Fujitsu P2120 Hard Disk Upgrade

Introduction

In the quest to maximize the performance of my Fujitsu P2120 sub-notebook, I decided to take a look at upgrading the hard disk to a speedier model. In general, systems based on the Transmeta Crusoe chips can be a little sluggish when starting up. Part of this is due to the Transmeta Code Morphing doing its thing and part of it is a sluggish hard disk. During my experience with my P2120 I’ve noticed that the hard disk gets a lot of usage even with plenty of memory. My P2120 came with a Toshiba MK4018GAP 40GB hard disk running at 4200 RPM (2MB cache). While being fairly quiet and decent in performance, I figured a 5400 RPM model would help make a difference and so I set out in search of a better hard disk. Additionally, I also wanted to know if the drives were quiet and how they affected battery life. So, here’s my story and what I found out.

The Candidates

The P2120 will only take 9.5mm 2.5” laptop hard disks so that limits the number of choices one has when upgrading. After much searching, I found several 5400RPM models available but ultimately settled on the two most feature-filled drives with very interesting specifications.

IBM/Hitachi Travelstar 40GNX Series

• 5400 RPM
• 8MB cache
• Fluid Bearing Drive mechanism
• ATA-5

Toshiba MK4019GAX Series

• 5400 RPM
• 16MB cache
• Fluid Bearing Drive mechanism
• ATA-5

On the surface, both drives are impressive. Toshiba’s 16MB cache is certainly eye-popping. I found an article at Barefeats (http://barefeats.com/fire32.html) that tested the hard disks in Firewire enclosures. It’s very informative and I suggest people take a look at it. However, running the hard disks in external enclosures and testing on the Mac platform didn’t help me very much in telling me how it would perform in the P2120. So, I decided to get my hands on both and test them inside of my P2120. If you’re interested in learning how to change the hard disk, you should check out the tutorial at Leog.net.

Testing the Drives

Unlike benchmarking desktop hard drives, I realized that testing laptop drives requires more than testing the hard disks in some benchmarks. Laptop users are a little more discriminating than just performance. Battery life, heat, and noise are other factors that play a huge importance in buying decision. So, I’ll be looking at the overall performance of the drives, battery performance, and heat issues and noise so you know what you’re getting into.

Benchmarks

I first took a look at the overall performance of the drives. The stock drive (MK4018GAP) is included as the baseline. The drives were tested on some well known industry used benchmark applications.

• Veritest Business Winstone 2003 v1.01
• Veritest Winbench 99 v2.0
• SiSoft Sandra 2003 File System Benchmark
• PCMark 2002 Disk Test
• HDTach 2.61

The system specs for the P2120 were as follows

• Transmeta Crusoe TM5800-933 MHz
• 512MB of SDRAM
• 10.6” 1280x768 LCD Screen running at 16-bit color
• ATI Radeon Mobility 8MB DDR SDRAM
• Windows XP Home SP1

The hard disks tested were the following:

• IBM TravelStar 40GNX 40GB - 5400 RPM, 8MB cache
• Toshiba MK4019GAX 40GB - 5400 RPM, 16MB cache
• Toshiba MK4018GAP 40GB - 4200 RPM, 2MB cache (original drive)

The system had all unnecessary services disabled and the drives were cloned identically. All tests were run 3 times with the scores being an average of the 3. All drives were defragmented before the tests were run.

Although I doubt the drives themselves can approach the theoretical max of the ATA-5 specification, I should also note that the ALI M5229 PCI IDE chipset found in the P2120 only supports up to ATA-4 and can’t take full advantage of the drives theoretical transfer rates.

Running Veritest’s Business Winstone 2003 produced the following result:

As you can see, the 5400 RPM drives pull ahead of the original drive although not by much. The IBM drive comes in at 0.1 points higher than the Toshiba. Toshiba’s 16MB cache doesn’t help it much in these tests. The one thing to note during the tests is that the 5400 RPM drives complete the tests noticeably faster than the 4200 RPM drive.

WinBench 99 version 2.0 is a very popular benchmark for testing the low level performance of hard disks. While a synthetic test, it still provides some useful information to look at.

As you can see, the 5400 RPM drives show a lot more performance potential than the slower 4200 RPM Toshiba MK4018GAP. In the High-End tests, the Toshiba MK4019GAX flexes its muscles and edges the IBM drive by a few points. Both far outperform the 4200 RPM model due to the higher rotational speed and the larger cache sizes. Although not noted on the graph, the CPU utilization during the tests was nearly identical for all drives averaging around 20% utilization during the Disk Transfer Rate tests. For the High-End Diskmark tests utilization for all hovered around 7%.

SiSoftware’s Sandra 2003 shows the Toshiba’s drives performing quite well. The IBM drive’s performance was lower than I would have expected since it was stronger in other benchmarks. Still, this is just a synthetic result and doesn’t necessarily translate into real world performance.

FutureMark’s PCMark 2002, another synthetic benchmark, also displays the superiority of the 5400 RPM drives with the IBM/Hitachi Travelstar 40GNX performing very well.

HDTach 2.61 also shows the Toshiba MK4019GAX showing admirable average read performance. Despite that performance, the Toshiba MK4019GAX trails the IBM and the slower MK4018GAP model in read burst performance. It seems as if the 16MB cache isn’t really helping out in this case.

So far, we’ve seen the some interesting results. In general, the 5400 RPM drives are helping to increase performance especially in the synthetic tests.

Battery Considerations

I was extremely curious about the battery life since I frequently use my P2120 away from a power source. Based on the stated specifications of the drives, it looks like increased power consumption should be negligible. I wanted to test this firsthand so I thought of a worst case scenario to run to see how all the drives performed. So, what was my worst case scenario? Run a MPEG file continuously until the system has just enough juice to standby. For each run, I charged a standard battery to 100%, opened Windows Media Player 9, and played a MPEG-1 file continuously until the system reached 10% battery life. The power settings were set to Always On (meaning full CPU speed) and LCD brightness was set at full. I monitored the battery life using Passmark’s Battery Mark. The results were quite intriguing.

It’s true, 5400 RPM drives do require more juice than their 4200 RPM brethren. Of course, that’s to be expected. I didn’t expect the Toshiba MK4019GAX to perform so poorly. A loss of 17 minutes seems pretty significant in this test. IBM’s 10 minute loss is noticeable as well but not nearly as striking as the MK4019GAX. Of course, this isn’t an everyday activity and most people will be using power saving features (lower brightness, lower CPU clock) but people should know that there is some loss of battery life with these newer drives. If I get a chance, I’ll revisit the tests with longer more “real” tests.

Heat and Noise

During my tests I was not surprised by the fact that the 5400 RPM drives ran much warmer than the standard 4200 RPM model after prolonged usage. This is not really surprising even though the drives are equipped with Fluid Drive Bearings. The drive is running nearly 30% faster so it’s only natural for the drive to get warmer. Although I didn’t take scientific measurements for the change in temperatures, I noticed that the 5400 RPM drives were warmer and after prolonged use, they would trigger the internal fan of the P2120 much sooner than the 4200 RPM drive. As an example of this, during the MPEG drain test above I noticed the following results:

So, if you’re running the hard disk continuously, there’s a good chance the fan will trigger more often than normal. Your ambient temperature and laptop positioning will also affect how often the fan activates. Most people probably aren’t going to be running their hard disk all the time so this shouldn’t be an issue for most people. However, if your activities include lots of disk access you should know that the 5400 RPM drives will likely activate the fan.

In terms of drive acoustics, I ran into some very interesting issues. First of all, generally speaking, the 5400 RPM drives are fairly quiet and the increased rotational velocity is not noticeable. However, the seeks are an entirely different issue. The issue at hand is that during my tests, the Toshiba MK4019GAX clicked annoyingly even when the system was idle. The IBM/Hitachi 40GNX did not exhibit this issue at all. I was extremely puzzled by the clicking of the Toshiba MK4019GAX since every other Toshiba laptop drive I’ve ever bought did not have this issue. I immediately RMA’d the drive under the assumption that there must be a physical problem with the drive since clicking is normally a sign of drive problems or imminent failure. My second drive exhibited the same problems even though a check of the S.M.A.R.T. status indicated no problems whatsoever.

An exhaustive search of newsgroups and Google searches resulted in some relevant information. The problem is that by default, the Toshiba MK4019GAX drive automatically parks the drive heads in order to conserve the battery. So, whenever the system goes idle, the drive parks the head resulting in a click. Apparently, this feature is documented and the solution is to install Intel’s Application Accelerator and to change the power setting to full. This forces the drive to keep the head from parking. So, if you’re thinking what I’m thinking then you’ll know that the Fujitsu P2120 doesn’t have an Intel chipset and this fix does us no good. Contacting Toshiba Technical Support resulted in no solution other than a confirmation of this “issue”. Toshiba does not provide any sort of drive utility to change the drive settings. Oddly enough IBM provides a utility, the Feature Tool v1.80, to end-users to tune the drive for Battery or Performance with various steps in between.

The key feature of this utility is the ability to adjust the APM (Advanced Power Management) settings of the IBM drive from Battery Life (0) to Performance (254). By default, IBM drives ship with the setting in the balanced setting (128). From my informal testing, the Max Performance setting allowed the IBM drive to score a few more points higher in most of the benchmarks although, it most cases, it was negligible. The drive acoustics didn’t seem any different regardless of the settings. Of course, since the IBM drive was quiet to begin with, this wasn’t a big deal. My benchmark results were performed with the Balanced Setting. 

Final Thoughts

After weeks and weeks of testing and mixed emotions, I finally came to the conclusion that a 5400 RPM is a worthy upgrade for the P2120 if you’re thinking that you could use a little bit more performance. Due to the problems I experienced with the Toshiba MK4019GAX, I can simply cannot recommend it due to the very noticeable and annoying clicking issue. The specs are impressive and the 16MB cache is sexy, but overall, the IBM simply performs better overall. This is not to say that the Toshiba MK4019GAX is a bad drive it’s simply not a good match for the Fujitsu P2120. Subjectively, although both drives were noticeably faster than the stock drive, the IBM/Hitachi 40GNX drive was noticeably quicker in everyday use. Of course, that’s just my opinion but this is based on my experience.

If you’re happy with your performance then you probably don’t need an upgrade. If you’re used to using a faster desktop system then you may find the P2120 to be pokey. Since I’m generally multitasking, a more responsive system is what I was looking for. With the new 5400 RPM drives, I get more acceptable performance and it’s definitely worth the price. Of the two drives, I simply have to recommend the IBM/Hitachi model since it doesn’t suffer from some of the issues that I brought up earlier in the article. At around $150 online, the IBM/Hitachi Travelstar 40GNX is a great buy.

• Care to comment? [0]