EFE-40 HF SSB QRP TRANSCEIVER

This article is also available in the Turkish version.

The EFE-40 is a 7 MHz SSB QRP transceiver derived from Luc Pistorius’s Forty QRP design and further developed by TA3OM Mustafa Bey. In this first part of the series, the design origins of the device, the circuit modifications, and important technical details related to its construction are discussed. 

The Turkish Brother of the Forty QRP Transceiver: EFE Is Born 

It was December 16, 2013 (yes, I remember!). Late in the afternoon at work, I was secretly reading the discussions on the forum then known as Bitlik when I saw a message from TA3OM Mustafa Bey and practically jumped out of my chair. He was explaining that he had taken F6BQU Luc Pistorius’s Forty transceiver, combined it with a DDS signal generator he had built himself, and—with some modifications—created a higher-performance QRP single sideband (SSB) transceiver. In an instant, my mind went back ten years…

It was the summer of the year I had finished my master’s degree in France and started my PhD. I had no chance of setting up a station or working on electronics (except for a FOXX-3 that I had assembled). My financial means were very limited, yet I still couldn’t resist subscribing to Megahertz magazine. It was a high-quality publication edited by F6GKQ Denis Bonomo. In fact, partly to break some of the prejudices about Turkey and to introduce our country, I also wrote a few articles for that magazine. I first came across F6BQU Luc Pistorius and the circuits he designed thanks to Megahertz. These circuits were introduced through articles in the magazine. If you wanted to buy them as kits, you could obtain them from Dahms Electronique in Strasbourg. Unfortunately, neither Megahertz magazine exists anymore nor does Dahms Electronique sell those kits; but thankfully, detailed descriptions of all of Mr. Pistorius’s circuits are still available on his website. There are very useful circuits ranging from simple to more advanced ones—I recommend taking a look.

That summer, when a heatwave swept across Europe, Luc Pistorius published the details of a QRP single sideband transceiver operating on the 40-meter band in the July 2003 issue of Megahertz: the Forty. I remember carrying the magazine with me and studying it until morning at the hotel where I worked nights (yes, I completed my PhD while working night shifts at a hotel!). I especially never forgot the last sentence of the article: “Despite its deliberate simplicity, this transceiver is not a toy.” A year earlier, in the August 2002 issue of the same magazine, Mr. Pistorius’s Toucan CW transceiver for the 30-meter band had also been published, and I had been quite tempted by that one as well. Anyway, in the end, amid all those financial constraints, the rumbling of my stomach prevailed; I buried both the Forty and the Toucan in my heart (!) and carried on with my life.

Time passed, and when I returned to Turkey, I learned that a group of amateurs, led by TA1AC Çetin Ağabey, had prepared kits of the Forty and a frequency counter that could be used with it, and that these were being sold as kits at Başak Elektronik in Abed Han in Karaköy. I obtained one of each. I built the frequency counter, but alas, I couldn’t even touch the Forty because it was time for my military service. My year as a reserve officer came and went. But my mobility never stopped. I went to Cyprus for work, and our Forty board was left orphaned again. At one point I intended to finally build it and even carried it there with me, but somehow my mind drifted to my first love, the Toucan. I left the Forty aside and built a Toucan instead, but since I didn’t really know CW properly, I never quite boxed it up and made it fully operational. In the meantime, I changed jobs and returned to Istanbul.

It was during this period, when I had finally begun to devote serious time again to amateur radio, that Mustafa Bey’s message appeared on the forum. You can imagine how happy I was. A week later, the boards for both the transceiver and the VFO were in my hands. Naturally, the first thing I did was to start removing components from the Toucan and the old Forty boards. I used most of those parts in the construction of the EFE-40 (and later the EFE-20).

Technical Specifications – Differences Between EFE QRP and FORTY QRP

Now that we have covered the historical background (!), let us move on to the technical part. I mentioned that the EFE is an improved version of the Forty. We should talk a bit about the differences between the two transceivers. First of all, it should be emphasized that Luc Pistorius continued to develop the Forty after August 2003. You can see this evolution by clicking the links on the relevant page. For example, the earliest versions did not even have a digital frequency display. Later came the PLL-synthesized Forty 2. As far as I understand, the version Mustafa Bey used as the basis while developing the EFE was the Forty 1B. What he essentially did was to eliminate some of the shortcomings that amateurs had complained about in the Forty’s circuit and to combine it with a variable frequency oscillator similar to the one in the Forty 2—but using DDS instead of PLL. Fortunately he did, and we owe him our thanks. There were amateurs in Turkey who had built the Forty 1 and Forty 2, but thanks to Mustafa Bey’s detailed explanations, his support, and the well-prepared circuit boards, we really experienced a QRP boom. 🙂

As for the points that were improved:

• Luc Pistorius designed the Forty according to the components he could find in France. Two of these parts are really difficult to obtain, and even if you can find them, they are very expensive. The first is the SSM-2165 integrated circuit from Analog Devices, which he used for microphone compression. Mustafa Bey solved this problem by redesigning the microphone input to work with a capacitive microphone and by using a TL072 op-amp to provide both pre-amplification and compression, thus eliminating the need for the SSM-2165.

• The second issue concerns the Neosid 5164 intermediate frequency (IF) transformers, four of which are required in total. It is possible to find other miniature transformers with similar values, but almost all of them measure 10×10×10 mm, which makes them somewhat bulky. This particular series from Neosid, however, measures 7.5×7.5×10 mm. For this reason, Mustafa Bey designed the PCB with traces and holes for both 7 mm and 10 mm transformers, making it possible to use whichever suitable transformer is available. I will provide more detailed information about this topic below.
• To increase the somewhat insufficient audio level of the Forty, he amplified the signal in the receive section after the second mixer and the demodulator, again using a TL072 op-amp.
• In the Forty, automatic gain control (AGC) occurs when an LED conducts only during the negative portion of signals whose peak-to-peak voltage exceeds 0.6 V (for very strong signals, a second diode comes into play, further reducing the signal level entering the mixers). In the EFE, however, the signal taken from the low-frequency preamplifier is amplified by a transistor and used both for the S-meter and for gain control. Naturally, an S-meter output circuit—absent in the Forty—was also added to the design.
• In addition to these changes, Mustafa Bey placed a higher-gain transistor at the output of the crystal filter in the receive section, moved the aluminum heat sink onto the PCB, and provided a VFO input with an adjustable resistor. He also added screw terminals—which I find very practical—to make the connections between the PCB and external components easier.
• Of course, for all these changes, he also redrew the PCB layout and the circuit diagram. The dimensions of the EFE boards are: RF board 120×95 mm, DDS board 115×54 mm.

Construction of the EFE-40: Some Observations, Tips and Suggestions

Now that we have also touched on the differences between the two devices, we can move on to notes about construction, adjustments and operation:

• As I mentioned, I obtained the EFE PCBs from Mustafa Bey. I started by assembling the DDS board. Apart from the SMD AD9833 integrated circuit and a few SMD capacitors around it, there is nothing particularly difficult here. When mounting the DDS board on the inside of the front panel of the enclosure, I had to move some components that were originally on that side to the back because of space constraints, but that was a matter of personal preference. When connecting the encoder to this board, make sure that the A and B terminals are wired correctly. If you do not like using a push-button encoder, you can use a separate button for menu operations; if I had had enough space, that is what I would have done.
• A 16F84 microcontroller controls the DDS chip. Mine came with the program already loaded when I received the boards. You can contact Mustafa Bey regarding this.
• The AD9833 integrated circuit is a DDS device capable of generating signals up to 12.5 MHz. Like all DDS circuits, it cannot generate signals at frequencies higher than half of the reference frequency (clock oscillator), so by using a voltage-controlled oscillator (VCO) designed for 25–27 MHz, you can make use of the AD9833 across this entire range. With a 25 MHz reference, it provides a resolution of 0.1 Hz, which is more than sufficient for this type of device. The circuit worked flawlessly with the 28.672 MHz VCXO that came with the boards.

Component Selection, DDS VFO, and Crystal Filters

Before starting the assembly, I carried out a careful effort to source the components. In all his explanations, F6BQU emphasized that using multilayer ceramic capacitors is important for good performance. In our country, however, it is very difficult to find places that sell such capacitors except for the larger ones designed for high voltage. For this reason, I ordered a kit containing various values from China via the internet, and additionally asked for plenty of commonly used values such as 100 pF, 1 nF, and 10 nF (note that not every capacitor labeled “multilayer” actually is one. Some are fake. It is best to buy from sellers you have previously purchased from). Before assembly, I measured the values of all passive components and discarded those that showed large deviations from their nominal values. All the cores I used were originals (Amidon and Micrometals) that I had also purchased from abroad. I avoid buying the binocular or toroidal cores found on the market in various colors but with unknown specifications, because if they do not have the required characteristics they only waste time.

Another component that required special care when sourcing for this circuit was the miniature transformers. As mentioned earlier, F6BQU used transformers from the Neosid company in all his circuits because he could obtain them in France, but acquiring these was a major problem. In Istanbul, many amateurs used transformers with pink and green cores that were available at Elmadağ Elektronik in the Karaköy İşhanı (unfortunately this wonderful shop closed in 2018 and turned into a “LED shop”). While working on my first Forty assembly, I had also bought and kept some of these. However, since there was no clear information about their exact values, I preferred to purchase TOKO KACSK3892 transformers from a seller in England instead of using the parts I already had. Of course, any other intermediate frequency transformer produced for 10.7 MHz with equivalent pin layout, winding ratio, and characteristics could also be used. Some friends even wound their own transformers. Later, when I built a precise LC meter, I measured these different transformers and took notes about them. Since this information might also be useful for other projects, I wrote a separate article on this subject, where you can find more detailed information.

For the single sideband filters in both receive and transmit, I used 4.9152 MHz HC-49U crystals that I was able to obtain at a reasonable price. Another suitable frequency between 4 and 6 MHz could also be used. Since it was mentioned on forums that they might have narrower bandwidth, I did not want to use the shorter HC-49S type crystals, although I also saw people who reported using them without problems. Do not forget to enter the crystal frequency into the DDS as the intermediate frequency. To do this, hold down the encoder button while applying power to the circuit. On the screen that appears, press the button again to move step by step through each digit, and rotate the encoder to change the number at that position. Since the intermediate frequency must be entered either as negative or positive, do not forget to change the sign as well. If you leave the number displayed there as “00000000”, it means that no intermediate frequency has been entered. In that case, the signal at the DDS output will remain at the frequency shown on the display.

After Construction: Adjustments and Tests

A few days after roughly adjusting the output power to about 4 watts (PEP of course) and completing the basic tuning, I managed to connect the battery that I was using to power the circuit with reversed polarity and burned almost everything from the power supply section and the output stage—from nearly the fourth mixer all the way to the final transistor! You may ask, “Why were you powering it from a battery?” In the apartment where I live, as in many buildings in our country, the mains grounding is problematic, and my simple power supplies cannot fully suppress the noise coming from the mains.

Anyway, as a result, I had to rebuild the output stage twice from scratch. In the meantime, finding the output transistor, the 2SC1971—or rather a genuine one—is very difficult. Even the ones I ordered from Greece for a high price turned out to be fake (the seller was someone I trusted; he was very surprised and immediately refunded my money, and I still continue to buy from him). Fortunately, TA3OM Mustafa Bey helped me with this issue, and the device was finally completed using the transistor he sent me. I also tried other transistors such as the 2SC2075, 2SC2078, and 2SC5739, but the highest output power was again achieved with the 2SC1971. Later, I found a seller in China who salvages these transistors from old equipment (probably medical devices), and I managed to obtain ten of them at a good price. Nine of the parts that arrived were in working condition.

After completing the receiver section of the device, I found a strong station and made the rough adjustments. I listened to traffic for a few days. Then I started working on the output stage, but because of three mistakes I made, finishing this stage took a long time. The first was that one of the intermediate frequency transformers I had removed from the Toucan and used in this build was defective. Assuming that the transformer could not be faulty, I kept struggling with the other components. Secondly, I had used a BN43-302 core instead of a BN43-202 in the L9 coil; I noticed this only while searching for another component and looking through the ferrites in my box.

After finishing the construction, I followed this procedure to make the adjustments: once I made sure that the BFOs were operating properly (using both an oscilloscope and a frequency counter), I soldered a wire to the relevant pins on the relay and shorted them so that the receiver would receive continuous DC supply voltage. After connecting a dummy load to the antenna connector and a pair of headphones to the audio output, I keyed the transmitter and spoke into the microphone while listening to my own voice. Using this method, I adjusted the microphone input level with P3, the single sideband balance with P4, and the BFO frequency with C63, leaving the settings where the sound seemed best to me. In the following days, when I had the opportunity, I connected the device to an antenna and further improved the adjustment by listening to my own signal this time with a commercial HF receiver (to make this receiver slightly less sensitive, I used only a 30 cm piece of wire as the antenna).

To house the circuit, I used the enclosure of an old power supply that I had on hand. Since its dimensions were close to those of the PCBs, making the front panel both attractive and functional was not particularly easy. I did not want to separate the display and the encoder from the DDS board. However, in order to position the board properly on the front panel of the enclosure, I had to solder the capacitor and the terminal block at the DC power input to the back side of the PCB. I chose a chassis-mount four-screw BNC connector for the antenna and a 2.1 mm connector for the DC power input. I designed the front and rear panel labels in MS Word and had them printed on suitable paper at a copy shop, then glued them onto the panels. After that, I installed the display, knobs, switches and connectors in their places and made the final wiring connections.

Next Stage: Improving the EFE-40 QRP

At first, I had prepared an old CB microphone to use with this device, but later I changed it because of a small issue that was bothering me. You can find my explanations about this and other modifications in my article titled “Improving the EFE-40.” Yes, of course there is more to come—did you think it was finished? The pursuit of perfection is an endless journey, my friends. :)

Icom IC-2AT Restoration and Modifications: CTCSS, Li-Ion Battery and USB Charging

This article is also available in the Turkish version.

The Icom IC-2AT is one of the iconic handheld transceivers of the 1980s, and even today it still enjoys a loyal following among amateur radio operators. In this article I take a closer look at the radio using my own unit as an example. I discuss the different versions of the IC-2AT, with particular attention to the differences between the North American and European models, as well as the role of the DTMF module and the process of adding a CTCSS circuit to the radio.  Restoring an older handheld transceiver and adapting it for modern repeater use turned out to be both an educational and an enjoyable project. If you are interested in vintage amateur radio gear, Icom IC-2AT modifications, or restoring classic handheld radios, this write-up covers both the technical aspects of the work and some personal observations from the process.

ICOM IC-2AT after modifications

Stumbling Upon an ICOM IC-AT at a Flea Market

If I remember correctly, one day in the fall of 1997 I ran into Erhan Ağabey (TA2DJ) in front of the Beşiktaş ferry dock in Kadıköy — the person I would later describe as the one who “infected me with the amateur radio bug.” We were both heading across the Bosphorus to attend an AKUT meeting that was going to take place at Galatasaray University. In those early days the association did not yet have its own venue, so activities such as general assemblies and training sessions were often held in rooms we arranged at Galatasaray University, where I was a student at the time. After greeting Erhan Ağabey, whom I already recognized from one or two earlier events, we started chatting. The handheld radio he was carrying caught my attention. I asked what it was and what it was used for, and as he explained, my curiosity kept growing and I continued asking questions. I remember that by the time we reached the other side he had told me quite a lot about amateur radio. He also made sure to briefly explain the features of the radio he was holding. The device in question was an ICOM IC-2E. Since I couldn’t see an LCD screen or any other kind of display, I asked him how the frequency was changed. He replied, “This radio uses mechanical switches,” and explained how the BCD switches used to set the frequency worked (the parts you push to rotate are shaped like small gears — they have little teeth).

As far as I’m concerned, that was the day I first became interested in amateur radio. I must have associated that day very strongly with the IC-2, because the radio never really left my mind afterwards. Over the last three or four years, whenever I saw the prices of these radios drop to almost ridiculously low levels in listings on eBay and similar sites, I kept thinking to myself, maybe I should buy one and at least keep it as a collector’s item. At this year’s flea market, I finally saw one of “Erhan Ağabey’s handheld radios.” There was a gentleman sitting at a table looking rather bored, with seven or eight old handheld radios and various pieces of radio scrap laid out in front of him. As I approached, I noticed a fairly clean ICOM IC-2AT among the pile — though it had neither an antenna nor a battery. In place of the battery there was an adapter attached that allowed the radio to be powered with 13.8 volts. I thought to myself, I’ll find a battery sooner or later, or in the worst case I can convert this adapter into a battery pack. As long as the radio worked, that would be enough. The case wasn’t cracked or damaged; it was simply very dusty and a bit dirty in places. After turning the radio over in my hands for a moment, I reached into my pocket and pulled out about eight dollars’ worth of loose change. I asked the gentleman if that would be acceptable for him, and when he said yes, the radio immediately ended up in my bag. By sheer luck, just a little further away I came across a cardboard box filled with thick handheld antennas from the 1980s and 1990s — the kind Americans used to call “rubber duckies.” The box had a sign on it that said “free,” and inside I found an antenna that fit the radio perfectly.

Thus the IC-2AT returned home with me. For a while it simply sat on the workbench. Those of you who are married amateur radio operators with children can probably imagine that one cannot just dive head-first into any project at any moment. Anyway, once the “quarantine” period was over, I picked up the radio again and began examining it with the help of the service manual I had downloaded from the internet. Before moving on to the results of that “inspection,” however, let me first give a little more information about the radio itself.

A Brief History of ICOM IC-2 Series

ICOM introduced the IC-2 family in 1979. The radio continued to be produced until the late 1980s and became very popular. The IC-2 series was designed for the amateur 2-meter band, the IC-3 series for the amateur 1.25-meter band, the IC-4 series for the amateur 70-centimeter band, and the IC-M series for the marine band. The IC-2A was the North American version, while the IC-2E and IC-2N were the European and Japanese versions respectively (presumably ICOM produced three different versions corresponding to the three IARU regions). According to KT2B, around 500,000 units of the amateur version were sold worldwide between 1979 and 1987 — yes, half a million.

In fairness to other manufacturers, it should also be noted that when the IC-2 was introduced, synthesizer-based handheld radios such as Henry Radio’s Tempo S1 and Yaesu’s FT-207 had already been on the market for a year or two. In fact, from a user-interface perspective the IC-2 resembles the Tempo S1 quite a bit — you set the frequency using BCD switches just like on that radio. The FT-207, on the other hand, allows direct frequency entry through a keypad, which makes it technically more advanced than the IC-2 in that respect. However, unlike those models, the IC-2 came with a removable battery pack rather than an internal one, and ICOM also offered a variety of battery options with different capacities and sizes, along with other accessories that appealed to amateur operators. In terms of size and weight it also had a slight advantage over its competitors. Perhaps these factors explain why it became so popular.

One of ICOM IC-2's precursors: Henry Radio TEMPO S1

Top panel of TEMPO S1

Top panel of ICOM IC-2AT

The unit I bought is the North American version, but not the 2A — it is the 2AT, which means there is a DTMF keypad on the front panel of the case. Naturally, there is also a DTMF module inside. Interestingly, in North America the non-AT version, which we could call the plain A version, is actually much harder to find. In Europe and Japan the situation seems to be the opposite: I have seen many 2ATs online, but I have never come across a 2ET or 2NT. The European version also has an interesting feature: it includes a 1750 Hz “tone burst” option, and the volume potentiometer is switched so that it can trigger this circuit.

As far as I can tell, amateur operators in North America at the time generally preferred to buy this radio with the DTMF option installed. This is not surprising, because during the 1980s, when the radio was most widely used, many repeaters in North America — especially in rural areas — were equipped with telephone patch connections. Many operators, particularly in mobile installations, wanted their radios to include DTMF capability. In a time before mobile phone technology existed, being able to access the telephone network from miles away via radio must have been an incredible convenience. Even today, at the flea market I mentioned earlier, I still see phone patch circuits and DTMF keypads built from parts salvaged from push-button telephones being sold every year.

ICOM IC-2AT: Technical Specifications

The 2AT version operates in the 144–148 MHz North American 2-meter amateur band. With its standard 8.4-volt battery, the output power is around 150–200 mW in the low-power setting and about 1.5 W in the high-power setting. Yes, you read that correctly — 1.5 watts — because the output stage consists of a single humble 2SC1947 transistor, whose metal case is soldered to the radio’s internal metal chassis to help with cooling. Of course, for those of us who have become accustomed to modern handheld radios equipped with lithium-ion batteries, DC-DC converters, and FET output stages delivering 7–8 watts, 1.5 watts may seem quite low. But in those years 5 watts in a handheld radio could only be achieved with a 13.8-volt supply. If you flip through amateur radio magazines from the early 1980s, you’ll notice that one of the most common advertisements is for small amplifiers designed specifically for handheld radios, and that’s no coincidence. That said, I always say the same thing: in most situations output power is not nearly as important as the operator’s own skill. In my view, the real art is communicating with the lowest possible power. From that perspective, 1.5 watts is perfectly sufficient for local communication.

On the top of the radio there are the earphone and microphone jacks, the volume control, the squelch control, the BCD switches used for frequency selection, a power switch, and another switch that determines whether the last digit of the frequency is “0” or “5.” On the back there are three additional sliding switches. The top one selects the output power level. Below that is the simplex/duplex switch, and the last one sets the repeater offset, either plus or minus 600 kHz. In short, everything is extremely simple, and even a radio operator seeing the device for the first time could probably get used to it within a few minutes. On the front panel, under the speaker, there is the DTMF keypad I mentioned earlier. Setting the frequency is very straightforward. For example, if the BCD switches display “727” and the 5 kHz switch is in the zero position, your frequency is 147.270 MHz. If you flip the 5 kHz switch to the other position, the frequency becomes 147.275 MHz. Of course, nowadays with our touch screens and digital interfaces this system may appear somewhat primitive, but every technology has its own advantages. Here’s an example: have you ever seen what happens to LCD screens at –20°C? I see it quite often here in Canada (!). This old IC-2, on the other hand, doesn’t suffer from problems such as the screen freezing in cold weather.

ICOM IC-2AT Back

A Broken Volume Potentiometer

After this bit of technical background, we can return to the repair and modification work. As I mentioned earlier, after some time had passed I applied power to the radio and checked whether it was working. As far as I could tell, it was transmitting properly on the frequency that had been set, but there was no reception. Fortunately, I usually work on these things late at night after everyone in the house has gone to bed, so the surroundings were very quiet. In that silence I noticed a very faint hiss coming from the speaker — so faint that you could hardly detect it without pressing your ear right up against it. When I keyed the other handheld radio I had nearby, the hiss stopped, and when I released the PTT it returned. At that moment I realized that the squelch circuit was working, and the receiver section was most likely functional as well, but there was clearly a problem somewhere in the audio path. As I continued experimenting, I discovered something else that encouraged me: if I turned the volume potentiometer all the way up and pressed down on it slightly with my finger, the speaker suddenly produced sound — admittedly at a rather unpleasantly loud level — and when another radio transmitted, I could hear it perfectly clearly. At that point I began to suspect that the problem was related to the potentiometer itself.

When I opened the radio and examined the volume potentiometer, I noticed a crack in its phenolic base. This crack had most likely interrupted the carbon track, meaning the contacts were not making proper connection until the control reached the extreme end of its rotation. To test this theory, I first soldered another potentiometer to the board using wires without removing the original one, and the result immediately confirmed my suspicion: reception worked perfectly, just as transmission did. The next step was therefore to remove the faulty potentiometer. However, the internal layout of the radio’s case and circuit boards made this somewhat difficult. In order to remove the top cover, you would normally have to detach the BCD switches from the board to which they are soldered. But that board is connected to the PLL board through flexible film-type traces, and removing those without melting the film is not really feasible. So I chose another approach: I cut the shaft of the potentiometer with a small hacksaw. Then I grabbed the small board that held both the volume and squelch potentiometers with needle-nose pliers, and by bending it carefully from the other side with another pair of pliers I managed to snap the board cleanly in the middle. After pulling the defective potentiometer downward and removing it, I installed another potentiometer of the same size in its place and completed the wiring. This way, the connections for the squelch potentiometer remained intact.

At left, the broken potentiometer 

After sawing the shaft of the broken pot

Broken phenolic base of the pot and the removed half
of the circuit board

After the replacement is complete

Replacing the Electrolytic Capacitors

As a second step, I did something I almost always do when working on equipment that is 40–50 years old: I replaced the electrolytic capacitors. The metal chassis inside this radio opens like a book, with the PLL board on one side and the main board on the other. A few of the electrolytic capacitors on these boards were located along the edge and were soldered directly along the line where the board meets the chassis. Removing those would have required heating the circuit far more than I was comfortable with using the tools I had available, so I left them alone. However, I replaced all the remaining electrolytic capacitors with new ones.

Main PCB of IC-2AT: Before the old caps are removed

After new caps are installed

The view from the (inner) back side of the PCBs

Building a new Li-Ion Battery with Internal USB Charger

The third thing I tackled was building a battery for the radio. The 13.8 V adapter that came with it was large enough to hold two or three AA cells. Since I had no intention of running the radio from a 13.8 V supply, I removed the circuitry inside the adapter housing. Then I stopped by a shop near where I live that sells Amazon return items at very low prices, and there I bought a children’s toy containing two AA-size lithium-ion cells for about two or three dollars. I also ordered from China a two-cell BMS (battery management system) suitable for lithium-ion cells, a few USB-C sockets, and an automatic charging module capable of boosting 5 V to 8.4 V. Just to give an idea, the total cost of all these parts probably did not exceed five dollars. After assembling everything, I ran several charge–discharge cycles and confirmed that the system worked properly. The charging module charges the series-connected lithium-ion cells up to 8.4 V and then stops at that voltage (and of course the BMS connected to the cells already protects them by cutting off the circuit in cases such as undervoltage or overvoltage).

Aside from fitting all these parts neatly inside the battery housing, the only modification I made was removing the LEDs on the charging module and replacing them with red and green LEDs mounted on the back of the battery case so that they would be visible from the outside, wiring them to the circuit with short leads. In this way I ended up with a battery that can be charged with almost any USB-C charger. Charging takes about 90 minutes. While charging is in progress the red LED lights up, and when charging is complete the green LED turns on. When fully charged, the cells provide 8.4 V, which is the same voltage as the radio’s original battery pack. When the voltage drops slightly below 7 volts, the BMS cuts off the output. The capacity of the two cells is 600 mAh. That may sound small, but the radio draws about 40 mA in standby (including the CTCSS modification I added), roughly 100 mA while receiving, and around 500 mA during transmission (at 1.5 W). That means with a reasonable usage pattern — say 10 minutes receiving, 10 minutes transmitting, and 40 minutes on standby per hour — the radio consumes about 120 mA per hour, which corresponds to roughly five hours of operation. Considering that these days the repeaters are almost silent most of the time, in practice this battery provides something like 8–10 hours of use. Of course it would be possible to increase the capacity by finding a larger battery case and adding more cells, but I’m not sure I’ll bother with that…

All the components before the final assembly: Li-Ion cells,
BMS (fixed to the cells), USB-C port, and the "smart" charger

After the completion: The difference is minimal 
compared to original (USB-C port)

Charging...

Charging complete!

Adding a CTCSS Circuit

The most important addition I made to the IC-2 was a CTCSS circuit. To accomplish this, I followed the same approach I had previously used when adding CTCSS to a KDK mobile radio: I programmed an ATtiny85 microcontroller to generate the twelve tones I needed, then connected it to the radio through a low-pass filter and an amplifier. Let me try to explain the details step by step:

• I chose the ATtiny85 for this job because the modules built around it are extremely small. I had worked with these modules before, and they can easily be programmed using an Arduino Uno. Although they are limited in terms of input/output pins and memory, they are perfectly adequate for tasks such as generating CTCSS tones.

• I didn’t write the program for the microcontroller myself — ChatGPT did. Don’t laugh :) My programming knowledge is quite limited, and if I had tried to write the code myself it probably would have taken me more than three months. That said, it didn’t take five minutes with ChatGPT either. It took about three weeks, working on it two or three nights each week. The reason is that working with so-called artificial intelligence does not end with writing two sentences describing what you want. You also have to manage what it produces with the correct commands and definitions, and correct it when it starts producing nonsense. During those three weeks I repeatedly uploaded the program generated by ChatGPT to the module and checked the circuit I had built on a breadboard, verifying with a frequency counter and an oscilloscope whether the functions I wanted were working properly.

• Beyond the functionality of the program, it was also necessary to check the accuracy of the generated tones. At the beginning the module was producing tones that were about 4–5 Hz off. I noted what the microcontroller was generating for each tone, and then ChatGPT added a correction factor in the program for each tone based on the difference between the expected and the actual frequency. Out of curiosity I later tried three different ATtiny85 modules, and each one required its own separate calibration. The correction factors that worked for one module did not work perfectly for the others.

• What does the program inside the ATtiny85 do? When the radio is keyed, the voltage arriving at pin 0 of the module rises. The microcontroller detects this and generates the selected CTCSS tone as a PWM signal on pin 2, based on the tone that was previously chosen using the buttons connected to pins 2 and 3. In addition, when the buttons connected to pins 2 and 3 are pressed simultaneously for more than three seconds, the tone generation function toggles on or off. This status change is indicated visually by the LED connected to pin 4, which flashes three times.

• In principle it would be possible to include all existing CTCSS tones in the program. However, I looked up the list of repeaters in my area and included only the twelve tones that I use most frequently. I divided them into two groups of six, assigning one group to each button. For example, if you press button 2 three times, you select the third tone of the second group. The LED connected to pin 4 then flashes to indicate which tone has been selected. In this example, the LED flashes twice, then after a longer pause flashes three more times, confirming the chosen tone.

• Once the tones were being generated at the correct frequencies, I added an RC low-pass filter to shape the square-wave signal into something closer to a sine wave. I set the –3 dB cutoff frequency slightly above the highest tone I intended to use, and chose a five-pole filter. The FFT function on the oscilloscope is extremely useful for this kind of work. Looking at the FFT display, I could see that with only two or three poles it was impossible to sufficiently suppress the harmonics and obtain a clean sinusoidal signal.

That gave me a CTCSS circuit with the functions I wanted. The next step was to remove the DTMF module on the main board and use the contacts there to connect the CTCSS output to the radio’s tone input. I have no real need for DTMF tones — frankly, I don’t know who still does — and in the worst case, if I ever desperately needed them, I could always just hold my phone up to the radio’s microphone... Anyway, I connected the circuit the way I had planned. But as often happens, things did not work out quite as neatly in practice as they had on paper: the signal coming through the filter was greatly attenuated, so it was not making it to the other side properly. I tried increasing the amplitude using a transistor. At that point the repeaters near me began to open, and the other radio I was using for testing could hear the CTCSS tone and open its squelch as well... but not every single time. Since I needed an amplifier that would operate reliably and consistently, I started looking at op-amps. Then I got a bit clever and tried to use an LM386 audio amplifier, but when that failed to give good results, I was brought back to reality by dear Cem Ağabey (TA1S), whose “go take a look at the output impedance range of that IC!” slap woke me up. Following his advice, I went back to the TL082 (“bias it at half the supply voltage and it’ll work beautifully!”). And indeed, with the TL082 I ended up with a nice, clean signal of sufficient amplitude going into the radio’s tone input.

First tests of the CTCSS circuit on a breadboard

More tests with the IC-2AT connected

Finally a clean and strong tone with TL082

At this point it may also be useful to briefly mention where the CTCSS signal is injected into the IC-2. Of course, this can easily be found by looking at the schematic, but a little explanation doesn’t hurt. Assume you are holding the radio with the antenna pointing upward. On the main board, once the DTMF module has been removed, you will see four contacts. From left to right these correspond to the tone input, mute, DC supply for the tone circuit, and ground (chassis). If you happen to have an IC-2E, you will instead see empty pads in that location, which can be used for soldering. As far as I can tell, if the 1750 Hz tone burst circuit is removed, the remaining contacts can also be used; this is shown in the schematic. The important point is that the tone must enter the “limiter amplifier” section of the main board at the point labeled “2.” At the same time, we do not want the tone to enter the audio amplifier, because unlike DTMF tones, we do not want to hear the tone we are transmitting ourselves. For that reason, the connection points marked “1,” “4,” and “5” on the main board are not used in this case.

The DTMF board that turns the IC-2AT into the ‘T’ version.
I reluctantly cut this flexible ribbon cable on the keypad side
and archived the board together with it

DTMF section from the schematics

The white arrow indicates the 4 pins
I mentioned above

At this stage the task became physically fitting the ATtiny85 module, the filter, and the TL082 amplifier inside the radio. There is truly very little spare space inside this device for adding extra circuitry. In fact, I must admit I pushed the project forward somewhat out of sheer stubbornness. As I tried to show in the photographs, there was a small amount of free space on the PLL board above the TC9122 integrated circuit. On the main board, space only became available after the DTMF module was removed. Even so, there still wasn’t enough room on either side to fit both the ATtiny85 module and the filter and amplifier circuitry together. In the end I decided to place the microcontroller on the PLL board side, and install the filter and amplifier on the other side.

In order to mount the filter and the TL082 op-amp properly, I prepared a small piece of perforated phenolic board sized to fit the available empty space. Then, to determine how to arrange the components in the most suitable way, I made a scale model of this little board and used it to lay out the integrated circuit along with the resistors and capacitors that make up the filter, finalizing the placement that way. I also did the soldering according to this model. After that, I soldered this small board to the contacts left behind by the DTMF socket. Here I used three contacts: tone input, ground, and 6 V DC supply. This 6 V only appears when the radio goes into transmit. I used it to power the TL082, and by means of a voltage divider I also derived 3 V, which I used to bias the TL082. By the way, in order to fit all of this onto a board measuring only 2 cm by 2.5 cm, I had to abandon through-hole components and use 1206-size SMD resistors and capacitors, as you can see in the photographs.

Optimizing the component layout on a paper model

Soldering the components on the "real" board

"All-up" testing

The board carrying the filter and the amplifier is finally
soldered to the main board of IC-2AT

On the other side, I prepared the wiring for the ATtiny85 module and glued a tiny board carrying a voltage divider made with SMD resistors to the module using acrylic adhesive. In this way, when the PTT is pressed, not only does the module receive the radio’s supply voltage, but half of that supply voltage also appears on pin 0, making that pin go logic high. After covering the module with Kapton (polyimide) tape, I connected it to the DC supply line that becomes active during transmit. I routed the other wires through the opening at the bottom of the chassis to the main board side, connected the button inputs to the 1 and 2 buttons of the DTMF keypad, which I had otherwise disabled, and connected the LED output to a red LED that I mounted by drilling a suitable hole in the front case and fixing it in place with epoxy so that the light would be visible from the outside.

At the other side of the ATtiny85 module, the voltage divider
(made of two SMD resistors)

After insulating the module with polyamide tape
and tucking it into the PLL board area…

Final Tests and the Result

Once all the connections were complete, I monitored the transmission with another radio and adjusted the tone level to the lowest possible setting using the R77 potentiometer on the main board, and then (finally) closed the case. On the underside of the radio there is now a list of the 12 available CTCSS tones. By pressing buttons 1 and 2, I can change the tone when needed, or disable it entirely.

After all the connections are made...
The red SMD LED is below the speaker, at right

With that, I completed another project that taught me quite a few things along the way. I managed to bring back into service a handheld radio that I like very much and that could now almost be considered an antique. I also enjoyed the process immensely.

If you have read this far, thank you very much. If you have any questions, feel free to ask.

ICOM IC-2AT as I found it at the flea market