As the cable and the disks work fine at UDMA Mode 4 in the SIS530 motherboard, I was wondering if the problem is actually a hardware problem on the MVP4 Board. There is a series of resistors - and some Zeners just beside the IDE Pin Headers on the motherboard. Could this be a place to investigate?

One more thing: The ATAPI CDRW Drive actually is set to UDMA Mode 4 - only the Hard drives are not. Could it be a BIOS problem?

This morning, I replaced the 80pin ribbon cable to the SSD with a 40pin as this prevents UDMA modes above Mode 2. With this cable, Windows XP reports UDMA Mode 2. So it seems to be a problem with the Chipset or Bios when UDMA Mode 6 is detected (maybe Mode 5 two - I have no way to tell because I have no UDMA/100 drives. Benchmarks confirm that the drive is no longer limited to 16mb/s

So a question to BIOS Specialists among you - is there any way to tell the BIOS to never select a UDMA Mode higher than i.e. 4 for a drive? The Bios for my Tekram p5m4-m+ has only UDMA Options "Disabled" and "Auto"

adegn wrote on 2025-06-19, 08:13:

This morning, I replaced the 80pin ribbon cable to the SSD with a 40pin as this prevents UDMA modes above Mode 2. With this cable, Windows XP reports UDMA Mode 2. So it seems to be a problem with the Chipset or Bios when UDMA Mode 6 is detected (maybe Mode 5 two - I have no way to tell because I have no UDMA/100 drives. Benchmarks confirm that the drive is no longer limited to 16mb/s

So a question to BIOS Specialists among you - is there any way to tell the BIOS to never select a UDMA Mode higher than i.e. 4 for a drive? The Bios for my Tekram p5m4-m+ has only UDMA Options "Disabled" and "Auto"

Hi adegn,

This bad IDE performance on your Tekram P5M4-M+ can be both a Windows and a BIOS issue.

Windows
The Atapi.sys driver in WinXP (SP1) will degrade the DMA mode to a lower level if more than 6 transfer timeouts occur, to prevent data errors. If the transfer timeouts continue, then it will change to PIO mode for the IDE device. The only way to re-enable DMA is to remove the device or controller in Device Manager, so it will redetect the device automatically at reboot and erase the timeout log for the device.

With Windows XP SP2 and Windows 2000 SP3 a less aggressive policy was introduced, so instead of remembering all DMA transfer timeouts throughout history, then it would reset the timeout counter if a successful transfer was performed.

So if you use WinXP SP2 or SP3 (or the KB817472 Hotfix), you have the updated Atapi.sys driver and Windows should not be the problem.

BIOS
Because your MVP4 board uses the VIA 686 southbridge, it supports up to UDMA mode 4. When a UDMA 5 or 6 drive is attached, the BIOS should report such a drive as UDMA 4 or UDMA 66. If you see UDMA 99 on the BIOS summary screen, then there is definitely a BIOS problem.

I will look into the disassembly listing I made, when I patched the v1.06 BIOS for K6-2+/III+ support back in 2002, to see if I can find a UDMA mode issue there.

Jan

Jan - you are the best. Amazing what you do for us all. Thanks 🙏

By the way - i am on xp sp3 - and the same issue exist when i am on win 7 sp1. UDMA 4 drives - like my DVD drive - are reported as UDMA 4 (66) by bios summary and in Windows. I have only UDMA 133 hard drives in my stock and SATA drives with an adapter. I tried replacing my adapters with a new Marvell one - as this is suggested in other threads here. I also tried to set the UDMA transfer mode using some DOS tool, but this is not persisted to the drive electronics - so it doesn't help at all. Only changing cable to the 40 pin makes a change and the drive is working in UDMA 2 mode - double as good as mwdma 2 but just half as fast as UDMA 4 which should be achievable. I am very grateful for your help.

I tried to experiment with the masterdevicetimingmodeallowed setting in XP registry. No matter which max UDMA setting is set, XP reverts to MW DMA 2 - unlesse I use a 40 pin 40 conductor ribbon cable - then it is set to UDMA 2

Here you see the ssd detected as UDMA 99. The TRAP error occurs when I create two primary partitions on the same drive - when the drive is bigger than 32 GB.

Okay, I found the routine in the P5M4M+ v1.06 BIOS that does the UDMA mode enumeration on the installed drives. This code indeed doesn’t limit the UMDA mode to 4 on UDMA 5, 6, or 7 drives, hence the UMDA 99 indication on the BIOS summary screen on your mode 6 drive.

I’m now working on a patch to limit the UDMA mode to 4 on UDMA 5, 6, 7 drives, while still reporting the correct mode on older UDMA 0, 1, 2, 3, and 4 drives.
I will send you a patched interim BIOS to try by e-mail, hopefully later today.

Jan

Awesome Jan - thank you

Jan created a patch for the BIOS. Now UDMA 4 is correctly detected for UDMA 5+ drives. These now show correctly as UDMA 66 in the Bios Summary and as UDMA Mode 4 in windows XP/7 Device Manager.
Thank you so much Jan
/Anders