Light is not on when debugging the 88e1111

[fengye2009]

Here is a board designed by myself, the power supply and oscilator is normal ,but when i connect the cable,the mode led is not on, what's the problem? And another strange thing is ,the MDI pin has no output  after power on.(In ML507 board the MDI has output ) .

[gszakacs]

This sounds like a pretty hard thing to debug over a web forum.  There are a number of reasons

why you might not get link connection, especially on a new board design.  For example, there

could be problems with the signal routing between the 88E1111 and the magnetics  and jack.

If you have more than one board to try, you might find it is an assembly problem.  Which package

is the 88E1111?  Many of these are QFN style which can be hard to see if they are soldered

properly.  What are you using for magnetics?  Is it an integrated "mag/jack" or separate

transformer plus connector?  Did you copy the design from an eval board schematic?  If not

did you follow the recommendations of the chip and magnetics manufacturers?  Did you

check all of the supply voltages to the 88E1111 and the magnetics?  The more details you

provide, the better help you will receive.

-- Gabor

[eteam00]

Did you copy the design from an eval board schematic?

My favourite new-board-design method of shooting myself in the foot is:  incorrect layout 'dolly' (i.e. board layout footprint).

Are there any components on this board design which have not been used before (e.g. RJ45 connector, e'net PHY, FPGA)?  These are the likely candidates for unproven layout dollies.

-- Bob Elkind

[joelby]

If you've set sensible defaults using the configuration pins, the PHY should light up LEDs when you plug a cable in, even without programming any registers. Check the RESETn pin - it is active low, so you must hold it high during normal operation. If you hold the 88E1111 in reset, it will not light any of its status LEDs.

[fengye2009]

I'm really clear that it's a pretty hard thing to debug over a web forum. But i have spent over a week to solute this problem.So i have to post in the web forum.

The Ethernet chip package  on my board is 128-PQFP, and i have tested all of the power and gnd pin ,all are normal (comparde with the ML507 board). but i want to make sure :

1、if the MDI have  waveform ouput after power on .

2、 what's peak_to_peak value of the waveform?

3、if this waveform is used to synchronize with the RJ45?

[joelby]

First of all, what voltage level do you measure on your RESETn pin?

The MDI signals and levels will depend on the speed of your network (10/100/1000). You could refer to the appropriate standard (e.g. 802.3ab) for details about what to expect.

If the PHY is auto-negotiating properly, you should expect to see a clock on the RX_CLK output. If the PHY is running at all, you should see something on 125CLK.

[gszakacs]

If the PHY is running at all, you should see something on 125CLK.

Actually for the 88E1111, the 125CLK output can be disabled as a strapping option.

In fact the strapping resistors / connections are a good place to start looking for

differences between your design and the ML507.

-- Gabor

[fengye2009]

first,the voltage level measured on RESETn pin is 3.3V, this is the first difference between ML507(2.5V),but the datasheet has such words “3.3V tolerant ”,so i think it doesn't mattter. About pin TX_CLK,RX_CLK,125CLK, the clock frequency on these pins are 2.5MHz,125MHz,125MHz,i think they are normal. Actually,the biggest diffence between ML507 and my board is : the peak_to_peak value of ouput on pin MDI is 2V ,but it's only 100mv on my board, i guess these waveform is used to synchronize with the RJ45,to make sure which mode should be used ,but 100mv is so small to CMOS ,so it can,t be detected. Now I'm puzzling why this output on my board is so small.

[adityag]

boss if u could found the solution for this issue then kindly post it