An Improved Superhet for WRMI reception
This, hopefully, will be a "new and improved" version of the simple superhet for reception of WRMI that was previously shown here.
Improvements will include wider range frequency coverage, better image rejection, and improved selectivity, achieved principally by changing from a single down conversion to an intermediate freque3ncy of 455 KHz, to a dual conversion up converting to 10.7 MHz as the first intermediate frequency.
With the use of a simple low pass filter as the front end, reception from the broadcast band to seven MHz or so should be readily accomplished. Images will be in the range above 22 Mhz so unlikely to present any issues, and a crystal filter at 10.7 Mhz will establish the initial selectivity ahead of the second IF.
Construction has begun. The Front end filter, mixer, and second conversion local oscillator are complete.
A TDA1072 again is used as the main device of the receiver. Unlike version 1, if successful, this second revision will use the onboard TDA`1072 RF amplifier as the first intermediate frequency amplifier, and the TDA1072 balanced mixer as the second mixer. The first mixer continues to be a mini circuits ADE-1+.
The crystal oscillator gave me fits. The TDA1072 data sheet says next to nothing about the onboard oscillator. No schematic of the chip is given, and I have been unable to locate one. Ultimately, I got the circuit shown in the data sheet to work, and work well. Prior to that I spent considerable time trying to get a crystal to emulate the tank circuit shown in the data sheet for a variable oscillator. This was utterly unsuccessful and I believe the data sheet crystal oscillator is likely the only circuit that will work. The second local oscillator frequency is 10.7 MHz.
My notebook pages for the work thus far are attached. I usually do not draw schematics as I tend to think of things in small blocks of circuits.
The low pass filter is a three pole circuit, with a measured -3 dB point of 7 Mhz. 50 ohm -2.5 dB attenuator pads at the mixer input and output establish a broadband 50 ohm termination for the mixer. A binocular core transformer will be used to step up the 50 ohm output impedance of the mixer to the higher impedance typical of narrow band monolithic crystal filters. Sweeping the transformer shows a relatively flat response across the 10 MHz band, with about an 8 dB rolloff beginning in the image range, so it will additionally provide some contribution to image rejection.
The low pass filter shows a bit of a bump around 23 -24 Mhz before recovering and beginning to attenuate again. I suspect this is related to the self resonant frequency of the 1.5 uH inductors in the filter.
Most of the parts are 0805 smd, but the oscillator uses a mix of leaded parts, and 1206, 0805, and 0603 smd parts. Some smd parts are located under the through hole parts and are not visible.
73,
Win W5JAG
oops: I just noticed I put in my notebook the attenuator is a T arrangement. Obviously, it is a Pi arrangement with the 15 ohm part the series element, and the 300 ohm resistors the shunt.
Also, the oscillator page had a serious error. The attached version is now the correct version.
I did produce a working receiver; in fact it worked very well, but unfortunately it is now on the scrap heap and I do not know if, how, or when it will be pursued further.
A couple of pictures of the working receiver, and the final back-end schematic are attached.
During burn in, the device would randomly start outputting static, somewhat like T Storm noise, but much sharper and harsher, with the additional symptoms of there not being a T Storm within five hundred miles, and it was present with the antenna disconnected.
I mistakenly attributed this to a failing ceramic filter, as my filters were really old, and had been stored in poor conditions. Also, I installed the first filter without blocking caps, despite Murata guidance to the contrary, because the voltages at the filter are really low (<4 vdc> and I had gotten away with it before. After replacing the first filter, with a replacement with proper blocking caps, the noise went away, giving me a confirmation bias that this was the problem. Except it wasn't. It came back again, and the bias was strong enough that I chased it through several filters even though it seemed implausible that so many could be bad.
And of course they weren't. The true failure point and diagnosis was when the receiver went from very sensitive to stone cold deaf and I was able to diagnose 0 volts at both of the RF amplifier input pins, when they should be 3 - 4 volts. The RF amp was dead. This was apparently because a 0603 100 nF bypass cap had shorted, probably from overheating and sloppy technique when I installed it, or maybe just bad luck and a bad cap, and pulling everything to ground. Because of the tight construction, I managed to cook the TDA1072 in the process of trying to repair it. I don't think the shorted cap did the chip in, I'm pretty sure I did it.
So anyway, I just decided it was best to start over and I removed everything from the matching transformer back, and cleaned the board off. I'm actually not too upset about this. In hindsight, it seems a bit silly to have limited the device to 0 to 7 MHz or so; it would have been about zero extra difficulty to just have made it a full 0 - 30 MHz, so it might have wound up as a teardown anyway.
I also learned that just because you can make something pretty small, does not necessarily make it a good idea to do. In making it a goal to cram everything into the one 50 x 80 mm board, I made it very difficult and ultimately effectively impossible to repair, resulting in the loss of the project.
Win W5JAG