65XX SBC general help and color display help needed

[backspace119]

Also, I stopped by my fiancee's parents place today and picked up the oscope, I've not tested it yet or even read any info on it, but it looks very capable, much like the unit I had in a lab in high school. Also, while I was there, he lent me a manual on a Ms. Pac-Man machine, that was based on a z80. I've found a schematic in it that I've attached below, and I believe I've isolated the video section, possibly useful for when I want to add video? Or is it completely useless in today's world?

This is the full schematic (sorry for low quality, my phone wouldn't focus on it correctly)

This is the video section (notice Red Green and Blue outputs to the right)

[Arlet]

It would be worth it to have the schematic scanned/photographed properly.

Edit: There's a better scan in the PDF manual here:

http://www.fixyourpacman.com/pacman-manuals/

Edit2: there's a custom chip generating video RAM signals.

[backspace119]

Arlet wrote:

It would be worth it to have the schematic scanned/photographed properly.

Edit: There's a better scan in the PDF manual here:

http://www.fixyourpacman.com/pacman-manuals/

Edit2: there's a custom chip generating video RAM signals.

I saw that chip, I hadn't studied it enough to know it was for video signals, so I assume then that I'll need to just stick to the solution that I had talked about earlier, and figure it out. Still though, it's an interesting schematic, and there may be a few useful parts of it

EDIT: Oh I guess it says right there on the schematic.

[backspace119]

Arlet wrote:

It would be worth it to have the schematic scanned/photographed properly.

Edit: There's a better scan in the PDF manual here:

http://www.fixyourpacman.com/pacman-manuals/

Edit2: there's a custom chip generating video RAM signals.

Sorry to quote again but didn't want to make another edit to my post. On a site I'd already found from googling some of the chips I found this which offers boards which are discrete component copies of the custom ICs.

Probably not all that useful in the long run, but in case anyone else wants to see about using part of this schematic for something. I'll probably stick with the video chip that was mentioned earlier in the thread.

[backspace119]

So, the very next page of the manual has the custom vram addresser circuit laid out, here it is (scanned this time):

EDIT:

Here's the actual schematic, it's on the next page,

[GARTHWILSON]

backspace119 wrote:

So, I've started picking out a few parts since I've decided on the base build requirements, I've found this ram which looks plenty fast and is 32KB, I suppose if I want more I can always add more chips, or try and find larger sizes that are also this fast (a cursory look came up with no results on mouser).

Does this ram look like it will work out alright? Looking at the data sheet everything seems to be ok, I'll just be forced into 5v since I don't want to use a level shifter for ram. I've decided to get the WDC processors, since they're on mouser and they're well documented, same with the IO chips.

I wouldn't use that one, because it is not guaranteed to be able to pull its outputs up to a valid CMOS logic '1'. It has a TTL-level output, specifying only 2.4V for a logic '1'. Look for one that can pull up close to 5V with a light load. This will probably result in a lower operating current too. The one you have there would be ok for the 65816 where you would have the bus transceiver anyway; but if you use an '816, you might as well use a much bigger SRAM. I can give you a 512Kx8 (16 times as much memory) for the same price you see there for the 32Kx8, except it will be in a 36-pin SOJ package which you might not like for your PCB milling machine. It's what I use for my memory modules shown on the front page of my site. (It will also be twice as fast, but that apparently is irrelevant for you.) If you keep looking, you should be able to find a 512Kx8 in a DIP and 5V with adequate speed.

[backspace119]

GARTHWILSON wrote:

backspace119 wrote:

So, I've started picking out a few parts since I've decided on the base build requirements, I've found this ram which looks plenty fast and is 32KB, I suppose if I want more I can always add more chips, or try and find larger sizes that are also this fast (a cursory look came up with no results on mouser).

Does this ram look like it will work out alright? Looking at the data sheet everything seems to be ok, I'll just be forced into 5v since I don't want to use a level shifter for ram. I've decided to get the WDC processors, since they're on mouser and they're well documented, same with the IO chips.

I wouldn't use that one, because it is not guaranteed to be able to pull its outputs up to a valid CMOS logic '1'. It has a TTL-level output, specifying only 2.4V for a logic '1'. Look for one that can pull up close to 5V with a light load. This will probably result in a lower operating current too. The one you have there would be ok for the 65816 where you would have the bus transceiver anyway; but if you use an '816, you might as well use a much bigger SRAM. I can give you a 512Kx8 (16 times as much memory) for the same price you see there for the 32Kx8, except it will be in a 36-pin SOJ package which you might not like for your PCB milling machine. It's what I use for my memory modules shown on the front page of my site. (It will also be twice as fast, but that apparently is irrelevant for you.) If you keep looking, you should be able to find a 512Kx8 in a DIP and 5V with adequate speed.

Alright, the machine can do SMT stuff, just not very reliably, so I would prefer to stay with through hole components if possible. I'll see if I can find one like that. I was thinking of starting with 32k so I wouldn't have to get the top 8 address bits from the data pins, but if it's better to just start out with them in use and use larger memory than I'll go that direction. When you say the bus tranceiver, are you talking about the latching mechanism mentioned earlier? and 8 bit latch with RDY and !phi2 as it's enable?

Sorry if I'm overlapping questions here that have already been answered, trying to soak in as much information as I can with each one.

[backspace119]

GARTHWILSON wrote:

I wouldn't use that one, because it is not guaranteed to be able to pull its outputs up to a valid CMOS logic '1'. It has a TTL-level output, specifying only 2.4V for a logic '1'. Look for one that can pull up close to 5V with a light load. This will probably result in a lower operating current too. The one you have there would be ok for the 65816 where you would have the bus transceiver anyway; but if you use an '816, you might as well use a much bigger SRAM. I can give you a 512Kx8 (16 times as much memory) for the same price you see there for the 32Kx8, except it will be in a 36-pin SOJ package which you might not like for your PCB milling machine. It's what I use for my memory modules shown on the front page of my site. (It will also be twice as fast, but that apparently is irrelevant for you.) If you keep looking, you should be able to find a 512Kx8 in a DIP and 5V with adequate speed.

I may be reading the datasheets wrong (I'm looking at output min voltage) but finding chips that have a 5v output seems to be rather difficult for through hole components, especially for anything that's fast (most of what I'm seeing is about 50ns, that 12ns one I posted earlier was actually the fastest I found in through hole)

[drogon]

backspace119 wrote:

GARTHWILSON wrote:

I wouldn't use that one, because it is not guaranteed to be able to pull its outputs up to a valid CMOS logic '1'. It has a TTL-level output, specifying only 2.4V for a logic '1'. Look for one that can pull up close to 5V with a light load. This will probably result in a lower operating current too. The one you have there would be ok for the 65816 where you would have the bus transceiver anyway; but if you use an '816, you might as well use a much bigger SRAM. I can give you a 512Kx8 (16 times as much memory) for the same price you see there for the 32Kx8, except it will be in a 36-pin SOJ package which you might not like for your PCB milling machine. It's what I use for my memory modules shown on the front page of my site. (It will also be twice as fast, but that apparently is irrelevant for you.) If you keep looking, you should be able to find a 512Kx8 in a DIP and 5V with adequate speed.

I may be reading the datasheets wrong (I'm looking at output min voltage) but finding chips that have a 5v output seems to be rather difficult for through hole components, especially for anything that's fast (most of what I'm seeing is about 50ns, that 12ns one I posted earlier was actually the fastest I found in through hole)

FWIW: That's the exact same RAM chips I'm using in my 65C02 system. Oddly enough I thought I'd gone for the 10ns variant, but I have that one - the 20ns version. It's working fine at 16Mhz, but as I keep thinking; I got lucky. I only got 2 just to test, but they worked fine on the breadboard, stripboard and my PCB version. The 2 of mine are vertically stacked with the chip-select on the top one brought out.

https://unicorn.drogon.net/IMG_20181115_164041.jpg

-Gordon

[GARTHWILSON]

backspace119 wrote:

I was thinking of starting with 32k so I wouldn't have to get the top 8 address bits from the data pins

If you use a SRAM that's any bigger than the memory-map space you want to give it, you can tie the extra address line(s) high or low. For example, IIRC, there's no 64KB SRAM (it goes from 32KB to 128KB to 512KB); so if you wanted more than 32K (for example 48K), you could use a 128KB and one of the address bits could just be tied high or low to get you your 64KB SRAM.

Quote:

When you say the bus transceiver, are you talking about the latching mechanism mentioned earlier?

See page 46 of the '816 data sheet at http://6502.org/documents/datasheets/wd ... 1_2018.pdf . You're thinking of the '573 or '373. The bus transceiver is the '245 below the '573 or '373. The '816 does not need it for drive strength; it's there to prevent bus contention. (It's still not the perfect solution. We talked about this in the topic "Managing the 65816 multiplexed bus," with what I think is a pretty good solution.)

[backspace119]

drogon wrote:

I may be reading the datasheets wrong (I'm looking at output min voltage) but finding chips that have a 5v output seems to be rather difficult for through hole components, especially for anything that's fast (most of what I'm seeing is about 50ns, that 12ns one I posted earlier was actually the fastest I found in through hole)

FWIW: That's the exact same RAM chips I'm using in my 65C02 system. Oddly enough I thought I'd gone for the 10ns variant, but I have that one - the 20ns version. It's working fine at 16Mhz, but as I keep thinking; I got lucky. I only got 2 just to test, but they worked fine on the breadboard, stripboard and my PCB version. The 2 of mine are vertically stacked with the chip-select on the top one brought out.

https://unicorn.drogon.net/IMG_20181115_164041.jpg

-Gordon

After some quick math ((16m/1)/2) * 1B it comes out to 31.25ns of time for accessing memory at 16Mhz if we only have half the clock to do the access in (someone please correct me if my math is wrong here) so I think that 12ns or even 20ns is just fine as long as the glue logic is fast enough.

[backspace119]

GARTHWILSON wrote:

backspace119 wrote:

I was thinking of starting with 32k so I wouldn't have to get the top 8 address bits from the data pins

If you use a SRAM that's any bigger than the memory-map space you want to give it, you can tie the extra address line(s) high or low. For example, IIRC, there's no 64KB SRAM (it goes from 32KB to 128KB to 512KB); so if you wanted more than 32K (for example 48K), you could use a 128KB and one of the address bits could just be tied high or low to get you your 64KB SRAM.

Quote:

When you say the bus transceiver, are you talking about the latching mechanism mentioned earlier?

See page 46 of the '816 data sheet at http://6502.org/documents/datasheets/wd ... 1_2018.pdf . You're thinking of the '573 or '373. The bus transceiver is the '245 below the '573 or '373. The '816 does not need it for drive strength; it's there to prevent bus contention. (It's still not the perfect solution. We talked about this in the topic "Managing the 65816 multiplexed bus," with what I think is a pretty good solution.)

wow I had actually pulled up that exact thread when I was googling about the transceiver, I didn't read all of it yet though, so I'll definitely use it as a reference.

Any thoughts on finding RAM that hits 5v well? or can I opt for ram that'll work at 3.3v and run everything at that level?

[Arlet]

How about running everything at 3.3V instead ?

[backspace119]

Arlet wrote:

How about running everything at 3.3V instead ?

This may be the best route, and use a level shifter for any devices that need 5v (and maintain 3.3v for all critical components that a level shifter may cause undesired delay on). I know that the video chip I looked at at least needs a few voltages, I think it needs 5v for logic and 12v for video driving, but that won't be in the first build, so I can figure it out later (will probably need a simple power supply circuit)

[Arlet]

In some cases, you don't even need level shifters. Compare the datasheets for output levels and input thresholds. A 3.3V device that outputs close to 3.3V will work well with most 5V inputs, for example. And there are also many 3.3V devices that are tolerant for 5V inputs.

And if you do need level shifting, often the simplest solution is just to insert a suitable buffer that does have appropriate output levels and/or input threshold voltages.

Another advantage of 3.3V is that it leaves the door open for more advanced video, based on programmable logic.