How do you use a multi-slot expander (motherboard)?

[Schlowski]

Only Switch 3 hard wired...

Björg

[saundby]

Something to note when using multicart expansions is that it's pretty easy to shut down the Vic by having too many carts pulling on the expansion port lines at once. Some of those lines are driven directly by the 6502, and they're only good for one TTL load--anything more will result in soft failures, if it's only overloaded a bit (occasional errors and such), and a heavy overload will keep the Vic from booting normally. This is particularly true of the simple boards that put all carts on at once, but any unbuffered boards will have this problem when several carts are switched on at once.

See the discussion on buffering the expansion port for the solution to this.

Basically the rule of thumb is that you can count on the unbuffered port driving any one cart (when properly designed), possibly two, and perhaps three but never expect it to drive more than three without buffering, and even then be prepared for odd errors and hiccups with three and possibly two, especially if one of those two is a multi-purpose cart (e.g. a cart that has RAM, ROM, and video would count as 3 carts all by itself.)

-Mark G.

[Boray]

I've never had any problems like that with the vic-1020, but it takes power directly from the PSU...

[saundby]

It's not a power supply problem, actually, it's a problem with the logic signal levels. The 6502 (and the other chips that drive lines) have a limited "fan-out" of how many other chips they can drive. The 6502 itself can drive one TTL chip input with one of its outputs. Put any more chips on that 6502 output line and you're likely to start getting odd errors, add some more and the 6502 won't be able to put a readable signal on the line.

This isn't affected by the power supply itself. You can have several cards that each draw only a few milliamps and come nowhere near stressing the Vic's power supply, but the various logic inputs that the 6502 is trying to drive simultaneously on different ICs in the different carts will cause problems even though the chips are getting plenty of power on the power supply leads.

-Mark G.

[Boray]

There are many people here that uses expansion port expanders all the time and I have never heard of any problems like that. I have always 4 carts plugged in. Sometimes 5.

Tell me, if it is impossible to drive more than one cart on the expansion port, then why did Commodore make these expanders? Or are you only talking about some simple no-name third party expanders that doesn't have any logic chips on them?

[6502dude]

Multiple devices may be on the address & data bus at the same time if they are tri-state devices (RAM, ROM, EPROM).

When /CS and /OE enable are not selected the cartridge it does not present a load to be driven by the processor.

Presumably, with dip switches on cartridge expander, /CS and /OE are only made active on one cartridge at a time.

As cables are used to extend expander further from the processor, capacitive affect on data and address bus will cause pulses to round off. Also, these cables are subject to (and cause of) RF interference. Using driver chips (74xx245) will help reshape TTL pulses, overcome minor loss in cabling, and allow more devices to be driven at same time.

Reason for Commodore not to use buffer chips in their multi-cart expanders = poor design and/or cost control.

I would include driver chips if building expander module from scratch, even if expansion cable was not used.

[Boray]

Reason for Commodore not to use buffer chips in their multi-cart expanders = poor design and/or cost control.

Which Commodore expander models doesn't use driver chips?

/Anders

[6502dude]

Which Commodore expander models doesn't use driver chips?

I don't know. I have never seen a Commodore expansion unit.

My comment was on speculation why Commodore (or any other vendor, that matter) would opt not to include buffer / driver chips for an expansion module.

[Boray]

The vic-1005 (probably the most sold Commodore expander (3-ports)) has two 74LS244 chips.

Looking down one of the slots of my vic-1020 (6 port expander), I can see 5 chips and a lot of other components. And the vic-1020 also takes power directly from the PSU.

/Anders

[Jogi]

Hi,

the 1020 Schematic.

http://www.cbmhardware.de/vc20/vic1020.gif


Regards VIC20


Ups Sorry

[Boray]

I was about to edit that picture into a link (because it's so big and messes with the page formatting), but realized I'm in the Hardware forum where I don't have these rights...

[Boray]

Thanks for the edit!

[saundby]

There are many people here that uses expansion port expanders all the time and I have never heard of any problems like that. I have always 4 carts plugged in. Sometimes 5.

Tell me, if it is impossible to drive more than one cart on the expansion port, then why did Commodore make these expanders? Or are you only talking about some simple no-name third party expanders that doesn't have any logic chips on them?

It's not impossible to drive more than one cart, clearly, and I didn't say so. However, a port expander that doesn't buffer the signals while allowing more than one active cart at a time is asking for trouble. Why do they do it? Because:
It's a heck of a lot cheaper (the boss says the connectors already cost too much.
If something goes wrong the user probably won't blame it on the board being loaded with a bunch of carts.
If the user does realize it, they'll just kick themselves for not buying one of the uptown models, or just pull a cart and try to remember not to stack the ones that seem to cause problems (typically the power-hungry ones.)

Anyway, I think the point's been well covered. Building one with buffering or building an inline buffer board for one without buffering is pretty simple, as electronics projects go. Use regular TTL chips if your soldering skills aren't the greatest.

-Mark G.

[moggep]

Threadlift of the Century!

I just bought "Special Cartridge by audio computers", which (in my case) is a switched 32K expansion. It also has a passthrough connector for an extra cartridge, four dip-switches to control memory and a ROM installed.

When expirmenting with the dipswitches, only no 2 and 3 seems to Control the amount of memory (i tried all 16 combinations):
2-on 3-off = 19967 bytes free
2 on 3-on = 28159 bytes free
2 off 3 off = 3583 bytes free

Switch no 1 and 4 seems to have no impact on memory config.

But my major question is about the ROM that is installed. It has a handwritten sticker that says "Screen ROM". Anyone knows what it is, and how to activate it?

[Mike]

Threadlift of the Century!

I was forewarned by your post in Forum64.

[...] my major question is about the ROM that is installed. It has a handwritten sticker that says "Screen ROM". Anyone knows what it is, and how to activate it?

Please take a look at the first program, EASYDUMP, in this thread.

The EPROM should be located in BLK5, but may not always be visible, depending on how the DIP-switches (especially 1 and 4) are set. The program stores a memory dump (5 corresponds to the range $A000..$BFFF, where most cartridge ROMs are located) on floppy, or SD-card with a SD2IEC, etc.

If you can provide the ROM image I will take a look at the code.