter enameled wire on top of the primary winding and secure the winding in place as shown in the picture below. Notice the winding direction of the wire, as well as it’s position relative to the primary winding. This is the feedback of the transformer, the one that is connected to the collector of the transistor.

Step 6:

Wind 14 turns of 1mm diameter enameled wire on top of the primary winding, starting from just next to the 2 turns one and secure this winding in place as shown in the picture below. Notice the winding direction of the wire, as well as its position relative to the primary and the 2 turns windings. This is the secondary (output) of the transformer, the one that is connected to the antenna. At this point do not worry about the taps yet.

Notice in the picture below, the way the windings are secured in place onto the pipe. The wire ends are passed through the pipe using small holes and then bent towards the ends of the pipe and once more to the surface of the pipe, where the connections will be made.

Step 7:

Wind 1 turn of 1mm diameter enameled wire onto the pipe and secure the winding in place as shown in the picture below. Notice the winding position relative to the other windings. This 1 turn winding is placed about 1cm away from the other windings. This is the RF pick up winding, the one that is connected to the incandescent bulb.

Step 8:

Use a sharp cutter (knife) and carefully scrap the enamel of all the windings ends. Do not worry if you cannot scrap the enamel at the bottom side of the wire ends (that touches to the pipe). We just want enough copper exposed to make the connection.

Step 9:

Tin the scrapped wire ends, taking care not to overheat them much.

Step 10:

Now it’s time to make the taps on the secondary winding. Use a sharp cutter (knife) and very carefully scrap the enamel of the wire at the tap points (number of turns). Take much care not to scrap the enamel of the previous and the next turn from each tap point. Do not worry if you just scrap the enamel at the top of the wire (external area). We just want enough copper exposed to make the connection.

Make each tap, a bit offset from the near by taps, like shown in the pictures. This will avoid any short circuits (especially at the 4, 5 and 6 taps) and it will allow for easier connections, especially if alligator clips are used to connect to the taps.

Step 11:

Tin all the tap points, taking care not to overheat them.

Step 12:

This step is optional and it depends on how you decide to do the connections to the taps. You may solder wires directly to the tap points, but in my case I wanted to use alligator clips, so I did the next: I took a piece of a component lead and soldered it’s one end to each tap point. Then I bent the component lead to U-shape and cut it accordingly. This created nice and rigid tap points for the alligator clip.

Step 13:

This step is optional and it depends on how you decide to mount the transformer to your enclosure. In my case, I wanted to create three small legs for the mounting. I cut three pieces of aluminum straps and made holes at both their ends. I made three small holes onto the transformer pipe end and mounted the aluminum straps using screws. After mounting them, I shaped the straps to L-shape. Then I used three more screws to mount the transformer to the enclosure.

The completed transformer is shown in the pictures above and below. The 6 connection points at the bottom of the pipe, are the low voltage points, whereas the 2 points at the top of the pipe, are the high voltage points.

If you have built the transformer as described, the bottom connections are as follows (from left to right):

Wire end 1, connected to the incandescent bulb
Wire end 2, connected to the incandescent bulb
Wire end 3, connected to the current meter
Wire end 4, connected to the current meter
Wire end 5, connected to the GND (ground)
Wire end 6, connected to the transistor collector

The top connections are as follows (from left to right):

Wire end 1, connected to the 25pF variable capacitor and the Cy fixed.
Wire end 2, is the 14th secondary tap and it is left unconnected, or tapped to the appropriate impedance antenna.

Videos of the EMTX in operation

I have made two small videos of the EMTX in operation.

The first 13.5MB video (right click to download), shows the operation when the transmitter is set for a bit less than 10W of output power.

The second 3.5MB video (right click to download), shows the operation when the transmitter is set for about 5W of output power.

EMTX chirp analysis

Every self-exited power oscillator (and even many multi-stage designs) exhibits some amount of chirp. Chirp is mainly considered as the sudden change in frequency when the power oscillator is keyed down. Apart from chirp, there is also the longer term frequency stability that may be considered. The chirp in the EMTX is surprisingly low, if it is built properly. Hans Summers, G0UPL has performed a chirp analysis on my EMTX (PDF) and the EMTX built by VK3YE and presented on YouTube. Hans, performed the analysis from the video/audio recordings of both transmitters. I sent him two videos, one with the EMTX set for an output power of 10W and one where it is set for 5W. The chirp at worst case (10W) was about 30Hz and at 5W in the order of 10Hz or so. Being so small, the chirp is almost undetectable by the ear and it surely poses no problems when passing the tone through narrow CW filters. This is an amazing accomplishment from a transmitter so simple and so powerful.

EMTX harmonics measurement

Every unfiltered transmitter will excibit harmonics at it’s output. This means that the output waveform has some distortion in comparison to a pure sinewave. Many of the transmitters I have seen, present a very distorted output waveform and absolutely need a LPF if they are to be connected to an antenna. I can’t say that this is true for the EMTX, because surprizingly, it has low distordion, despite the high output power it can achieve. Although a LPF is always a good idea, it is not that much needed on the EMTX. However you have to use one to comply with the regulations.

The image above, shows the measurements on the output of the EMTX, when it is set closely to 10W at 50 ohms. The main carrier is exactly at 9.9W and all the harmonics are less than 50mW! Also, the harmonics, do not extend into the VHF region.

The image below, shows the measurements on the output of the EMTX, when it is set closely to 5W at 50 ohms. The main carrier is exactly at 5.17W and all the harmonics are less than 9.6mW! Again, the harmonics, do not extend into the VHF region.

These small harmonics levels aren’t going to be heard very far at all, compared to the powerful carrier. This means only one thing. A LPF, although a good practice, is not mandatory in this transmitter. But you should better use one so that you comply with the regulations.

Many HAMs use just a watt meter to measure the output of their homebrew transmitters. This is not the proper way of doing it, because the watt meter is a non-selective meter. It will measure both the fundamental carrier and the harmonics, without being able to distinguish them. So in an unfiltered transmitter, or in a transmitter with a simple (often non measured) LPF, this way will give a totally false reading of the output power of the transmitter at the set frequency.

The proper way of accurately measuring the output power of a transmitter and the harmonics levels, is a spectrum analyzer. The FFT available in many modern oscilloscopes, having a dynamic range of approximately 50-55dB, is adequate for this purpose as well. A 50 ohms dummy load must be connected at the transmitter output and then the high impedance probe of the scope, is connected to the output of the transmitter as well. This was the way that the above measurements have been performed.

WebSDR tests

Here are some test transmissions, to determine how far one can get with such a transmitter. I have to say that there is an antenna tuner between the EMTX and my inefficient short dipole (not cut for 40m and not even matched to the coaxial). However I could still cover a distance of more than 2500Km even on the 5W setting.

A screenshot of the transmitter signal, as received on a WebSDR 2500Km away and when the EMTX is set for an output power of 10W.

Below, is a picture and an audio recording of the transmitter signal, as received on the same WebSDR and when the EMTX is set for an output power of 5W.


Pictures of the finished transmitter. You don’t have to build it that nice-looking if you don’t care.

EMTX prototype built on a breadboard. Yes it worked just fine onto a piece of wood.

This is a phenomenal project, Kostas. Thank you so much for sharing it with us. I love the simplicity of this design–truly form following function. With a little patience, anyone could build this transmitter.

Check out this project and numerous others on Kostas’ excellent website.

Spread the radio love

How to install a mechanical SSB filter on the Yaesu FRG-7

Many thanks to SWLing Post contributor, Kostas (SV3ORA), for sharing the following guest post which originally appeared on his radio website:

How to install a mechanical SSB filter on the Yaesu FRG-7

by Kostas (SV3ORA)

The Yaesu FRG-7 is a general coverage MW/SW receiver that uses the Wadley Loop system for stabilizing the frequency tuning. The receiver has a good sound on AM mode, that reminds me the tube receivers sound. However, on sideband mode, it is pretty much useless. The IF ceramic filter that is used, does not have enough selectivity to reject the opposite sideband. No matter if the front panel mode selector switch states USB/CW and LSB, these just shift the BFO, nothing more. The receiver is a DSB set not SSB. A cheap way you can accomplish single signal sideband reception with the FRG-7 is described in this link. Whereas it works, it increases the audio bandwidth of the signals to the high pitch.

A better approach is to install an additional mechanical filter to the receiver. This of course requires expensive 455KHz mechanical filters, but if you have one in hand or if you are willing to pay for the improvement in performance, then this is the recommended option. But you can’t just desolder the ceramic filter of the receiver and solder a mechanical filter in place. On AM mode, you need wider bandwidth, but on SSB mode you need narrower. So both filters must be in place and a selection must be done in each mode. Thankfully, this modification is pretty easy on the FRG-7 and it does not require any modification of the external appearance of the radio.

The schematic of the FRG-7 is shown above. Everything with red color, are part of the modification. The modification is pretty straight forward. You have to desolder the original ceramic filter from the FRG-7 PCB and install it on a separate PCB along with the new 455KHz mechanical filter. To select between the two filters, a 9-12v DPDT relay can be used and it must be connected as shown in the schematic. The power for the relay coil is derived from one section of the mode switch (S3d). On USB or LSB modes, the BFO is energized and this power is also used to energize the relay, which in turn switches to the narrow mechanical filter on these modes.

A good place for the new PCB that accommodates the filters, is just below the main tuning dial of the receiver. There is a hole there and three screws, which can be used to also hold this PCB in place. I needed to replace these screws in mine with longer ones, because I used spacers to prevent the PCB from touching the chassis. But this is optional.

Two small pieces of coaxial cables are used to connect the new PCB to the pads of the ceramic filter, that has been now removed from the original PCB of the receiver. Ground these cables on both ends.

The power cables for the relay coil (shown with red and black in the picture above), are passed below the PCB to the chassis opening and through a hole to the bottom of the original PCB of the receiver. The ground wire is soldered to the filter ground point and the red wire is soldered to the mode selector switch S3d. S3d is the outer wafer onto the switch. Use a multimeter to find the contact of the switch that has VCC when the mode is switched to USB or LSB. This is the point where you want to connect the red wire.

After installing everything, you should perform an alignment of the TC404 and the T406 in the BFO section as described in the manual. This requires a frequency counter, but I did my alignment by simply adjusting the two controls by ear, until I got roughly the same pitch on LSB and SSB audio bandpass. These controls interact, so you have to do a bit of back and forth in both of them. It is very easy.

After installing the modification and aligning the receiver, the result is pretty obvious. No more DSB reception, SSB signals are received just once in the dial and their bandwidth is limited as it should on SSB. The mechanical filter I had, was a bit narrow (2.1KHz) so I can also hear a bit os “seashell” sound on SSB, but SSB voice signals are perfectly understood. It is interesting that the audio volume between the ceramic filter and the mechanical filter was just about the same, which indicates that there is no additional loss in the newly installed filter. Another interesting thing is that there was no need for any impedance matching using active devices or transformers on the mechanical filter. It worked just by directly connecting it. Neither it’s loss, not it’s response seems to be affected by any possible impedance mismatches.

Note that Collins produced both symmetrical and asymmetrical mechanical filters (yes they used two filters, one for USB and one for SSB in some of their gear). My filter is a symmetrical one (same roll-off response curve on both sides of the filter passband). If you use an asymmetrical filter, expect a bit different pitch when switching from LSB to USB and vice versa. Not a huge problem, but just a note.

By performing this simple modification, you will end up with an FRG-7 receiver that is trully selective, allowing for real SSB reception. Most importantly you do not ruin the appearance of your precious FRG-7, but just improving it’s performance. This modification would probably be appreciated much when deciding to sell your FRG-7 to someone else.

Thank you for sharing this practical and affordable project with us, Kostas!

Post Readers: Check out this project and numerous others on Kostas’ excellent website.

Spread the radio love

How to build an automatic remote antenna switch

Many thanks to SWLing Post contributor, Kostas (SV3ORA), for sharing the following guest post which originally appeared on his radio website:

How to build an automatic rig/antenna switching system

by Kostas (SV3ORA)

When I started collecting vintage rigs, I ended up in a line of rigs on my bench, that were sitting there, disconnected from any mains cables or the antenna. I wanted these rigs to be ready to fire at any time I wanted to, without having to connect/disconnect cables all the time. I also wanted to be able to compare different rigs performances at the flip of a switch, which is the only way this can be done on the HF quick fading conditions. For power cables, the solution was to leave them connected in the mains plugs all the time. My rigs that have an internal PSU, have mechanical switches, so they are isolated from the mains when they are switched off. The rigs that are powered by an external PSU, depend on the external PSU main switch for isolation (in case they haven’t mechanical switches on them), which in my case is mechanical and switches off the mains power, when the PSU is switched off.

However, for the RF cables, this was a different story. Having only one antenna and multiple rigs, means that you have to connect each rig to the antenna every time you want to operate each rig. This is not only boring and time consuming (you have to reach the back of the transceivers to connect/disconnect the connectors), but eventually causes the connectors of the coaxial cable and the rigs to wear out. I decided to make things better and make an RF rig selector for my rigs. This RF rig selector has been described in this link.

The current antenna I use is fine for transmitting, but in the noisy neighbourhood where I live, it picks up a lot of noise. I have tried many solutions, without significant effect in the noise level. This is why I decided to use a separate antenna for receiving, from that used for transmitting. This antenna will be some kind of loop probably, so as to be immune to noise or insensitive to the direction of the noise. It will be placed in a different location than the transmitting antenna, a location which will be less noisy. Unfortunately, the space I have for the TX antenna lies in a very noisy location in my property. So a separate RX antenna, in another physical location is a must. This means that a separate coaxial for the RX antenna must be used. Thankfully, the RX coaxial can be very small in diameter, passing easily through the sides of the windows, without extra holes.

To satisfy all of my requirements, I developed the circuit shown above. The circuit is able to switch a common antenna to four different rigs. Why four? Because this was the capacity of my switch and the number of connectors I had available. If you have a greater capacity switch and more connectors, expand the circuit to your needs.

The circuit of the shack switch, allows for 4 separate rigs to be selected, and two antennas, one for RX and one for TX. TX/RX antenna selection is being done automatically (split antenna operation) and controlled by the PTT of any of the rigs connected. This feature can be bypassed by the switch, so that the TX antenna can be used for both RX and TX. The same switch allows also RX operation with passive RX antennas of active ones. When in the active RX antenna position, power is passed to the remote RX preamplifier through the RX coaxial cable, using a bias-T circuit. The values of the bias-T circuit have been chosen very large, so as active RX antennas that operate at LF and lower could still be used. The RF relay defaults in the TX antenna, so that if there is a power failure, or if the circuit is not supplied with power, you can still receive (and transmit) with the TX antenna. The other way around, would be fatal for both the transceiver and the RX antenna (If you transmitted accidentally into it).

The PTT circuits are based on my transceivers. Unfortunately, there is no “standard” for the PTT circuits, each rig has its own way, so the PTT circuits must be thought for each of them. I followed an “inhibit” approach for the PTTs. That is, all the PTT switches are connected in series and DC is passed through them. If any of the rigs transmits, the PTT switch is opened and the circuit switches to the TX antenna. For the rigs that do not have an internal relay but output DC on TX instead, an additional small relay is used (for greater isolation and lossless switching). The only drawback of this “inhibit” topology is that the PTTs of all the rigs must be connected to the circuit simultaneously. If you want to exclude a rig of course, you may short circuit it’s PTT connector in the circuit. The PTT circuits as I said, are non-standard, so you might want to change the circuit to your needs, but anyway you got the idea.

Notice the connections in the circuit. One section of the RF switch (on the left) is used for the positive wire (central conductor of the coaxial) and another for the negative (braid of the coaxial). Why is that? This is because I canted to add a special feature to the switch. That is, the ability to disconnect the antenna from any rig when the rigs are not used. Previously, I used to disconnect the antenna coaxial from the transceiver when I was away, so as to protect the transceiver from antenna static discharges and possibly destroy it’s front end circuits. Now, with a single flip of the switch, I am able to do so. Because I wanted the switch to operate on different types of antennas (balanced or not) I decided to short circuit both poles of the antenna at this position, to equalize their charges.

But equalizing their charges was not enough. I had to find a way to let these charges go to the ground, so that the antenna is discharged. Directly grounding the short circuit, did not seem a good thing to do, because the whole TX wire antenna on the roof would be grounded. Whether this is a good idea to avoid lightings or not, I do not know. So I decided to keep the short circuited antenna floating and instantly discharge it only when adequate static charge is built upon it. For this purpose, I used a neon tube, permanently connected to the switch NC (not-connected) position. When the switch is in the non-connected position, the tube lights up and discharges the antenna (both poles) if an appropriate amount of static charges has been built upon it. When the switch is in any of the selected rigs connections, the tube is disconnected, preventing it from lighting up when you transmit into the antenna. Note that this configuration, requires that the output (antennas) coaxial connectors must be isolated from the metal chassis of the RF switch!

Isolation of the output antenna connectors has been done with a PVC sheet and isolated screw rings. Also note the usage of BNC connectors on TX and SMA on RX. I used BNC connectors for various reasons. They are excellent connectors with quick lock/unlock features. You do not need to screw them (and wear them out) and once fit in place they are not unscrewed. Once fitted in place, they allow for rotating the connection without unscrewing the cable or bending it. They can handle 100W easily. Despite all these features, they are much smaller in size and lighter. Their reduced size fits easily to reduced diameter cables like the RG-58 and similar. In an RF switch where there are lots of cables connected, this does make a difference. They are also very common and very cheap. There are even types that do not require soldering at all to fit a coaxial to them. I use BNC connectors even at my antenna side, as they have been proven to be quite waterproof. The types of BNC connectors I choose are not silver plated. Despite silver plated connectors are better, in the long term they are corroded by humidity and become much worst than the nickel plated connectors. The connectors I used are nickel plated with gold plated central conductors. I have found these types to be much more durable over the years, despite being cheaper. The same goes for the RX SMA connector, but I used an SMA connector there so as to accommodate thinner coaxial cables for RX.

The BNC connectors used, are the square flange types. I used this type of connectors because when they are fitted onto the chassis, they cannot be unscrewed, unlike the single-hole types. For the RX though, I used an SMA connector because it is even smaller and it can accommodate smaller diameter cables. The coaxial cable used for the internal switch connections on TX, is the RG-223. This cable is silver-plated (both the central conductor and the braid), it has double braid for increased shielding, it is of the same diameter as the RG-58 and it has a bit lower loss. The cable loss is negligible though for such small pieces of cable. The same type of cable has been used for the internal switch-relay connections as well as for the connections of the selector to the rigs. Appropriate lengths of RG-223 cables were cut and fitted with BNC connectors at one side and the appropriate rig connectors at their other side. For the RX antenna, you may use the thinner diameter cable you can find. I used a small piece of very thin coaxial (taken out of the WiFi card of an old laptop) and passed this piece through the side of the windows of the shack and through the mosquito net of the windows. No extra holes are required that way! For the rest of the RX cable, you can use whatever cable diameter you want, but I tried to use the smallest diameter I could find, so that the cable is as much phantom as possible.

All the coaxial rig cables are grounded at the connectors side. I used a piece of coaxial braid and fitted it to the connectors screws. Then I soldered the braids of the coaxial cables onto this piece. Notice the black ring screw isolators at the antenna connector, to isolate it from the chassis. Speaking about the chassis, do not use a plastic chassis for the RF switch, use only a metal one! The picture below, as well as all the next pictures, show the RF cables arrangement, but note that the circuit in these pictures is not complete yet.

The coaxial cables are soldered onto the switch contacts. Where a ground connection is required, a piece of braid accomplishes this. Do not use thin wires, the device has to allow for at least 100W of HF RF power to pass through it. I have tested the switch with 200W of power and there were no problems at all. The neon tube directly connects to the appropriate switch contact and to the chassis.

The most important part of an RF switch is of course the switch itself. For 100W of HF RF power, I would suggest you to use a porcelain switch. I had a 5-positions 4-sections small porcelain switch, which I used. I connected two sections at each side in parallel (adjacent pins connected together). That is, two sections in parallel for the positive wire and two sections in parallel for the braid. I did that for various reasons. First, by using two contacts for each connection instead of one, you increase the power handling capability of the switch. Then, you ensure a sure-contact throughout the years. Any corrosion or wearing on the switch contacts would cause contact problems eventually. By using two contacts for each connection instead of one, you double the probability for a good contact. After all, I had a switch with more sections, so why not make a good use of them?

The completed selector is shown above. The relay was been taken out of an old CB radio. Use the best quality relay you can afford, as this will be switched quite often and it must handle at least 100W of RF power.

The results from the RF switch operation are quite satisfying. The overall construction is kept small and low profile. The switch makes a good contact despite being small. The automatic discharger seems to work well. On receive, there is some RF leakage, as I expected, in the near by cables, which is noticed in the higher HF bands or in very strong signals. The very sensitive receivers we use, are able to detect that. This RF leakage occurs even when the switch is in the NC position, where the antenna is disconnected and floating. So, to be honest I have not figured out if the leakage is from the switch or from the external cables in the shack. On TX, there is of course severe leakage from the transmitting coaxial to the rest of the ports. This IS expected. There is leakage even without using any selector at all, in the nearby receivers, when a transmitter operates at such high powers. There is nothing you can do about it really, unless your receiver has a mute capability, which I did not bother to take care of.

The TX/RX switching is taken care automatically and this is very useful and relaxing for the operator as he does not have to worry about anything. The active or passive RX antenna selector and the feature to disable the auxiliary RX antenna are really useful and you can do many antenna and rigs comparisons on-the-fly with it, by the flip of a switch. Depended on the noise level and the sensitivity you want to achieve, the switch will provide you the most optimal RX conditions instantly!

The most important thing though, is that the goal of this project was achieved. I am able to switch the antenna to whatever rig I want at the flip of a switch. And before I go away, at the flip of a switch I can isolate and automatically discharge the antenna when needed. This is so much more convenient than having to connect and disconnect cables all the time. I can also now use a separate antenna for RX, which greatly improves reception in my case. This antenna is automatically switched by any rig I have and I do not have to worry about anything. I can also do comparisons between different antennas on RX, which is crucial in deciding which antenna is better for receiving. All these features make this little simple to build circuit, so useful and an integral part of the shack.

Thank you for sharing this practical and affordable project with us, Kostas!

Post Readers: Check out this project and numerous others on Kostas’ excellent website.

Spread the radio love

Kostas’ Yaesu FRG-7 adjustment that improves opposite sideband rejection

Many thanks to SWLing Post contributor, Kostas (SV3ORA), who shares the following video and writes:

In this 7.8Mb video (attached) is my solution for “converting” the Yaesu FRG-7 for single signal reception on SSB:

Not a mod actually, no additional filters, no soldering of any kind. Just tune the BFO on USB and on LSB a bit far away from the 455KHz ceramic filter (using the transformer for LSB and the capacitor for USB, as the manual states). As the video shows, this provides the near to
carrier selectivity to cut off the unwanted sideband.

The price you pay is more high frequencies (but in the wanted sideband) and a bit attenuated low frequencies as the filter is effectively shifted to higher frequencies. Very high frequencies cut-off is helped by the tone control of the receiver to some point.

This is a cost-free mod and requiring even no soldering skills, neither any mod to the receiver. Now as you tune the bands in SSB and CW, you do not hear the same signal twice. On AM mode nothing changes, since the BFO is switched off in this mode.

Many thanks for sharing this, Kostas! This seems like a simple adjustment for one of my all-time favorite receivers!

Post readers: Check out Kostas’ website for more modifications, ideas and radio projects.

Spread the radio love