G
Guest
Hello
I am having a problem that quite frankly has me baffled and livid.
Configuration
First of all we are an OEM that is using W2K with a National Instruments PCI-6527 digital Input/Output PCI board. We have developed a new product using this NI board within the past 6 months. The problem with our system began to occur when we went into production on the new product. We have found that the latest hardware revision of the NI board will only reliably work when we install 2 more, non-specific PCI, boards into our backplane. The development NI boards were an earlier hardware revision.
Scenarios
1) Motherboard 1, W2K, SP3, 1 PCI-6527, no other PCI boards. Not working reliably – I/O Board will function briefly and then stop. Sandra tests` indicate system loses sight of board on the PCI bus
2) Motherboard 1, W2K, SP3, 1 PCI-6527, 2 other PCI boards installed. Working reliably. It does not matter what type of additional PCI boards are installed. I/O board works reliably in all case
3) Motherboard 2, W2K, SP3, 1 PCI-6527, no other PCI boards. Working reliabl
For scenario 1, when I run the Sandra tests on both the old and new hardware revisions of the NI board, the only significant difference item I observed is the PCI Latency of the boards.
The new hardware revision of the NI board will always follow the PCI latency number that is set in the BIOS. i.e. If the PCI latency in the BIOS is set to 32 PCI clock cycles, the Sandra test shows the board at a setting of 32. Identical for 64 and so on
The old hardware revision of the board will always show a PCI latency of 8 in the Sandra tests no matter what PCI latency setting is set in the BIOS
So you say the solution is simple, Ask NI for a fix since it is obviously their problem. Here is where the livid part comes into play. NI is very very slow in attempting to reach a solution and insensitive to how critical this is to our company! Their engineering is supposedly working on this. However their only solution offered after a few weeks is to sell us some used old rev boards to fix the problem. What kind of solution is that!!!!
Therefore, does anybody have experience or have a copy of the pci.sys file detailed in the following tech note? I really hate to spend $245 dollars on a phone call if I do know that the file will fix my problem
Microsoft Knowledge Base Article - 824395 Interrupts That Come from a PCI Device That Uses a Windows NT 4.0-Style Driver Are Ignored (http://support.microsoft.com/default.aspx?scid=fh;en-us;Prodoffer20a&sd=GN#faq1400
Does anybody have any comments or suggestions? Any help would be appreciated
I am having a problem that quite frankly has me baffled and livid.
Configuration
First of all we are an OEM that is using W2K with a National Instruments PCI-6527 digital Input/Output PCI board. We have developed a new product using this NI board within the past 6 months. The problem with our system began to occur when we went into production on the new product. We have found that the latest hardware revision of the NI board will only reliably work when we install 2 more, non-specific PCI, boards into our backplane. The development NI boards were an earlier hardware revision.
Scenarios
1) Motherboard 1, W2K, SP3, 1 PCI-6527, no other PCI boards. Not working reliably – I/O Board will function briefly and then stop. Sandra tests` indicate system loses sight of board on the PCI bus
2) Motherboard 1, W2K, SP3, 1 PCI-6527, 2 other PCI boards installed. Working reliably. It does not matter what type of additional PCI boards are installed. I/O board works reliably in all case
3) Motherboard 2, W2K, SP3, 1 PCI-6527, no other PCI boards. Working reliabl
For scenario 1, when I run the Sandra tests on both the old and new hardware revisions of the NI board, the only significant difference item I observed is the PCI Latency of the boards.
The new hardware revision of the NI board will always follow the PCI latency number that is set in the BIOS. i.e. If the PCI latency in the BIOS is set to 32 PCI clock cycles, the Sandra test shows the board at a setting of 32. Identical for 64 and so on
The old hardware revision of the board will always show a PCI latency of 8 in the Sandra tests no matter what PCI latency setting is set in the BIOS
So you say the solution is simple, Ask NI for a fix since it is obviously their problem. Here is where the livid part comes into play. NI is very very slow in attempting to reach a solution and insensitive to how critical this is to our company! Their engineering is supposedly working on this. However their only solution offered after a few weeks is to sell us some used old rev boards to fix the problem. What kind of solution is that!!!!
Therefore, does anybody have experience or have a copy of the pci.sys file detailed in the following tech note? I really hate to spend $245 dollars on a phone call if I do know that the file will fix my problem
Microsoft Knowledge Base Article - 824395 Interrupts That Come from a PCI Device That Uses a Windows NT 4.0-Style Driver Are Ignored (http://support.microsoft.com/default.aspx?scid=fh;en-us;Prodoffer20a&sd=GN#faq1400
Does anybody have any comments or suggestions? Any help would be appreciated