hard1k wrote:

In addition to the previous discussion about 16Mb of onboard RAM:
IIRC, I've heard some reports about rare compatibility issues when using 16Mb with some applications. Something like "the software is unaware of such amount of installed GUS RAM leading to errors and malfunction". If anybody can confirm this statement, then we'd better choose the SIMM72 variant allowing the end user to install less RAM if he encounters any such problems.

Maybe the best way to do it is put the memory on the board but add jumpers to disable it? Like, having 16MB + 512KB onboard, and having jumpers to do the following:

• Enable the 512KB mapped to banks 0 and 1
• Enable 8MB of RAM mapped to banks 0 and 1
• Enable 8MB of RAM mapped to banks 2 and 3

That way you could easily switch between 512KB only, 512KB + 8MB and full 16MB, and you'd just have to be careful to not enable the 512KB chip and the first two 4MB banks at the same time.

Good idea, btw!

Awesome project! Can't wait to see what come of this.

alexanrs wrote:

Maybe the best way to do it is put the memory on the board but add jumpers to disable it? Like, having 16MB + 512KB onboard, and […]
Show full quote

hard1k wrote:

In addition to the previous discussion about 16Mb of onboard RAM:
IIRC, I've heard some reports about rare compatibility issues when using 16Mb with some applications. Something like "the software is unaware of such amount of installed GUS RAM leading to errors and malfunction". If anybody can confirm this statement, then we'd better choose the SIMM72 variant allowing the end user to install less RAM if he encounters any such problems.

Maybe the best way to do it is put the memory on the board but add jumpers to disable it? Like, having 16MB + 512KB onboard, and having jumpers to do the following:

• Enable the 512KB mapped to banks 0 and 1
• Enable 8MB of RAM mapped to banks 0 and 1
• Enable 8MB of RAM mapped to banks 2 and 3

That way you could easily switch between 512KB only, 512KB + 8MB and full 16MB, and you'd just have to be careful to not enable the 512KB chip and the first two 4MB banks at the same time.

One megabyte is the optimum for GUS.

awgamer wrote:

One megabyte is the optimum for GUS.

I am not saying that you are wrong, but some explanation concerning that affirmation seems to be necessary.

Synoptic wrote:

awgamer wrote:

One megabyte is the optimum for GUS.

I am not saying that you are wrong, but some explanation concerning that affirmation seems to be necessary.

Unfamiliar with the GUS I see. If the people doing such a project needed an explanation for such basic info on what they're working on..

I assume it is because the classic GUS could have up to 1MB of RAM, and thus any GUS-enabled software should be able to deal with that. With that in mind I'd replace the 512 onboard RAM from my idea with an 1MB one, and keep the option to go up to 16MB.

Let's not talk this to pieces.
Primary aim is to replicate the Philips card - to get a 512kb GUS PnP ... all the rest are secondary targets and/or options.

You baffle me arguing for doing a cut down version of the GUS.

Well for now shock__ is doing the basic stuff. I think it's more than enough. If he decides to share the sch, anyone is free to make their own changes. I just think it's a bless to have a gus pnp clone with 512kb of ram . It's good for those who didn't have the chance to experiment a GUS back in the old days. 😀

awgamer wrote:

One megabyte is the optimum for GUS.

The Man knows what he says 😁

hard1k wrote:

http://quotes.lifehack.org/media/quotes/quote-Bill-Gates-640k-ought-to-be-enough-for-anybody-89027.png […]
Show full quote

awgamer wrote:

One megabyte is the optimum for GUS.

I've been following this interesting topic and if it helps development, I would like to offer one of my GUS PnP cards to help keep things moving. I do not have the time or skills to help create the new card, but maybe I can contribute to the project in this way.
In my opinion, the card should come with (at least) 1MB RAM as standard.

Yesterday I made some attempts to find GUS PnP cards and it seems I've found one dead beyond repair in St. Petersburg!
Current owner is willing to donate it to the project free of charge, but:
1. He won't deal with shipping, it'll be my part.
2. I have to get it from him somehow (from SPb to Moscow).
3. The owner says the card has taken part in "a donor program" leaving it as almost bare PCB. But I guess even a naked PCB will help with some certain aspects of the project 😀

So don't rush to solder other GUS PnP that's still alive and I'll keep you informed.

awgamer wrote:

You baffle me arguing for doing a cut down version of the GUS.

You baffle me asking for features without having contributed anything, if you read the thread I've written what the short- and longterm goals for this project are (1MB on board, 16MB via SIMM, 1 or 4MB ROM, optional SMD and DIP footprints for various parts in short - all this might still change as this project unfolds) 😀 While the redrawing of the Philips card is closing in on being finished this project is still in a pretty early phase.
So yeah, I could design a fully featured monster GUS, but what would be the point if I can't confirm the basic design as working? DOSBox didn't start out with Win98 support either - milestones being the keyword.
To calm your worries a little ... another project I did (RAM expander for the Atari ST) was updated to support 256MB as a maximum compared to 64MB on the original card along with various other small improvements.

CHiLL72: Wrote you a PM.

hard1k: great stuff 😀

General news: Seems like this is the perfect time to get a hot air gun ... didn't find anyone who could lend me one for a day or two and having such a tool will make soldering the 160pin QFP a lot easier.

Maybe it is enough to blow out everything from GUS pcb and make a high res scan in SPb?

I guess that having the PCB on hand would allow to understand the ROM pinout (both SMD and DIP) much more easily.

BTW, for great justice, it was Hyperborean who gave me a clue where to look for this dead GUS PnP 😀

GUS PnP is a 3 or 4 layer design, with the inner layers appearing to contain more than just GND/VCC so having physical access would be better - but for the sake of the project, hi-res scans might work as well (I'd most likely "copy" the RAM portions of the GUS PnP which seems to be fully on the upper/lower layers).

Hyperborean's name will be added to the PCB as well 😁

EDIT:
.gif showing some worksteps so far - http://i.imgur.com/nPyIefw.gif

Holy cow! A friend of mine (Igor aka White, who holds the ModLabs.net website) also responded to my appeal, he can donate his fully working GUS PnP to the project as well.
Just in case you need one extra, let me know 😀

I'd say get it in any case 😉
Waiting for CHiLL72 to reply to my PM right now, if it works out you can keep that GUS PnP for yourself, otherwise I might be interested (also in the "waste" PnP from SPb).

Kinda goes with my experience, when it comes to cloning/redeveloping hardware (I've really dug into ZX Spectrum clones in the past) there are no better supporters than folks from Russia 😀