I am trying to drive a Broadband Synthesizer chip ( ADF4351 ) with RP2040 using micropython language

[Timeless74]

I need to manage the Broadband Synthesizer chip ( ADF4351 ) and I want to use an RP2040 using the micropython language
DataSheet ADF4351

I've written a whopping 14 different versions but I can't get any 55Mhz waveforms to generate and the last few versions I've had the AIs help me but even getting the AIs to do the code doesn't work.

the chip works with SPI and I have to load 6 32bit registers and I have at my disposal a data analyzer and I can see that from the ports the packets come out.
I tried the chip with a closed arduino code and the chip works so I am doing something wrong but I don't understand.

Do you know if there is any example on RP2040 in micropython for ADF4350/ADF4351 ?

Ps. I connected a TM1637 diplay on pin 14 (clk) and pin 15 (data) and this works
and I connected a ADF4351 on pin
spi_id=0
sck_pin=2
mosi_pin=3
le_pin=4
ce_pin=5
mux_pin=6

MyTest_Rp2040_Generator_Dispaly.zip

[davefes]

I guess "closed" Arduino code means you can not see how it was done. Is it not a sketch?

I wrote a driver for the ADF4111 in C about 10-15 years ago. Unfortunately, I chucked that stuff out last year.

With your data analyzer can you not see lack of activity on any of the pins between Arduino code and your Micropython code?

[Timeless74]

I guess "closed" Arduino code means you can not see how it was done. Is it not a sketch?

I wrote a driver for the ADF4111 in C about 10-15 years ago. Unfortunately, I chucked that stuff out last year.

With your data analyzer can you not see lack of activity on any of the pins between Arduino code and your Micropython code?

Can you tell me if you remember at what speed you had set the data communication , in the datasheet it is not specified among the operating values

[davefes]

Sorry, I don't remember. I would have though any reasonable SPI speed, ie 100kHz up to several MHz should be OK. I see a comment about a max of 20MHz.

[KnorkFox]

Page 11:
To operate with PFD
frequencies higher than 45 MHz, VCO band select must be disabled by setting the phase adjust bit (DB28) to 1 in Register 1.

[Timeless74]

Page 11: To operate with PFD frequencies higher than 45 MHz, VCO band select must be disabled by setting the phase adjust bit (DB28) to 1 in Register 1.

thank you very much I will check and if necessary correct the code , but allettale state I have no output signal

[Timeless74]

main_fix_phase.py.log

News: I saw that I was setting the phase adjustment bit (DB28) on Register 1 to the on state, I corrected it, and now I see a ~200 mV signal at the output.
The waveform is strange and arbitrary, but when I make a frequency change, I see no change, and when I give the command to turn off the generator, the waveform disappears.
ps. the Spi port is set to 10Khz but I don't know if it is right

[Timeless74]

I'm doing something wrong and I don't understand where. To check, I wrote some clean code from scratch, but the output waveform continues to be arbitrary and I don't understand what's happening.

I am attaching my code with comments in English, the terminal output with debugging enabled, and the oscilloscope bitmaps.

main13g.log
main13g.py

full code
main12b.log
main12b.py

[davefes]

Provide a link to the Arduino code that you have run without problems. I found this:
https://github.com/dfannin/adf4351

[davefes]

How about trying to read the registers back to confirm?

[davefes]

I have asked my synth ex-workmate to have a look to see if he can suggest anything.

Have you got a link to the schematic?

[Timeless74]

for now there is no schematic , they are in the design phase and it is all connected by wires ... if you see the code there are connections with the GPIOs .

Power supply via USB C cable connected to PC (4.895 V)
I then take the positive 5V from pin 40 (VBUS) and the ground from pin 38 (GND) and connect them to the ADF4351 development board, which has a 3.3V AMS1117.

I then take the positive 3.3V from pin 36 (3V3OUT) and the ground from pin 33 (GND) and connect them to the TM1637 development board, which has 6 digital digits.

Then, the ADF4351 development board is connected with a connector
sck_pin = PIn4 (GIPO2) SPI0_CLK
mosi_pin = PIn5 (GIPO3) SPI0_TX
le_pin= PIn6 (GIPO4)
ce_pin= PIn7 (GIPO5)
mux_pin= PIn9 (GIPO6)

and the TM1637 development board is connected with a connector
clk = Pin(14) SPI1_CLK
dio = Pin(15) SPI1_TX

[ekspla]

More than 10 years ago, I also wrote a software on PIC/AVR for ADF4351 and have been used it
in a system.

I would recommend you to do the followings.

If your Arduino code works as expected at the specific frequency, you can read the register
values and timings of pins (e.g. latch) with your analyzer.

Then, you can try to send the same sequence by a simpler code in MicroPython.

You can download a software for ADF4351 evaluation board from Analog Devices's web site.
Although writing register values to your board is not possible, the software can be used
for simply calculating the register values. Use this feature to calculate the register
values for your specific case and compare the values with those obtained by your MicroPython
code.

Good luck.

[ekspla]

I could dowload a file named ADF435x_v4_5_0.zip and could install the software. The link of the zip file was shown in EVAL-ADF4351 page. As shown in the snapshot below, you can obtain register values without connecting to the evaluation board.

[Timeless74]

Thank you very much
That one I hadn't lost sight of since he was talking about their scorecard ...

[Timeless74]

Now I'll see what I can do .. knowing that the registers are those I can do some tests without having too many variables....

[davefes]

Sorry to report that my syth-friend last worked on the ADF4111 as well.

[Timeless74]

thank you very much , I am working on it and will let you know ... it is strange to me that there is no library already written

[KnorkFox]

I am not sure, if the problem is in your code. It could be output impedancy, a defective chip, cableing, a missing capacitor...

[Timeless74]

I have two development boards that I purchased on Amazon.
I tested the boards because I had some doubts, given that they are manufactured in China.
I have a board that mounts AMTEL ATMEGA8L and ADF4351, which I purchased a few years ago, and, excluding the board chip, I connected GND, 5V, CLK, MOSI, and LE, and the development boards generate the signal.

Ps. I noticed that AMTEL ATMEGA8L, for which I don't have the code, transmits many packets when I activate the ADF4351 chip, but I don't understand what they are. Then, when the frequency changes, the 6 recorded packets are sent.

Now I am passionate about the RP2040 and, for my purposes, AMTEL chips are too limited.

[davefes]

You marked mine as the answer, but I don't see how that solved your problem.

[Timeless74]

ops

[Timeless74]

Okay, it works. I took the register values from the Analog Device ADF435x software for a frequency of 55.5 MHz and copied them into my clean code main13.py. I tried setting the SPI speed to 1Mb/s, and the commands were accepted. I can see the 55.5 MHz 1.2 Vpp waveform. Then I added a filter, and the signal improved.

main13g.py

After inserting an RC filter

[Timeless74]

I tested the SPI communication speed settings at 100000, 1000000, and 10000000, and the commands were always accepted. The problem lies in some function that creates the registers, so I am investigating.
So far, I have discovered that my code generates slightly different values on the registers. Checking the code step by step, I cannot find any errors compared to what is written in the technical data sheet.

I will let you know the news as soon as I find out where I went wrong... for now, I am dismantling the code and testing each formula.

[davefes]

Are you sure lines 84,85 and 86 in main12b.py are doing what you want? Just a suggestion.

[Timeless74]

Yes, I already have doubts about those lines and am examining them. The results of the functions are the same as those visible in the ADF4351 software.
The calculation of frac_val is not incorrect, but the value returned is not the same, and so far I have not understood where the problem lies.

I have a version in which I am mapping all the variables to understand where the problem is.

[davefes]

There have been some recent discussions on float accuracy in Micropython. I'd try printing those values out to many decimal points to see how they compare with the ADF software.

[Timeless74]

Yes, I am checking every calculation, I am entering all the accounts into LibreOffice Calc and comparing them... as soon as I have sorted it out, I will produce the final versions.