Sciencemadness Discussion Board - Radio receiver design

Radio receiver design

Rainwater - 11-2-2023 at 09:23

Im requesting comments on the attached paper.
Specifically on the interpretation of the concepts involved,
and the math used.
It seams everytime I read over it I find something needing correction
Am I even headed in the right direction?

Attachment: Nmr build_230211_113330.pdf (2.7MB)
This file has been downloaded 360 times

Keras - 11-2-2023 at 12:11

Quote: Originally posted by Rainwater

Im requesting comments on the attached paper.
Specifically on the interpretation of the concepts involved,
and the math used.
It seams everytime I read over it I find something needing correction
Am I even headed in the right direction?

Do you have any specific questions?

Rainwater - 11-2-2023 at 12:25

In section 4.5.1
Originally I converted the signal strength into a voltage, then
applied the gain factors, resulting in a value of 7v/pW of the
amplified signal. Then ran the ADC calculations off that. But
quickly realized it was to good to be true, and started searching
for the error. After finding it, I reworked all the equations again
to ensure the figures generated are accurate.

Really just wanting another set of eyes on it. I havent used most
of this math in years, and I have been making simple mistakes

Keras - 12-2-2023 at 07:05

7 V/50 Ω (that’s the standard impedance in radio) is around 1W (49/50). So that would mean a power gain of one trillion (120 dB).

The noise power in a given bandwidth is -174 dBm/Hz. To get an appreciable signal/noise ratio (20 dB), that means the noise power at the end of the amplifier cannot exceed 10 dBm, so that translates to -110 dBm at the input, therefore a BW of ca. 1 MHz.

This is achievable in theory, but would need extreme precautions, because the slightest feedback from the output to the input would lead to uncontrolled oscillations. The best way to achieve this result is to use a heterodyne type of receptor (you change the frequency along the way using mixers).

Sulaiman - 13-2-2023 at 02:37

I know very little about NMR, a little more about rf signals.

NMR
Unless you are going to create your own database of spectra
You will need access to a database (download to ensure future availability)
I'd aim to use the same rf frequency as used to create the database.

RF prototyping - some hobby experience

External man made rf noise is everywhere

Small functional blocks in individual metal boxes using sma connectors is a conventional approach

Where possible buy ready made modules to more quickly get a proof-of-concept prototype working

I'd have the receive antenna signal go through a 'filtering' module first,
block all frequencies that are not required
(hpf, lpf and maybe a notch filter if fm radio transmitter nearby etc.)
Signal limiting/clamping and switching.

Next would be a commercial LNA module, and a couple of spares

Then comes your main electronics prototyping board(s),
protected by the (disposable or repairable) front end modules.

60dB gain in one enclosure is possible, but not recommended - good shielding helps.

Similarly, for the rf power you can copy application notes or buy commercial rf amplifier and preamplifier modules.

For a one-off project I'd use as many ready-made modules as affordable. Saves a lot of time.

Edit : an Rtl-Sdr dongle (or similar) could be useful?
Mine only has about 25MHz bandwidth,
but it does have mixers and clever stuff and a fairly good (0.5ppm/^oC) tco.
Only 8-bit but better are available.
Some SDR's come with separate Tx and Rx connections.

If you could base your design around something like an SDR with various modules, it may be easier to make your work more widely available?

[Edited on 13-2-2023 by Sulaiman]

Rainwater - 13-2-2023 at 04:54

Quote: Originally posted by Sulaiman

I know very little about NMR, a little more about rf signals.

NMR
Unless you are going to create your own database of spectra
You will need access to a database (download to ensure future availability)
I'd aim to use the same rf frequency as used to create the database.

From the reading material i have seen, the base frequency of the spectrum changes based on the strength of the magnetic field and target nuclei, but the frequency shifts remain constant in the ppm scale. Is this incorrect?

Quote:

External man made rf noise is everywhere

Yes. Noise will require its own chapter, and have to be measured experimentally.
Its something I deal with everyday, unless you burry it 3ft deep, fm and wifi broadcast will overpower many signals.

Quote:

Where possible buy ready made modules to more quickly get a proof-of-concept prototype working
....
Next would be a commercial LNA module, and a couple of spares

The chips I mentioned, I had laying around. Their standard components for a lot of the repairs I do. That doesnt mean im an expert at what they do or how they do it. Its like swapping parts on a car.

Getting 60 db gain isnt to bad. The differential inputs are a lot less influenced by common ground interference than say an audio amp would be. The problem lies in limiting the power that reaches the LNA, any more than a few microwatts and she is toast. But again. Thats a different chapter.

My main goal now is to correctly assume the abilities and requirements of the proposed design before construction.
No real time constraints as I plan on writing chapters for every section of the apparatus, as Ive done the receiver. I want to ensure it will work on paper before spending money.

Sulaiman - 13-2-2023 at 05:30

"From the reading material i have seen, the base frequency of the spectrum changes based on the strength of the magnetic field and target nuclei, but the frequency shifts remain constant in the ppm scale. Is this incorrect?"

That's how I understood it also from the little reading that I just did

Good luck with the project, keep us updated.
I hope that other members can better contribute.

Texium - 13-2-2023 at 11:12

Quote: Originally posted by Rainwater

Quote: Originally posted by Sulaiman

I know very little about the physics/engineering of radio receivers, but I know a good amount about NMR. Yes, the frequency used depends on the strength of the magnetic field and also on the nucleus being observed. For example, an instrument with a field strength of 11.74 Tesla would have an operational frequency of 500 MHz for protons, but only 126 MHz for carbon-13 nuclei. Typically, instruments are referred to by their proton NMR operational frequency. You can determine the operational frequency from the magnetic field strength and the gyromagnetic ratio of the nucleus being observed.

I’m skipping over Fourier Transform stuff because frankly it confuses the hell out of me.

When it comes to ppm shift, that is found by setting a reference point (tetramethylsilane is used as a 0 point because its protons are extremely heavily shielded, aka respond weakly to magnetic field). Then the value in Hz that a proton responds to relative to TMS is divided by the instrument’s operational frequency. Hz cancels out, leaving you with a normalized, unitless value that will be the same regardless of the instrument’s field strength/operational frequency. This number is multiplied by 1,000,000 to get it to fall in an easy to digest range of ~0-12, and that’s what the “ppm” term is derived from! So the signals will fall at the same shift regardless of your field strength, but the resolution will be different. For instance, a 300 MHz instrument running proton NMR would have 1 ppm = 300 Hz, while a 600 MHz instrument would have 1 ppm = 600 Hz. This means you’re effectively getting double the data points for 600 MHz and it will be easier to see finer details. It’s like comparing two copies of the same image but one is in 360p and the other is in 720p.

I also feel I should point out that you don’t need a database. It’s not like IR, where you won’t be able to confirm the exact compound to a high degree of certainty without a reference spectrum. With even a proton NMR spectrum alone, you should be able to interpret it and know what molecule you have, especially if you have a general idea of what to expect. That’s why it’s so good. Anyway, I hope that info is helpful.

Twospoons - 13-2-2023 at 12:08

You would do better posting technical questions like this over on https://www.eevblog.com/forum/index.php . There are many RF engineers there, and a wealth of knowledge.