Back in October 2022 I posted details on the blog of a of a "High Performance AM, CW & SSB Modulator for The G6LBQ Irwell HF Transceiver" this module was somewhat experimental and for me a practical introduction to the MC1496 modulator IC. The module contained no microphone amplifier stages as it was my intention to add these as a separate module at a later date.
To view my 2022 Modulator blog post article click here.
In
this blog post I get to grips with the microphone pre-amplifier and share some of my design methods and thought processes along the way. Note It is not my intention to provide a step-by-step guide to the design but more of an insight and reference.
The success of this module is crucial in order that I can continue development on the final transmit stages and achieve finished project status.
As a starting point I built a simple microphone pre-amplifier based on a Texas Instruments TL071 FET Op-Amp IC using spice simulation software. This approach makes it very easy to analyse the behaviour of the circuit and make changes to component values and therefore speed up the development process.
The circuit for the software simulated pre-amplifier is shown below.
With the component values shown I used the simulation software's Bode Plotter to analyse the frequency response and gain of the pre-amplifier and produced the image shown below. Note the 1k resistor and 10K preset that form part of the gain and equalization circuit were set for a combined value of 3K.
My plan is to use an Electret Condenser
microphone with the transceiver, this type of microphone produces
between 10-20mV at a normal speaking level. The application examples in
the MC1496 data sheet suggest a modulating input signal of 300mV-rms so
in terms of voltage gain the simulated circuit is ideal.
As can be seen from the above image the gain of the pre-amplifier circuit is 32.7dB and the -3dB bandwidth is 9.15KHz. In terms of voltage gain the circuit when fed with an input signal of 10mV produces a peak output voltage of 430mV into a 10K resistive load.
The circuit does require some attention to the frequency response, the lower frequency drop off below 100Hz is acceptable for now and can be easily tweaked (more on that later) however the higher frequency response is flat up-to the -3db point of 9.15KHz which is far higher than is needed for radio communication purposes. For SSB operation the audio speech bandwidth should be in the region of 200Hz to 3KHz so work needs to be done to address this.
Before continuing development of the microphone pre-amplifier, I decided to build the simulated circuit on a breadboard and carry out some real hardware tests, whilst software simulation is ideal for a base design and testing it should be noted that the simulation assumes everything is ideal so it will be interesting to make a comparison with actual hardware.
Here
is the microphone pre-amplifier built up on my breadboard, I will use this in conjunction with the simulating software so I can develop
it further and easily swap and change components or add additional stages. Again, the 1k resistor and 10K preset that form part of the gain and equalization circuit were set for a combined value of 3K.
Next is an image that shows a sweep of the bread-boarded circuit using a Digilent Analogue Discovery 3 USB multi-function test and measuring device.
 |
| Click image to enlarge to full size! |
As can be seen from the scan, the microphone pre-amplifier built on breadboard produces an identical scan to the one created in the simulation software however there are some differences.
The -3dB bandwidth of my bread-boarded circuit is 8.37KHz compared to that of the simulated circuit which is 9.15KHz.
If you look at the lower frequency -3dB points there is only a small 2.4Hz difference between the simulated and hardware circuits, the main difference is in the high frequency -3dB points where there is a difference of 780Hz (9.29KHz minus` 8.51KHz).
The bread-boarded circuit produced a gain of 33.23dB which is within 1dB of the simulated circuit, this means the voltage gain is also going to be near identical.
It was an interesting experiment to compare the software simulated audio pre-amplifier against the same circuit built with real hardware and it proves how useful simulating circuits in a software environment can be.
My next task was to look at improving the overall frequency response of the amplifier and in particular filtering out frequencies above 3KHz, this can be achieved by adding additional op-amp stages so I changed the single TL071 Op-Amp IC to a dual TL072 device.
The TL072 dual op-amp has a different pin configuration to the TL071 but fundamentally the TL072 is just two TL071's in one IC.
The second op-amp in the TL072 was configured as a unity buffer stage and low pass filter as per the schematic below.
 |
| Click image to enlarge to full size! |
The
low pass filter stage was designed with the aid of Texas Instruments
"Filter Pro" software version 2.0, this version was superseded by V3.0
but I personally prefer the version 2. The software was originally available as
a free download from the TI website however it appears the software is
no longer supported or available.
Here is a screen capture from the Filter Pro software taken from my computer.
 |
| Click image to enlarge to full size! |
From the above image you can see that I defined my low-pass filter as a 2 pole Sallen-Key circuit configured as a Butterworth filter type. Note! that I set the Cutoff Frequency to 4.5KHz as I found the software's suggested component values when set to a Cutoff Frequency of 3KHz actually produced a filter with a -3dB bandwidth of 2.7KHz.
Using the component values suggested by the TI software I updated my breadboard circuit and carried out another scan with the Digilent Analogue Discovery 3 test device which produced the following plot.
 |
| Click image to enlarge to full size! |
The -3dB bandwidth has now been reduced from 8.37KHz to 3.07KHz which is exactly what is required for voice communication so all in all this is a great result.
When you compare this scan to the previous one without the LPF you can see that the filter Is doing a great job at removing everything above 30KHz. For clarity I added a third marker to the scan at a frequency of 10KHz and this shows that the signal is down 18dB from the wanted frequency range.
While good progress has been made at filtering out the audio frequencies above 3KHz I feel there is still some room for improvement by reducing the angle of the slope beyond the 3KHz point.
The filter built so far has two poles and by increasing the number of poles in the circuit the angle of the slope should become steeper and the unwanted frequencies further reduced in amplitude.
Once again I used the Texas Instruments "Filter Pro" software to produce a 4-pole Sallen-Key circuit configured as a Butterworth filter with a Cutoff Frequency of 3.5KHz, this provides a bit of head room for fine tuning component values if needed.
Here is a screen capture from the Filter Pro software after I configured it for the 4-pole filter.
 |
| Click image to enlarge to full size! |
Next is an image of the breadboard populated with the additional 4 pole filter components, note the circuit now contains a second op-amp IC (TL071) which is required for adding the additional 2 poles of filtering.
Here is a plot of the bread boarded microphone amplifier with the 4 pole LPF filter.
 |
| Click image to enlarge to full size! |
The scan shows that the higher frequency curve is now much steeper and at 10KHz the signal is down to the 0dB point, the 2-pole filter only managed to reduce the 10Khz signal down to 15.27dB so clearly the extra 2 poles of filtering were worth the effort.
The overall -3dB bandwidth has now been reduced to 3KHz which is exactly what I set out to achieve.
I had to fine tune some of the component values to optimize the frequency bandwidth and gain which resulted in the following circuit diagram.
 |
| Click image to enlarge to full size! |
The TL072 has the same pin-out as the popular NE5532 op-amp so whilst I had the opportunity, I decided to repeat my test using this IC. The test revealed that the NE5532 has an identical gain and frequency curve to that of the TL072 thus concluding that the NE5532 could be used as an alternative device in this application.
When I built the experimental modulator in 2022 and carried out some preliminary tests, I noted that the optimum audio level required to modulate the MC1496 varied depending on the mode of operation selected (AM/SSB) so something needs to be done to manage this.
The simplest way to regulate the microphone gain would be to include a potentiometer on the output of the pre-amplifier's op-amp stages but after some experimentation I opted for a more effective solution that also provides a means of fine tuning the frequency equalization for each mode of operation.
The circuit I developed is shown below and provides individual gain adjustment for AM, FM and SSB modes.
 |
| Click image to enlarge to full size! |
Note
by changing the value of the capacitor in series with the 1k resistor
and 10K preset the lower frequency response can be fine-tuned which is particularly beneficial to AM mode where additional bandwidth can be utilized to create a warmer sounding audio.
Whilst the microphone pre-amplifier includes provision for FM mode this requires further development and will be added as an upgrade once the rest of the transceiver is complete.
Now that I have developed the microphone pre-amplifier to support the different voice modes the next job was to look at interfacing my digital modes interface to the pre-amp circuit.
To view my 2022 Digital Interface blog post article click here.
To connect my digital interface to the microphone pre-amplifier I added an additional TL071 operational amplifier to act as a line input stage with the proviso it could also be used for other external audio devices if needed. The additional op-amp operates independently to the main microphone pre-amp stages and includes a pair of solid- state analogue switches for audio routing.
The final schematic for the microphone pre-amplifier is shown below.
 |
| Click image to enlarge to full size! |
To conclude the microphone pre-amplifier is now complete and tests on the breadboard look good. On my next blog post I will bring the microphone pre-amplifier and modulator together as one on a new PCB and carry out some tests before moving onto some TX amplification.
Once the modulator is finished and tested 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...
Thanks Andy for another well executed transceiver module design and evaluation. The Irwell modules use readily available and proven parts and are easily reproduced. Appreciate your attention to the detais (such as gain setting for each mode and output switching) that most homebrew desgns omit, these elements can make he difference between acceptable and high performance in a homebrew transceiver project. I built your product detector, audio muting, routing, filtering and amplifier module into one of my CW rigs, it worked a treat, particularly the receiver muting and sidetone for pleasing break-in. Paul VK3HN.
ReplyDeleteHi Paul, it is always a pleasure to hear from you and I appreciate you taking the time to comment on the blog post.
Delete