Help me settle an argument...

[LivinLaVivaPollo]

Ok, my friend and I are arguing over how to setup some IDE channels. I always thought an IDE channel is as fast as the slowest device on the channel. Let's say I have an ATA100 HD as the master and a cdrom as a slave. The cdrom would be AT MOST UltraDMA33 right? So that would effectively reduce the bandwidth of the channel to 33 correct? He is saying that the slave is solely dependent on the master, so if the master was an ATA100 HD, the cdrom could be the slave but the HD would still run at ATA100. What I am saying is that if I have a cdrom and an HD on the SAME channel, the HD being the master, and the cdrom being the slave the fastest that the channel would run is 33MBps burst. Help me settle this!

 

[Rent]

Depends on cable and what is using the cable at the time. If the CDRom is idle, then the HD should be able to get its 100mb bursts provided that you have the right cable. When the CDRom is using the channel, it would froce the HD to go along with its ATA33.

Please correct me if I'm wrong.

 

[yoda291]

By no measure am I an expert at this, but I was always told that the master HD will set the speed and run its ata100 burst fine. The CDROM will likely be working off and competing with the HD for priority to the controller. I think it has something more to do with placement on the cable than anything else, but even this shouldn't really be all that big a deal.

 

[veryape]

Yea,you're right,the cd will slow you down if its on the same ide channel.

 

[LivinLaVivaPollo]

Yeah, I'm still not quite sure. I need a definite answer.

 

[bot2600]

If you have the proper cable, the first connector will run @ATA100 the second will run @ATA33 assuming devices can live up to those speeds. The CD ROM on the end of the cable will NOT slow down the harddrive. That USED to be the way it worked, but on anything new enough to have ATA100 on the controller and the hard drive side, it is NOT an issue. There is you answer.

Bot

 

[NDGeist]

the cdrom will bring the entire bus down to ata33

Chip

 

[LivinLaVivaPollo]

Anyone have some rock solid empirical data?