Problem with IDE devices detection during boot

[Wade]

Hi.

I seem to be having a problem with my TP600. Each time during boot-up it doesn't see neither hard drive, nor cd-rom, only floppy drive. I thought this to be a fatal problem, but suddenly tried 'initialize'. It worked!

And it works. Though now I have to 'initialize' it each time before the actual boot. Then everything seems to be working. Even in the case of reboot I have to do the same thing.

I have measured the voltage of my BIOS battery, that's not the case, it's 3V, nominal voltage.

Maybe somebody had such a problem before?

Btw, it seems almost allright, despite the fact that I cannot get a wheel mouse's wheel working without disabling the trackpoint, because the latter change needs rebooting to take effect.

[pkiff]

I haven't seen this or seen a post about it before. Are you booting up with a floppy disk then? I don't understand when each thing happens (i.e. how can you boot if it doesn't recognize anything but the floppy, and how do you know only the floppy is recognized if you haven't booted up).

And how did you get an operating system installed on the hard drive?

What happens if you change your boot order? Can you set it to hard drive first, or does the hard drive appear greyed out in the list of items to choose from?

Have you got another hard drive you can try to do a fresh install on? Was it working before with a previous hard drive or have you never had it working?

Not very helpful, I guess, just a bunch of questions!

Phil.

[Wade]

Phil, those are good questions, I guess I haven't been straightforward enough in my explanations.

1) I know that it doesn't detect IDE devices by the following: first, you can hear those two (hdd and cdrom being initialized immediately after switch on since they give away some sounds (like spinning up sound). In case of non-detection they don't. Second, if nothing appropriate was found, the diskette is drawn on the screen (invitation to insert one into a floppy drive) and invitation to press F1. In this case, if I insert the boot floppy I may boot from it, therefore I know the floppy is recognized. And from that floppy no cdrom devices are found (during the cdrom driver start on win98 boot disk) or hdds (fdisk report).

2) I installed windows 98 because it haven't always been like that. I believe something happened spontaneously. Before this BIOS problem, I could change the boot order and install OS as usual.

3) Like I said, after I do initialize I can change boot order. It's like that:

I boot my TP600. If I do nothing, I get this 'diskette mode'.
If I enter bios, do initialize and make reboot, next immediate boot (which follows after I click on 'reboot') I have everything working. The problem is that by initializing it sets some defaults (like trackpointer enabled, volume level etc.), and I can't get the wheel working, that I believe you know why Since to disable trackpointer I need to reboot, and upon reboot I get the diskette mode, and if I do initialize (my only option) trackpointer is enabled again by default.

4) Yeah, I have 2 more hard drives which I can use, but that is not an option, reason mentioned above.

That is something to do with BIOS. It doesn't remember... Like an old lady. It doesn't remember it has those 2 neat devices unless reminded, and after reminded, it remembers that for 1 session, after which it has to be reminded again.

I guess I have cleared the situation a bit

Oh, btw, I have found a post somewhere where the guy talked about the same problem, IBM told him to replace a CMOS battery. But in my case, the battery is fine, I wrote about that in my first post.

[Wade]

UPD: surprisingly enough, everything works fine again.

I am clueless about what happened. Nevertheless, I'll watch it and write about it if I come across anything that deserves attention of the respected comunity

Phil, thanks again

[rkawakami]

Oh, btw, I have found a post somewhere where the guy talked about the same problem, IBM told him to replace a CMOS battery. But in my case, the battery is fine, I wrote about that in my first post.

Don't be so fast to dismiss the battery even if it reads a "nominal" 3.0v. What is more important to the proper operation of the backup battery is current. I just had a CMOS battery from a T23 that read 3v but supplied next to nothing in the way of current (approx. 20mA). It continued to have problems with remembering the BIOS setup until I swapped it out with a fresh one (which measured over 600mA when connected directly to my meter). This doesn't make that much sense to me because the current draw of the BIOS chip should be in the microamp region, but that's what I saw. I would suggest you measure the battery's current capability and if it is under 50mA or so, try a fresh one. Everything you've said about the BIOS forgetting the setup points to the battery as being the cause. If that doesn't fix the problem, then perhaps the issue is with the motherboard.

[Rick Aguinaldo]

It continued to have problems with remembering the BIOS setup until I swapped it out with a fresh one (which measured over 600mA when connected directly to my meter)...

Ray, I suppose you don't mean to put the meter leads across the battery with the meter set to mA/A range, right? A fresh coin battery may develop an internal resistance high enough to limit the current below the internal fuse's 2A rating of most commercial multi-testers, but shorting it is not advisable. A better way would be to load the battery to a few or even 1ma since the operating load would be in the microamps. A good battery should have an onload voltage greater than the minimun open circuit voltage specified in the HMM.

In your case with the dud cmos battery, shouldn't the date and time cmos data were lost as well, triggering the 161/163 errors? If the errors were not triggered, then the tip is another valuable info to keep. BTW, second time bumping into you Ray . First time was something related to locating the 24rf08 in the 600x (in another forum). Thanks for the great resources you have provided here and elsewhere.

[rkawakami]

Ray, I suppose you don't mean to put the meter leads across the battery with the meter set to mA/A range, right?

Yes that's what I meant and what I do, but you're right, I should have prefaced it with "quickly measure". I put my meters on the 2A or 20A range and touch the leads down on the battery very quickly. Don't leave the leads on the battery for anything longer than a half second. With the coin battery there isn't too much potential for blowing out the meter's fuse. Do this same thing with a freshly charged, NiMH D cell and be prepared to open up your meter and replace a smoking fuse .

A better way would be to load the battery to a few or even 1ma since the operating load would be in the microamps.

Agreed, but not many people have an assortment of resistors handy.

In your case with the dud cmos battery, shouldn't the date and time cmos data were lost as well, triggering the 161/163 errors?

Yes, but my T23 generated a couple of other error codes as well. I don't recall exactly what they were at the moment, but it was something along the lines of "CMOS configuration invalid, resetting to defaults".

BTW, second time bumping into you Ray . First time was something related to locating the 24rf08 in the 600x (in another forum). Thanks for the great resources you have provided here and elsewhere.

Hmm.... (Ray going off to check another forum...), Ricardo, I presume?