Buffer issue with Plextor 716A on MSI Neo4 Platinum SLI



Hello all,

I’m needing some help in solving my buffer underrun issues with my Plextor 716A drive (bios 1.08).

Here is my configuration.
AMD 3200+ (Venice)
MSI Neo4 Platinum SLI (bios 3.00)
2x250 Maxtor Maxline III SATA on NVRaid 0
OCZ3200 EL Platinum Rev 2.
The system is OC to 2.7Ghz (270x10) on 2.5-4-3-8 1T, 1:1
ATI X700Pro 256MB
Plextor 716A
Pioneer 106
Antec TruePower 2.0 550W
SATA, PCI Spread Spectrum is disabled.

What’s happening is BPRec will kick in at least 10 to 20+ times during DVD burning from 4X ~ 16X. I can hear the drive stop, wait for the software buffer to fill up and then spin up again. The drive buffer also drops to 0 after the software buffer refills, but refills very quickly. I’m using FujiFilm TYG02 discs rated at 8X but burns at 16X (Avg 11X) and some generic cheap discs to test 4X.

This drive is an RMA drive from Plextor with manufacturer date of Jan 2005. My original drive died :sad:

I’m running WinXP SP2 with Nvidia SWIDE drivers. DMA mode is set to Ultra DMA 66. The problem also exists with standard Windows IDE drivers set to Ultra DMA 66. WinXP is a clean install, nothing installed except for some system monitoring/testing tools such as CPU-Z, Clockgen, Plextools, Nero etc.

This is not an OC issue as I have resetted the bios to factory default and the problem still exists.

I’ve ran the Plextor self test and the test passed.

I’m almost out of options. I have not yet increase the voltage to the NB chipset which may increase the stability of the SATA devices. (I’m doubtful on this one)

I have not upgraded the bios yet as any update from 3.00 will not allow me to OC to 2.7Ghz…

Nero’s Burst Test worked once with this drive and the result was 44MB/s. other times, nothing shows up (Nero software bug with Plextors?). The Pioneer I have in this system is reporting 24MB/s.

Any ideas/solution? Does anyone else have this problem?


why do people use cheap components with high end hardware and expect it to always work as if it was designed to on a propperly made high end system?

msi mobos are known to perform at the same wave length as ecs mobos, as they are both entry level mobo makers. your problem is you don’t have enough bandwidth on the motherboard to send the needed data fast enough causing the buffer to suffer in some special cases even stall. About the other drive dying check the antec psu since antec psu’s suck… sorry to say but they are nowhere close to being good… they aren’t even true voltage all gimic… weight matters. Also the 2.0 spec starts at 660watts on high end psu’s. As you may not know o\c a system can also kill comonents and degraded the bridges and why the hell would you have PCI Spread Spectrum is disabled. That would greatly kill the ability to provide BANDWIDTH to your devices. Sereing how you have many probs in setup alone your prob also overusing your allocated ram slots by using all 4. Which NO home level MOTHERBOARDS can HANDLE to date. MAX is 2. To benefit from performance standpoint all 4 channels actually make cas 2 ram act as cas 3.5 - cas 4.

Also why did you pick AMD?? they are values for a reason they aren’t faster. It’s all AMD marketing in a dying company who has to resort to suing the competition to stay afloat. AMD’s are overpriced for the performance / price they offer.

Intel’s offer stability (this also includes providing the PROPPER bandwidth to devices). It is very rare for a intel motherboard to have bandwidth issues bc of its bus, unless it was the user using commpents not designed for that generation ex. 2700 single channel ram on a 800fsb dual channel chip…

Learn from your mistake and move on.

Also overclocking doesn’t make the computer go faster, thats all in your mind. The cas speeds is what matters.


Do a read simulation in Plextools, is the average read spead around 2.4x? If yes we’re most certainly looking at a drive running in PIO mode instead of UDMA. Please read the FAQ on this: http://club.cdfreaks.com/showthread.php?t=49767#general-dma

(sound of major alarm going off)
Is that SWIDE an IDE driver? Please uninstall ANY third party IDE drivers installed and use plain windows IDE drivers only. That measure alone solves >90% of all problems.

Unfortunately the mode set does not necessarily corespond to the mode in use. Use device manager to unistall the device and redetect under windows. If that does not help uninstall the whole controller the drive is connected to and redetect.

Good boy! :iagree:

Me too. The stabilisation might occur in carbon state.

Seriously. Forget about OC your MB and get the most stability with the newest FW. OC the MB often kills PCI timings and does other nasty stuff that hampers proper working of components.

45MB/s is what i’m getting too with my 716a in CD/DCD-speed in UDMA mode. Every time i do the test, no but in the software.


Wow. I think I’ve managed to break every single one of your “rules” and still have a stable, running system for the last three years. But I guess I’m just lucky.


amd are anything but stable


My guess would be a bug in Windows and the current moon-phase.


I didn’t know MSI was cheap, I normally buy ASUS products, but their A8N board, had alot of issues with SATA devices, plus I wanted the creative onboard sound chip. In hind sight, this may have been a mistake as MSI bios updates really suck.

You’re probably right on this one, but is it a hardware or software issue?

Sorry, my mistake, I should have been more clear, the other Plextor died on my old computer, not the new one. As for the PS, you’re probably right again, but every review I read online says its pretty good.

I’m not sure if the PCI Spread Spectrum has anything to do with my buffer problem. My understanding of it is, MB makers put this into the bios to pass FCC regulations. With it disabled, my PCI bus is locked at 100mhz, rather than floating between 99 to 101 etc. But I will turn it on tonight to see if it solves the problem. I should have been more clear… I only have one Gig of RAM on Dual Channel.

I dunno… 2.7Ghz for $230 (CDN) is pretty good performance/price to me. I’ve only got a buffer issue with my Plextor drive. Plus everyone is sueing someone these day, Intel is going to counter sue anyways.

I agree, Intel is more stable, but there are always issues when you are playing with the latest technologies. Its not like Intel has never had any problems in the past.

Yes, I don’t think I’d buy another MSI MB, but still I’d like to find out what exactly the problem is. BTW, if my memory serves me correct, running at default speed with timings of 2-2-2-5, SuperPI 1M took 40s and OC with timings of 2.5-4-3-8, it took 30s seems faster to me, but your right, it doesn’t matter when your surfin the web which is what I do most :iagree:


Nonsense, not to mention trollish.

More nonsense. Stampedo’s specs show a powerful machine with more than 5X the power of my AMD, and mine (ECS mobo, too) has no trouble burning at top speed without buffer underruns on quality media.

Buy a new PC to solve a minor burning problem? Let me guess - you’re an MSCE, right?



Does your Pioneer have buffer underruns, too? If not, try switching the IDE channels on each. And make sure your preferred burning drive isn’t on the same IDE channel as the hard disc you’re burning from.

If you have packet-writing software installed (InCD, Drag2disc), you might try removing that. My experience with the packet writers showed heavy interference on the IDE channel being caused by that software. My read transfer tests looked like an EEG…


I’ll try out some of your suggestion tonight… thanks for the info.


The Plextor is on the secondary IDE (Master) and Pioneer Primary IDE (slave). The Plextor was on Primary IDE (Master) before I added the Pioneer. I will test out the Pioneer tonight, hopefully it has issues too, then I can narrow it down to MB + drivers. My HDs are on SATA 1&2, are the SATA channels shared with the IDE channel?? The only optical drive related software I have on the system is DVD Decrypter, Nero and Plextools.


With hard drives on SATA and the two burners on separate IDE channels, it looks like you have it set up as good as can be with none of the devices “crowding” a particular channel.

A friend of mine had slowdowns on his BenQ, but it shared the secondary IDE with a Pioneer 106. When he unplugged the Pioneer, everything was fine … but that likely isn’t the solution for you. hwp suggested running transfer tests and that’s still a good suggestion. A read-transfer graph should show a more or less smooth line as the drives revvs up to full speed.

BTW … you’re not burning “on the fly” from one DVD to another are you? I’m assuming you’re burning from hard disc images/files and not directly from another DVD drive, ie. Pioneer -> Plextor. If you are burning directly from another DVD drive, that would certainly explain the buffer level dropping to zero so often.

The only optical drive related software I have on the system is DVD Decrypter, Nero and Plextools.

As long as Nero’s InCD isn’t part of the install, that setup sounds just fine. Some software like Sony Sonic has been known to cause problems…

Please let us know how you get on tonight and good luck!


agreed…dragger’s posts don’t make much sense at all…


I had issues with my Plex 716A being on the same IDE cable as my other burner. Either as master or slave. I finally had to put the Plex on my second IDE port. ALso, I believe the PLex 716 users manual suggests that if you have to have two burners on the same IDE cable, the plex be set up as master, subject to it being ‘compatible’ somehow, with the second burner (in my case it wasn’t).

You must also verify, in device manager, that the plex is connected as UDMA mode, not PIO mode. The symptoms you describe appear to be related to this.

Best Regards


umm no i’m a programmer i actually understand hardware and what happens when cheap components come into the picture yes intel has issues of its own but they are nowhere near amds… intels newest heat issues shouldn’t be an issue to most users and i’d recomend a socket 479 anyway of the weak… they are faster even tho atm they are 400mhz slower.


btw any issues with bandwidth output the issue is always hardware


First of all, the problem is not a problem with the CPU (AMD). I’ve own Intel and AMD processors for over 10 years now and never had an issue with either brand of processors.

Second, if MSI is cheap, what is better ASUS? ok and? Intel MB? or Dell? Don’t make me laugh. DFI is in the same ranks as MSI IMO.

I’m going to assume that you saying AMD is 400mhz slower to some P4 chip.
As a programmer, you should understand optimization. AMD chips are simply more efficient and optimized. Take the Intel Centrino for example, slower in mhz, but just as fast as the P4 except in some encoding benchmarks, why is that… hmmm.

Do not make any assumptions without proof… as a programmer myself… I can write any software that will eat up all your bandwidth. You can run it on your Intel system and blame it on the hardware.

BTW-I did enable the PCI Spread Spectrum which made no difference as I expected.


dragger, stop posting nonsense information. Your replies are not helpful and quite frankly, just not true.


The pioneer burns fine on the primary channel (slave). I’ve uninstalled the Nvidia ide drivers also. Maybe its the NVRaid drivers + Plextor drive = buffer issues? I’m going to move the raid to the silicon raid and if that doesn’t work, I’ll run it without raid. The reason is HD Tach tests show the NVRaid’s transfer rate drop down to 10MB/s, but then shoots back up to 110MB/s fairly quickly.
Anyone with my configuration have a plextor working without any buffer underrun issues?


A HD transfer-rate issue would definitely translate into a buffer underrun issue esp. when the rate is as low as 10MB/sec … you need at least a sustained 11MB/sec just to keep up with 8x burning. However, if that really was the source of the problem it would also happen with the Pioneer. A write-transfer test would show how an 8- or 16x burn progresses, but even if it turned out nominally all it would tell you is that the drive is physically capable of burning at that speed, yet tell you nothing about buffering.

Other than trying to burn from an IDE HD (thus ruling out SATA/RAID issues), I’m outta ideas Stampedo.