Wasn't that an old Depeche Mode song or something? Anyway... I need help with IDE devices.

In addition to upgrading the car player, I also got a bigger hard disk for my PC. Before the upgrade, my configuration was as follows:

IDE channel 1: 20gb hard disk master, 10gb hard disk slave.
IDE channel 2: 4gb hard disk master, DVD/CDRW slave.

I remove the 4gb and put the new 60gb in its place. It works, but any drive operations make the whole PC behave sluggishly (mouse movement hitches and skips), and it's reeeeeely slow for all operations. Disconnecting the DVD/CDRW and making the 60 gig drive all by itself on the secondary controller makes everything nice and fast.

For all I know, it could have been this way all along with that 4gb drive, too, I just didn't notice because I didn't put anything speed-critical on that drive.

Now, I always thought Conventional Wisdom was something like this (quoted from a web page about IDE troubleshooting):

Drive running too slow? Remember that with IDE the speed of the master governs the speed of the slave. If you buy a new fast hard disk and slave it to your old slow disk, you will have 2 slow disks. For this reason CD and DVD devices should always be slaved wherever possible, and hard disks should be masters.

which is exactly what I did. But it was still slow. Removing the CD slave sped it up.

I agree that one option would be to copy all the stuff from the 10gb slave and just get rid of it. But I want all the space I can get at this point.

So does anyone have any idea how I can get all three disk drives AND the DVD drive working properly, with only two IDE channels and without spending any more money? Is there some magic trick I can do to make the DVD play nice with a disk drive on the same cable?

Is it possible that the DVD is disabling DMA on the whole channel? Isn't there a way to set that stuff in Windows? At the very least you should be able to check the DMA setting both with and without the DVD attached and see if that's the issue, whether or not there's a way to solve it.

Oh yeah, forgot to say:

First thing I did was look to see if the little "DMA" checkbox was checked for the disk drive and the DVD drive in the Windows Device Manager. Check box was checked for both of them, and is still checked with the DVD disconnected.

Not that I trust Windows to tell me the truth about it... So I'll concede that it's possible that's the very problem. But how would I know for sure?

Wasn't that an old Depeche Mode song or something?

Yep, 84 or 85 I believe. Unfortunately I have no answer for the hard drive issue.

Hi Tony,

This may be caused by improper termination on the IDE cable. I think most IDE devices are "sort of" self terminating, but there still may be some ringing going on. You might try and switch the device order, (it doesn't matter if the Overlord or Subject are in order on the cable) and check to see if the devices have any termination jumpers that need setting.

L

You might try and switch the device order, (it doesn't matter if the Overlord or Subject are in order on the cable)

So do you mean, switch it so that the DVD is the master? Or do you mean physically switch the positions of the drives so that disk drive falls at the end of the cable chain (currently it's in the middle)?

Heh, I guess I could try both.

I didn't realize IDE had "termination". I would have expected this with a SCSI drive, but not an IDE drive.

Your new harddisk probably needs an 80 pin IDE cable to function properly. Are you using one of those?

Also with an 80pin cable, it's necessary to put the master device on the end of the cable and the slave device in the middle. This wasn't so the case with 40 pin cable.

If that doesn't work you can try switching devices, but that probably won't work.

Overall, I've found that it's generally a bad idea (now especially with the 80 pin cables) to mix harddisks and non-harddisk IDE devices (cd-rom etc.) on the same IDE channel (cable).

Your best bet is probably to buy one of those ATA/133 PCI cards that adds an entirely new IDE channel to your PC. You know, like the ones Promise makes. They aren't too expensive and function well.

Speaking of that, one of your old drives is probably of a slower speed (not UltraATA-133 or whatever). You could try moving devices around so that the CDROM is on the same chain with one of those slower drives. Of course, this causes problems under Windows, since you can't really change drive letters, and I don't know how to deal with that problem (maybe keep the drives in the same relative order?), so you'd have to figure that one out.

Of course, this causes problems under Windows, since you can't really change drive letters...

In win2k you can (although I believe the boot drive still has to be c:), and if I'm not mistaken in winXP too. But I believe Tony's using win2k anyway, so that shouldn't be a problem.

That might indeed be a solution, but like you say with all the trouble that *may* arise because the drives are not exactly where windows expects them to, it might be more trouble than it's worth.

Your new harddisk probably needs an 80 pin IDE cable to function properly. Are you using one of those?

Hm. The existing wires and plugs "just fit". However, there was a cable included with the drive, perhaps I'll try hooking that one up.

Also with an 80pin cable, it's necessary to put the master device on the end of the cable and the slave device in the middle.

I'll try that, too, thanks.

The 80 wire ribbon cables are terminated with the 40 pin connectors just like the 40 wire ide cables. The 80 wire cables have a ground wire between each of the original wires to provide shielding. This allows the drive to operate at the higher transmission rates. At least thats how I understand it.

Yip. And as to the Master selecting the speed of the chain, BS. They both have to run at the same speed regardless of which comes first, just like RAM. IDE devices do NOT have terminators. All devices are self terminating. 80 pin cables DO have 40 pin ends, and still make a difference.

I meant to physically switch them such that the disk drive is at the end. I wouldn't try and make the DVD the master.

Each IDE device has it's own termination (sort of) for the same reasons that SCSI busses have it. The SCSI system of end termination is more reliable, but requires a bit more thinking to set up.

Oh, yeah, if it is ringing causing your problems, you can slow down the ide interface clock and it should work.

AND THE WINNER IS....

Archeon, with:

Your new harddisk probably needs an 80 pin IDE cable to function properly.

All I needed to do was to install the 80-conductor cable that came with the drive and all is well. However....

Also with an 80pin cable, it's necessary to put the master device on the end of the cable and the slave device in the middle. This wasn't so the case with 40 pin cable.

This didn't seem to be necessary. My IDE disk master is in the middle and the DVD is on the end. Seems fine to me now.

Oh, forgot to say, in that last message:

THANK YOU, ARCHEON!

And also, thank you everyone else for your suggestions!

This BBS is so cool.

Also with an 80pin cable, it's necessary to put the master device on the end of the cable and the slave device in the middle. This wasn't so the case with 40 pin cable.

Just getting the motherboard end in the right place is good enough, usually. Indeed, "backwards" devices also have always worked for me.

If you put the "motherboard" end on the furthest device, you end up losing the benefits of the 80 conductor cable.

If you put the "motherboard" end on the furthest device, you end up losing the benefits of the 80 conductor cable.

Ah, well I at least got THAT part right.

I'm considering getting another 80c cable to put on the drives on the IDE0 channel to see if that makes *them* work faster. That'd be cool...

I guess I bought too many big Maxtor drives. I don't seem to have any spares these days that aren't 80 conductor cables.

In the vein of "things you might not guess" my Pioneer DVR-A03 burner has this problem where when I try to enable (any) DMA to it, I get underruns, but if I do PIO it works fine.

Who would have guessed...

In fact, I've heard you need to do that with some burners... Don't know exactly why...

It's documented for the A03 that you "may" need to do it, but they explain what you might see in Windows. This was of course about as useful to me as Windows itself.

The official word for ATA (IDE) drives is.. the MASTER must be on the end of the cable. The SLAVE, if present, goes in the middle of the cable, unless it is the only drive (no MASTER).

40/80 wire cables don't change that. However, 80W is required to utilize transfer speeds faster than 33MB/sec (the cable type is detected in s/w, and fast speeds are disabled if an 80W cable is not found), and 80W generally improves reliability in marginal setups.

-ml

I was curious about this stuff, all the master and slave talk, and wondering why the 80 conductor cable was an issue, and after some digging found these links:

http://www.pcguide.com/ref/hdd/if/ide/confCable-c.html
http://www.pcguide.com/ref/hdd/if/ide/conf_Cable80.htm

My PC's BIOS is smart enough to detect and display (at bootup) if an 80-wire cable is present or not. It must have been deliberately throttling the drive because of the lack of the 80-wire cable, hence the issue.

I can't find anything in there saying that the drives in a master/slave configuration "must" be physically positioned a certain way. What I do see is that if you're using cable select then the master/slave configuration will come out that way but I can't find anything stating that two drives on the same cable must be that way.

One thing they do say is that if you're running one drive all by itself, then you shouldn't put it on the middle connector. Leaving the end connector dangling makes it pick up EMI interference like an antenna. Hmm... our beloved Mk2 is that way by default...

All I needed to do was to install the 80-conductor cable that came with the drive and all is well

Wow, didn't even think of that, I haven't seen a 40 conductor cable in ages. I remember working desktop support when ATA-66 started shipping, and having to ask that question every call that began with "I added a new drive onto the Promise controller, and it's not at 66 speed."

try putting in one of these. I've heard they make the drives go faster.

THANK YOU, ARCHEON!

No problem Tony! You've helped me out on more than one occasion, so I'm happy I could finally return the favour.

Also with an 80pin cable, it's necessary to put the master device on the end of the cable and the slave device in the middle. This wasn't so the case with 40 pin cable.
This didn't seem to be necessary. My IDE disk master is in the middle and the DVD is on the end. Seems fine to me now.

Oh, but you've good a very good chance that it will work, and will continue to work.
It's just that the official way of doing it it -as Mark also states- is master drive on the end, slave in the middle. God knows I've done it backwards myself on many occasions and most of the times it worked just fine too.
I just said what I said because I wanted to be absolutely sure that that little thing had nothing to do with your problem (because it could have), but now all is well, so : hooray!

>I can't find anything in there saying that the drives in a master/slave >configuration "must" be physically positioned a certain way.

Here: I'll attach two pages from the ATA6 standards.

For the now standard 80W cable:


But for the 40W cable, oddly enough they still suggest connecting drives the other way around, probably cuz the interface speed is slow enough to tolerate it and it is more convenient for in-case connection of a CD/DVD drive (slave) and an internal hard disk on one cable.



-ml

holder for second page attachment

It's just that the official way of doing it it -as Mark also states- is master drive on the end, slave in the middle.

I looked for evidence that this was really the case, and specifically why this would be the case, and haven't found any. Does anyone have any evidence to back this up?

See, I don't think that "official way" really exists. Not in the way it's been stated in this thread. I think that what people are seeing is a second- or third-hand retelling of the concept of cable select, and this misinterpreted version is getting retold and spread like an urban myth. I think that if the two drives are deliberately jumpered for master/slave, then it doesn't matter what their order is, and it never did matter.

I'd be happy to change my tune on this, if someone can show evidence above and beyond "I heard it was supposed to be done that way." Of all the people on the BBS who I'd expect to have this evidence on hand, it'd be Mark (remember, he's the Linux IDE guy ).

Here I was, composing my post at the same time mark was posting evidence.

Only... apart from labeling those connectors as "device 0" and "device 1", I don't see anything stating that the drives have to be in that order or why. That could still just be a retelling of the cable select order. Am I missing something?