Plug and Pray is probably the culprit if the card does PnP, since you are obviously are not doing that initialization. Some cards had utilities that would allow you to force a fixed configuration on them. D-link web site would have them in the archives section (or I can try to find my DE-220 disks). Stupidly they called the program 'setup' so it conflicted with every other 'setup' program out there). You could also try asking D-link tech support for the 'real' part number, since it's an obsolete card. I found them to be most helpful, but that was 5 years ago and times may have changed. One of the tricks you have to watch out for is that some cards expanded their register set by aliasing the extra registers to high addresses. e.g. x330 & x730 looked the same to the bus (only 10 bits were allocated for I/O addresses on the early ISA bus) but would address the shadow registers. That may be why the card is pulling up the high address bits. In an old computer, they would have floated high by default. And the DE220 definitely stores it's configuration on board. Are you toggling the reset line to force it to read the config out of EEPROM? Robert Eric Christensen wrote: > > I've run into this similar problem. I believe the problem is caused by the > network card being used in a jumperless configuration. Many NE2000 network > controllers have additional non-NE2000 defined features that allow base i/o > addresses and other configuration info to be stored in the controller itself. > > For instance, the Realtek 8019AS chip has a whole series of configuration > registers in Bank3 of the non-volatile register file. If the controller is set > to this mode, you must determine (trial and error...) the base i/o address and > then modify the config info. One chipset that I looked at had some very high > base address that you could write to 3 times to change the configured base > address. I don't remember what chip that was, I want to say it was an SMC chip, > but I'm not sure. > > You may be able to reset the base address with a utility created by the card's > manufacturer or an OS. I tried putting mine in a windows machine and it didn't > actually save new config data. I don't like to shutdown my linux box, so I > didn't try it in there :) > > Anyways... I think to figure it out, your going to need to see the datasheet > for whatever chip that card has. I'm not sure how you would find that out > exactly. You could try using it in a linux machine and see what modules it > uses. Typically ne2k cards use the 8390 module for the core ne2k support and > then some other module to get at the special features of the given card. > > Hope that helps, > Eric > > On Tue, 9 Mar 2004 16:25:23 -0500 > Herbert Graf wrote: > > > > > I suppose it's possible they've been rebadged, I don't > > > > know. One is on a > > > > DE-220 NIC and the other is on a DE-250 NIC. I haven't been able > > > > to find any > > > > info on them. Thanks, TTYL > > > > > > I don't know the chip, but if you want to interface your PIC to > > > the ethernet > > > using the cards (I am guessing here), you could connect directly > > > to the ISA > > > connection on the card. There is info on the ISA bus at > > > http://www.techfest.com/hardware/bus/isa.htm. Otherwise, I guess > > > you could > > > trace the tracks back to the IC to find out which signals are on > > > which pin. > > > If you need to repair one, then it probably wouldn't help. > > > > I am interfacing to ISA NICs using the code from the picnic project. > > > > The problem is the DE-250 card is working but my DE-220 cards act like > > they > > aren't there, every register read comes back 0xff. I've confirmed the > > address being addressed and all looked fine until I noticed that bits 11, 12 > > and 13 were being pulled up high by the card all the time?? > > > > I need the datasheet of the chip to figure out why it's in this state, > > perhaps it's not 100% NE2000 compatible, or perhaps there's a part of the > > NE2000 initialization was being skipped. My theory so far has been perhaps > > ISA PnP getting in the way: i.e. the cards are waiting to start the PnP > > process, but further research seems to point to no on that theory. > > > > The DE-250 card is working, but poorly, again I'm working under the > > assumption that there is some slight deviation in either the PIC's routines > > or NE2000 compatibility of the cards. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.