(This article is from 2001, when 100GB was a big drive and the G4 Cube was new.)
Maxtor was the first company to announce a 100GB IDE/ATA hard drive in spring 2001. The DiamondMax 536DX 100GB drive features an ATA/100 interface (backward compatible with most older IDE controllers), 2MB of cache and 5400 RPM spindle speed. (Western Digital recently announced a 100GB, 7200 RPM drive for the same price, but it was not available at retail as I'm writing this.) Although originally thought to have a 40GB/platter density, the 100GB Maxtor actually is 33.3GB per platter. Current list price of the 100GB Maxtor drive is $299.95. OEM versions (drive only/no retail package) may be less. This drive was purchased for a G4 Cube, which can have only one internal drive - so capacity and low-noise were major factors in choosing this model. (BTW - Who would have thought a year or so ago you could have an 8-inch cube sized computer with 100GB of storage and 1GB of RAM, with prices of drives and RAM as affordable as they are now.)
Cube Install Notes: Apple has a page with PDF and Quicktime Movie guides on Customer Installable Parts for the Cube. (link from 2001 MIA now. Here's a page with some G4 Cube manuals and parts replacement PDFs and G4 Cube related articles here also.) Their PDF file and QT movie on the hard drive does not show removing/replacing the black plastic guide channel (bar) on the lower edge of the hard drive. It's secured by two Torx T9 screws. Also I noted my Cube had T9 sized screws on the heatsink also, although the Apple PDF file noted T8. (If there's enough Cube owner interest, I will post my photos taken of the Cube drive upgrade.)
What's Included in the Maxtor Retail Box:
- 100GB DiamondMax 536DX drive
- Installation guide (geared at PCs) with jumper settings (also on drive label)
- Maxtor Installation floppy disk (for PCs - not needed for Macs)
- UltraATA dual drive cable (not used with the Cube)
- Metal 3.5"->5.25" mounting bracket
(may be useful for some Mac clones - not needed for Apple Macs)
- $50 Rebate (expires 9/30/2001) on Maxtor ATA/100 IDE card (won't work in Macs)
DiamondMax 536DX 100 GB Drive Specs: (per Maxtor)
- 5400 RPM spindle speed
- 2MB cache
- 33.3 GB per platter (6 Heads/3 platters)
- <11ms Avg. Access time
- Avg. latency 5.55ms
- Seek Times: 1.0ms Track to Track/10.5ms Avg/18.0 ms full stroke
- Start time (0-drive ready) <10sec typical
- Up to 29 MB/sec sustained transfers to/from platter (outer diamter),
Up to 100 MB/sec from interface (with ATA/100 controller)
- Component design life: 5 years minimum (3 year warranty noted on retail drive box)
- Start/Stop cycles: >50,000
- Annual return rate: <0.5%
(based on similar drive design stats apparently since this drive is new)
- Drive height: 26.1mm
- Shock (2ms duration): 30 Gs operating (no errors), 300 Gs non-operating (no damage)
- Noise (bel): 3.0 avg/3.2 max (idle), 3.4 avg/3.6 max (seeking)
For a complete list of drive specifications, see Maxtor's 536DX Specs PDF file. (2001 link removed as it's 404 now.)
Important Note on Maxtor ATA/100 Drives and Write Performance:
As noted in our FAQ's IDE Hard Drive section about a year ago, new Maxtor ATA/100 drives usually have "write verification" enabled from the factory until the drive has been power cycled 10 times. (Power Cycle means shutting down/powering the drive off.) Therefore the drive will have much lower write performance than normal until this feature is disabled. To illustrate this effect, I ran ATTO Tool's test before and after the write verify feature had been disabled. The graph below shows the results. (Tests were run on the 2nd partition in both cases, with about 6GB of files installed on that partition.)
You can see the dramatic difference in the Sustained Write speeds after write verify was disabled. (Peak write rates are usually writes only to the HD's cache, so sustained writes are shown.)
Test System Hardware Summary
- Apple G4/450 Cube
- 384MB SDRAM (CL2)
- Original OEM Western Digital 20GB drive (5400 RPM/2MB cache)
- DVD ROM drive (OEM)
- OEM ATI Radeon AGP graphics card (original model w/fan)
- OS 9.1/OS X 10.0.4, QT 5.02 Pro, standard extensions set (Multiple Users not enabled)
- Memory Control Panel disk cache was set to automatic (8160KB cache size with the amount of installed RAM in the system)
OS X and OS 9.1 were installed on the same volume with the Western Digital original 20GB drive. I used Apple drive setup's Custom Initialization options on the 100GB Maxtor drive to create 3 partions - an appx 1.1GB dedicated Swap file partiton for OS X (first partion), a 2nd partition for OS X (appx. 46.8GB) and a 3rd partion for OS 9.1 (appx 46.5GB). The first partition/outer tracks of a drive are the fastest, so the results of tests with OS 9 basically used the 2nd half of the disk. (Rates would have been a bit faster had the OS 9.1 partition been the first on the drive. I've usually included all 3 partitions in benchmarks below to show this effect. Since the 1st partition is so small [1.1GB] relative to the total size of the disk, there's not a lot of difference in the partition 1 vs partition 2 results.)
Formatted capacity of the Maxtor 100GB is appx. 94.4 GB. (This is primary due to the decimal (1000=1K) vs Binary (1024=1K) numbering system. (This is true of any drive - SCSI, IDE, etc.)
Disk Driver Used:
Apple's Drive Setup drivers from the OS 9.1 CD was used on both drives used in this article. (As of this date no other ATA/IDE or SCSI disk driver I'm aware of is compatible with OS X.)
Real World Test Results:
Boot and Load Tests (OS 9.1/OS X 10.0.4) The table below shows results of
several tests like Booting (timed from the end of the boot chime until system ready to use), and how long applications or files took to load. Time to load and save a 18MB Photoshop file is also listed. Scores in bold are better (the 100GB drive was faster in all these tests as expected).
The next tests cover duplicating a single large file and a folder with thousands of files.
Duplicate 102.3MB File:
Note: The 100GB Maxtor drive was not optimized for the above tests, the 20GB Western Digital OEM drive was. The OS 9.1 volume was reported as less than 1% fragmented, but the OS X partition (as usual after an install/update) was listed as having "severe" fragmentation. Repeating these tests after optimizing the OS X volume however had minimal effect on performance. (less than a few tenths of a second difference, easily within the human error of stopwatch timing.)
Duplicate Folder (Thousands of files):
Benchmark Test Results:
MacBench 5.0 Results:
The image below shows the MacBench 5.0 Disk Tests Suite results from the B&W G3. All drives were empty (no files present) and formatted just before the test. As with all tests, a clean reboot was performed before executing the benchmark.
Althought the Disk/Pub Disk mix tests showed only 10-20% improvement with the Maxtor 100GB drive over the OEM 20GB, the Random R/W tests showed a more significant advantage (in some cases dramatic).
ATTO Tools 2.5 Benchmark:
Rates in MegaBytes per Second (longer bars are faster).
As you can see from the sustained rates, this drive is not as fast as say an IBM 75GXP or 60GXP (7200 rpm) model. However for this application (G4 Cube), I wanted the largest drive possible and thought the 5400 rpm drive might be more quiet and run cooler than a 7200 rpm model. I'm very satisfied so far with the Maxtor 100GB.
One thing not shown in a bar graph of just the sustained/peak numbers is the graph across I/O sizes. The hard drive's cache is responsible for the higher rates (taller curves) at the smaller I/O sizes, but as size increases the graph levels off to show sustained (to the platter) rates. (The graphs were all very similar for the 3 partions, I'm showing only partition 2 to reduce the number of images on this page.)
The Maxtor 100GB DiamondMax drive seems very quiet. Since the drive was purchased for a G4 Cube this was an important factor. (Although some 7200 RPM drives are also quiet.) You can hear when the drive is seeking (head movement), but it's a muffled sound in the Cube and not objectionable to me. When duplicating files on the same volume, the drive was literally silent. Copying files from one partition (volume) to another, or when booting you can hear the head movement, which is normal.
The Cube's top vent temperature didn't seem to be any hotter with this drive than the original drive. (Subjective test however.) I installed an OEM Radeon graphics card (which has a fan - I hear the latest versions don't), so the graphics card fan, although low volume, is the only noise heard from the Cube at idle. Although it's not totally silent, the Cube is so much quieter (even with the Radeon's fan) than any desktop Mac I've ever used except for the iMac DV. (However that iMac DV SE developed some drive noise at idle after several months that is about as audible as the Radeon's fan.)
The primary reasons I bought this drive was for its huge storage capacity and the 5400 rpm spindle speed that I hoped would make the drive run quiet and cooler (important factors for a Cube). For $299.xx list, I've increased storage space by nearly 80GB over the original drive and gained better performance to boot. Since the Cube can't have more than one internal drive, I'm a happy camper with the 100GB Maxtor. (But if the WD 100GB/7200 RPM drive had been available, I may have been tempted to go for that model...)
The 1.1GB OS X "Swap" file partition is really larger than needed, which I found out after the fact (see the article here on OS X Swap File Sizes/Info). I used a reader's custom swapfile startup item for OS X noted in the article on OS X Dedicated Swap File Partitons.
FYI: Remember you can reuse your original IDE drive either in another Mac (if a Cube or iMac where only one internal HD is possible) or as a slave drive (if your Mac supports them) or in a Firewire external case kit. See the Firewire articles page for case kit build guides that show how to install an IDE drive in them. For info on 2nd (Slave) drives in other Macs, see the IDE/ATA articles page.
Other Apple G4 Cube related articles/mods/upgrades are on the Mac Mods/Upgrades and CPU upgrades pages.