Hi all,

Has anyone tried to use the LPC2368 with DP83848 phy runing at 100Mbits at a two layers PCB? Did it work? Anyway, has anyone successfully used the LPC2368 with the DP83848 phy runing the ethernet at 100Mbits? At work we just cant put the LPC2368 (with DP83848) to run at 100Mbits, excpet for a forbidden PLL frequency output (198MHz, if doesnt fail me) when the ethernet runs perfectly at 100Mbits with less than 1% packet loss.

Thanks,

Fernando Almeida.


---------------------------------
Novos enderes, o Yahoo! que vocconhece. Crie um email novo com a sua cara @ymail.com ou @rocketmail.com.


Start a New Thread

Hi Fernando,



I have a board with the LPC2368 and another one with the LPC2366, both with
the

DP83848I (industrial version) running almost without no problem.

My board is a 2-layer board, with length-compensated tracks, 0.5oz copper
and

8 mils spacing in the high density areas.



I said almost, because I faced an estrange error with the LPC2366, at the

transmitter side (receiving packets is perfect).

The problem is, randomly, after some packets are sent, the MAC simply

stalls. Let me explain: the MAC is supposed to send packets as long as

the descriptor array has some data, and the latter condition is defined as,

whenever the TxProducerIndex is different of TxConsumerIndex.

So the MAC stalls with the TxProducerIndex having some data and there

is no apparent reason for that. No pause frames received, no underrun,

no collision errors. What I can say is that the frames where 'produced'

very fast.

With the LPC2368 however, I put the same code, and this problem didn't

show. I changed the LPC2366 for a newone and same behavior.



Belive me I wasted a lot of time trying to figure out why this happens but,

the thing is that, I put a work arround and now it's working.

I contacted NXP about this issue but I didn't get an answer so far.



So, regarding your question, as long as you got a good layout,

and supply right power rails, and use the proper oscillator (50MHz),

it will work fine.





bye

Jorge







-------Original Message-------



From: Fernando Augusto

Date: 6/26/2008 15:21

To: l...

Subject: [lpc2000] LPC2368 ethernet 100MBits



Hi all,



Has anyone tried to use the LPC2368 with DP83848 phy runing at 100Mbits at a
two layers PCB? Did it work? Anyway, has anyone successfully used the
LPC2368 with the DP83848 phy runing the ethernet at 100Mbits? At work we
just cant put the LPC2368 (with DP83848) to run at 100Mbits, excpet for a
forbidden PLL frequency output (198MHz, if doesnt fail me) when the ethernet
runs perfectly at 100Mbits with less than 1% packet loss.



Thanks,



Fernando Almeida.



---------------------------------

Novos enderes, o Yahoo! que vocconhece. Crie um email novo com a sua
cara @ymail.com ou @rocketmail.com.














Start a New Thread

Hello...

i build a board with LPC 2368 and DP83848C with some SMD-PCB-Adaters on a breadboard. It works with 10Mbs... but: I have the same problem like Jorge. I build some UDP packages in a sending-function and send them as fast as possible. To prevent a buffer problems i check the state of TxProduceIndex and TxConsumeIndex. If TxProduceIndex is TxConsumeIndex - 1 (i noticed wrap around) i return 0 and put nothing in the ring-buffer. But if this condition occurs at TxConsumeIndex=2 and TxProduceIndex=1 nothing happens anymore :( Normally i expect the DMA engine to send packages - so i could send further packages after a small wait-time.
Greets

Sven
--
Psssst! Schon vom neuen GMX MultiMessenger geht?
Der kann`s mit allen: http://www.gmx.net/de/go/multimessenger

Start a New Thread

I have LPC2368 and DP83848 on double sided board with no length compensation. I haven't noticed any problems. I have 50 MHz oscillator connected to the PHY. Clock for the MCU can be directly from the oscillator or from the PHY; it is working in both cases.
PCLK should not be very important. I'm almost sure that LPC2368 cannot run at 198 MHz.

Start a New Thread

Hi,

thanks for the answers guys, as I feared I'll have to redraw tha board. In the previous board (the one with the 100Mbit problem) I've used 10mils tracks width with 10mils spacing, the ethernet tracks all run in parallel (but did'nt check for length match), neither used termination resistors and the LP2368 is about 3 inches from the DP83848I. Have you guys used any termination resistors? As for the track width I've seen 10mils in some document before I decide to use 10mils, but I also remember seeing 8mils width, what width did you guys used? Well, I'll appreciate any hint anyone can give me.

Thanks in advance,
Fernando Almeida.

"Glazar, Bostjan" escreveu: I have LPC2368 and DP83848 on double sided board with no length compensation. I haven't noticed any problems. I have 50 MHz oscillator connected to the PHY. Clock for the MCU can be directly from the oscillator or from the PHY; it is working in both cases.
PCLK should not be very important. I'm almost sure that LPC2368 cannot run at 198 MHz.








---------------------------------
Novos enderes, o Yahoo! que vocconhece. Crie um email novo com a sua cara @ymail.com ou @rocketmail.com.


Start a New Thread