I just tested this Cirrus Logic CL-GD510/520 based card again. This time, I altered the dip-switches on the back, so that dip-switch #6 was set to Off. This means that the card is in CGA mode. I've never tried this before with an old video card, so it was really great to see the monitor display its text and graphics in an "old n crusty" chunky CGA resolution. I was using an LCD, and it looked fine. I tried some random "CGA Comp" tests, and they all worked. Previously, with the card in its default VGA mode, some of these tests failed.

However, I tried running several games, and when the game is supposed to start, the monitor unexpectedly and abruptly "aborts" and then switches over to the image displayed from my other computer. Is this because the CGA game's signal is "out of range"? There is no error message to tell me this, or what the problem is. I wonder if there is a DOS TSR that I need to run, to set the refresh rate to 60Hz? Or perhaps I just need to run these games on an old CRT based monitor, and not an LCD?

Edit:
Or, perhaps I need to alter the dip-switches that relate to the monitor type. The choices are:

MONITOR TYPE
Type SW1/1 SW1/2 SW1/3 SW1/4
PS/2 Off On On Off
Monochrome On On On On
Color Off On On On
Enhanced color On Off On On
TTL multi-frequency Off Off On On
Analog multi-frequency Off On On Off

Edit 2: Correction. Sorry, I have made a mistake. Earlier, I put the "TTL setting" above in a dark blue colour. I meant to put the Analog multi-frequency in dark blue. End Edit 2.

Currently, the card is set to use Analog multi-frequency, which is highlighted above in dark blue. Perhaps I need to alter this setting, and change it to something else, for example the Color setting?

What LCD monitor are you using? Does it have SCART inputs perchance?

Monochrome, CGA and EGA, differ from VGA in that the latter is analog and all the others are digital TTL, in addition, they all have different horizontal sync rates. CGA is 15kHz and EGA is about 22kHz, VGA is 30kHz.

I believe the way it works is, when set to monochrome, CGA or EGA it will stick to the one frequency no matter what - so, when you set to EGA mode, and try to run a CGA app it will still be using the EGA horizontal sync - this is because the old monitors were all only designed for one sync rate. Correct me if I'm wrong 😀

I'm not sure what's happening with the TTL multi-frequency to allow it to show up on your LCD but I'm guessing it must initially run at a high enough sync rate for your monitor to pick up. Try disconnecting the other computer and see if your monitor throws up a message, probably 'out of range' or something, as when you try to run an EGA or CGA game the sync drops too low for t he monitor to pick it up.

The 'color' and 'enhanced' jumper settings are for EGA and CGA respectively. Analog multi-frequency is VGA. See if your monitor works with either of those settings, if it doesn't then I'd say that what I've described above is the issue.

This is a big problem with Amigas as they work on a 15kHz ANALOG signal, not compatible with VGA, there are some modern LCD's that will work with them (usually with SCART) so I'd be interested to know what monitor you have 😀

Correction. Sorry, I have made a mistake. Earlier, I put the "TTL setting" in my post above in a dark blue colour. I meant to put the Analog multi-frequency in dark blue. I have edited my post above, to reflect this update.

The monitor I am using is: 19" FLATRON L1960TR, by LG. On the back, it has 2 monitor type connectors: DSUB and DVI-D. When I ran the video card in dip-switch #6 mode (CGA), the monitor automatically ran at 70Hz. I am familiar with SCART sockets, as they appear on appliances such as TVs, DVDs, and satellite set top boxes. Please note that there is no such "SCART" socket on the back of the monitor.

Tomorrow, I will carry out your suggested tests. Thanks a lot for those ideas!

that makes more sense then 😀

What happens if you try one of the TTL modes? I'd expect the monitor to lose sync. It won't cause any damage to a modern LCD (at least it didn't to mine when I tried it 😀 ) as they are smart enough to drop the signal rather than try to display something that is impossible for them - but old CRT's are often not so forgiving!

> are often not so forgiving!

Got anything to back up that rumor?

I tried setting the video card's dip-swtich monitor type setting to color mode, and also to enhanced color mode, but when I powered on the 386 mobo, the monitor displayed an "out of range" error. However, the TTL multi-frequency mode works OK, just like the Analog multi-frequency mode.

When I run games, it seems that some work, and some do not. Eg Round42, Janitor Joe and Flightmare works, but Striker does not. I'm sure that Striker is a CGA game. It was released in 1986. Please note that I haven't yet tried disconnecting the DVI-D monitor cable that is used for my other machine.

h-a-l-9000 wrote:

> are often not so forgiving!

Got anything to back up that rumor?

not me personally but the word on the street is to be very careful with TTL monitors because if you feed them the wrong type of input you can blow components... unless they are multisync.

have you evidence to the contrary?

I'm not going to sacrifice my only EGA monitor 😀

DonutKing wrote:

This is a big problem with Amigas as they work on a 15kHz ANALOG signal, not compatible with VGA, there are some modern LCD's that will work with them (usually with SCART) so I'd be interested to know what monitor you have 😀

That's why you buy a good upscaler:

http://junkerhq.net/xrgb/index.php/XRGB-3

I'm actually waiting for the Indivision AGA mk2 to come out next year, so I can get a native VGA output on my A1200 - plus it will include Graffiti chunky framebuffer emulation, which is used by the Shapeshifter mac emulator 😀

An upscaler will make ALL my consoles look purdy though! 😉

I'm curious as to why the old 1986 DOS game called "Striker" does not display correctly, when I set my CL-GD510/520 based video card to CGA via its DIP switch settings. If I set the video card back to VGA via its DIP switch settings, it works OK, so there's nothing wrong with the rest of the system, generally speaking. When the vid card is in CGA mode, the game does not display its main menu. The screen goes blank, with no error message. I'm using an LCD.

Does anyone have a CGA monitor, and also any CGA capable video card? It's not difficult to locate this freeware game on the net, if someone could please try and run it on CGA hardware. Thanks a lot if possible.

DonutKing wrote:

This is a big problem with Amigas as they work on a 15kHz ANALOG signal, not compatible with VGA, there are some modern LCD's that will work with them (usually with SCART) so I'd be interested to know what monitor you have 😀

I use for an Amiga a Targa TM1480 Multisync Monitor with digital/analog switch. There is a Adapter from CBM that maps the Amigas Video Out to a VGA connector and then I use a cable from an old Samsung Multisync monitor that has VGA on one end and 9 pin DSUB on the other.

Alternativly I use my VGA Grabber card. Since it detects the signal as interlaced I have to back-convert the signal on the fly using DScaler with filter Old Game. I set the PCs graphics cards output to the TFT or LCD-TV to 50 Hz then.

Wow... So much to comment on...

First, the CL-GD510/520 card.
Thanks rg100 for testing the card in CGA emulation mode. So it passes all the CGAcomp tests? Looks like this card has potential to be the holy grail for hardware CGA emulation! 😉

Second, the "Striker" game.
From first glance, it doesn't seem to use any weird CGA tricks, as it looks about right on a standard VGA. The incompatibility with the CL-GD510/520 in CGA mode could just be a specific problem with this game.

Third, modern displays, upscalers and old multisync monitors.

The simplest way to get CGA RGB output onto a modern TV or LCD would be a cheap CGA to SCART adapter. There are some problems with this kind of adapter however: You need a display with SCART or 15kHz RGB inputs, color 6 (brown) will be displayed as dark yellow and you have to rely on the scaling abilities of your display, which are probably not that good.

The external upscaling boxes surely look nice, but has anyone actually used one of these with PC CGA/EGA games? I wouldn't expect the cheap ones to work much better than the usual crap built into TVs, and I'm reluctant to sink hundreds of Euros into an expensive one, which rather seem to be designed for old consoles/arcade setups.

Elianda's awesome VGA grabber is another possibility, but what about input lag?

That leaves the old multisync CRT monitors. I've been looking for one of these since my Sony Trinitron CPD-1302 died, which was ages ago. So far I've been under the impression that finding a working multisync monitor is even more difficult than finding a real CGA monitor.

BTW, I'm still looking for CGAComp test results, including the cards that have already been tested, just to confirm the results.

5u3 wrote:

BTW, I'm still looking for CGAComp test results, including the cards that have already been tested, just to confirm the results.

Thanks a lot for your additional information. I have just rerun the CGA compatibility tester program, with the 16-bit ISA CL-GD510/520 video card. I put the card in to CGA emulation mode, using its DIP switch settings. I ran all of the tests that appear in 5u3's test results table on page 1 of this thread, and all of the relevant tests passed, with these extra notes:

* The Vertical retrace detect test was unhappy with my LCD running at 70Hz, and printed a message saying that 70Hz was too far away from CGA's default 60Hz value. This is to be expected.

* The Horizontal retrace detect test showed the coloured bars going down to where the bottom of the "testing text" is displayed, but it went no further. However, on the reference video shown on the oldskool website, this demo (at 3:19 minutes) shows the coloured bars moving past this "testing text", and more towards the bottom of the screen. I'm unsure about the validity my result - if it's a problem or not. I'm inclined to suspect that it may be a problem, but I don't know for sure.

* The display positioning test showed the screen attempting to reposition itself to the centre of the screen, and I think this test worked OK. Sometimes it was hard to determine what this test was trying to achieve. My LCD frequently needs to have its "magic recentering button" pressed, because its display is not automatically centered correctly anyway!

* The start address reprogramming test looked a little bit odd. When the test pattern scrolled upwards, the text was chopped in to 2 sections. Unfortunately, I couldn't see this test performed on the oldskool reference video, so I was unable to check if this test was performing correctly. The next test, which was the sideways scrolling test looked OK because the text was not broken up in to 2 parts, but kept together as a whole image. If any of my tests could be considered a failure, then it may be this vertical scrolling test, because of the way in which the text was chopped in to 2 pieces.

Edit: For the test mentioned directly above, when I use the phrase "chopped in to 2 pieces", I really meant to say that the text was "wrapped around" the screen. Here's an example, below. Imagine that the "test text" consists of the line of digits: 123456789. I can't remember what the actual test text was when I was running this test, so I'll use this "9 digits string" to illustrate my point. When this test text of 123456789 was scrolling vertically, sometimes it would appear normally, as:

123456789

But other times, it was "wrapped around" the screen, like this:

567891234

Each time the test text had finished scrolling upwards, the next time the test text would scroll again, the "wraparound position" would alter, and so another time it would change and look something like this:

789123456

Thanks for running all these tests again! I have updated the table on the first page.

retro games 100 wrote:

* The Vertical retrace detect test was unhappy with my LCD running at 70Hz, and printed a message saying that 70Hz was too far away from CGA's default 60Hz value. This is to be expected.

Um, actually it's the other way round. The card fails this test, because it doesn't run at 60 Hz in CGA mode like it's supposed to. The LCD monitor has no say in this, it can either sync to the signal, or not (that's when monitors display their distressed OUT OF SYNC message).

retro games 100 wrote:

* The Horizontal retrace detect test showed the coloured bars going down to where the bottom of the "testing text" is displayed, but it went no further.

This is interesting, didn't notice that before. It's the same with my cards, the copper bars only move within the upper half of the screen.
Considering the way these bars are generated, I'm pretty sure it's because of double scan (the mechanism that doubles the 200 lines of CGA to 400 lines, so it can be displayed by VGA monitors).

retro games 100 wrote:

* The display positioning test showed the screen attempting to reposition itself to the centre of the screen, and I think this test worked OK. Sometimes it was hard to determine what this test was trying to achieve. My LCD frequently needs to have its "magic recentering button" pressed, because its display is not automatically centered correctly anyway!

Yep, that's normal. I've never seen a LCD that was not annoying in that respect.

retro games 100 wrote:

* The start address reprogramming test looked a little bit odd. When the test pattern scrolled upwards, the text was chopped in to 2 sections. Unfortunately, I couldn't see this test performed on the oldskool reference video, so I was unable to check if this test was performing correctly. The next test, which was the sideways scrolling test looked OK because the text was not broken up in to 2 parts, but kept together as a whole image. If any of my tests could be considered a failure, then it may be this vertical scrolling test, because of the way in which the text was chopped in to 2 pieces.

I'm not sure about this either. My guess was that the "chopped up" horizontal alignment was normal, and the things to look out for were empty or obviously corrupt sections in the lower parts of the scroll. This usually happens after two pages have scrolled through and the third one begins. After this point, all my cards just scroll through empty screens when in VGA mode. With the CGA emulation turned on, the Trident fills that space with white and pink vertical stripes. The Tseng scrolls through four and one empty pages, then abruptly jumps to the first page. The WD card scrolls through five pages, smoothly wrapping around to the first one, giving the appearance of an endless scroller. I believe that is what the test is trying to achieve.

Regarding the vertical retrace detect test, ah what a pity it won't pass that test, otherwise it would pass everything with flying colours. Extremely vaguely related tidbit of info: when I run anything in real DOS, such as booting in to DOS 6, the monitor always defaults to 70Hz, with a resolution of 720x400. I guess that refresh rate & resolution is the closest match that it has available.

Regarding the start address reprogramming test, the scrolling did appear to be smooth, without any empty or obviously corrupt sections, or any unexpected jumps or gaps. This test did give the appearance of "endless scrolling".

Very minor observation: for the 510/520 column in your table of results on page 1 of this thread, I believe that the [9] symbol can be removed, because there were no problems with that particular test. I wonder if that [9] symbol has accidentally "slipped over" from another adjacent table column, in to the 510/520 column? Thanks a lot for double-checking this.

retro games 100 wrote:

Regarding the vertical retrace detect test, ah what a pity it won't pass that test, otherwise it would pass everything with flying colours. Extremely vaguely related tidbit of info: when I run anything in real DOS, such as booting in to DOS 6, the monitor always defaults to 70Hz, with a resolution of 720x400. I guess that refresh rate & resolution is the closest match that it has available.

Usually the refresh rate info given by a monitors OSD is accurate, but the resolution is often detected wrong.
It seems you are stuck with 70 Hz on this card as long as you have a VGA monitor connected and the DIP-switches set to one of the "multisync" positions.

retro games 100 wrote:

Regarding the start address reprogramming test, the scrolling did appear to be smooth, without any empty or obviously corrupt sections, or any unexpected jumps or gaps. This test did give the appearance of "endless scrolling".

Let's write it down as "ok" for now, and correct it later if it turns out that the pages are supposed to be aligned in the vertical scrolling test.

retro games 100 wrote:

Very minor observation: for the 510/520 column in your table of results on page 1 of this thread, I believe that the [9] symbol can be removed, because there were no problems with that particular test. I wonder if that [9] symbol has accidentally "slipped over" from another adjacent table column, in to the 510/520 column? Thanks a lot for double-checking this.

Thanks for pointing out the error. The symbol was still left from Vlasks test, which was done in the "VGA" setting.
Also, the ET4000's result for the start adress programming test is corrected to "fail", because of the missing last page in the vertical scroll (did not notice this detail during my first test).

I wonder if there is a test to determine EGA compatibility, for testing old VGA cards? If not, are there any EGA games that are known to "break" VGA cards?

Edit: I found an old thread on Vogons, here. In it, Vogons user h-a-l-9000 mentions that -

VGA is *mostly* EGA compatible 😉
There are some games that have issues, but most work.

Incompatibilities I know about are:
- panning handling and timing (flickering screen)
- IRQ2 not activated on VGA (Gauntlet hangs)
- double scan register incompatibility (screen layout messed up)

Of my cards, only the Tseng ET4000 and the Trident 8900C offer an EGA compatibility mode.
Gauntlet hangs with a black screen on the ET4000, but the Trident actually manages to display something. Unfortunately the screen is "chopped up", so that the status bar runs right trough the middle of the screen. 😒

For testing EGA cards I can recommend two scene demos:
Megademo by The Space Pigs, which is one of the earliest demos on the PC, and Mega-EGA, written by a programmer who unfortunately was still stuck with an EGA machine in 1994. 🤣
The latter demo brought out a flaw in both my cards: The EGA emulation runs at 70 Hz instead of 60, which screws up the timing of the demo.

Thanks a lot! 😀 I can't wait to test them. I've downloaded them, and because my retro hardware is temporarily packed away for a short period of time, I tried both demos on DosBox - just to check them out! I set the DosBox graphics configuration to = ega, but neither demo worked. Curious and very interesting...