RE: Debug DP83848

Miroslav,

Is your measurement being taken on the same design as the previous posts?  If it is a different design, would you mind if I split this thread into a separate post and referenced the current post in order to avoid any confusion between the posts?

Patrick

Hi  Patrick,

I  am using as a basic design Microchip's  Starter Kit schematic, but with 64pin PIC32MX795F512H  chip. DP83848 must be in RMII Mode, so i strap pin39 to 3.3V through 2,2k pill-up resistor. After powering up my design without DP83848, pin39 is 3.3V. When DP83848 is connected, i measure 0V at pin39. Pin 6 is strapped at 0V.

My general problem is that there is no communication between PIC32 and PHY. Both IC's are working separately. When i plug ethernet cable LED LINK gets ON, LED ACT is blinking from time to time but PIC32 can't read/write to PHY. So i started to check connections and only think what confuses me is pin39. I expect to read 3.3V on it, because pin39 is pulled-up by 2.2k resistor to 3.3V (RMII mode is what I need so pin39 is strapped to 3.3V and pin6 is strapped to GND). Instead of 3.3V i read 0. Because i don't know inside pin39 connection i am not sure what must be read from pin39. I check for short connections and did not find any, also when PHY is desoldered from PCB i read 3.3V on pin39.

I am not an expert (I am fighting with a similar 83848 problem like yours:)), so this is my opinion only:

Pin39 RX_DV/MII_MODE configuration strap is read after reset (this is when pull-up counts) for  short time.

Then pin39 switches to output mode as RX_DV.  This signal is "RMII Receive Data Valid" (active high), so 0V simply means no valid data received. You have your problem (not receiving), but this is a normal symptom, not the cause.

Miroslav,

The input from Tamas is quite correct.  In normal mode, RX_DV will report data.  If there is no data, the device will drive the pin low. 

If you want to confirm the strap value, you should hold the RESET_N pin low and measure the voltage at the RX_DV pin.  In that case, the 2.2 kOhm resistor should pull it high. 

When you note that you cannot read or write to the Phy, are you referring to register access or packet traffic?  Are you able to read and write the device registers?

Patrick

Thank you! The problem is more general and i can't find the reason. MAC gives init sequence to PHY, after that MAC tries to read address 0 from PHY (using MDIO and MDC) several times (that comes from software init procedure ) and because receives all ones error is returned in software (i saw this process on scope). I checked my design many times and i cant find any problem. My next step will be to redesign my PCB.

Hi, I am facing the same problem: the MAC send through MDIO/MDC the request for the status but the 83848 seems not replying, i.e. after the request  I see the MDIO always set to 1.

Have you solved your problem?

Hi Michele,

Yes i found the problem in my case. It was very stupid, as most cases. Design problem with PCB. Between TX and RX pins i found very "light" short connection at MCU side. I found it accidentally as i start making  things that comes from helplessness :) .

Hi I solved just yesterday, in the PCB the power feedback 1 and 2 and the power out wasn't tied together... the effect was the same you was experiencing, furthermore there was no pullup on mdio line...

Anyway thx again for your reply! I hope that if in the following I will need help I can still ask you!

Michele