The Si5351 boards really need to be calibrated for best results. This if more important for HF SSB than for VHF FM.
An old VFO board has been pressed into service for an Si5351 calibrator.
With the addition of a couple of standoffs and some Pogo Pins, a new Si5351board can be clamped on top then the calibrator sketch run.
Here is the output before calibration. It should be 10Mhz.
Load the Arduino si5351_calibrator sketch..
Then open the serial monitor at 115200 Baud.
On the screen, it tells you what keys to push for Up or Down each decade.
And, here are the results or running it for this board..
The number, in this case, 15600, is added into the Si5351 initialization setup.
Define the calibration number...
. int32_t cal_factor = 15600; // found from Si5351 calibrator sketch.
And in the setup routine...
. clockgen.init(SI5351_CRYSTAL_LOAD_8PF, 0, 0);
clockgen.set_correction(cal_factor, SI5351_PLL_INPUT_XO); // <<<<<<<this is found from Si5351 calibrator code.
side note: I was thinking about a low cost hackable sig gen this morning, and since i have been playing with the rp2040 my mind drifted there. It seems to me with a bit of code and a r2r dac it should be possible to get something like a 8-12 bit 50-100Msps(maybe more with overclock) of arb gen out of its pio and dma, it would have no power and would want a vga and offset amp to make it useful, but it would be fun. waveform would have to be in ram since i don't think the qspi flash or adc is quick enough to fill a 400Mbps+ buffer? I think it would leave one core and 2 pio state machines in the rp2040 free for display/ui for glitch free? operation.
I like this idea since jlcpba has them for $1.20USD ea in 1 off.
anyone know a good design for low cost vga?
end side note.
Putting the scope onto the crystal connected to pin 2 of the Si5351 then removing it makes the chip run correctly.
But, a new command to the IC will cause it to go funny again and the scope needs to be tapped on again to get it to work.
So, is it the crystal or the chip?
This is repeatable on a second board.
Yes, I will try that. But it is not an easy thing to add because of the size. View attachment 846
This shows where the cap may need to go.
It could also be the other side connected to pin 3 of the Si5351.
this will test my soldering skill!
Order today, ships today. CX2016DB25000D0FLJC2 – 25 MHz ±10ppm Crystal 8pF 150 Ohms 4-SMD, No Lead from Kyocera International Inc. Electronic Components. Pricing and Availability on millions of electronic components from Digi-Key Electronics.
Well, a new batch of 5 Si5351 boards arrived today. I ordered those that have a Silver/Gold crystal fitted, like that shown in post #51.
They all work ok.
The boards with the black crystal are suspect, so if you are ordering boards, look for the metal case crystal, not the plastic ones.
Replacement crystals are on the way and I'll post an update when they get here to see if it is a crystal or IC problem.