SDR Tranceiver - Part 5 More Troubleshooting

More Receiver troubleshooting (without much luck unfortunately)

0:00 Probing before Audio Switch
2:00 Adding more decoupling caps
3:30 Replacing the FST3253
5:26 Byassing the Audio Switch
6:55 Looking at the ESP32 Header
8:44 New header board arrive - does it fix the problem?

[superconductorchip]

This may cure it, it works for me. Put lots of vias between top and bottom PCB.
Only if your top and bottom PCB are at the same potential.
Great series of videos look forward to the next installment.
Keep up the good work.

[justme-n-gracie]

I'm leaning heavily towards ground loop. RF and digital grounds need to be separate and isolated. Noise from the ESP coupling through the ground and modulating the 5351. Many less diligent constructors may not have noticed these problems or just ignored them. It's great to see your thoughtful process of testing and diagnosis. As another has suggested, I would try separating the ESP from the board and then the ESP and 5351. As you demonstrate, it's just a process of isolation, substitution and elimination. I love this series! Although it can be tedious for you, there are a lot of important lessons to be learned for everyone following along! Great suggestions contributed by viewers/commenters also! Thank you again!

[semeykinpavel]

Hello. Thanks for your exciting videos. I myself had a problem with low frequency interference, and my noise source was the display, only it was OLED display, probably due to PWM. I was helped by electrolytic capacitors near 2200 mkf on the power line.

[chrisherd]

For the half 5v rail 6:22 try lower value resistors in the voltage divider ie more current through it. Add a low value inductor in series with 5v feed for just before the voltage divider, maybe an electrolytic across the half 5v rail for good measure. Yes still clutching at straws.

[MalayalamElectronics]

This is very interesting trouble shoot ❤. I know you might have tried, adding decoupling caps "on the" daughter/adapter board for the esp32a1s & some ground to ground bridging with copper strips across the board while probing the output noise. Also a huge electrolytic on the power rail of esp32, "on the" adapter board.

[justanothercomment]

Now that you mention it, AFAIK, and from what I recall reading, that PCB antenna should NOT be directly over a ground plane like that. It should extend beyond the ground plane as otherwise it can and will induce noise. Which I believe is what you're fighting.

[mgeo]

I did similar with a superhet receiver a few years ago. Turned out to be I2C bus to Display & Si5351.

How I found it. Keep the probe on the noisy output. Then poke around for signals that match on another scope channel. You'll see what it is when your scope signals correlate.

A better layout. Ground rings around modules to reduce ground plane noise, etc.

[MXFXUK]

Hi Ian, yes it certainly is a pain of an issue. I still think the noise could be coming from the clock generator. Why not supply the SI5351 with 5V and use a level shifter to 3.3V for the I2C to the ESP-32. I could be way off track but it may be worth a try. Loving the videos.

[TheEmbeddedHobbyist]

I would have put some decoupling capacitors on the esp32 board, unless they are on the other side. Quite a long path for the power to get to the chip pin and Mc are quite good at wanting large current bursts.

[justme-n-gracie]

I just picked up this little tidbit when researching voltage regulator generated noise... from Analog Devices AN-144... try removing the ground lead with alligator clip from the scope probe(s) and make as short as possible connection from probe grnd to grnd closest to your measuring point(s). Seems that the scope probe ground lead is a known source of noise pickup....

[vontas]

My guess would be that these are the spurious tones that are generated by the PLL. Maybe it's worth a while to check the purity of what comes out of that synthesizer board at particular frequencies.

[vincei]

w2aew: #84: Basics of Ferrite Beads: Filters, EMI Suppression, Parasitic oscillation suppression / Tutorial

[Trainwreck]

Yes. try "star" connecting the + and - to the power supply pcb. High gain audio coax should only be grounded at one end only.

[rfdave]

HI Ian, Enjoy watching your videos. Have you power up your board without the SI5351 and ESP boards in just to confirm. I know you did say the noise source was the SI5351 board. Have you tried using a EMF probe and your Spectrum Analyser? Just wondering, I know how it bugs at you. I've been there. Thanks 73

[vincei]

Digital noise. No shielding. Almost definitely the microprocessor. Does the esp32 share the same analog ground as the rf deck? Do you have ferrite beads on the power lines leading to the esp32 ? Are you powering the systems with a switched mode power supply ?

[daveevans]

Er, is the antenna on the ESP32 board active? Can you stop the code looping once it's set a frequency? Does it have a sleep mode?

[wodowiesel]

like your videos, keep up the cool projects! Not quite sure but maybe something wrong in the kicad schemetics?

[paulh]

Measure the input signal without esp32, or power the esp32 total seperatly use audio transformer to seperate the ground. Also in the software switchoff the wifi, that takes a lot of power. Maybe this give some pointers

[youian]

Hi Ian, I wonder if you could remove the esp32 and program the SI5351 from something off-board to eliminate the esp32 as the issue? Not sure if you tried? Anyway good luck with it. I really appreciate your videos - very interesting indeed. 73, VK7IAN