Tested: DVD±RW Endurance Round 2 – Part 3: DVD-RAM Comparison & A Dying Drive

Now, I know DVD-RAM is not a type of DVD±RW, but hear me out – I thought it was worth including in this round-up for a few reasons. The first is that it uses a phase-change recording layer, just a different formulation, meaning they share the same recording technique. The second is that my recent acquisition of DVD-RAM2 (12x) discs, I now own the lowest-endurance DVD-RAM type, often stated at 10,000 cycles rather than the 100,000 cycles of the former types. That, combined with its high rewriting speed, would make wearing out the disc potentially feasible.

As a result, in parallel, I quietly assigned one of the drives to test the DVD-RAM2 in parallel. The set-up is basically the same, but in order to maximise speeds and replicate the way that DVD±RW discs are used, hardware defect management was disabled and the full area of the disc was simply written and read back sequentially in much the same way as an ordinary rewritable disc. Unfortunately, because of the way DVD-RAM works, quality scans are occasionally interrupted as it seems the OS accesses the disk randomly, resulting in the occasional quality scan with a gap. This test was done with the same disc that previously featured in this post, thus is not a fresh disc. But given the long claimed endurance, I figured that probably wouldn’t be a problem.

So lets see if DVD-RAM2 manages to demonstrate its superior rewritability compared to DVD±RW discs.

Attempt 1 – DVD-RAM at 12x

I started with one drive and a dream to run my tests at 12x. We already know from prior experience that 12x burns are a little finicky towards the outside diameter, but I thought I’d try it anyway. Unfortunately, the drive seems to have different ideas. It decided to randomly flip-flop between 12x and 6x burns with an increasing proportion of 6x burns as it was going on. Under this regime, 1078 cycles were completed.

Unfortunately, the first verify-error occurred at 495 cycles. This is not clearly better than DVD±RW, but it seemed to flicker on and off …

It seems that dips in the TRT always occurred in the 12x burn zone. This suggests that perhaps 12x is really pushing it too far.

It seems most likely that the verify error is related to the outer diameter where there is a error “hump” that disappears every time the disc is rewritten at 6x. This suggests to me that the elevated errors towards the outer is likely a result of limitations in the mechanical rigidity of the disc and the ability of the tracking servo to compensate for tracking, focus and tilt error.

This gave me an idea – let’s reduce the speed to 6x to match the high-speed DVD-RW format. At least, in this mode, operation will be in CLV and that would relax requirements on the disc’s rigidity. Hopefully this will preserve the higher-quality write and reduce the possibility of an early verify failure.

Attempt 2 – DVD-RAM at 6x

Now, we’ve locked it down to 6x burns. This added another 2064 cycles to the disc.

Unfortunately, the first verify error now occurred at 230 cycles, but I continued to let it run to see what would happen.

Errors in the TRT are few and far-between but now cluster around 0.5GB to 1.5GB.

Looking at the quality scan, we can see a spiky region of elevated error around 0.5 to 5GB. Initially, the errors only show up in PIEs, but by 170 cycles, it was consistently showing up in PIFs which will affect readability. The spike does sometimes reduces and comes back … this doesn’t seem to be a problem with the disc but a problem with the drive or the power supply.

I replaced the power supply, but the noticed the drive was noisier than usual. I suspect mechanical failure so I set the drive aside and try another.

Attempt 3 – DVD-RAM with a New Drive at 6x & 12x

Swapping to a new drive, I try a mixture of speeds – initially 6x, then 12x for a little before returning to 6x because of the same kind of elevated errors on the 12x zone. A total of 2,055 cycles were added to the disc, for a total of over 5000 cycles throughout this part.

The first verification error now occurred after 801 cycles. Again, this is not heads-and-shoulders above DVD±RW but is more acceptable. But considering this is exactly the same disc that failed earlier twice, being able to do another 801 cycles before failing a verify suggests the disc is just fine.

Aside from the single error at cycle 801, no verify error occurred through to cycle 2055 suggesting that it managed another 1,253 cycles with no error -which would be above the rating for DVD±RW. I did terminate the test even though the disc remained working as I think I’d done enough cycles to prove that the disc itself is surprisingly durable and I didn’t want to waste excessive amounts of energy.

The TRT is mostly stable with the cycle #801 read error seemingly being the anomaly.

The quality scan instead seems to suggest the spiky error region may have been inherited from the other drive and it takes almost 300 cycles for the spike to be pared down and the error rates to become more stable. But it still seems to come back, so maybe there’s still something happening (bad firmware strategy? power supply oscillation?). While the PIFs are stably low to the point of being even better than a lot of the best-burned recordable discs, the PIEs bounce up and down like a yo-yo, which may be somewhat representative of the recording layer response changing and perhaps the power calibration is not ideal.

A Dying Drive

It seems a foregone conclusion that the drive I was using to run the endurance test is dying. As to why? I suspect it’s a worn out spindle motor bearing – the drive got louder over time and changing the power supply and even swapping to a different computer didn’t seem to resolve it. The final check-out before the drive is laid to rest is to try and get a few ordinary burns out of it to diagnose it.

The first was some decent quality MCC-004-000 – requested burn at 16x but the drive was very noisy and simply couldn’t calibrate at 16x, falling back to 8x. There is a very large error spike and error hill … that’s not what a good burner would do. The error clumps are also very unexpected. Perhaps the laser isn’t doing great either.

I gave the drive the benefit of the best media I had on hand – the MXL RG04. This one it did burn at 16x …

The burn quality was excellent – I did also test an Sony NEC Optiarc AD-7200A on a quality scan and there’s definitely a correlation, although the middle block between 1.5GB and 3.0GB looks elevated on the AD-7200A likely because there’s some variation in beta that the NEC is sensitive to.

This seems rather conflicting – I burned another disc but the drive froze-up mid-burn and didn’t finish. It’s definitely sick.

Two different types of CD-Rs, both failed the burn although at slightly different positions. I think it’s more and more likely a worn spindle motor that results in run-out (wobble) the servo simply can’t compensate for.

Unfortunately, with these results, the drive was disassembled and disposed of, as it was no longer reliable or healthy, thus no longer useful to me. It was a foreseeable outcome from using drives to do endurance tests and a key reason why I tried to standardise on using a model of drive I could readily amass in case failures were to hit so I can swap to another one to continue the experiment. Thankfully, all other drives still seem to be in good health for now, but as to how much health they have remaining, I can’t be sure.

Conclusion

While DVD-RAM is not a DVD±RW format, they are relatives as both are phase-change based recording layers. The DVD-RAM recording layer has always been advertised as having a higher endurance than that used by the DVD±RW formats. The 12x variety of DVD-RAM is advertised with the lowest endurance of around 10,000 cycles, so I decided to use it for this test.

On the basis of first verification failure, it seems that DVD-RAM did not excel far beyond DVD±RW when tested on the basis of a direct-write with verification and hardware defect management disabled. But when we consider the longest stretch of no verify failure, we achieved 1,253 cycles when the test was aborted to save energy. In all likelihood, a longer stretch of error-free cycles is indeed possible.

However, it is worth considering that this is an uncommon way to use DVD-RAM and more closely matches DVD±RW usage. However, in spite of this, any failure appeared to be a soft failure – error spots were cleared on a rewrite and the overall error rates did not escalate consistently unlike that of DVD±RW format discs. The spikes in error rate appear to be a drive-specific anomaly – one drive being bad created consistently bad spikes, but the replacement drive which is better still did exhibit an occasional spike near 0.5GB-1GB. It might just be a specific issue with the firmware or perhaps the load causes power supply oscillations … it’s hard to know for sure.

As a result, the DVD-RAM disc that I have has completed over 5000 cycles, but the final write isn’t clearly any worse in terms of quality scan result compared to what might be seen from a fresh disc. The PIF rate is surprisingly low, to the level of being better than some of the best-burned write-once discs while the PIE rate does bounce around a little, suggesting maybe there is some wear happening. When used with hardware defect management and proper formatting, bad areas would be remapped to other areas, extending the usable lifetime of the disc. That being said, DVD+RW can also support such operation in Mount Rainer (DVD+MRW) mode, although this is not mandatory or commonly used.

With that, and the loss of one of my iHAS120-6 drives, this brings Round 2 of this endurance test to an end. But even as I post this, Round 3 is in progress. Why? Well, I’ll leave that for you to guess … but this one might take a year or even longer to complete.