Network card and RTC being troublesome

[dongfeng]

I've come across this a couple of times, when putting network cards in my 8088 systems.

I have three "NE1000" cards and one WD8003. I also have a 3Com Etherlink II, but it seems horribly complicated and I haven't sussed it out yet

For example, tonight I was playing with my "bitsa" XT (the "slow BIOS" one ) and I put in the 8-bit NE1000 network card I had working in my 386. I had it working in this XT about a month ago, but since then I had installed a SMT no-slot-clock (a little gadget that fits under one of the BIOS chips to provide an RTC function). Network card seems to initialise okay, but cannot connect to anything.

Tomorrow I will pull the no-slot-clock and see what happens.

I had exactly the same problem in my 5150. That had a QuadBoard QuadRAM expansion card with RTC. The NE1000 network card WOULD NOT connect, although appeared to initialise properly. On disabling the RTC on the card, it worked without fault. I put in the WD8003 card instead, and it is now happy.

Having said that, the same NE1000 card works fine in my other XT with AST RTC card, and also my Amstrad PC1640 with it's weird Amstrad RTC.

And I have checked the card interrupts!!!

Is it a common problem?

 

[modem7]

Conflicts

Conflicts

Conflicting ISA cards were fairly common. Not only do you need to look our for common interrupts, but you also need to look out for cards that are using the same I/O ports, same DMA channels and same memory addresses.

And note that just because your XXXX card has jumpers/switches to control the used interrupt and I/O address range doesn't mean that the card isn't decoding a range of memory addresses.

Sometimes the user manual that comes with a card has the required 'usage' information, but sadly I found that to be rare. Even if the manual for a simple RTC card indicated that the card used I/O ports XXX through YYY, it would be handy.

Some diagnostic software (I think Checkit may be one) can show interrupt/DMA/IO usage in your ISA based computer. That can aid you, but it is only an aid because such diagnostic software cannot determine everything. For example, I can engineer my own ISA board that counts triggers from an external device. The diagnostic software can in no way determine that my card is going to use interrupt 3. It should be able to determine that there is TSR named HAPPY.EXE that is hooked into interrupt 3, but it won't be able to make the connection between HAPPY.EXE and my card.

 

[Trixter]

Also, not to scare you, but some cards need to be configured using their software, which gets written into nvram somewhere on the card, and then you can use them.

I have an Intel EtherExpress 8/16 in mine; it's a 16-bit ISA card but will operate in 8-bit mode if you configure it that way. Only problem is, the damn install software only works on a 386 or higher!! So I had to put it into an ISA-bus machine with a 386 or higher (don't remember which one), boot DOS, run the software, configure the card, then I could rip it out and THEN the packet driver worked.

I've never had a problem with the 3-com cards; their packet drivers usually work fine. Xircom parallel-port adapters work perfectly too (which is not something I ever would have tried until Mike Brutman convinced me otherwise). I have never tried a "generic" ne2000 card because I've never been able to find one.

BTW, also in my experience, some (most?) 3-com 16-bit ISA cards work in 8-bit mode as well.

 

[dongfeng]

Thanks for the info, very interesting!

I have used CheckIt to help me find the ideal interrupts and I/O settings, but unfortunately it seems not to pick up the card. It's nearly always interrupts 2 and 3 that are free (sometimes 4). Without the detailed hardware manuals it is quite difficult to get things working properly.

I've had all the cards working in different XT class hardware at one time or another, so I know it is just my XT being picky.

I'll see if it will run later, with the RTC removed

 

[dongfeng]

I have finally managed to get it working now

I took the card from the Amstrad PC1640 (8MHz 8086), and although being of the same design is a second revision. It works in the XT! The Amstrad also seems quite happy with the first revision card that was giving me problems in the XT.

What was confusing me yesterday was that the card would not work regardless of whatever IRQ or I/O settings I gave it

This is using an 8-bit card with a transceiver on the AUI port.

 

[mbbrutman]

The 'no slot' clock solutions that fit under a ROM chip don't use IRQs or DMA, and generally don't intefere with anything. Look at http://www.brutman.com/PCjr/DS1216E.html for a description on how they work.

(I've written driver software for the Dallas versions of the chip. It is a clever solution.)

 

[dongfeng]

In that case, I don't know what was causing the problems, other than a fussy card. The XT has a memory expansion, ST-412 hard drive controller and floppy controller installed - all parts one would assume wouldn't cause problems.

The No-Slot-Clock seems like a great addition, although my version was made in 1990, I am wondering how much life I can get out of it before the battery dies. I found a Chinese supplier who can supply them new, but the minimum order quantity was 1,000! I don't think my collection can top that!

I have the SMT drivers disk for the IBM PC and Apple II.