3
\$\begingroup\$

What. I'd like to build a multi-channel tuner that will pick up, say, two [known] channels, one at 156.3 MHz that's 50 kHz wide, and one at 855 MHz that's 25 kHz wide. Demodulation will be done in a Software-Defined Radio such as GNURadio.

How Rather than have one broadband tuner that picks up and downconverts this wide spectrum, I'd like to have two tuners - one that will pick up and downconvert the first channel to one low-IF, say 1MHz, and a second tuner that will pick up the second channel and downconvert it to another distinct low-IF (using a different LO of course), say 2MHz. Then I want to merely sum the two IFs, and then pass it through an ADC before feeding it to GNURadio where I have the channel selection and demodulation pieces. The advantage will be that the sampling rate can be kept much lower than if I used a wideband tuner to pick up everything in the 150-900 MHz range.

Questions (1) Do the RF experts see any problems with this idea? (2) The Elonics E4005 was a tuner chip that allowed user-selectable IF. Is there any similar variable-low-IF chip I could use since that chip is no longer available? (I'd use one such chip for each channel to be received).

Alternative approach. After reading up on some modern chips available it does appear that the chips available today are pretty much geared towards Zero-IF direct-conversion architectures and it would be simplest to build something with one of these. If I used one of these, then, would it be a valid approach to convert BOTH of the above channels to zero-IF, then UPconvert to two low-IFs, say 100kHz and 200kHz, and THEN sum the IFs before the ADC? That way much of the complexity is handled by the one chip, post which I only have to deal with low-frequency design and layout issues.

\$\endgroup\$

2 Answers 2

1
\$\begingroup\$

I think that most of the Zero IF receiver IC's that are available have ADC's build in, so you actually have a digital output. Even without this, ZIF receivers have IQ baseband outputs, where you have 2 analog output's for a single receiver path, which would require a IQ modulator to upconvert this into a Lo-IF approach. It's my impression that "wideband" receivers that cover 150-900 MHz have multiple conversion stages, as well as IF filters, so that they would only work over a subset of that frequency range at one time, perhaps 2 to 10 MHz within that 150-900 MHz.

You might want to look at the Lime LM6002 transceiver part, which would give you a Digital I/Q output, which would possibly be easier to deal with in a signal processing point of view, as you have a single channel of digital information to deal with for each channel.

\$\endgroup\$
0
\$\begingroup\$
  1. No, no problem with this idea.

  2. Usually, the only frequency restrictions built into a tuner chip are related to the range that the internal VCO(s) can cover. The exact input and output frequencies (and bandwidths) are generally determined by external components. I would think that you could use pretty much any modern chip, unless its datasheet specifically says that it has tighter constraints.

\$\endgroup\$
1
  • \$\begingroup\$ Thanks Dave. #1 is reassuring to know. On #2 it does seem that for most chips, varying the VCO to vary the IF can work within small ranges (+/- 10%), but not for, say, a range of IF from 0-5+ MHz, which is what I would need eventually if I accumulated not just 2 but 20-30 channels in the above manner which is my ultimate goal. One of the limitations appears to be in the fact that I would also need a narrow bandpass filter to isolate each channel and the VCO control per se does not provide that flexibility. Would I be right? \$\endgroup\$
    – Ram
    Commented Oct 4, 2012 at 12:35

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.