24/10/40A on Promise Ultra66

Hey All,

First time poster…
MAN!!! What a forum…
Wealth of info here…on to the problem…

We have a Plex 24/10/40A at work running on a Tyan Thunder2500 motherboard. It has a wierd chipset (Serverworks…aka Reliance) and has problems with IDE/DMA mode. For this reason I got an old Promise Ultra66 card I had and installed it. I turned off all the motherboard IDE ports and cabled the cdroms as follows:

Ultra66-IDE1 to Plex 24/10/40A
Ultra66-IDE2 to Acer 52MAX CDRom

I then create a Disc image to take the source cdrom out of the picture. I have tried Nero and CDMate, but all I can seem to achieve without underruns is 16X. I am using a disc with about 680MB of plain data. On 24X I get about 60 underruns, 20X about 30, and on occasion I will get 1 or 2 with it running 16X

I took the CDR’s home and they ran 24X fine on my TDK VeloCD.
I am running F/w 1.03…

all help appreciated :bow: …thx


ps. when I run task manager and check the CPU utilization, I am hitting about 65% in the middle of writing at 16X…is that right???

Connect your harddisk to the Promise controller,and try the cdr as master on your primary ide and the cdrom as master on your secondary ide…
Connecting cdrom,dvd or burners to a Promise ide card is asking for problems…too much incompatibility issues…

i have that drive too
the drive cant burn faster then 16 speed in the begin
in the middle of the burning proces it turn up to 24 x
but there is an new firmware for that drive the

click here for the european firmware

and click here for the american version

if you install this firmware then it burns faster and it support more cd®’s

sorry for my bad english i am dutch :slight_smile:

slight update:

I guess I should have mentioned 2 other things:

  1. My boot drive is a SCSI drive…
    the onboard IDE controllers are a bit screwy…thus the Promise controller

  2. I have the Nero cache set to 80mb. when I go to burn an image file, the first 80 mb writes at 24x with the buffer staying at 98-99%. After the 80mb runs out is when the underruns start occuring…

:rody- I have a CD in that is big enogh that it should be kicking into higher speeds…

thx for advice…


Originally posted by Southpaw
slight update:
the onboard IDE controllers are a bit screwy…thus the Promise controller

Search for updated busmaster drivers for your mobo,or reinstall yours…the Promise card won’t make things better…:slight_smile:

The thing is, it runs at full 24X for the first 80mb of the CD(nero cache)…that tells me the burner is running fine…

seems like the transfer from the SCSI hard drive to the cache/cd is causing the problem…(bogs down after the 80mb cache is depleted)

would an on-the-fly give me another data point?? since it is only going through the Promise card…taking the hard drive out of the loop??

I ran an HDTach on the hard drive and get about 12MB/s



ps…why wont the Promise card work??? just not good with CDRW in general??


It is probably the SCSI controller…
I tried to copy about a 50MB file between partitions…
it took 20 seconds…

thats only 2.5MB/s about the equivalent of 16X…

looks like my problem is getting the info off the hard drive…

geez…I hate SCSI



I cannot enable dma on my system for my plextor 241040a - it went away and I cannot convince Win98SE to give it back.

But, even at x12, I occasionally get a ‘underrun was avoided 1 times’ - this is with a 2MB (or is that 4MB) buffer on the drive and nero always staying at 98% full - I believe the drive cannot reliably store the data despite what is claimed. It would take several seconds of data stoppage to give a buffer empty scenario which I do not believe is happening - nothing else is running at all.

This occurs with many makes of drive not just plextor - I feel there is more to it than large OS ram buffers and large buffers on the cd-writers.


Originally posted by icey
I occasionally get a ‘underrun was avoided 1 times’ - this is with a 2MB (or is that 4MB) buffer on the drive and nero always staying at 98% full
This is because your drive switches between write zones (Z-CLV writer) from 20x to 24x. When the drive switches, its BURN-Proof technique prevents the buffer underrun and thus you’ll get the message: ‘underrun was avoided 1 times’ … So this is quite normal normal :slight_smile: