Ethernet issue

Hello Marco,

I cannot locate the data sheet for the device mentioned. If the TM4C129 already has the PHY, then why did you put another PHY. Or is the RJ45 and Magnetics that is different. Please check if the MOSC is 25MHz and RBIAS is installed if using the on chip PHY.

Regards
Amit

Hi Amit.

The schematics is the following attached.

 

May I didn't described well my board but the PHY is the one of TIVA. The only difference between the development board and our hardware is the interface (see my schematics).

 

Hello Marco

Why is the Pin 8 on the integrated magnetic connected to Ground via a 0Ohm resistor in parallel with a Cap, instead of being connected to GND directly?

Regards
Amit

we want to have the possibility to put or not the shield on earth. this is the only reason.

But from your point of view, the connector is compatible with the schematic on development board?

Regards

Another concern is the tamper module init at boot time. Is it true that if the signal on tamper pins are in a such way , the ethernet controller could not be initialized properly?

Regards
MarcoC

Hello Marco

No. Tamper initialization is not going to affect the Ethernet. However there is one debug point which is still missing from your response, which is that of the Crystal Frequency. Can you check if it is 25MHz and oscillating?

Regards
Amit

Hi Amit.

Yes, it's 25MHz and running. It's spread spectrum +/- 25 KHz.

Below my schematics:

Another consideration:
if I connect the board ethernet directly to my PC (either with cross or direct cable) the blink leds are all off.
if I connect the board ethernet via network switch and then to my PC i see link led on and activity led blinking but the board doesn't respond to ping.
The same firmware put into development board works fine without any error (also connecting directly to my the ethernet of my PC).

Do you have any suggestion?

Regards,
Marco

Same update:

during my tests I see that the register "Ethernet MAC Receive Frame Count for CRC Error Frames" increment its value when I'm doing ping from PC, and no EMAC interrupt occurs into EMACSnow_hwiIntFxn().

Marco

Hello Marco,

I also did note that this is a singled ended clock. Do note that for single ended clocks the MOSC has to be configured correctly as well.

Regards
Amit

Hi Amit.

I've configured the MOSC with the following:
// Init peripherals
Board_InitDMA();
Board_InitGPIO();

//SYSCTL_MISC |=
(*((volatile uint32_t *)SYSCTL_IMC)) |= SYSCTL_IMC_MOSCPUPIM;
SYSCTL_MOSCCTL_R |= SYSCTL_MOSCCTL_CVAL;
SYSCTL_MOSCCTL_R |= SYSCTL_MOSCCTL_NOXTAL;

#if TI_DRIVERS_UART_INCLUDED
Board_InitUART();
#endif

#if TI_DRIVERS_EMAC_INCLUDED
Board_InitEMAC();
#endif

and moreover I followed step by step the init of ethernet and seems to be ok the low level init. May be there's another way to check?

Regards,
Marco

Hi Amit.
If I want to check the following register:
EPHYBMCR (0x000), EPHYBMSR (0x001), EPHYSCR (0 x011, bit PSMODE), EPHYSTS (0x010), EPHYCDCR

How can I access to these registers?

Regards,
Marco

hi Amit,

I attach you a screen shot from my oscilloscope.

I capture an init sequence from a ping command from my PC. Are you able to tell me if the waveform are correct?

 

Thanks in advance,

Regards,

Marco

Hi Amit,

I attach you another screenshot related to the link status signal from the board.

 

Hello Marco

EMACPHYRead and EMACPHYExtendedRead API's are to be called to read the registers. Since the registers are below 31, then it should be EMACPHYRead API

Regards
Amit

Hi Amit.
Thanks for your support.
Do you know also the register where I can find frame error bit? I'm looking into memory mapped register but I can't find it.

Regards,
Marco

Hi.

Thanks for your support.

As a matter of fact, we have changed the oscillator with a new one without spread s. and now ethernet works.

Thank you all again.

Marco

Hello Marco

Chester is the true hero of this post.

Regards
Amit