Cube-436_mini oscilloscope

60M @ 12bit, 256pit depth mini oscilloscope

Upgrade tip: ADC can be replaced with AD9226 (65MSPS, supports replacement with AD9220 pin 2 pin)

The theoretical sampling rate can reach 70MSPS @ 280Mhz or 63MSPS @ 252Mhz

The main frequency depends on the highest frequency (constitution) of FALSH. It is recommended to use 133Mhz flash. After testing, W25Q128BF can run stably at the main frequency of 280Mhz.

Thanks to the group friends for providing AD9226 board test. The core voltage is 1.35V, the main frequency is 252Mhz, FFT is turned off, the actual sampling rate is about 60M (theoretical 63M)

Signal source 2Mhz duty50% PWM

It is easy to see from the figure: 256 sampling times are about 4.27us, a single sampling is 16.67ns, and the frequency is 59.98M

AD9220 solution, core voltage 1.2V, main frequency 280Mhz, FFT enabled, actual sampling rate is about 46M (theoretical 46.66M)

Signal source 2Mhz duty50% PWM

It is easy to see from the figure: 256 sampling times are about 5.53us, a single sampling is 21.6ns, and the frequency is 46.29M

ground

2Mhz Sin

1MhzExp-log

1Mhz multi audio

Back surface temperature after running for 10 minutes (room temperature 27)

Code open source repository: https://github.com/panyihang/Cube-436_miniOSC

Being updated, the code is not perfect yet, though. . . It’s not that it can’t be used (everyone is welcome to suggest PR)

If you need cooperation, please contact the email root@an-ye.top

Q group 565264047 Welcome to play~

The measured AD9220 can be overclocked to 25M+, and the ultimate performance is being tested

The MCU, ADC, and screen are all running overclocked. It is normal for the board to be a little hot.

ps: RP2040 can run up to 420Mhz, but the 1.3V chip I soldered is above 280Mhz, and the qspi frequency is divided to 8 and it cannot run. I suspect that the frequency dividing setting does not take effect.

ps1: Be careful to enable the fast Fourier transform in the source code to find the frequency. It is currently very unstable.

0x00: Preface

There are several DRC errors reported on the PCB. The packaging problem does not affect the use, so just ignore it.

The hardware part of this project was drawn in public in the classroom on the last night of self-study before the college entrance examination. Please forgive me for any shortcomings in the hasty drawing.

0x01:Update instructions

ver0.2: Change the LED driver power supply, the original power supply scheme has insufficient current.

0x02: Hardware part

Main control MCU:

• RP2040 with Raspberry Pi
• Core voltage 1.2V
• System frequency 280Mhz
• After testing, it can run stably (without turning on FFT)
  

ADC:

• AD9220
• 12bit@25Mhz (40.2ns)
• Input voltage range -3.3V-6.6V adjustable
• Overclocked, a little hot

Screen:

• 8-bit parallel port driver
• Resolution 240x240
• Due to lack of small money, I used the stock screen, model HL015Z1E
• The binding chip is st7789, please find products/replacements by yourself.
• Turn on the inverse color for the initialization parameters of the substitute, and the display effect will be better~

storage:

• 128Mb SPI Flash
• Used to store firmware
• Can store waveform files

0x03:Software framework

Compile:

• Compiler c sdk for raspberry pi
• Develop using the PlatformIO extension for vs code
• The PIO development package used is WizIO-PICO

In order to improve efficiency, the c sdk used for compilation has been modified. In theory, it can be compiled normally using the official SDK (now it has basically been replaced and can be compiled directly)

Display frame:

• The display frame uses lvgl8.2
• The screen driver mode is gpio batch operation

adc collection:

• The collection depth is 256pit
• Sampling accuracy 12bit
• Accurately control sampling frequency using assembly
• In the future, pio will be used with idle cores for sampling.

Automatic control:

• Calculate the maximum value and variance of sampled data
• Automatically control the ADC’s ​​reference voltage and sampling range
• No need to worry about missing small waveforms

The new project "Cube-443_oscillometric tweezers" is under preparation. . . Common components will be used instead, and 65M @ 12bit ADC is expected to be used

New project "Cube-443_oscilloscope tweezers" has been released