"Frequency out of range" when monitor goes to sleep

[billtodd]

I'm generating a new system for my wife, who has some software that runs only on Win9x. Since it will multi-boot with Win7, I'm using an nForce3 board (the most recent hardware we have that still runs Win98 well).

In the past I had a GeForce MX 4000 AGP card on this board and had a minor issue when running Win98SE with the CRT refusing to power down when the power management time-out told it to: it would display the message "Frequency out of range 00.0 Khz / 186.8 KHz" (the second number may vary a bit) after trying to power down, cycle through this a few times at a few seconds each, then return to normal awake status. This problem did not occur when running Win2K on the same hardware, and I eventually made it go away on Win98SE by changing the card's refresh rate from 'optimal' to any specific value supported by the monitor ('Adapter default' may have worked as well).

nVidia doesn't list any Win7 drivers for the MX 4000 (though any Win7 nVidia driver might work) and in any event that's a fairly old card to be running Win7 on, so I tried a (marginally newer) FX 5200 we had lying around (for which nVidia does list a Win7 driver). Unfortunately, not only does it have the same problem running Win98SE but I haven't been able to find any combination of refresh rate, screen resolution, and color depth that will make the problem disappear (Google provided some hits that suggested this screen message could be caused by over-taxing the card's bandwidth capability, but the situations being discussed involved active use of the monitor rather than going to sleep - when I wouldn't expect bandwidth to be an issue - and the Google hits also did not refer to 00.0 KHz as the first number).

I'm using the unofficial 82.69 nVidia driver since my wife's monitor is wide-screen (I think I was using an earlier driver version with the MX 4000): I'm just getting the system up and running on the CRT that's having the problem, but it would be nice to resolve it anyway (even if it doesn't occur on the LCD display, which I haven't tried it with yet). There's 1 GB of RAM in the box and the AGP aperture is set to 128 MB, but I do have MaxFileCache limited to 393216 so would not have thought this could be caused by a lack of available system memory. Incidentally, before installing any nVidia driver, when the default Windows driver was handling the card, this strange behavior did not occur.

Anyone familiar with this issue, or have any ideas what might be causing it?

Thanks,

- bill

[Offler]

Globally the Windows 98 had some issues with power management. I dont use power management for many reasons, but the bugs are part of the problem - the windows does not sleep or wake correctly with all devices. This may wary on used hardware - newer devices with WDM drivers may run more correctly, but only in case you are using Windows 98SE or Windows ME.

[billtodd]

I never did get Win98SE to awaken from standby with this board but the problem here does not involve that - just letting the monitor power down after some period of inactivity (which I've never had any problem with in Win98SE elsewhere, and which it would do even on this board with the MX 4000 as long as the refresh rate was not set to 'optimal').

- bill