IDE Interface Features
The main requirement of the recording device to the interface is the ability to continuously transmit data to it during the recording process. Requirements for the speed of receipt of data by modern standards are very modest. For recording at 4th speed, a continuous data stream with a speed of only 600 kb / s is required. But this is deceitful modesty. After all, the flow must be continuous! And here there are nuances.
The IDE interface was originally designed so that at any time you can access only one device. Moreover, having started the exchange operation with one device, the interface cannot contact, at least temporarily, to transfer any command to another device. Therefore, the program providing the recording process must read a certain portion of data from the data source device, end the exchange with this device, and only then transfer the received data to the recording device. Fundamentally, this is not a problem for high-quality recording. Initial data can first be prepared accordingly (create an image – Image). Then, for the data source, the operating system opens the file once, at the right time, the CD writer reads the required number of bytes and transfers them to the writer. No problems! The speed and standard features of the interface are enough for this. But it turns out that nobody is interested in potential abilities! The owner of the recorder does not want to waste time creating an image. He wants to use all the capabilities of modern recording devices: select the files necessary for recording in the window interface directly in the source disk directory and give a command to write selected files (necessarily at 4th speed!). And here everything will depend on the capabilities of your machine.
Let’s not delve into the technical details. Let’s just see how it turns out in practice. Without any hesitation, we assembled the machine in the configuration most often encountered by users: a 430TX board, a Pentium 100 processor (by the way, practically nothing depends on the processor in this case), a main disk (PIO Mode 4 mode is supported) and an IDE recorder ( supports PIO Mode 3) connected to the second channel of the IDE.
During the experiments, the second device connected to one or another IDE channel: an old ST3850A disk (PIO Mode 4), a ST34311A disk (UDMA-33), or a SAMSUNG CD-ROM (PIO Mode 4). Of course, the result could be foreseen in advance, but we were not too lazy to systematically check various combinations of connecting devices. Moreover, through trial and error we selected such a mixture of files (a kind of “Molotov cocktail”) for the IDE recording device, which not every combination of devices could “overpower”. At 4th speed, this “cocktail” could only be recorded from a hard drive operating in UDMA mode, due to the fact that one of the extensions of the standard IDE interface for this mode is the ability to work simultaneously with two devices. If only temporary files of the burning program are placed on a disk working in UDMA mode, and the source files are placed on a disk with PIO Mode 4 (connected to another channel), then in this case it was possible to record this “cocktail”, at least in Easy CD Creator. True, this could not be done in the WinOnCD program (these programs use temporary files in different ways). From a disk that does not have UDMA mode connected to the same channel as the recorder, most often it was not possible to record not only this test mixture, but also any other. The explanation is simple: the data does not have time to arrive at the required pace. To illustrate this fact, we launched two tests simultaneously: a CD-ROM performance test and a hard disk performance test (simulating continuous reading from a disk).