HP Pavilion (ze2000) ZE2005EA 104-Unsupported wireless network device FIX
HP Pavilion (ze2000) ZE2005EA
104-Unsupported wireless network device detected FIX
>> NEW HP/Compaq V2000(verified ok)/M2000/dvXXXX/zeXXXX style BIOS 104/FIX info
>> NEW HP/Compaq NC4200/NX8220/NX6110/NC6000/NX6125 style BIOS 104/fix? info
>> NEW HP ZV6170US/zv5415ea(verified ok)/ZVxxx/R3000(verified ok) style BIOS 104/FIX info
Bought a (new) miniPCI card and been greeted with this dreaded screen?
104-Unsupported wireless network device detected.
System Halted. Remove device and restart.
I imagine if you arrived here you probably did.
This guide will explain how to get ^ this ^ Gigabyte (Atheros AR5005 GS) card working!
Gigabyte GN-WIAG02 miniPCI card on Atheros AR5005GS chipset information
UPDATE 20/May/06 Sebastien Laflamme successfully modded his HP ZD7000, using this method, with most recent BIOS (F.35) to take a Toshiba Intel 2200BG WM3B2200BG PA3362U-1MPC Wireless 11b/g mini pci, Toshiba P/N: G86C0000X310
UPDATE 11/March/06 spilventinjsh successfully modded his NX9020 (same BIOS also in NX9030, NX9040), using this method, to take same Atheros gigabyte card as me!
Here is the full 2 day story, I am sorry it's not concise if you are keen to get it done.
It is also badly written and probably contains mistakes and typo's. I wrote it quickly in the middle of the night, after I got it working before I forgot everything I had done - as much as a record for myself than any other reason.
Well I read several guides to solving this issue in general.
This guy's was very interesting and ultimately without this I wouldn't have known where to begin. .
It is for an IBM Thinkpad by Paul and inspired me to do this
I strongly suggest reading it first as it contains good background to this, that I won't repeat here.
I later on came across some specific guides about HP's & Dells, there are generally 3 existing methods I found;
1)This method involves changing the card's ID's rather than the BIOS to make BIOS think its an acceptable card.
2)This involves even more complications of hot glue guns and SMT soldering switches...hmmm
3)Finally was firing up the laptop with an accepted card in, going into Standby, hot swapping the card and coming back out of Standby. Not exactly practical.
[ I actually ended up doing this to check my card's ID''s were correct later on]
Anyway, I was sure if the codes were in the BIOS, they were changeable, and these work around methods were probably not necessary, so I thought.
The (LONG) Story of Failure and Success
I must have read and read the Paul Sladen guide about 10 times, its alot to get your head around if you never had to deal with it all before.
To start with I poked about in the BIOS (3080F22.wph - How to extra the .wph file from the HP softpaq, in this case SP31067.exe) from HP with and couldn't find anything that looked hopeful.
Then on another re-read I saw about their bios being compressed. Well there was some clear text in the bios i was looking at so I assumed it wasn't, [but in the end I was wrong.
It is compressed (its about 515k).]
So I started messing about with phflash to try and get an uncompressed version. I tried about 4 different versions of it in DOS.
It kept complaining I was running HIMEM when I wasn't. This was using the Windows ME boot disk XP creates. So I decided it was worth a shot with older versions, so got a Dos6.22 image and tried with that.
Well no more complains about Himem. Great I think, soon have this going. I was wrong. To be an image out it seems to have to read in an existing ok bios. Problem was it read the image in, but when it came to writing out old bios it made PC power off instantly.
I tried setting the /BU the /RO and without and loads of other things. Each time a power off. I then thought it was maybe DOS6.22 so i tried a win95 image and same result. Great.
Next I thought maybe it was crapping out because not enough space on floppy to write out uncompressed image. Unfortunately I didn't have a FAT partition to start messing with DOS on so I thought it was a shame this laptop only supports booting off USB floppy.
I thought I was out of luck because I'd tried booting off other USB stuff (that was correctly set up with active partition, MBR etc etc.) and failed.
This is when I had lie #1 told to me by HP
( I was john smith )
what usb devices should boot?
..as in USB HDD, USB-ZIP, USB-floppy etc
BIOS will not detect any other external devices, other than USB Floppy drive.
Well this is bull shit as I discovered to my great surprise. I had some DOS USB utils (panasonic 2.20 thingy) and I thought I could boot from USB floppy, and by using my Sandisk Cruzer Micro, try running phflash from that as it had enough space.
Spent ages pissing about with this and it always resulted in locking up laptop or crashing or not working. I was about to give up when for some reason I cannot remember was looking in BIOS and noticed the Cruzer had appeared under hard drives in the boot options. Now this was odd because my 3.5" to USB adapter made a hdrive always appear under floppy.
Interesting I think, so I am looking at this and wondering...
It's a 1gb Sandisk Cruzer Micro and tried a few 'memory key boot utilities on it' but they didn't like it. Anyway to cut an even longer story short, I found the HP util and it formatted it and made it bootable(does need a DOS floppy to read command.com/io.sys etc from). Its free on on their site.
Well any, I'm back in the BIOS boot order menu I hit F5 to change the boot order and to my amazement it swapped places with hdrive. Quick F10 and holy shit it booted from it, so much for HP's techs.
So now I'm sitting in Win95 with about 980meg free and thinking phflash is going to work.....but not. It just powers off laptop again.
Ok screw that, back into windows.
I then came across this page about using Winflash to dump an image, I was sceptical and I was right, it just dumped another compressed bios image.
Then I went back to Paul Sladen's guide again for another re-read. (At this point I still wasn't all that sure my BIOS was even compressed I only refer to it above as compressed with hindsight)
Next thing to try was phnxdeco.exe which is meant to dump the bios out of the file. He hadn't had much luck with it and for me it also dumps abit and stops with an error.
?-=PhoenixDeco, version 0.31 (DOS)=-?
Filelength : 809B6 (526774 bytes)
Filename : 3080F22.wph
PhoenixBIOS hook found at : 78110
System Information at : 7818A
BootBlock : 7000 bytes
BankSize : 512 KB
Version : DEVELCT1
Start : 78BB5
Offset : 70000
BCP Modules : 32
BCPFCP : 7B573
FCP 1st module : 9B45 (79B45)
Released : 08 August 2005 at 13:36:32
/* Copyrighted Information */
/* ----------------------- */
BootBlock ... O'k
G.0 DECOMPC0.rom ... O'k
U.1 USER1.rom ... O'k
B.0 BIOSCOD0.rom ... O'k
Well back to messing about in DOS, at least a command window this time. After a good fiddle with its options (nothing like a good fiddle) the best I could get was a list of the BIOS parts and a dump of some of it.
Well I figure I'll look at these bits and see if I can find anything interesting.
Now this is where my main breakthrough came. I searched for the (little endian) reversed ID string for a card I knew worked, Intel Pro/Wireless 2100B with code 8086:1043/8086:2551 - I started searching for
the whole lot to begin with, no joy, then 86 80 43 10. This was when I got my first hit. This occured several time in the files.
I then tried 8086:4220 (86 80 20 42) and again got a hit. This was the first time I really thought I was getting somewhere, however it was in more than once place in different files and the sub-system bit wasnt there.
I thought maybe it wasnt subsystem specific so kept going. Then another idea came, to try for the broadcom cards I knew also should work (Drivers for Intel and Broadcom ones on HP''s site) I looked in the .inf again for some ID's and after a few goes hit upon 14E4:4320 (E4 14 20 43).
Now this helped a great deal as it was only in one file (phoenix_.B1). Remembering the output table from the phnxdeco v0.31 I concluded it was part of the bioscode file 01, not that it meant anything much to me.
B.01 ( BIOSCODE) 0A400 (41984) => 0EAC0 ( 60096) LZINT ( 69%) 22B93h
Well this showed it was compressed with LZINT which then reminded me of Paul Sladen's thing that mentioned that. Ok so another re-read and I was satisifed I was dealing with a comrpessed bios....but if I cant fully extract it I'm screwed right?
So after a good google I came upon a site with alot of people into flashing and had a hunt, alas there were mentions of the wireless locks but no solutions. I did happen upon information though, a util called Pheonix bios editor.
A while later i find it, file is BEDemo.zip - of course that is also mentioned on Paul Sladen's guide but it hadn't registered. It really has alot of info but you only realise the half of it as you start getting a clue.
Installed the demo, open the .whp bios and suddenly I have errors. Well to cut another long storey shirt, there is doc's with it from the author saying its normal some bits wont work and you will 'probably' be ok. Very reassuring for BIOS editing, not.
Well I had a good poke around in all the files, I got quite excited with the MultiBoot III thing, I hoped I could change the BIOS to allow selection of removeable storage....wierd thing was it was allready selected. Network boot was deselected which was odd as thats avaialable. So I get sidetracked and decide to change a couple of these options and see what happens.
Anyhow this demo doesn't let you save so fat lot of use that was.
A quick search on an unmentioned P2P service furnished me with finding something more helpful in letting me save. (hint, ZWT)
It was time to call it a night and hope it was downloaded by morning.....
So I am fiddling in Multiboot III menu and change (*) the Network boot and somehting else as a test. Now comes the save, well it shows you a log and
then makes you a nice freash COMPRESSED image, arse. No options to write out a full one. Well can there be any more problems to defeat me?
So I decided to give my altered bios a go in winflash, but first I thought i best reopen it and make sure its still ok. Well the od thing is it came up with a new error...hmm.
(I first checked my warranty, 2 months left.) Well I like to take risks with things like this so I tried it anwyay. Laptop reboots and I go into bios, no change.
Great so I dont know whats failing now, not writing it correctly, not flashing right , who knows.
Well back to looking at the file from PhoenixDeco...I came to the conclusion it must be just not blocking the subsystem, just the main vendor/device part hence the rest of the code wasnt there.
So I remember it was Bioscode 01 file and start looking around in Phoenix editor (v2.0.18E, [latest versions is 126.96.36.199 but seems buggy])..and low and behold there is a section with 5 bioscode chunks in...I think I am onto somehting, I think I can possibly replace the 01 code with the version from PhoenixDeco, hopeing its same and not screwed up as it was from crashed program.
Would you belive it the load/save part is greyed out for Bioscode parts, doh.
Ok, so what next? Well I figure it must temporarily store the stuff somewhere and do a search for .ROM (extension of these bioscode modules) in case I can get to what its working with.
No luck, fek. Then decided to have a poke around in the Phoenix bios editor's dir to see if there was .ini's worth altering stuff in to maybe alow it to save out uncompressed... luckily I had accidently left the program open and to my amazement was a TEMP dir with all the loose files in, yay!
I assume from phoenixdeco the file will be BIOSCOD1.ROM. Which as it turns out it was, on a roll now. Now I did a compare with the one dumped by phoenixdeco and one from Phoenix official editor, they were slightly different lengths, thank god it was greyed out otherwise I am sure I would have killed the laptop importing the slightly wrong sized one!
So why didnt I find them doing a search, well (just this moment while writing this realised while searching for a file i knew existed) turned out i had 'case sensative' still left on from when I was finding something earlier, doh.
Well now, I had a bioscode file with dev/vend id's in, a hex editor and a way to save them back into a bios........ things were finally looking up again.
Time for a play!
I very carefully hex edit the dev/vend code for the Broadcom card whith Poenix bios editor, save the file, go into the editor and make a Multiboot III menu change to allow it to save a change....
So I am now sitting with my new .wph bios trying to decide if I want to risk my laptop.
Is the hex change correct? Will not having a correct checksum kill it? Will having changed the wrong thing kill it? I was most worried because of the Intel codes being in other files, maybe it was just coincidence I could find the dev/vend id's and im completely off-track.
Well I thought about it for a good while and explored the HP site/goole to see what the chances were of getting it RMA'd if I killed it. Not good - but then again I'd got this far, if it went tits up I'd eBay this as spares, get a laptop with a damn wifi card in this time and make sure it wasn't locked out!
Is the elephant going to make its water?
Winflash sets off, painfully slowly (I'd enabled the extra checks and verify's in the Advanced menu)
It reboots - fingers crossed.....it boots, yay!
Stick the Giga-byte GN-WIAG02 in....
104-Unsupported wireless network device detected.
System Halted. Remove device and restart.
Ok, well back on my main pc I have another think, maybe that Broadcom card was on the list bit not enabled, I'll try changing the other one, being as there are 2 and I know Broadcom work, seems like good logic.
ROM file edited then to change an option in the code to allow saving. This time I opted to change the 104 message for a laugh. Ok so I have a new WPH file.
Flashing went ok, so I restart laptop,
EEEK its dead, shit.
Ok I killed my laptop. Brilliant. I know that AWARD and AMi bios had some bootstrap recovery thing you could do with a floppy so maybe is something similar for Phoenix, I am not very hopefull as laptop doesn't even have a floppy, but luckily I had just bought a 2x USB floppy from fleaBay. I didn't imagine a USB one would work in this way though.
I'll cut this bit short as its not really relevant to this, suffice to say I found out about this Phoenix BIOS Crisis Recovery Diskette, eventually tracked one down on IBM's site. I replaced the BIOS file on this disk with mine (must be called bios.wph). It did't seem to work and now after about 4 hours of a dead laptop was about to give up :(. Then I came across a post that mentioned pressing Windows Key-B instead of Fn-B and amazingly the floppy started reading, laptop started beeping at an incredible volume, fans came on full and after what seemed like minutes, it powered off. To my great relief it then rebooted OK and WORKED!
Now the odd thing about this is it was using a fairly old phflash on the floppy yet it managed to reprogram bios, yet my attempts with phlash always failed in DOS 6.22 (as described above!)...computers can be very strange.
With my new sense of bios invincibility I decided to get going again on this problem. I repeated my previous excercise editing biosfile, but without changing the 104 line, and this time didn't have a failure. I broke the rule of changing one thing at a time, so I really don't know if that really killed it but decided I wasn't that keen to break it again to find out for sure!
However it still didn't take my Atheros AR5005GS Gigabyte card. At this point im wondering again if im still editing the right thing and if anything is really changing - but how to determine whats going on? I hit upon a good idea and decided instead of trying to enable my new card, I'd try and disable my old Intel one it liked, a Intel 2100 mPCI 3B - MoW (VEN_8086&DEV_1043&SUBSYS_25278086).
I replaced 8086:1043 with my Gigabyte code 168C:0013.
Now the 3rd great bit of success, my Intel card now gets a 104 Error and so does Gigabyte! We are cooking with gas now!!
But why can this be disabled and not the other one enabled?? Just to eliminate a few other things I replaced one of the Broadcom address with the Intel one and it started working again, what does this mean? At least I know they dont seem to be place orientated in the code.
I figure there must be more to it, but all I have is a load of meaningless hex code in front of me.
Then I wondered if maybe the Susbystem code was perhaps there afterall but in a different place. So I open up all the files again and start hunting for the Subsys 2527:8086 (as 27 25 86 80). Nothing, damn.
Ok well now I am running out of ideas....I repeated in vain hope with Broadcom code but no joy either.
I am a great beliver in keeping everyhting on the pc and not printing stuff out, but for some reason I decided to print a page of this hex code out and look at it for any intelligence amongst the ASCII set next to the hex.
After wasting more time trying to fathom out why it wouldnt f*cking print i discovered the parallel plugs half out the back, doh.
Am now staring at this sheet of numbers ... meaningless to me.
I am looking at this sheet and thnking I have no idea here.... there are 4 characters (8 hex) between each DEV/VEND ID and I am just looking at the pattern trying to see something before or after the Intel/Broadcom ones.
I can see in the ASCII that this seems to be a sort of semi defined start (two chunks of 00's) and a kind of end as it says in the ASCII 'BIOS version'. In the chunk betweek this in the ASCII are = signs which correspond to the gaps between the DEV/VEND codes.
So I am looking in this chunk specifically for anything likely but there are no subsys codes :(
Now came the 4th and final great bit of success, I suddenly noticed 3C 10 below the DEV/VEND codes in a chunk. Now that is part of the Broadcom Subsys ID I had somehow got in my head alot of these HEX numbers by now from staring at them from the .inf files.
next to it is F5 12, so it reads 3C 10 A5 12, no way that is the subsys!!!!, in little endian or whatever. [I later realised I was searching for the wrong thing earlier when I looked for 27 25 86 80 instead of subsystem reversed to 86 80 27 25 on the Intel one - all these numbers make you go loopy after a while]
BIOSCOD1.ROM extract in Editpad
VENDOR/DEVICE ID , SUBSYS ID
- NEW! - Thanks to Solomon Bessire for pointing out missing European ‘RoW’ 2200 3C10F612!
- Just verified the ZE2202 has similar format to this ZE2000(ZE2005EA), ZV6170US has it in ROMEXEC0.ROM instead.
I quickly discovered I think 9 other Broadcom ones and 2 Intel ones. I am now gaining in confidence!! The only strange thing is I couldnt find the subsys for the Intel 2200G card (DEV 8086 VEND 4220), even though DEV/VEND was present...well I am not 100% sure this laptop came with that so not to worry.
Ok, time to replace something and try it. I chose a Broadcom one. I replaced E4 14 20 43 with 8C 16 13 00 (Device/vendor) and 3C 10 F8 12 with 58 14 11 E9 (Subsys)
[remember these new values were taken from Atheros info VEN_168C&DEV_0013&SUBSYS_E9111458]
Flashed the bios, IT WORKED, SUCCESS! No more 104 on my Gigabyte at boot time!!!!!
- BTW. These Atheros cards are quite hard to come by at least in the UK, the best deal (by a mile) I have found on one is from DigitalKind Link for the Atheros GN-WIAG02/(AR5005GS) good price and fast service.
Only trouble was now it had the damn wireless hardware switch off, and HP in their wisdom decided to save about 1 penny by missing out a SMT switch,resistor and LED from the board where that should have been with the on/off and audio buttons.
Fortunately I had allready gone though this getting my Intel one to work to begin with so I knew the drill.
Remove the battery, remove the two screws marked with a keyboard symbol in the battery compartment, lever the plastic off thats in front of keyboard (stick a knife/fine ended screwdriver under the small hole in front of F11), pull this up abit(no need to remove it totally) put the screwdriver/blade across 1&2 on SW10 (see pic), the blue LED at the front should then may come on if its not allready one (depends on card - from memory i think on Intel card it comes on and Atheros it was allready on).
Now it works (in windows now!)
One happily working Gigabyte SuperG mini pci card in Client Utilities (I soon changed to Wireless Zero Config though!)
So there you have it, how to frig your BIOS to take any card, just replace the ID's and reflash. It will take about 30 seconds to open bios, 1 minute to replace HEX if you are carefull, 30 sec to resave it out, couple minutes to flash. So all in all a quick 5 minute procedure, and if you do change the wrong hex, you most likely can recover it anyway.
Screen Antenna Mod - make sure you get ones with long enough (super high loss) cables!
Please drop me a note if this helped you, I am interested in the amount of people who think HP and IBM etc. practise of crippling machines is important. I think this is especially stupid in my case as mine doesnt even have antenna's in it or a wifi card to nullify any FCC regs!
In writing this I want to thank Paul Sladen's guide without which this woulnd't have happened and a guy called Snyder at HP for his complete incompetance.
Lie #2 told to me by HP
( I was john smith )
do you know if there is a BIOS restriction in the ze2005ea that will only let certain wifi cards work?
Please go ahead.
miniPCI ones sorry
No, there is no restriction on the wireless cards in BIOS, You can use any wifi/wireless card.
Whitelist for ZE2005EA
Intel PRO/Wireless 2200
PCI\VEN_8086&DEV_4220&SUBSYS_27618086 (NO subsys info?)
PCI\VEN_8086&DEV_4220&SUBSYS_27628086 (NO subsys info?)
PCI\VEN_8086&DEV_4220&SUBSYS_12F5103C (NO subsys info?)
PCI\VEN_8086&DEV_4220&SUBSYS_12F6103C (NO subsys info?)
Intel PRO/Wireless 2100 3B Mini PCI Adapter
%NIC_MPCI3% = Install_MPCI3_W2K, PCI\VEN_8086&DEV_1043&SUBSYS_25278086 ; Intel 2100 mPCI 3B - MoW YES
%NIC_MPCI3% = Install_MPCI3_W2K, PCI\VEN_8086&DEV_1043&SUBSYS_25228086 ; Intel 2100 mPCI 3B - RoW YES
%NIC_MPCI4% = Install_MPCI4_W2K, PCI\VEN_8086&DEV_1043&SUBSYS_25238086 ; Intel 2100 mPCI 3A - RoW NO subsys
%NIC_MPCI3% = Install_MPCI3_W2K, PCI\VEN_8086&DEV_1043&SUBSYS_25288086 ; Intel 2100 mPCI 3B - RoW2 NO subsys
%NIC_MPCI3% = Install_MPCI3_W2K, PCI\VEN_8086&DEV_1043&SUBSYS_25298086 ; Intel 2100 mPCI 3B - France NO subsys
%NIC_MPCI4% = Install_MPCI4_W2K, PCI\VEN_8086&DEV_1043&SUBSYS_252B8086 ; Intel 2100 mPCI 3A - MoW NO subsys
%NIC_MPCI4% = Install_MPCI4_W2K, PCI\VEN_8086&DEV_1043&SUBSYS_252C8086 ; Intel 2100 mPCI 3A - RoW2 NO subsys
%NIC_MPCI4% = Install_MPCI4_W2K, PCI\VEN_8086&DEV_1043&SUBSYS_252D8086 ; Intel 2100 mPCI 3A - France NO subsys
Broadcom 802.11b/g Adapter
%BCM430B_DeviceDesc% = BCM43XX, PCI\VEN_14E4&DEV_4301&SUBSYS_12F3103C NO
%BCM430G_DeviceDesc% = BCM43XG, PCI\VEN_14E4&DEV_4320&SUBSYS_00E70E11 NO
%BCM430G_DeviceDesc% = BCM43XGT, PCI\VEN_14E4&DEV_4320&SUBSYS_12F4103C NO (no subsys)
%BCM430G_DeviceDesc% = BCM43XG1, PCI\VEN_14E4&DEV_4320&SUBSYS_12F8103C YES
%BCM430G_DeviceDesc% = BCM43XG2, PCI\VEN_14E4&DEV_4320&SUBSYS_12FA103C YES
%BCM430G_DeviceDesc% = BCM43XG3, PCI\VEN_14E4&DEV_4320&SUBSYS_12FB103C YES
%BCM430M_DeviceDesc% = BCM43XM1, PCI\VEN_14E4&DEV_4324&SUBSYS_12F9103C NO
%BCM430M_DeviceDesc% = BCM43XM2, PCI\VEN_14E4&DEV_4324&SUBSYS_12FC103C NO
%BCM430M_DeviceDesc% = BCM43XM3, PCI\VEN_14E4&DEV_4324&SUBSYS_12FD103C NO
%BCM430G_DeviceDesc% = BCM43XG1, PCI\VEN_14E4&DEV_4318&SUBSYS_1355103C YES
%BCM430G_DeviceDesc% = BCM43XG2, PCI\VEN_14E4&DEV_4318&SUBSYS_1356103C YES
%BCM430G_DeviceDesc% = BCM43XG3, PCI\VEN_14E4&DEV_4318&SUBSYS_1357103C YES
%BCM430M_DeviceDesc% = BCM43XM1, PCI\VEN_14E4&DEV_4319&SUBSYS_1358103C No
%BCM430M_DeviceDesc% = BCM43XM2, PCI\VEN_14E4&DEV_4319&SUBSYS_1359103C NO
%BCM430M_DeviceDesc% = BCM43XM3, PCI\VEN_14E4&DEV_4319&SUBSYS_135A103C NO
My Gigabyte / Giga-Byte / GN-WIAG02 / Atheros / AR5005GS
%GIGABYTE.DeviceDesc.E911% = GIGA_DEV_E911.ndi, PCI\VEN_168C&DEV_0013&SUBSYS_E9111458 YES NOW :) !!!
Subsystems for device 1458:e911: GN-WIAG02
- FULL HP W400/W500 Atheros based list
- FULL Intel based List
- FULL Broadcom based list (& HP W450)
If anyone has been inspired to
edit their bios more and change their boot splash logo, click here
for a forum post on Borg No. One's forum.
It relates to Dell rather than HP but they use same Phoenix BIOS. (however dumpvgabios didn't work on my bios though, so your mileage may vary)
Mail me at anything to this domain - please keep any initial questions short and to the point.
Please dont mail asking for a shorter version as I have had alot of requests allready - I may get round to it oneday.
Due to nature of the strings/code being different and in different files, I am reluctant in case anyone goes in blind and screws up.
Please do mail me if you do or don't have success with your laptop model/bios version/what you did.
Incidently I have got this laptop into a state when LILO went funny once [had slightly wrong partition size/info] so that it actually corrupted the BIOS when booting from Linux. It stopped the laptop booting at the LOGO screen (which went larger size than normal and off the screen at one side). Flashing the BIOS DIDNT fix it and eventually fixed it by getting into the BIOS and loading the default values, very odd. Booting again from LILO/Linux after resetting BIOS caused this repeatable problem until I fixed LILO.