The video is integrated on the mainboard?
Your issue almost sounds like your VGA BIOS is in some sort of backwards compatibility mode. I'm not sure if Tseng has nonvolatile memory on their W32 boards but if it has there may be a configuration tool.
Are there DIP switches somewhere near the video? Have you tried powering up without a monitor connected once and again with it connected?
I doubt hotplugging VGA is dangerous, at least not if the protective earth cirquit with the two devices is okay. I do it all the time.

Thanks for your reply.

Yes, the ET4000 chip is soldered to the motherboard along with the Music RAMDAC that is placed close to the VGA port and its VRAM. The VGA BIOS does display the output it usually does in all cases. The motherboard has only one BIOS ROM for both VGA and System, which seems oddball?

There are no DIP switches or configuration jumpers in regards to video modes or anything. The only thing in regards to video are two jumpers that disable the VGA if both are removed together as instructed. I did try and all that makes the computer do is beep on reboot as expected.

I have not yet tried powering it up and then down without connecting monitor. I'll try that tomorrow as its getting late over here. But the weird thing is that it did work with the Sony Bravia LCD tv plugged in from the start.

I'll try to get the machine back together and run MSD.EXE and see what it says.

Goodnight from GMT+10

PS: Oops, I only just noticed that there is a hardware forum. Sorry, feel free to move if necessary.

Okay, I've managed to get hold of another VGA monitor, a generic CRT. I also got hold of the IBM 9515 to check out.

VGA worked perfectly with this generic monitor, but nothing changed at all with this HP vs17. This leads me to think there is something wrong with the crappy HP LCD. But even so, the other two machines video cards (Radeon 7500 and Geforce6 based) I use with it work perfectly with it.

One thing I did discover is that an EDID reader cannot read the DDC information from this monitor. However, since this computer is 1993 vintage, this was just before DDC was standardized (from wiki)??

I also tested the IBM 9515. Whats with XGA-2 being incompatible, I am not sure, as I have gotten a stable image with 1024x768 or 640x480 at 75 Hz (using Radeon 7500), even though it is skewed to the right.

I researched video card damage from monitor. This seems rare save for a few cases, the converse seems more prevalent (and more understandable.) What has me worried is that the ET4000 does produce some strange artifacts in its reduced mode (noticed some flicker in the colours in Skyroads) but that could be due to poor implementation of legacy modes?

What happens to you guys if you boot your computer that has an ET4000 without the VGA connected, then connect the VGA after BIOS has loaded?

EDIT: ahh, I see now. VGA pins 11, 12, 4, 15 are used as ID pins before DDC. But do these really effect whether VGA capability or not? There are things like B&W VGA monitors used in older POS machines at the supermarkets....

What I do want to ask seriously is, should a VGA card (not only ET4000) default to VGA-colour if it is capable of doing so?

un3nabled wrote:

What happens to you guys if you boot your computer that has an ET4000 without the VGA connected, then connect the VGA after BIOS has loaded?

Works normal for me. I have a taiwanese noname ET4000 AX.

un3nabled wrote:

What I do want to ask seriously is, should a VGA card (not only ET4000) default to VGA-colour if it is capable of doing so?

Yes, but maybe the ET4000 gets fooled by the HP monitor (wrong ID pins or whatever) and switches to a different mode.

Download the VMODE utility (see attachment) and try to switch the card into VGA mode (VMODE VGA) before starting a game.

Welcome to the forums btw...
I had a ProLinea 3/25zs once, it was the slowest model of the series. Mine had an onboard WD Paradise 512k VGA.

Hello, 5u3. Thank you very much sir, for posting VMODE, really nice utility.

Sadly, using "vmode vga" didn't solve the problem for me, even after a "warm reboot". But yes, I am able to easily switch modes. At least that shows something is working properly.

Let me clarify the problem:
-When I use the HP LCD, VGA does work correctly as long as the program enforces it (even the VESA driver loads properly as wfw 3.11 is all good.) It FAILS in that the programs which look for it, in other words, they CANNOT DETECT VGA. Why is this the case just because I'm using a HP vs17????
-I can get VGA detectable by plugging in the generic monitor whilst the monitor is powered down and unplugged from AC to the computer, then booting computer and hotswapping the VGA cables after the system has booted. It continues to work even while warm-rebooting machine. Odd eh?
-Hot plugging a known working monitor after BIOS has loaded has symptoms of point 1, that is VGA is not detected by games, but something like VGALEMMI.COM and wfw3.11 still work...

Here is the output of "vmode tell". It is same for both "HP vs17-mode" and "normal-mode". This is output is normal right?

1Display type(s): Variable-frequency
2Emulation mode: VGA
3Crt mode: 03
4Chip info: ET4000/W32
5Board info: setup-1
6VMODE version: version 2.0
7VMODE date: 06/08/94
8Translation ROM: Not present
9640x480 modes: 72 Hz
10800x600 modes: 60 Hz
111024x768 modes: Intl

MSD.EXE tells me a pretty different story in video-tab (and yeah, I don't have VESA tsr running):

"Normal-mode":

1Video Adapter Type: VGA
2Manufacturer: Compaq
3Model: <blank>
4Display Type: VGA Color
5Video Mode: 3
6Number of Columns: 80
7Number of Rows: 25
8Video BIOS Version: <blank>
9Video BIOS Date: 06/17/93
10VESA Support Installed: No
11Secondary Adapter: None

"HP vs17-mode"

1Video Adapter Type: None
2Manufacturer: Compaq
3Model: <blank>
4Display Type: <blank>
5Video Mode: 3 <what? This is VGA!!>
6Number of Columns: 80
7Number of Rows: 25
8Video BIOS Version: <blank>
9Video BIOS Date: 06/17/93
10VESA Support Installed: No
11Secondary Adapter: None

Go figure... I guess I'll invest in a new LCD monitor (HP vs17's colours are very bad anyway, the late Sony Trinitron produced much better colours in its heyday...)

If anyone knows, how does VGA detection work in DOS? I have a very rudimentary understanding of programming (if playing with LOGO and BASIC, with a touch of C++ and Python did much...)

EDIT:
By the way, the IBM 9515 is done for. I tested it and used for it 13 hours, and then the screen faded, and the power light started to flash on and fade out continuously, each time with a nasty crackle coming from something inside the monitor. My guess is something in its power supply is cooked, or was it me? I'll... somehow explain this to my friend (but he's long lost the IBM PS/2 56 it came with.)

Tseng uses a different scheme to detect the monitor type than most other manufacturers, they actually check the ID pins that were defined by XGA. Did you maybe use a different cable on the non-working monitor after the incident with the fixed-frequency monitor? Maybe you could make/get an adapter that only passes the vital signals on the connector.

I always thought VGA was hot-pluggable as long as you do it with the monitor powered off, and the video card set to a resolution that is supported by the monitor(s) being attached. Better/newer monitors will even protect you from incompatible resolutions.

Maybe the fixed-frequency monitor toasted part of the ET4000 chip that games are querying for VGA support?

You could also try clearing the motherboard CMOS and reconfiguring the BIOS from scratch. Maybe some hidden setting got activated to try to help it cope with that XGA-2 monitor?

Thank you very much guys, for your insights.

@h-a-l-9000; good point. I haven't actually checked the pins of the monitors, I'll do so tonight. I seem to remember both the generic and the XGA-2's pinouts missing several pins, but am not 100% sure. I used the same standard KVM VGA cable with the HP LCD monitor and Sony Bravia (all straight through) so obviously it has something to do with HP LCD's internal wiring. And also, in general, DDC pinout does not affect ID pinout.

I don't have any spare VGA cables to make into a breakout so I'm going to test this by grounding the pins on the VGA connector directly (wish me luck, and don't try this if you don't know what you're doing.)

I'm starting to get less concerned about hotplugging damage since I was playing the thing just fine with the generic monitor (which is slightly out of focus when warmed up.) I did try hotswapping the 9515 connector (whilst 9515 was powered down) to the Radeon 7500 previously and the computer it was running in ground to a halt (with funny screen colours) so probably hotplugging some cards is not a good idea. But Radeon continued to work fine after power cycle.

I have yet to work out a mechanism to "detect" VGA, will get there when I have the time. IRT HunterZ, I did clear the BIOS by lifting "password clear" jumper. This disconnects battery soldered on motherboard. It had no effect on the VGA, but did clear all CMOS settings (computer needed to remember its memory size again, and floppy parameters, and time and date.)

I read that some of these fixed-frequency monitors were made "from the best names" so I hope it had some level of protection (but it does try scanning unsupported rates.) I don't know what this one went through, but it died after 13 hours...

There were warnings that fixed-frequency monitors may display wrong rates seemingly correct but overheat and die after a while. However your description with the crackling sound seems more the high-voltage unit having broken isolation which is completely 'legit' to happen on any CRT monitor. Eizo (best name?) monitors died like flies that way...

This sounds quite a lot like what my old 286 used to do back in the day, some twenty years ago - some programs wouldn't be able to find the right video mode until certain other specific programs were run. I have no idea if the problem was dependent on the monitor, though.

Ok guys, sorry for going AWOL without completing this topic. I've been pretty short of time last week.

The culprit is definitely the stupid HP monitor. As far as I can tell, from even a DDC capable monitor (Sony Trinitron 200ES and Sony Bravia), Pin 11 (ID0) on the VGA should always be grounded. This (at least on the ET4000) enables VGA colour. The HP monitor's Pin 11 seems to be floating. Even the DDC ground was not working properly. A monitor from 2005 should not be like this. Great, what a lemon.

My friend offered his Visiontek Radeon x1300 card to test. That one caused the whole system to fail to boot when used with the HP LCD. Case solved, and sorry for all the hype.

On the other hand, got any recommendations for a good monitor? I am currently using my trusty old Sony Trinitron 200ES but it takes up a lot of desk space...

PS: the IBM XGA-2 had another weird pinout. It was missing pin 11 (ID0) but had pin 4 (ID2) grounded, which is not in any of the VGA ID combinations...

XGA has its own definitions of the ID pins.

You could maybe find a way to ground those pins so that you can use the HP monitor. Probably the safest way would be to disconnect the monitor's connection to the pins (if any) and then route the connector pins to ground. That way the monitor would think they're still floating and the video card would think they're grounded.