# Issues with frequency conversion

I am planning on building a device that will perform a frequency conversion of a band of RF in approximately the region of 650 MHz down to 520 MHz. The bandwidth will need to be approximately 50 MHz.

It is a "repeater" style device, where the device will receive 650 MHz and then re-transmit on 520 MHz.

What are the likely issues with this? I am assuming that there will be filtering challenges and I will likely need to use a much higher Intermediate Frequency, i.e.

650 MHz -> 900 MHz
900 MHz -> 520 MHz

In order to overcome filtering challenges that I assume I would run into if attempting to do a direct conversion.

This will be the first time I have attempted something like this - are there any other issues I might need to consider before planning my design?

For readers new to the topic: This device doesn't need to be invented. It's called a transverter, and its central component is a mixer.

You'd drive it with the difference frequency (i.e. (650-520) MHz = 130 MHz), and filter out the unwanted sum frequency around (650+130) MHz = 780 MHz with a simple low-pass that has its cutoff somewhere above (520 + 50/2) MHz = 545 MHz and has sufficient attenuation at (780 - 50/2) MHz = 755 MHz.

I don't see why you'd first mix up; the transition width of 210 MHz should allow for picking good enough "classical" LC/RC filters. (Gut feeling: 9-pole passive Chebyshev LC (that's 4 stages) should do this, and might still be "manageably complex" at these frequencies, i.e. possible to manually fine-tune if parasitic effects aren't eliminateable).

If you decide that e.g. a three or four stage RLC doesn't attenuate enough (and "enough" is up to your definition! You didn't define that in your question.), then, indeed, find a higher IF that you can mix onto, at which there's an SAW filter of required quality, and then mix down. But: this would require multiple oscillators, and each of these will introduce spurs into your system.

In essence: every solution begins with a precise description of your problem, in your case, which flatness in passband, which attenuation in stopband, which circuit complexity, temperature stability, and cost effectiveness you need to achieve.

• The device I have described is NOT a "mixer", it is a frequency converter. A "Mixer" is but one component of a frequency converter. It is necessary to combine a "mixer" with oscillators, filters and amplifiers in order to achieve the required frequency conversion. A "mixer" alone will not do the job.
– Mark
Nov 25, 2017 at 13:50
• well, I'll agree with you here; the mixer is the part that actually does the conversion, but you need more than that. I'll fix that. Nov 25, 2017 at 13:51
• @Mark, Better now? Nov 25, 2017 at 13:52
• I'd suggest one more improvement. This is also not a transverter. We are not attempting full-duplex frequency conversion here either. It is simply just re-transmitting whatever we hear on one frequency, on another. We are not expecting anyone to call us back. "transverter" implies bi-directional comms.
– Mark
Nov 25, 2017 at 21:12