Herbert Graf wrote: > On Sat, 2006-07-29 at 18:33 +0200, stef mientki wrote: > >> hi All, >> >> I've been playing around with the ENC28J60 for a few weeks, >> got ARP, IP [PING, TCP [HTTP] ] working, >> but unfortunately not errorless. >> >> I test the psychical connection with a 1000 times ping test, >> and whatever I do, even in the best connection, >> once every 400 pings the ENC-chip hangs. >> >> So my first question, is this behaviour seen by others ? >> > > Nope. Mirroring my experience with most MChip products, the ENC28J60 has > been rock solid for me, probably hundreds of thousands of packets > without a glitch (that I've seen). > > This question just came up (after Firefox crashed): on what kind of network did you uses the ENC-chip ? I just found out that the ENC-chip performed much better, when Firefox was not running. Even better performances occurred, after I shut down the other computers in my network, thunderbird and some other network programs. For now it seems that the ENC chip has a lot of problems, with large data stream (not meant for the ENC-chip) ??? >> Now I've to still find out what causes the problem, >> PHY, MAC or one of the higher level subsystems. >> There are a lot bits/bytes to test, >> and a lot of sub systems that can individual be reset. >> So what's the general protocol to test and/or reset the subsystems >> of the ENC chip ? >> > > By the way you describe things the problem COULD be anywhere. That said, > it's always best to start with the hardware and work from there. Have > you followed ALL the recommendations for layout and component placement > per the datasheet? I couldn't find any layout recommendations / component placements :-( > A noisy power supply, or insufficient or incorrect > decoupling could easily cause the problems you describe. Put a scope on > the power pins of the ENC28J60, send some packets and see how noisy > things are. Adding decoupling can't hurt. > > If all that checks out try swapping the ENC itself, the ethernet cable, > the switch, the computer, see if the problem changes. > > Only after going over everything hardware would I then start looking at > software. Based on your description of the problem though I'd say that > while checking the hardware out you'll find something. > > See above, I all did that, but I couldn't get reproducible results, until I close as much network activities as I could. thanks, Stef Mientki > TTYL > > -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist