Next is a bit more of a common card, I think.

It's an AST Rampage PC. This card has the full 2MB installed, but I cannot find the correct jumper settings or manuals or drivers or anything. The closest is the AST Rampage/2, but it only has one jumper - E2, which you set to tell the card to reconfigure with the driver.
This has "E1" "E2" and a block labelled B one through four.
My first guess is that "B1" through "B4" are for setting the card's I/O base address, but it could also be for the EMS window, too. I don't know what E1 and E2 are for.

Jumper settings would be helpful, I want to use this board in my XT when the time comes, I want to run Windows 3.0 on it with EMS, and this card having the full 2MB would be great for that. I hear some of these Rampage cards can map UMBs, too, so I would like to free up as much conventional memory as possible with it. A manual would be amazing.
Does this card need any special software or are there generic EMS drivers I can use?

It's part number 202138-002 8 if that helps. I'm going to guess that the "8" is how many banks. It also says "<1M>," but i figure that is because it looks like this board came with 1MB from the factory - The first 4 banks are GoldStar GM71256-12 ICs and they all have similar manufacture dates to the card. The second 4 banks are Oki M41256A-10s.

A label on the back of the card reads "RAMP PC-1000 S/N 4061027 MADE IN HONG KONG BY AST RESEARCH (FE) LTD"

Here are some pictures of the card:

And finally, last but by no means the least...

This unusual Cheetah card. I bought it in-box on eBay, but it came with no manuals or disks or any information. The box doesn't even give any information on the card itself, just phone numbers and the Cheetah logo.

On the card, it says "Cheetah Card FAST 640K/2.5M PC-AT Memory"
There are 5 rows of 18 DRAMs each. My quick math says that with all of these filled with 4164 DRAMs, you would get 640K, and if you fill them with 41256 DRAMs, you will get 2.5MB.
Mine is filled (potentially from the factory) with Mitsubishi MN41256-10 DRAMs. I have the full 2.5MB. Pretty sure this is an XMS board.

This card has tons of switches and jumpers - I really do not want to go into this card blind like I've read some people have. Maybe it's fine if your switch block has 6 switches and 2 jumpers, or if you have 10 switches total... But I do not want to jump into an XMS card with 6 banks of 8 switches each, 2 jumpers, and 2x 4-position jumpers.
All is not lost right now, since some of these are labelled, albeit cryptically...
J1, 2 pins - "0 Wait States"
J2, 5 pins - "cas"
J3, 5 pins - "mux"
J4, 3 pins - "pen"
SW1 - "BANK 1"
SW2 - "BANK 2"
SW3 - "BANK 3"
SW4 - "BANK 4"
SW5 - "BANK 5"
SW6 - "mode"

The edge of the card says "Copyright (C) 1985 Cheetah International Inc."
I cannot find Cheetah on TH99, and all of the instances of Cheetah in magazines I can find on Google are of accelerator cards.

The box:

The card:

The switches:

The last one you will get running with the method here
ISA XMS/EMS Memory Extension / Expansion cards: Now Running without Driver / Documentation :-)

For the 2nd one you most likely need a configuration software as its jumperless.

Predator99 wrote:

The last one you will get running with the method here
ISA XMS/EMS Memory Extension / Expansion cards: Now Running without Driver / Documentation :-)

For the 2nd one you most likely need a configuration software as its jumperless.

Yes, I read that thread. Like I said, there are SIX BANKS OF EIGHT SWITCHES. There are 256 possible positions for each.
There are 4294967296*65536 possible combinations for JUST the switches. I cannot use this method, it will take me until the end of time.

One of the switchbanks is labelled "mode" and is probably not even used to map the memory - There is no way to tell which combination will put it into XMS.

And this is before looking at the 3 jumpers with nonsensical labels, "cas" "mux" and "pen"!

I know I need software for it, but none of the AST drivers I can find online list compatibility with that part number. I mentioned this in the post.

Remove all RAM except bank 5 (will not conflict with base RAM) and start with the "bank 5"-DIP.

Would also not a bad idea to take a look what the current configuration is? Put it in a 286 nad start testx...

Predator99 wrote:

Remove all RAM except bank 5 (will not conflict with base RAM) and start with the "bank 5"-DIP.

Would also not a bad idea to take a look what the current configuration is? Put it in a 286 nad start testx...

Installing it as it is in my 286 yeilds zero results. Nothing on BIOS memory count, and nothing in my XMS viewer. There is 0KB of XMS, my EMS driver does not detect it, and the board still reports 512KB of conventional. It basically does not exist as it sits.

Did you read the thread? What is the output of testx?

Predator99 wrote:

Did you read the thread? What is the output of testx?

I read the thread 2 hours ago, I have been at work at work, I have another 5 hours before I will have the chance to even try it.

Did you read my post? There are 48 switches and and 3 jumpers to set. 2^48 settings + boot time + several minutes for each run of testx = Days, weeks, months of flipping switches. I would rather not you tell me "just do it at random" like I said I did not want to do in my initial post. There has to be some kind of documentation out there for this card.

Another XMS viewing program I have (some generic diagnostics software, I would have to look at the name) displays the same information, address space 1M to 16M, and there was nothing. Totally empty. All Fs. I know it works, because with my AST Advantage! 128 card, I get 640K of XMS at the beginning of the address space and the BIOS will detect it. With both the AST and the Cheetah installed, I get 640K of XMS at the beginning and the rest is displayed as empty. It also does not backfill conventional or UMBs because Microsoft Diagnostics reports UMBs to have open address space and the motherboard only finds 512K of RAM.

What do you think how many combinations I tested before decoding the jumpers on my cards? Not 2^48. Maybe 20. Other people also did it that way. If you like to do it different or keep searching for the manual - good luck.
Frankly speaking, when reading what you did so far I have not the impression you know what you are doing and where the problem is...

the cheetah card looks impressive. 😀

Predator99 wrote:

What do you think how many combinations I tested before decoding the jumpers on my cards? Not 2^48. Maybe 20. Other people also did it that way. If you like to do it different or keep searching for the manual - good luck.
Frankly speaking, when reading what you did so far I have not the impression you know what you are doing and where the problem is...

There are 48 switches. That's where 2^48 comes from. I do not want to sit and play with switches the rest of my natural lifetime. If there is no documentation, I cannot just set them outright to be the correct pattern.

There is a bank of 8 switches labelled "mode." The card currently does absolutely nothing when installed in the machine it is to go into. That means that if it is not this combination, then it must be one of the other 255 combinations for the mode switch. That, or it could be one of the three cryptic jumpers. Or it could be any amount of the switches labeled "Bank" - Of which there are 40 to set.

Because I do not have a manual or any other information on this board, I am going to have to go through other combinations until I find one that works - The only jumper I can be sure is set right is the wait state jumper.
Because there are 48 switches with two positions each, that brings us to a total of 2^48 combinations possible across all 6 banks of switches. Again, this does not include either of the 4 position jumpers, nor the 3 pin jumper at the bottom of the board.

Claiming I do not know what I am doing is incredibly rude when I'm telling you I am not going to spend the next few weeks flipping switches on this card blindly until something happens. I have also already read the information you linked me to and I have also already made an attempt at this method

The thread you linked - Which I had already read before posting this thread, mind you - Is people blindly flipping switches because they cannot find manuals or other documentation for their boards. There is some method to it, I admit, as they can watch the memory window move in testx - but nothing appears in my XMS space with this board installed as it is.

Again, it does not backfill conventional memory. It does not map to UMBs. It is not detected by a generic EMS driver. The machine does not recognize any XMS memory, and an XMS viewer I have already tried does not show any memory in the XMS address space. This means that, in it's current configuration, this card is not functional at all. The method described in the thread relies on some kind of working configuration to be able to see the memory in the XMS address space and simply shift it until it is in a sensible position - This card is not showing up. Period. Which means I'm going to have to at least change the mode switches or one of those three jumpers. There are 255 more combinations of switch positions for the mode switch alone.

Do not be rude when you cannot (or simply will not) read the information I have already provided.

Warlord wrote:

the cheetah card looks impressive. 😀

Certainly is, beyond just the 2.5MB of RAM it would give me (which I believe it came out of the factory with!), it's full size!

I'm on a quest to fill my AT with full size cards. This would make 3 of 4 cards installed full size, and once I get my full length VGA card, I will have 100% full size cards in this machine. Ought to look impressive there, too!

...

What "XMS viewer" do you use? It will most likely be better than the testx I am using, it performs a random testing and it takes ages until a scan of 16 mb is completed...

Predator99 wrote:

...

What "XMS viewer" do you use? It will most likely be better than the testx I am using, it performs a random testing and it takes ages until a scan of 16 mb is completed...

I'd have to look. I just got home from work and I have errands to run yet.

This takes no less time, especially because you have no control over where it will scan - It will scan all of it.

I'm sure if you wanted to, you could write a program in C or something to scan the address bus quite quickly relatively, bring it down to just a minute or two.

EDIT: Now that I think about it some more... I may try to write a program to do this. Compiled, a program to check if XMS is present would be pretty quick compared to thorough testing of the RAM that may or may not be there. In fact, basic testing might be pretty quick that way, too. Depends on how far you would want to go with it.

Here is the combo card version of that Cheetah card:
https://books.google.com/books?id=vcqI6sY2e-k … %20card&f=false

Not having any luck finding a manual or driver for either though.

What speed is the CPU in the system you are trying to run it in? At least according to that web page, with 70ns memory, the max is maybe 12Mhz. Since yours is 100ns, I would guess the max CPU speed would be even less.

Edit: Here is a page that mentions the manual and the programs that came with it:
https://books.google.com/books?id=pDGnxFyejN4 … ettings&f=false

Scroll up a few pages to read more about that card.

cyclone3d wrote:

Here is the combo card version of that Cheetah card: https://books.google.com/books?id=vcqI6sY2e-k … %20card&f=false […]
Show full quote

Here is the combo card version of that Cheetah card:
https://books.google.com/books?id=vcqI6sY2e-k … %20card&f=false

Not having any luck finding a manual or driver for either though.

What speed is the CPU in the system you are trying to run it in? At least according to that web page, with 70ns memory, the max is maybe 12Mhz. Since yours is 100ns, I would guess the max CPU speed would be even less.

Edit: Here is a page that mentions the manual and the programs that came with it:
https://books.google.com/books?id=pDGnxFyejN4 … ettings&f=false

Scroll up a few pages to read more about that card.

The CPU is at 20MHz, but the chipset is at 16MHz, and the 287 is at 12MHz. The ISA bus is running at the stock 8MHz.

Apparently, this must be fine, because the 1Mbit DRAMs on the motherboard are 100ns as well, and the 64Kbit DRAMs on my AST Advantage are 150ns and 120ns, looking at spec sheets.

Predator99 wrote:

What "XMS viewer" do you use? It will most likely be better than the testx I am using, it performs a random testing and it takes ages until a scan of 16 mb is completed...

Just checked, it's NSSI in the XMS address window. If you tell it you have 16MB of XMS in BIOS, it will scan it all and show it to you in a window... With much nicer colors than TestX.

Regardless, here are the results using TestX with both the AST Advantage 128 and the Cheetah card. The AST has 128+512+512, and the first 128 is being used to backfill the motherboard to 640K, leaving 1024K for XMS.

As you can see, no other XMS is reported, like I was saying earlier. If the Cheetah card was interfering in some way with the AST card, removing the AST would cause it to detect XMS on boot (it detects none without the AST), and the XMS window in reported in both of these programs would be larger than the 1MB that the AST has for XMS. It would be 2.5MB.

Just for shits and giggles, I pulled the AST out and ran it again just to confirm what I'm saying.

Here's the BIOS screen and TestX results:

Yes I also tried nssi, but I found testx more useful and reliable, dont remember the reason.

So only explanation is that the card is currently configured as EMS and doesnt show any XMS. You need to switch to XMS-mode. Most likely this is only one DIP to be found.
Can you switch all DIPs from the "Mode - SW6" in the opposite direction and try again?

Did you notice the Bank 1-5 DIPs are all different?
B1 11x111xx
B2 11x11xxx
B3 11x1x1xx
B4 11x1xxxx
B5 11xx11xx
This makes sense as they are used to set the starting adress of each bank individually. But maybe one bit is also used for XMS/EMS selection. I would let them untouched for the moment.