A Multimode Radio Transceiver Modulator

Following on from the last blog post that I made in July 2024, I present the progress I have made with the microphone preamplifier stages for the Irwell HF Transceivers modulator.

As per usual time has moved on and I have been sidetracked by other projects but with that said, the Irwell HF Transceiver project is and has always been a long-term project.

Since my last update, I have produced and developed various prototype circuits in a bid to get the modulator finished, some of which are illustrated below.

During my time experimenting with different circuits for the modulator, my attention was drawn to the matter of how I was going to produce a narrow band frequency modulation and it soon became apparent that this was not going to be an easy task.

Ideally, it would be nice to modulate the signal produced from my VFO but as the SI5351 device I used does not provide a means of connecting to its internal VCXO that is not an option. It may be possible to modulate the Si5351’s xtal reference oscillator but that might not provide good FM modulation, especially given Xtals do not tend to pull evenly in both directions so this method is not really ideal. Other options include phase modulation and software generated FM but the latter is beyond my programming capabilities. In modern times this is where SDR technology is king. If you want to add another mode of operation, such as FM, you simply write some additional software code. No faffing about with PLLs!

It is at this point the words “Fools rush in where wise men never go” seem rather appropriate, in particular, with reference to my desire to include FM as a mode of operation. It could be argued that the inclusion of FM mode in a HF transceiver is not really a necessity but for the sake of completion and achieving my end goal, I want to include it. The inclusion of FM paves the way for transverting to higher frequencies or enjoying some 10-meter FM activity when the band is open so I think it is worth the effort and development time.

I spent a fair amount of time searching the internet for FM inspiration but It appears there are not many conventional homebrew multimode transceivers that include FM modulation. During my searching, Google led me to the web site of Paul VK3HN, who has built some impressive homebrew transceivers but what really caught my eye on Paul’s website was a statement he made and I quote,

“There’s a reason why most homebrew transceiver kits and scratch-built projects are monoband and single mode — there’s a chance you’ll finish it”

Reading Paul’s statement felt like a reality check and one that should be etched into the history of world-famous quotes!

Having got as far as I have with my transceiver build, I can unequivocally say that taking on the challenge of building a conventional multimode transceiver is not a simple task and it presents many challenges. The work I have done on the modulator is a prime example of this when you consider the following key points that had to be achieved,

• Different audio levels for each of the modulation modes.
• Tailored frequency responses for the different modes.
• Finding a suitable solution to create Frequency Modulation
  
• Pre-emphasis for Frequency Modulation.
• Audio clipping or limiting
• Audio Compression
  
• Switching for the different modes of operation.
• Interface path for my digital modes interface module.
• Achieving the design goals with minimal components.

I managed to fulfil all of the above design goals during prototyping though it was very much a stop-start journey taking longer to achieve than anticipated, this has therefore contributed to the long delay since my last blog post!

One of the key points that proved difficult for me to resolve was the FM modulation and, in the end, I opted for using a dedicated TSA6057 PLL IC to do the job. This was no easy task; understanding the IC's data sheet was a challenge and then there was the job of writing the microcontroller software required to programme the PLL, not an easy task when you have no software programming background!

Whilst my PLL prototype hardware and software work fine, I am mindful that the TSA6057 IC is actually obsolete but seen as I had some in stock it was an ideal candidate for experimenting with. The FM PLL will remain as a separate module to the main modulator that way It can be redesigned & replaced at a later date if needed.

There will be a dedicated blog post on the FM PLL sometime in the future, but for now I want to return to the main modulator and elaborate on some of the bullet points made above.

Different audio levels for each of the modulation modes & tailored frequency responses for the different modes.

In my previous blog post I presented a simple method that allowed me to fine tune the microphone amplifiers gain and low frequency respone for each mode of operation. The reference image below illustrates this.

The variable preset resistors PR6, PR7 & PR8 are used to set the microphone gain and capacitors C59, C61 & C63 set the low frequency audio response. A larger value capacitor increases the lower frequency content present in the audio, a lower value decreases it.

In my latest design the preset resistors have been replaced with fixed resistors and the gain per mode is now set later on in the circuit, everything else remains the same.

Pre-emphasis for Frequency Modulation.

In narrow band FM communication, pre-emphasis is needed to maintain a good signal-to-noise ratio. In essence pre-emphasis can be simplified as an increase in microphone audio level that is proportionate to an increase in frequency. From a technical point of view, it is defined as an increase that follows a 6dB per octave rate meaning as the frequency doubles the amplitude increases by 6dB.

Now, as we have already ascertained the low frequency response of the microphone preamplifier can be tuned with a single capacitor, my next task was to see if a suitable value capacitor would get me the 6dB per octave needed for FM pre-emphasis.

Referring to the image above, with a value of 22nF for C61 and a 3.3K resistor in place of PR7, I obtained the following bode plot in FM mode.

Using the above bode plot the dB per octave value can be calculated using the following formula:

dB/Octave = (End Level – Start Level) / log2(End Frequency / Start Frequency)

I calculated a value of 5.63 dB Per Octave which may not be a military spec but it is pretty damn close to the 6dB wanted. There is room for fine-tuning component values but for now I am happy.

Audio Compression.

Well, I thought I would save some development time when it came down to compression so I opted to do the job with an Analogue Devices SSM2167 which is a complete microphone signal conditioning system on an integrated circuit. You can buy the IC as a completed module off eBay, Amazon or Ali-Express.

PCB Shown bigger than actual size

I purchased the above module and set about integrating it into my chain of prototype boards but ran into issues.

In order for me to use the SSM2167 as a compressor, I needed it to co-exist and proceed with my existing microphone amplifier and low frequency mode tuning circuits. I made the assumption that some gain before the SSM2167 would be beneficial in operating the noise gate facility but in practice there was way too much gain from my existing microphone amp feeding the 2167 and any attempt to lower this played havoc with my low frequency tuning arrangement. I also played around with the value of the resistor used to set the 2167's compression ratio but failed to get acceptable audio.

My thoughts were drawn to the possibility that the actual SSM2167 device on the module could be a fake Chinese clone so I ordered a pack of genuine parts from RS Components in the UK but, in all honesty, they made no difference so I scrapped the idea.

I am aware hams have used these modules for BitX and other homebrew transceivers but in my experience, I think they could do more harm than good unless proceeded by a suitable limiter circuit.

In the end I decided that this part was just not for me or my project and I went with a more conventional compressor using a field effect transistor in the feedback loop of an operational amplifier. The circuit is simple and works well without over cooking the audio.

Compressor based on ETI Magazine December 1978 Article.

 

 Audio clipping or limiting.

There are various ways to carry out audio limiting but the simplest and most common method uses diodes to clip the signal. The problem with using diodes is that they are prone to producing harmonic distortion. I did try introducing diode clippers into my design mainly from an experimental & interesting point of view but, as expected they clipped at a higher level than I wanted and produced distortion.

In my quest for an alternative, I carried a web search to see if there were any dedicated audio limiter IC's available and the results led me to a device manufactured by Nisshinbo Micro Devices (former New Japan Radio Co.) the NJM2762.

The NJM2762 is a 2-channel device that senses both input signals and if either channel exceeds a preset limit, it will reduce both channels together so that the outputs are always balanced. A look at the parts data sheet revealed the following features:

• Wide Operating Voltage - 2.7V to 13.0V
• Variable Limit Level Set By A Resistor - 180mVrms to 1Vrms
• Low Output Noise - 90dBV max.
• Low Operating Current - 3mA max.
• Limiter On/Off  Control.

The device is in current production and available in two SMD type packages, MSOP10 & SSOP14. RS Components stock the part in the MSOP10 package and sell them in packs of 2 for £3.40 including VAT, I considered this inexpensive so I ordered a pack to try out.

Below is the schematic of my prototype NJM2762 audio limiter.

NJM2762 Audio Limiter circuit.

 

And next the prototype board I built on stripboard.

 

Audio Limiter Prototype Board Not To Size.

I fed both of the input channels of the chip with the same audio signal via two DC decoupling capacitors; I did this because the chip expects to see a signal on both channels at all times in order to balance the output; if only one input received a signal, I figured the chip would not perform as intended.

The next image is a test of the prototype board being fed with a 1-volt peak sine wave at a frequency of 1 kHz. The limiter was set to produce an output of 0.5 volt peak.

 

NJM2762 Limiter IC Scan One.

A further test was carried out to see how the limiter performed when the input signal level was increased to 2 volts peak.

NJM2762 Limiter IC Scan Two.

As can be seen, the NJM2762 does a fine job at limiting the output signal to 0.5 volts so I think it is safe to say this IC has great potential for speech processors and ham radio projects.

 

 Switching for the different modes of operation and managing audio levels for each of the modulation modes.

I have always had good results using the CMOS CD4066B Quad Bilateral Switch for switching audio signals; in fact, I used this IC in my audio amplifier module, where it has proven to be reliable and clunk free.

 

Audio path switching is therefore borrowed from my amplifier module and reworked to include level adjustments for each mode of operation. The completed circuit is shown below.

If you are wondering why there is an op-amp after the switch in the FM audio path, it is because the FM PLL VCO requires more audio drive than the MC1496, the IC responsible for producing the other modulation modes.

The test PLL VCO I built for FM required around 1.5 volts peak audio to produce sufficient deviation, now taking into account the NJM2762 IC has a limit range of 180mVrms to 1 volt RMS this leaves me short of audio, resulting in low FM deviation. The op-amp IC4B was included to ensure I could produce sufficient audio for FM by giving a little boost to the audio signal after it has been limited.

Bringing the prototype modules together.

Having worked my way through various prototype circuits and satisfied my design goals, the next job was to pull everything together and produce a schematic diagram and a suitable printed circuit board.

The final modulator is divided into two schematic diagrams, the first contains the microphone amplifier, AF filtering, processing, limiting and switching. The second schematic covers the MC1496 - SSB, AM & CW modulator.

NOTE:
Original schematics from the blog post have been removed and replaced with updated schematics, see project update lower down in the blog post.

The SSB, AM & CW modulator based around the MC1496 intergrated circuit was covered in detail back in 2022 so there is little point in covering it again here. If you do want to read that blog post, click on the link below.

 To view my 2022 Modulator blog post article click here.

From the schematics a custom PCB was designed and sent to the JLCPCB fabrication plant in China for manufacture. The PCB is a two-layer board that requires some wire links installed on the underside; I could have avoided the use of wire links if I had made the board with four layers but that would have increased the manufacturing cost.

Here's an image of the PCB's top layer after populating it with components. Note I used good quality Polyester Film Capacitors in the main audio signal path.

Actual PCB Size is 100 X 90mm.

And now an image of the bottom layer where you can see the wire links.

Actual PCB Size is 100 X 90mm.

The tiny SMD IC marked IC5 is the NJM2762, I decided to mount this on the underside of the PCB so that it could be removed and replaced in the future with ease if needed. Had I have mounted it on the top layer, it would have been heavily surrounded by other components, making future removal difficult.

It can be seen from the above images the PCB is not fully populated with parts and finished. I wanted to finish the board and carry out testing before doing the blog post but due to a personal injury and my right arm being in a sling, I have not been able to do this. The blog post is long overdue so rather than delay it even further, I decided to post where I am up to and I will update it as soon as I can, then I will move on to the next module.

Project update - 13th May 2025

 

Having recovered from a shoulder injury, I have now been able to complete the "Irwell HF Transceiver Modulator". 

The completed board worked the first time; I did, however, have some issues with the frequency responses produced by the different modes of operation. A fault was traced to a dodgy 2N7002 transistor in the FM mode tuning circuit that was skewing the waveforms in SSB and AM mode. I replaced the faulty 2N7002 transistor, the waveforms improved, but they were not identical to those produced by my prototype modules.

In order to fine-tune everything, I decided to build a second board and thoroughly tested each stage as it was built while fine-tuning component values as I went along. The end result was worth the effort, and the waveforms produced are now excellent. I updated the first board with the component changes, and I now have two identical boards that perform exactly the same.

I have removed the schematics from the original blog post that I made on the 17th of April and updated them to reflect the changes.

Here is the updated microphone amplifier schematic with updated component values.

Click image to enlarge to full size!

 

Note that some capacitors have been replaced with resistors; these are marked on the schematic. 0805 SMD Resistors fit fine across the pads previously allocated to the capacitors on the PCB. 

 

The updated schematic also contains a BAV99 dual diode pack that is connected to the  audio signal path before it reaches the MC1496 modulator IC.  Fitting this part is optional, but it may be useful as a secondary clipper/limiter to catch any overshoot from the NJM2762 limiter IC. The PCB does not contain a location for the BAV99 but it can easily be tagged onto the bottom of the board. A suitable location would be the negative pad of C15 along with an adjacent ground point that may require the removal of a little solder mask to create a solder joint.

Next is the updated schematic for the MC1496 modulator.

Click image to enlarge to full size!

The finished modulator produces good sounding audio and with the addition of the compressor and limiter the audio is boosted without being over the top. I will test the modulator further once I finish the transmit sections and can get some on air test reports.

Next I am going to work on lowpass filtering where I will be taking a look at solidstate high voltage switching circuits so keep an eye out for that.

Project files will be made available via the Groups.io platform by joining my G6LBQ community group, where you can discuss my projects, ask questions and help others.

Joining my group is free. Just click on the button below.

 

 

Until next time... 

 

 

73's From Andy G6LBQ

Its all about the Radio Ga Ga...