At these frequencies, the diode ring needs hefty transformers, and it gives no advantage over a Tayloe mixer (Google for it) built with opamps and an analog switch/multiplexer.
I'd still question the sense of using a DSB (double sideband suppressed carrier) approach due to the difficulty of decoding it properly.
With a suitably band-limited input, a voltage-to-frequency converter might be the best to feed the PC sound card with an ELF signal. It is relatively easily decodable in the digital domain.
bg, are you the original OP on this thread? If so, you asked aout 'shifting' and then described freuency MULTIPLICATION. Seems to have confused a lot of people.
Any multiplication you do on a voltage waveform in the time domain will not yield the results it sounds like you desired. Using mixers, etc will only SHIFT the spectrum, replicating it, and causing no end of confusion from those pesky sidebands.
If you get a chance to get a copy of Ron Bracewell's Fourier Transform book, it's a good read. I took the course from him while at Stanford, and his opening claim was, "When I'm done, you'll be able to do FT's WITHOUT all those confusing equations." He was right. The book has a lot of diagrams and 'verbal' explanations, so you can understand without all those formulas, or if you want, get bogged down in formulas, your choice.
His book has a few examples of what has been proposed as solutions for you. All of those proposed solutions would fail miserably UNLESS the original waveform is a single tone.
It's easy to create some sample waveforms using octave, free Matlab clone. Process those .wav by DOING what has been proposed, or better yet, do what you really want, then generate a new .wav, and either listen, or analyze the results. Simple simulations save you a lot of grief in generating useless hardware. Should say, disappointing hardware.
Did I miss somehting here? Any mixer will SHIFT the frequency upwards, and that includes the negative half not shown in plots, That is the lower half that now appears as the lower sideband. In other words, this is a shift, not a multiplication! Envision two tones in the baseband, one at 1Hz and one at 10Hz, 'shifting' the band up by using a mixer set at 1000Hz will create a new band around 1000Hz, that is, tones at 990Hz, 999Hz, 1001Hz, and 1010Hz, and a smidgeon of rality at 1000Hz. So the point is WHERE IS THE FREQUENCY MULTIPLICATION ASKED FOR???!!!
My interpretation of the original post was how to make a circuit that takes the 1Hz and the 10Hz and 'shifts' [should have said frequency multiplied] up to 1000Hz and 10kHz, now THAT'S multiplication!
Yeah, I remember those and contemplated them for an ultrasound system. Never used them on anything though because of $$$ and because I thought they were boutique chips that would quickly become obsolete. Which they did, often along with their mfg.
Didn't read until I was 7, and still don't have the hang of it yet!
*IF* the double balanced mixer [either solid state or passive transformertype] is driven by the sound card, there will be advanatage to the synchronous-ness. Then a simple window can be used, and the spectrum can be plotted. Takes something in the range of a second of time to get a spectrum resolution of 1Hz. Do the soundcard apps allow that kind of resolution?
Right now, I'm using the EMU1212 soundcard and its ASIO drivers to get down into the sub Hz ranges, but I write my own C/C++ code. Bragging, or complaining? A little of both, believe me!
I remember them. About 30 years ago i was using a "high-speed" digitizer built with some thing like those (more likely a similar older part). It did up to 25 Msamp/s at 6 bits resolution (it was old equipment at the time, the CA3308 had just came out). Digitizing and playback used similar tricks to change the effective time base. The actual digitizer was 7 bit sar and did only about 200 ksamp/s. And the DAC wasn't much faster.
refinements (like accuracy) would work. Just how accurate do you want this ?
Well at least quite similar. See the Armstrong method for generating FM/PM. The other side of the coin is that is going to keep the sidebands way down compared to the carrier, making the sidebands in the spectrograph a bit hard to see (20 to 40 dBc down).
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.