[libmaple] Black F407VET6

You can try the “Generic F407VE series” but as the name says, it is the VET version, not the ZGT, means different flash/ram size and board pinning.
But you could generate a new board variant, if you want.

I’d like to try this core with “Blue F407VET6” that has 25Mhz resonator. Steve, I wonder if you know what changes are needed for getting the 25Mhz resonator working ?

Image

Image

That is the generic F407 mini, see here: http://stm32duino.com/viewtopic.php?f=39&t=2804.

Select “Generic STM32F407V mini series” in the Arduino IDE menu board selector.

It only works with my repo: https://github.com/stevstrong/Arduino_STM32

Downloaded your repo Steve and cant get serial USB to work or get recognised, using a STM32F407 mini.

Hi all,

Thanks for all the work you’ve done for this board. I’ve been using the STMF103CBT6 for some time now and was looking to “trade-up” to something with more ram, found this board on ebay… and now I’m here.

I noticed that there was a post for the W5500 being ported, so this morning I freshly downloaded Arduino_STM32 and Ethernet_STM32 and then tried this sketch:

#include <Ethernet_STM32.h>

SPIClass mSpi(1);

void setup() {
Serial.begin(115200);
delay(500);
Serial.print("initialising ethernet... ");
Ethernet.init(mSpi, PA4);
Serial.println("done");
uint8_t mac[6] = {0x00,0x01,0x02,0x03,0x04,0x05};
Serial.print("initialising ethernet with DHCP... ");
if (Ethernet.begin(mac) == 0) {
Serial.println("failed");
} else {
Serial.println(" DHCP assigned IP ");
Serial.println(Ethernet.localIP());
}

}

Thanks for pointing to this issue, I just added that macro to /cores/maple/libmaple/gpio.h.
Please check it out again.

what changes are needed for getting the 25Mhz resonator working ?

Go through the code and search for HSE_Value. It can be changed in the code or via the ide preprocessor – st isn’t consistent on that.

You may also look through PLL settings to make sure that you don’t exceeded it’s max frequency.

Btw, did they really use a resonator and not a crystal? Shame on them if so.

Thanks for pointing to this issue, I just added that macro to /cores/maple/libmaple/gpio.h.
Please check it out again.

Thanks for the fast response, I’ve just done a fresh download and it now compiles.

Of course it doesn’t do anything else, ha ha, but I’m new to the W5500, so I’ll make sure I’ve got everything set up correctly.

Regards.

You could start with the WebClient example of my W5500 lib, using your own SPI instance.
Make sure that your router supports DHCP and allows new connections.
Also, be aware that not any MAC addresses are valid, so the router could refuse the connection based on that being invalid.

You could start with the WebClient example of my W5500 lib, using your own SPI instance.
Make sure that your router supports DHCP and allows new connections.
Also, be aware that not any MAC addresses are valid, so the router could refuse the connection based on that being invalid.

Thanks Steve.

My main project last year involved Christmas lights panels on the windows. The panels of WS2812B leds were controlled by STM32F103(C8T6) modules and the PC sent the control information via WiFi to a ESP8266 which then passed on the data to the STM32F103 via serial. (9 panels of leds with 5 ESP8266->STM32F103 modules.)

This worked very well, but this year I want to do more, and the 20kB ram on the STM32F103 was getting to be a bit of a restriction.

So I wanted a board that had more ram, but was still easy to program, preferably by Arduino IDE. I have used STMCube + IAR but it it WAYYYY more complicated.

This black F407VET6 looks like a very promising replacement, and so this may also be a good time to consider removing the Wifi ESP8266 and connect the F407 directly to the LAN. Hence my interest.

I will look into your suggestion. As I am new to the W5500 I think I will maybe go back to the Uno, get this to work with the W5500, then progress to the F407VE, possibly via the F103…

W5500 lib should work with my repo, check the web client example.

Hello again,

Firstly, the W5500 module works fine with the Uno and Ethernet library, so I know the pin connections are correct and the module functions.

Secondly, I ran this to check the SPI pin allocations:

#include <SPI.h>
#include <Ethernet_STM32.h>

void setup()
{
Serial.begin(115200);
delay(500);

Serial.println("SPI default");
Serial.print("nss: "); Serial.print(SPI.nssPin()); Serial.print(" sck: "); Serial.print(SPI.sckPin()); Serial.print(" miso: "); Serial.print(SPI.misoPin()); Serial.print(" mosi: "); Serial.println(SPI.mosiPin());
Serial.print("nss: "); Serial.print(PA15); Serial.print(" sck: "); Serial.print(PB3); Serial.print(" miso: "); Serial.print(PB4); Serial.print(" mosi: "); Serial.println(PB5);
SPI.setModule(1);
Serial.println("SPI 1");
Serial.print("nss: "); Serial.print(SPI.nssPin()); Serial.print(" sck: "); Serial.print(SPI.sckPin()); Serial.print(" miso: "); Serial.print(SPI.misoPin()); Serial.print(" mosi: "); Serial.println(SPI.mosiPin());
Serial.print("nss: "); Serial.print(PA4); Serial.print(" sck: "); Serial.print(PA5); Serial.print(" miso: "); Serial.print(PA6); Serial.print(" mosi: "); Serial.println(PA7);
SPI.setModule(2);
Serial.println("SPI 2");
Serial.print("nss: "); Serial.print(SPI.nssPin()); Serial.print(" sck: "); Serial.print(SPI.sckPin()); Serial.print(" miso: "); Serial.print(SPI.misoPin()); Serial.print(" mosi: "); Serial.println(SPI.mosiPin());
Serial.print("nss: "); Serial.print(PB12); Serial.print(" sck: "); Serial.print(PB13); Serial.print(" miso: "); Serial.print(PB14); Serial.print(" mosi: "); Serial.println(PB15);
SPI.setModule(3);
Serial.println("SPI 3");
Serial.print("nss: "); Serial.print(SPI.nssPin()); Serial.print(" sck: "); Serial.print(SPI.sckPin()); Serial.print(" miso: "); Serial.print(SPI.misoPin()); Serial.print(" mosi: "); Serial.println(SPI.mosiPin());
Serial.print("nss: "); Serial.print(PA15); Serial.print(" sck: "); Serial.print(PB3); Serial.print(" miso: "); Serial.print(PB4); Serial.print(" mosi: "); Serial.println(PB5);

PB4 on the schematic goes to the flash and the nrf24 socket, only!
you might have to play with remapping them.
stephen

PB4 is marked on the schematic as PB4/TRST (JTAG header) and as SPI1_MISO (connected to FLASH and NRF24).
PB4 is SPI3_MISO, not SPI1_MISO. Hence, SPI1 should actually mean (be) SPI3. I think we had already a discussion about this at an earlier time.
SPI3 pins overlap with the SPI1 alternative configuration, but the default declared SPI instance is 3, inline with your pin test.

this fails immediately
What does it mean?
The sketch will hang when the connections are wrong.

I will check this again.

Thanks guys, I have now found PB4 in the NRF interface, I will test for SPI3.

Steve, maybe apologies.

I have tested the SPI3 with the same results, But I then checked everything again and I have realised that the chip on my ethernet module is a W5100 not a W5500. It was an unused one that I had in a drawer, I must have got confused about which it was.

Apologies if this makes a difference.

No worries.
I was just thinking what else can be wrong in your case because everything works for me
But now it’s clear.
OTOH, my lib is restricted to W5500, no other variants are supported.

W5500 is more powerful than W5100, and has the same price, so it worth to have it.

OTOH, my lib is restricted to W5500, no other variants are supported.

Good! We now know what the problem is!

I have just ordered 2 x W5500. If I only ordered 1 it would be faulty!

Sorry for wasting your time.

Hello again,

I’ve been doing a bit of PWM investigation. According to STM32CubeMX there are 32 PWM pins available, and with a bit of jiggery-pokery, I’ve got them all working.

Image

The above image shows the pin allocations for each timer.
The colouring shows the default frequency when using pwmWrite. Timers 1, 8, 9, 10 & 11 are attached to ABP2 which runs at 84MHz, and the others are attached to APB1 which runs at 42MHz, When using pwmWrite the count runs from zero to 65535 and so the frequency should be 84,000,000 / 65536 for timers on on APB1, and 42,000,000 / 65536 for timers on on APB2. The equate to 1282Hz and 641Hz respectively, which are remarkably close to the measured values (or a complete coincidence).

Image

The above image shows those pins that work without changing anything (I basically set a different frequency for each timer by setting the prescaler value and the ARR, and then set a different duty cycle for each channel, so every pwn pin had a unique signature).

(Note that PB4 is accessible via the “bottom right” pin of the NRF24L01 interface taking J-TAG as the “top” of the board.)

Firstly, I think there’s a bit of confusion over PA5 (timer 2 channel 1) as this appears in various places as PA15. If PA15 is indeed an alternative, I could not get it to work. However, after changing the details in Arduino_STM32\STM32F4\variants\generic_f407v\pin_map.c from PA15 to PA5 the pin sprang to life.

Similarly, the entries for PB10, PB11 and the pins for timers TIM10 – TIM14 were simply missing from pin_map.c, after adding in the timer and channel for each of these pins, they all started working too.

Of course, I don’t know whether adding in these entries will adversely affect anything else.

Regards…

Can you do a PR?

PA15 is reserved for SPI3 NSS.
Timer 2 ch 1 is not mapped to PA5 to allow to be used as SPI1 SCK.
In the same context, PB10 and PB11 are reserved for (SPI2_SCK or I2C2_SCL or UASRT3_TX) and (I2C2_SDA or USART3_RX), respectively.

So Timer 2 was given lower prio versus I2C2/USART3, because there are plenty of other timers available on other pins.
Same applies for other pins you mentioned, in favor of SPI, I2C and USART.
Take a look into generic_f407v.h for reserved pins and what are they reserved for.

To activate TIMER2_CH1 on PA15, PA15 must be mapped to the respective alternate function AF1, which is not done in the current version of the repo.
You can use the function gpio_set_af_mode() using my repo.
gpio_set_af_mode(PA15, GPIO_AFMODE_TIM1_2);