Hi all, looking for advice on the architecture for an airband (118-136MHz) receiver with the following design goals:
- Digital Audio Output (don't care at what point in the chain it goes digital)
- Low Power (thinking something like a teensy)
- Minimal tuned analog components (don't have any experience/gear, and would prefer to avoid things that need hand-tuning)
The goal is something portable and low-power that you could use plane-spotting at the airport - not looking to push the envelope of sensitivity.
I've been primarily looking at a QSD/Tayloe mixer front-end, clocking with an si5351. I briefly looked at direct sampling using undersampling, but it seems like the filtering required for that would be prohibitive. With undersampling I think the ADC sample rate only needs to be on the order of the signal bandwidth, so ~25kHz, which seems very doable. Beyond the sample rate I'm not sure what other ADC requirements I should be thinking about.
Once sampled I think the rest of the signal processing is pretty straightforward. I have working code for demodulating AM using an RTL-SDR (basically the same as rtl_fm -
http://kmkeen.com/rtl-demod-guide/) that I think would work here.
I've never done any RF hardware projects, so looking for a sanity check that this plan makes sense, if there are other approaches that may be better, and pitfalls to watch out for.
Cheers
You have bellow topic on use of Si47xx
For example Si4732 with premixing as you described can be nice option.
With analog and digital outs.
All in case you like to monitor only one airband freq at a time.
If you like to monitor more than one channel than consider something
like SdrPlay or Lime SDR capable to monitor 10 MHz bandwith or more.
For example, during the first Covid lockup this was my ad hoc solution
https://hackaday.io/project/171415-sdr-limetogo
Have a look into AM/FM chips used in broadcast receivers as a start. Toko got crystal filters with bandwidth that is usable for AM demodulation.
Usually the IC dont go up to the aircraft band, but if you set the IC with a fixed frequency and use a DBM like SBL-1 or ADE-1 as a downconverter in fornt of the demodulator chip you can get a very wide frequency range.
A filter should be used after the mixer to kill spur and harmonics this can be a 50-100khz crystal filter.
Any IF amplifiers can be MMIC that gives you a simple way of getting gain. Look at the MAR-3 or some other dc-1000Mhz MMIC.
Downconverting and using a standard AM/FM IC is a great idea! I looked for an off-the-shelf chip initially and when saw they didn't support this frequency range I dropped the idea entirely. Never thought to just shift the signal down into the range they will work in.
The benefit of using a demod chip is that you get a full receiver with a minimum of components. The design is more or less already done in the datasheet if you select the correct chip.
This is making sense, here's the chain I'm looking at:
\ / -------- ------------ ---------- ---------
|____ | Mixer | ____ | Bandpass | ____ | LNA | ____ | RX |
| ADE-2+| | ??? | | MAR-6SM+| | si47xx |
------- ------------ ---------- ---------
|
---------
| LO |
| si5351 |
---------
Mixer: ADE-2+ or ADE-6+ (1dB less conversion loss for $6 vs. $3) look good
LO: Stick with the si5351
Bandpass: haven't been able to find a crystal or ceramic filter with a center frequency in the AM band of the Rx (520-1710kHz). I've found up to 500kHz, then they jump to 2.5MHz.
LNA: The MAR-6SM+ looks like an inexpensive low-noise amplifier (lower noise seems better, right?)
Rx: The SI47xx line looks good. Individual chips are rather pricey on digikey, but I found these pre-build modules on Aliexpress that are super cheap (questionably so, but might be a good starting point)
https://www.aliexpress.com/wholesale?catId=0&initiative_id=SB_20210311101541&SearchText=si4730. Questions I'm trying to figure out now:
- Should I have a filter before the mixer?
- Do I need to worry about the LNA overpowering the receiver? The only spec I really looked at was the noise figure.
- Where to find a filter
Feels like this plan is coming together, thanks for the help!
EDIT: Just stumbled upon this project that seems to do almost exactly what I'm trying to do.
https://github.com/LadyRoninEngineer/Si473x/blob/main/FrontEnd/Documentation/21.0222_FrontEnd.pdfInteresting that the LNA is before the mixer, with a variable attenuator in between, which seems to answer my question about how to avoid too large of a signal.
If you look at other reciever designs there are usualy some filtering up front of the mixer. Normaly a BP or LP filter to kill the harmonics. The SI5351 does have a lot of harmonics because it got square wave output. For your design i would probably just make a BP filter from about 115 to 140Mhz.
Then you could consider putting inn a BP filter after the mixer to clean the mixer output. Thisl filter can be made by modifying a TOKO 455khz canned LC filter or iven a TOKO transformer with tuning slug.
If the filter is a tuned transformer you can ajust how wide it is, how sharp the flanks are and move the center frequency wherever you want. This involves using trimmers as capacitors.