Reviews Leaderboard Database Reference Search StorageReview Discussion Reliability Survey Search About Contents

SCSI CD-ROM Drive Roundup

  June 2, 2000 Author: Tim Zakharov  
Special thanks to* for providing evaluation units.

* Remember, mention when ordering from HyperMicro and receive FREE shipping!

Part 2: SCSI CD-ROM Drives


In the previous article we examined five of the fastest ATAPI CD-ROM drives around, ultimately crowning Kenwood's 72X drive with our Editor's Choice Award. Today we look at some of the top SCSI CD-ROM drives available. With speeds ranging from 40X to 52X, and an average street price of $112.75, how will these drives compare to the ATAPI units, which ranged from 48X to 72X and averaged $68.40?

Before delving into our SCSI comparison, let's take a moment to examine some key differences between SCSI and ATAPI CD-ROM drives and how you might determine which is more appropriate for you. First and most obvious is the interface. Nearly all SCSI CD-ROM drives use a standard 50-pin connection. As of today, there is only one SCSI CD-ROM drive that uses a 68-pin Ultra Wide connection-Plextor's 40X Wide, which is included in this roundup. ATAPI drives, on the other hand, use the standard 40-pin IDE interface integrated into every modern-day motherboard. While an ATAPI drive seems the easiest, most logical choice, there are potential limitations. Most motherboards only have two IDE connectors. Each connector can have at most two devices attached. Only one device per connector can be active at a time. So there crops up two limitations-the amount of IDE/ATAPI devices you may use and possible performance degradation when maxing the interface out. The SCSI interface is much more flexible in both regards. A standard SCSI bus can handle up to 7 devices, while a wide SCSI bus can accept 15 devices. All this with only one IRQ taken, as opposed to one IRQ for each of the motherboard IDE connectors, plus an additional IRQ for each add-in PCI IDE controller card you choose to add. Additionally, SCSI devices can be accessed (in effect) simultaneously, losing little to no performance as devices are added to the chain.

The second key difference between SCSI and ATAPI CD-ROM drives is cost. As you may have noticed in the first paragraph, the average cost of the four tested SCSI drives is nearly twice that of the ATAPI drives tested, with lower average speed ratings. The reasons for this are somewhat similar to the reasons SCSI hard drives cost more than ATA hard drives. Mainly, SCSI sells at a lower volume than ATAPI. High volume usually equates to lower cost. SCSI requires the purchase of a host adapter or a more expensive motherboard with SCSI connections built in (additional cost above and beyond the extra cost of the drive itself). ATAPI drives, however, will work in any modern motherboard. Which is most likely to be purchased? ATAPI is the type that is most easily integrated into the vast majority of systems.

While there are other differences, these are the most important to the consumer. It can be concluded that ATAPI CD-ROM drives sacrifice a degree of flexibility and potentially some performance (in maxed-out configurations) for a much lower cost. Conversely, SCSI CD-ROM drives sacrifice cost for increased flexibility in configuration. Which is most important is entirely dependent on the individual and their needs. With that, let us proceed to our SCSI CD-ROM drive competitors!

The Drives

Kenwood UCR-415 - Firmware revision 100K:
From Kenwood Corporation comes the fastest-rated SCSI CD-ROM drive currently available (and Kenwood's only SCSI CD-ROM offering). Like the ATAPI 72X previously reviewed, the UCR-415 boasts True-X technology from Zen Research, where a specially constructed pickup splits the read laser into multiple beams to allow the reading of 7 tracks in parallel for nearly-CLV 52X performance. Like the 72X ATAPI model, the UCR-415 advertises lower RPMs and corresponding low noise levels compared to the competition. Kenwood specifies a 90ms access time, huge 2048kb buffer, and a standard 1-year warranty.

In terms of heat and noise, the 52X SCSI was indistinguishable from the 72X ATAPI. The drive got very warm during extended stress-testing, but sound levels were the quietest of the roundup. The same shaking during random accesses that we noticed on the ATAPI 72X Kenwood also was present on this unit. Only a 4-pin analog audio out is included on the drive, so SB Live! owners looking to exploit their digital CD audio inputs must look elsewhere.

ESP: $159

Plextor PX-40TSi - Firmware revision 1.11:
Plextor brings two drives to the table. First is the PX-40TSi, Plextor's mainstream CD-ROM offering. Rated at 40X Max, this drive (as well as its UW twin) comes with a 512kb buffer, with access times rated at 85ms and an industry-best 2-year warranty. Worth noting (you'll see why when we get to the benchmark results ) is Plextor's digital audio extraction specification of 24X Max, and 8X CD-RW read speeds.

The drive only became slightly warm to the touch after our stress test. Noise levels, on the other hand, were quite high. Not the highest we've heard, but definitely up there. However, this appears to be standard fare for most high speed CAV CD-ROM drives today. Truly quiet units seem to be the exception, rather than the norm. Our PX-40TSi came only with a 4-pin analog audio out.

ESP: $99

Plextor PX-40TSUWi - Firmware revision 1.02:
On paper, the only difference between the Ultra Wide and standard version of Plextor's 40X Max drive is the interface. Plextor advertises Ultra Wide burst levels (see the sidebar in the ATAPI roundup) and improved connectivity in a high-end SCSI environment with its 68-pin UW interface. Otherwise, the remaining specs are identical between this drive and the PX-40TSi, including the 2-year warranty.

The same can be said for heat and noise, as well as the CD audio out. Plextor's 40Wide is marketed for business/server environments, where single cable connectivity is desired. Although Plextor boasts of UW burst speed performance advantages, in reality this will likely never come into play. It certainly should not be the deciding factor when making a purchase.

ESP: $110

Toshiba XM-6401B - Firmware revision 1001:
Toshiba's entry in our SCSI CD-ROM roundup is a 40X Max SCSI unit, currently their fastest SCSI CD-ROM drive. Featuring a rated access time of 80ms and a 256kb buffer, the XM-6401B is the least expensive drive in our SCSI roundup. Toshiba backs their only current SCSI CD-ROM drive with a 1-year warranty.

During stress testing, the SCSI Toshiba was noticeably quieter than the Plextor units, with noise levels slightly below my personal "annoying" level. Heat was not an issue at all, as the drive was barely warm to the touch after extensive use. However, what I did find annoying was the non-standard 3-pin analog audio out connector. In fact, I had to go out and buy a universal cd audio cable in order to verify that it worked. If memory serves well, this type of connector was standard on most drives during the height of ISA sound cards. Perhaps Toshiba is counting on users of this drive (again marketed toward business use) to still be using older sound cards, which in a way makes sense.

ESP: $83

For an overview on methodology, click here.

Performance Results

Low-Level Measurements

 Optical Testbed I Low-Level MeasurementsDetails... 
Windows 98 SE using FAT 32
Kenwood UCR-415 (U-SCSI) - 7100|
Toshiba XM-6401B (U-SCSI) - 3070|
Plextor PX-40TSi (U-SCSI) - 2990|
Plextor PX-40TSUWi (UW-SCSI) - 2890|
Windows 98 SE using FAT 32
Kenwood UCR-415 (U-SCSI) - 7830|
Toshiba XM-6401B (U-SCSI) - 6300|
Plextor PX-40TSi (U-SCSI) - 6190|
Plextor PX-40TSUWi (UW-SCSI) - 6140|
Windows 98 SE using FAT 32
Toshiba XM-6401B (U-SCSI) - 80.9|
Kenwood UCR-415 (U-SCSI) - 91.2|
Plextor PX-40TSi (U-SCSI) - 94.2|

 Optical Testbed I CD Tach 98 - Burst Speed 16kDetails... 
Windows 98 SE using FAT 32
Plextor PX-40TSUWi (UW-SCSI) - 23618|
Plextor PX-40TSi (U-SCSI) - 14607|
Toshiba XM-6401B (U-SCSI) - 10724|
Kenwood UCR-415 (U-SCSI) - 7012|

When it comes to CD Winbench 99, it's all about access times. As we can see, the Toshiba holds the crown in our access time tests, coming in just a hair slower than its 80ms specification. This trend continues as the Kenwood, the next speediest, yields a 91.2ms measured access time. Plextor's PX-40TSi shows an even greater discrepency between expected and actual, a full 9.2ms slower than its rating. As you may recall from our previous article, 3 out of 5 ATAPI drives beat their access time spec, two by a significant margin. The UW Plextor, however, stumbled on our test CD, repeatedly erroring out and spinning down during testing. We could not get even one clean run, so we tried some other CDs in informal testing and did find one that the PX-40TSUWi could complete the test with (this indicates no mechanical problems). Its results were comparable to the PX-40TSi, but because of our strict testing procedures we could not include these results.

When it comes to transfer rates, Kenwood's 52X dominates, with a nearly flat transfer rate across our test CD. Speeds started at 47X at the inner tracks, and ended just a hair faster than 52X. Our next-closest competitor, Toshiba, was far behind at 20.5X-42X transfer rates. The Plextor brothers, both rated at the same 40X as the Toshiba, fell in a bit behind the XM-6401B. An interesting comparison throughout testing, the UW Plextor was a hair slower than the Ultra version here, 3% back at the inner tracks, but catching up and nearly equaling its brother by the outer edge of the disc.

So, do our access time rankings hold when examining CD-ROM Winmark scores? Read on to find out!

CD-ROM Winmark 99

 Optical Testbed I CD WinBench 99Details... 
Windows 98 SE using FAT 32
Toshiba XM-6401B (U-SCSI) - 1437|
Kenwood UCR-415 (U-SCSI) - 1408|
Plextor PX-40TSi (U-SCSI) - 1258|
Plextor PX-40TSUWi (UW-SCSI) - 1106|

Predictably, the drives finished in exactly the same order they did in the access time test. The Kenwood, despite a 12X advantage in transfer rates, finished 2% behind the Toshiba, which tested over 10ms faster than the UCR-415 in access times. Interestingly, the Toshiba could have finished much higher, but had real problems with one of our CDWB99 discs. We retested for verification purposes, but the read problems continued, creating an ugly 22% deviation (repeated in the retest) in scores-not pretty to a benchmarker, but nontheless reality.

Next up was Plextor's Ultra 40X Max unit, just over 12% back of the leader. Our UW Plextor again had read problems, appearing to favor some of our CD Winbench 99 discs, while taking a particular dislike to one of them. A retest with all four discs confirmed the PX-40TSUWi's pickiness with reading our test CDs. Because the Winmark test is application-level, we accept runs with read errors, as it is a good indicator of how each drive may react to a variety of discs in real-world use. The UW Plextor's scores varied by over 12% between the four test CDs (at least the deviation was consistent between retests!), in the end leaving it 12% behind its brother and 23% behind the Toshiba.

In our ATAPI roundup, we noticed a strong correlation between two low-level tests and our two real-world copy tests. The transfer rate test seemed a good indicator of how the drives performed in our sequential File Copy test; access time results seemed to play a strong role in our multiple-file Disc Copy test. Let's see if these trends hold for our SCSI units.

File and Disc Copy

 Optical Testbed I File Copy PerformanceDetails... 
Windows 98 SE using FAT 32
Kenwood UCR-415 (U-SCSI) - 02:07.6|
Plextor PX-40TSi (U-SCSI) - 02:33.6|
Toshiba XM-6401B (U-SCSI) - 02:44.0|
Plextor PX-40TSUWi (UW-SCSI) - 02:44.4|

If not for the Toshiba, all four drives would have placed in the same order as in the transfer rate tests. If you recall from our ATAPI roundup, the Toshiba offering there did the same thing. The SCSI Toshiba goes through the same read troubles at the outer edge of the CDTach disc that the ATAPI 48X unit did; speeding up and slowing down as if it needed to re-read the information...but only right at the end of the test. Again it is the Kenwood coming out on top here, finishing the file copy over 20% faster than the 2nd place PX-40TSi. The PX-40TSUWi and the Toshiba unit virtually tied, nearly 11 seconds slower than Plextor's U-SCSI 40X.

 Optical Testbed I Disc Copy PerformanceDetails... 
Windows 98 SE using FAT 32
Kenwood UCR-415 (U-SCSI) - 02:28.0|
Toshiba XM-6401B (U-SCSI) - 03:01.3|
Plextor PX-40TSi (U-SCSI) - 03:12.3|
Plextor PX-40TSUWi (UW-SCSI) - 03:19.8|

In the multiple file Disc Copy test, the outcome is again similar to what took place in our ATAPI roundup. The Kenwood, despite a disadvantage in access times, used its brute-force transfer rates to outrace the competition, finishing at over 22% faster than the next-fastest drive, the Toshiba 40X. From there, all is orderly as the remaining three drives (all rated the same speed), finish in the same order as their access times. Again, the UW Plextor is slower than the Ultra version, this time by 7.5 seconds, or almost 4%.

In taking a closer look at the two Plextors, we see that the UW is 7% slower in the sequential File Copy, while only 4% slower in the access time-weighted Disc Copy. This seems to indicate that the UW's problems are not strictly isolated to its difficulties in the access time tests. A bit of research on Plextor's website shows that the UW version has had significantly fewer firmware updates than their mainstream flagship CD-ROM drive, the PX-40TSi. Perhaps Plextor is placing more man-hours of development with their meat and potatoes unit. Let's see what happens as we delve into other, less traditional areas of CD-ROM performance.

Digital Audio Extraction

 Optical Testbed I CD Speed 99 - DAE Transfer RateDetails... 
DAE Transfer Rate - Average
Windows 98 SE using FAT 32
Plextor PX-40TSi (U-SCSI) - 30.29|
Plextor PX-40TSUWi (UW-SCSI) - 23.54|
Kenwood UCR-415 (U-SCSI) - 11.23|
Toshiba XM-6401B (U-SCSI) - 11.12|
DAE Transfer Rate - Start
Windows 98 SE using FAT 32
Plextor PX-40TSi (U-SCSI) - 18.53|
Plextor PX-40TSUWi (UW-SCSI) - 17.90|
Kenwood UCR-415 (U-SCSI) - 10.83|
Toshiba XM-6401B (U-SCSI) - 6.80|
DAE Transfer Rate - End
Windows 98 SE using FAT 32
Plextor PX-40TSi (U-SCSI) - 39.59|
Plextor PX-40TSUWi (UW-SCSI) - 24.24|
Toshiba XM-6401B (U-SCSI) - 14.54|
Kenwood UCR-415 (U-SCSI) - 11.36|

We see here a continuation of the dichotomy between the two Plextor drives. While both drives are rated as 24X Max audio extractors, the mainstream 40Max pulls emphatically ahead of the more niche-marketed 40Wide. We see nearly identical performance in DAE and CD-ROM read speeds, as the PX-40TSi starts at 18.5X and continues all the way to almost 40X extraction speeds at the outer edge of our test CD-DA. Average DAE is an impressive 30.3X, nearly 29% faster than its UW brother. The 40Wide does what Plextor says it should do, extracting digital audio at a minimum of almost 18X, maxing out slightly ahead of its rated 24X. We can see by the transfer rate graph that the 40Wide is a P-CAV digital audio extractor. Coming in a distant 3rd, Kenwood's 52X drive extracts audio at a merely acceptable CLV of 11X. Toshiba's entry mimics its ATAPI 48X sibling's performance, with inadequate 6.8X extraction speeds at the inner tracks. By the outer tracks, it's holding its own at 14.5X.

As for the quality of audio extraction, all of the SCSI units are acceptable. CDSpeed99 rates each drive at a perfect 10/10. Using our Grado SR60 headphones to listen to selected ripped tracks from our test audio CD, the .wav files played back sounding no different than directly off of the CD. Each of these drives should provide outstanding quality in audio extraction.

CD-R Media Compatibility and Performance

 Optical Testbed I CD-R Media PerformanceDetails... 
Windows 98 SE using FAT 32
Kenwood UCR-415 (U-SCSI) - 4587|
Toshiba XM-6401B (U-SCSI) - 3140|
Plextor PX-40TSi (U-SCSI) - 3060|
Plextor PX-40TSUWi (UW-SCSI) - 2950|
Windows 98 SE using FAT 32
Toshiba XM-6401B (U-SCSI) - 6350|
Plextor PX-40TSi (U-SCSI) - 6240|
Plextor PX-40TSUWi (UW-SCSI) - 6190|
Kenwood UCR-415 (U-SCSI) - 1657|

Here we see a first: The Kenwood, which places 1st at the inner tracks of our CD-R test, does a complete turnaround and is in last place by the outer edge of our test disc. When I first saw these results, I hadn't seen the transfer rate graph and believed that there was a read error at the outer edge, accounting for the odd figures. However, upon studying the transfer rate graph, I saw that this was indeed the shape-somewhat upside down from what one would expect! As the Kenwood reads from inner to outer tracks, sequential transfer rates steadily decline at an inverted P-CAV pace. While the drive was able to read our CD-R, overall it was the slowest of the four drives. The other drives followed the trend set by our ATAPI units: equal to or slightly better than standard CD-ROM transfer rates. Let's see what happens when we put our Verbatim CD-RW media into these drives...

CD-RW Media Compatibility and Performance

 Optical Testbed I CD-RW Media PerformanceDetails... 
Windows 98 SE using FAT 32
Plextor PX-40TSi (U-SCSI) - 2470|
Plextor PX-40TSUWi (UW-SCSI) - 1290|
Toshiba XM-6401B (U-SCSI) - 1150|
Windows 98 SE using FAT 32
Plextor PX-40TSi (U-SCSI) - 5010|
Toshiba XM-6401B (U-SCSI) - 2330|
Plextor PX-40TSUWi (UW-SCSI) - 1290|

Again, the two Plextors split paths, this time in CD-RW read speeds. The flagship 40Max takes the cake with read speeds comparable to Creative's 52X ATAPI unit, reviewed previously. At the inner tracks it's a 16X reader, finishing our test disc at over 33X! The Wide version, on the other hand "merely" reads along at a slightly faster clip than Plextor advertises-8.6X CLV. Rather than looking at it in the light that the UW is inferior, though, one must remember that it is performing no worse (actually, a bit better!) than Plextor advertises. The PX-40TSi has just had more time and development behind it, it seems. In their firmware section, Plextor has separate firmware updates for 1.1x (more recent) and 1.0x (original) versions of the 40Max. They explicitly state that the two are not compatible with one another. A search through some older reviews of the 40Max seems to indicate that the original 1.0x version performed similarly to our 40Wide. As for the other drives, the Toshiba again imitates its 48X ATAPI brother, with CAV read speeds approaching 8X-16X. Sadly, the Kenwood also emulates its ATAPI brother, failing to read our Verbatim CD-RW at all. At least the 72X had firmware updates attempting to address this; the 52X SCSI version did not have any firmware updates available! Hopefully Kenwood will address this issue soon. While it may not be an issue for many people, it has the potential of being a serious problem for those who rely on CD-RW media.

Editor's Choice: Plextor PX-40TSi's Editors ChoiceAlthough Kenwood's 52X unit was easily the fastest in traditional CD-ROM performance, its prohibitive sticker price, average DAE and poor media compatibility combined to pull it out of contention. Plextor's venerable 40X Max, on the other hand, supplemented adequate traditional performance with truly outstanding DAE, media compatibility, and CD-RW read speeds. Factor in the $99 street price and 2-year warranty, and it really is no contest. Remember, the Plextor is advertised as a 24X Max audio extractor and 8X CD-RW reader. The PX-40TSi performs so far beyond expectations in these two areas that it makes the price tag that much more attractive! Thus we feel no qualms about giving it's Editor's Choice Award for SCSI CD-ROM drives.

See also: ATAPI CD-ROM Roundup


Copyright © 1998-2005, Inc. All rights reserved.
Write: Webmaster