Archived from groups: comp.sys.ibm.pc.hardware.storage (
More info?)
alfredeneumann@madnet.org wrote:
>> alfredeneumann@madnet.org wrote:
>>
>>>> Rod Speed <rod_speed@yahoo.com> wrote
>>>
>>>>>> alfredeneumann@madnet.org wrote
>>
>>
>>>>>>>>>>>>>>>> Bios says drives are set for dma/level 2,
>>>>>>>>>>>>>>>> but under win98 dma is not enabled.
>>
>>
>>>>>>>>>>>> Thats normal, if its getting a lot of errors from a drive,
>>>>>>>>>>>> Win98 turns the dma off because its safer with it off.
>>
>>
>>>> I think that was the original setting when I first go the box.
>>
>>
>> That would be unusual.
>
> (sorry for the poor formating, trying to fix it)
> The bios was set to "auto" for the drives and the DMA/2 option was
> faded out (meaning will auto detect mode).
> I enabled DMA in windows and it took, but it has no effect on the 4-5
> second
> delay when it first, under explorer, reads the drive(s).
> The drive it killed I have already replaced with another WD 8GB,
> which tests
> good. Still geting 4-5 second delay (approx.) when first reading the
> drives
> with explorer.
> Maybe this is normal, cuz I do have alot of directories and
> subdirectories.
Yeah, 98 is pretty primitive in that area.
> Used AmiDiag for win95 which reported DMA controllers as good;
> nothing else
> wrong, except I get RTC Periodic Interrupt not coming, which I always
> get with AmiDiag even in other systems.
> Have not run another diag. program yet, still installing and finding
> others.
>>>>>>>> Thanks alot for your tips, Rod. I probably will buy some of
>>>>>>>> those 80 wire cables as I think the drives are capable of
>>>>>>>> ATA33
>>
>>
>>>>>> You dont need 80 wire for that.
>>
>>
>>>> Can you elaborate, several have said it is required for AT66?
>>
>>
>> Yes, that is why I broke your sentence there, it isnt for ATA33.
> Ok, is DMA/Level 2 same as ATA33 or 66?
It isnt directly related.
>>>> If I have to replace the cables, might as well get them.
>>
>>
>>>>>>>> or maybe even 66. The ribbons that are in there are all twisted
>>>>>>>> up and so short it's impossible to get them straightened out.
>>
>>
>>>>>> That doesnt matter. When a ribbon cable fails its usually
>>>>>> because the connector that bites the ribbon isnt making
>>>>>> good contact anymore. That can happen when the
>>>>>> ribbon is used to pull the connector off the drive.
>>
>>
>>>> Ok, they were twisted pretty bad. I straightened them
>>>> out a little and reconnected them, so maybe that's why
>>>> creating new folders now auto refreshes the new name,
>>
>>
>> Yes, if the connectors arent biting the cable solidly, rearranging
>> them may well have been enough to change things.
>
> It is quite possible this was the problem; room temperature might
> just be coincidental with my resetting the various jacks.
>>
>>
>>>> but I would think this is a video card (voodoo function?
>>
>>
>> Nope, nothing to do with the video card.
>
> Good to know.
>>
>>
>>>>>>>> The drives take forever to start up and spin
>>>>>>>> when they have been idle for just a few seconds.
>>
>>
>>>>>> That may just the be bad sectors again.
>>
>>
>>>> No, no bad sectors on either drive, ran DLG, I think on at least
>>>> one of them and scandisk on both. That is no bad sectors that I am
>>>> aware of.
>>
>>
>> So why did DLG claim that one drive had died ?
>
> That was two-three days ago when I first posted, already replaced it
> with another 8GB drive WD, good. It also was starting to click, but
> usually just once at
> end of scandisk; maybe this has gone away since I reseated everything.
> The thing that is so annoying is the 5 second delay when the drive
> first reads
> the root directory. After first reading it is normal/fast, but the
> first read
> takes forever.
> Also tried again to move directory from one drive to the slave; it
> was fast.
> I think maybe the reseating of the plugs might have solved the
> problem,except
> for the initial directory listing delay under explorer. I have the
> buffers
> and files setting set in my config.sys, would those settings effect
> speed of
> disk access under w98?
Yes.
>>>>>>>> Also noticed that the reason why new folders do not refresh
>>>>>>>> automatically to their new names when they are made and
>>>>>>>> instead show "new folder" is temperature related. Decrease
>>>>>>>> the temp 5-10 degrees, down to about 70 or so and they
>>>>>>>> refresh automatically like they should.
>>
>>
>>>>>> OK, it may be that its just the temperature that is the problem.
>>>>>> What does DLG say about the worst drive when its cool ?
>>
>>
>>>> Stupid me, I was talking about the ambient room temperature in 70F,
>>>> compared to about 80F. When I crank up the air cond. the folder
>>>> refresh problem goes away and the drives SEEM faster. The drives
>>>> themeselves are running very cool with cases off. Lukewarm to the
>>>> touch.
>>
>>
>> There is clearly something that is temperature sensitive.
>
> Might be coincidental to my resetting the various jacks.
>>
>>
>>>>>>>> Maybe sys memory is starting to flake out?
>>
>>
>>>>>> Its possible. The best test for that is to see what DLG says
>>>>>> about the drive that it claims is dead in a completely different
>>>>>> system, without even using the same ribbon cable.
>>
>>
>>>> I was getting multibit errors and pushed
>>>> on all the connections and they went away.
>>
>>
>> OK, I'd definitely try another ribbon cable.
>
> Not getting them any more. No errors in event log now.
>>
>>
>>>> Thanks again for your input.
>>
>>
>> No problem, thats what these technical newsgroups are for.
>>
>>
>>>>>>>> I will run some diagnostic programs and report back.
>>
>>
>>>>>> Yes please.
>>
>>
>>>> I've got about 5-6 programs for that now,
>>>> so have to choose which to use.
>>
>>
>> I'd use DLG for the hard drives and MEMTEST86 for the ram.
>
> Will look for memtest86. I think usually when ram goes bad it is
> final-there
> is no intermittancy about it.
You can however see older system attempt to guess the ram specs
instead of reading the spd on the ram sticks and guess wrong.
> One other thing I did was put some more shielding around a microwave
> that might
> be leaking slightly,so that might have had an impact on the memory
> errors?
Shouldnt have any effect.