Just started building my new 65C02 project (which appears in various threads on this forum ). As this subject seems to cause glassy-eyed syndrome in my friends and work colleagues, I thought I'd share my project here .

So far:
- I've attached the main wirewrap sockets
- Soldered the smoothing caps on the underside for each IC and added some larger caps (1x 33uF tant and 1x 1000uF electrolytic)
- Soldered pull up resistors for the CPU pins that need it
- Put in a star ground and star VDD plane to a central 14 pin wire wrap socket
- Implemented dual size fitting for SRAM - standard DIL plus another line for adapter with 12ns IC.
- Added Power regulation
- Wired up the power VDD and VSS/GND pins of the sockets
- Wired up the address and data buses
- Wired up VIA PA and PB external lines for both ICs to pin headers
- Created the first decoder GAL
- Fitted a D type flipflop to allow for better phi2 drive capability
- Implemented Reset and NMI buttons

To-do:
- Create the second decoder GAL for expanded (paged) memory banks and IRQ handling
- Wire up the control bus
- Write some firmware for the EPROM(!)

Later to-do:
- Add RS232
- Add RTC with battery backup

Here's a picy:

Looks good - given the special arrangement for the super fast SRAM, I assume you're going for speed. How fast do the GALs go?

Edit: found a datasheet

The GALs are rated at maximum 15ns pin to pin prop delay - so quite fast. The thing which will really slow things down is the EEPROM which is 45ns. I might see if I can wangle a ROM emulator in or the two speed setup that various people have mentioned.

Looking good...

I'm also in the middle of a new 65C02 SBC project. I'm looking at the Atmel ATF750C/CL chips... low power version is 15ns prop delay and normal version is 7.5ns. These are considered a superset of the 22V10 GALs, according to Atmel. Maybe something to look at. How fast are you planning to run the CPU?

EPROM emulation: One thing to consider is using the RTC and associated backup to keep an external SRAM alive and use this as your ROM. The Benchmarq BQ4845 can manage an external SRAM to maintain data integrity during power on/off. It also states CEin to CEout with less than 12ns prop delay, so it might work depending on the rest of the system timing.

http://www.mouser.com/ds/2/405/bq4845-404348.pdf

_________________
Regards, KM
https://github.com/floobydust

You might be better served by looking at the ATF1504AS instead of the ATF750. The former has something like twice the resources in a PLCC-44 package. Performance on the 1504 is 10ns pin-to-pin.


Or perhaps shadowing the EPROM into RAM and then mapping out the EPROM. A single wait-state per EPROM access would be necessary if the intention is to run the system above 14 MHz. BTW, I can run POC V1.1 at 15 MHz without the SCSI host adapter, and that is with a 55ns EPROM.

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

I was hoping to get "as fast as possible1" (:D) out of it, but realistically I was hoping for 12-14MHz.

Those Atmels...? What're you planning to program them with?

Hi BDD,

Yea, I looked at the larger Atmel CPLD... very nice, much more capable, but my programmer won't handle it, but then again, it "looks" like the ATF750 will support what I need, which is pretty simple glue logic, using the CLK and R/W to validate proper MR and MW signals, plus handle the EEPROM select and 6- I/O selects, each 32 bytes wide.

To Banedon,

I have a Dataman 40-Pro programmer which supports a pretty wide range of chips, sadly not the 1504 BDD mentioned, but for my current project the ATF750 "should" be fine. I'm making a single small (3.8"x2.5") 4-layer board which uses PLCC chips mostly and provides a 65C02, 32KB RAM, 32KB ROM, a 65C22 VIA, NXP SCC2691 with FTDI USB interface and an expansion connector with 4- I/O selects. Target clock is 8MHz. I'm also planning to use the NXP SCC2691 for the console UART, to avoid future potential problems with 65C51 availability and fully functional chips.

Nowadays, I'm too old and lazy to do wire-wrap. so I'm looking to do a 4-layer PCB up front.

_________________
Regards, KM
https://github.com/floobydust

Are you sure you want to use the 2691? That device is a lot like me: old and slow. Perhaps you would be better served by the NXP 28L91, which is the single channel version of the 28L92 that I am currently running in POC V1.1.

_________________
x86?  We ain't got no x86.  We don't NEED no stinking x86!

I've scoured UARTs till I'm blue... I'm looking to use a PLCC28 size UART, which for single channel, should be the norm. I really don't have enough space for another PLCC44 size chip. I only need a single UART for console and Xmodem loading. If I had the space, I'd likely use the 28L92 and add a single row header for the second port. I did manage to get a small quantity of PLCC 2691 chips so am hoping to fire one up soon... also have the 24-pin DIP version too. Just need more time.

_________________
Regards, KM
https://github.com/floobydust

Nice project banedon. It also showed me ,that what I was looking for for a long time really exists, a fast 128kbyte SRAM, I found even a source for the AS7C1024-20TPC which is a 300mil PDIP-32 package.

The SCC2691 is slow, although it works fine. Slow means the bus interface will works with a 6502 clocked with 5MHz but above YMMV. I myself had no luck. Another problem is that baud-rates above 19200 are somewhat "hidden" (BDD provided a document in one of the threads that explains how "hidden" they are). I like them because they are available in PDIP-24, a real space saver.

As for the ATF1504AS, they have two programming modes one is using a programmer and then you have 36 I/O pins for the PLCC/TQFP-44 package. However if you only need 32 I/O pins then a rather cheap (60USD) ATDH1501USB-K is all you need to program them via the JTAG interface. That's how I'm using them as you can see in http://forum.6502.org/viewtopic.php?f=10&t=3276. TQFP-44 with 0.8mm spacing is very easy to solder (sufficient solder flux is your friend) and as you can see in the thread I have even found some PCB adapters to convert a TQFP-44 into a DIP-48 which is breadboard friendly. The breadboard version of my current 6502 project runs at 11MHz using a ATF1504AS-10AU44 that has (almost) the complete GLUE logic. Soon I will convert it into a vero-board version and check how fast it will work.

Good lord! I'd need to sell a kdney or two to fund a Dataman 40-Pro

Keep in mind that if you already have SPI or I²C, you can save a lot of board space on certain things that don't have to be as fast as the less-than-single-cycle parallel memory accesses. The 14-pin-DIP MAX3100 UART is nice, having SPI and 8-byte transmit and receive buffers. The MAX3110 is the same thing with the RS-232 line drivers and receivers built in, so you don't need an additional IC for them. There are I²C RTCs in 8-pin DIP or SOIC, and other small things you might need.

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

Going to look at saving for a decent CPLD programmer after I pay off the two new bits of equipment that I've just bought (purchased for upcoming 6502 setups):
Tenma 72-10495 (dual PSU, 30V @ 5A max each) - already arrived - and a Rigol DS1102E 100MHz 'scope (arriving tomorrow) . I'm now officially poor .

Here's the latest pic of my project. Everything except the memory bank select latch system and the ACIA are hooked up. All lines and buses look as they should (no odd voltage levels), so I've removed both the oscillator and the D-type flip flop clock/PHI2 driver and am now starting to single step test to see if the right values are being read from the EEPROM (both are sitting unconnected off on the right side of the board).


BTW that green LED only has approx. 3mA going into it and it's remarkably bright! Whole project so far (in quiescent state) consumes 120mA.

Hoping you wouldn't sell a kidney that cheap... someone would take up on it Good test equipment is key... I've been buying it for decades cause I'm old. Picked up a Saleae16-channel logic analyzer some time ago. This was highly useful in finding the recent W65C51 Xmit bug. Also have a few scopes, many analogue meters, signal generators, a distortion analyzer and more.

At 120ma of power, I'm guessing the GAL chips are taking most of that. Atmel are making some lower powered 22V10 chips which might help, if it's an issue. Also, did you solder into the wirewrap socket for components? I found some nice DIP header pins that plug into the IC socket and have heavy duty pins that can accept wirewrap. Saves a socket

I measured the current draw for my small 65C02 system at idle, which has a RTC IRQ loop running at 250 times per second. The board draws just over 35ma, with 4ma being a pair of LED indicators.

My 65C02 board setup is:
WDC 65C02, WDC 65C22, WDC 65C51 (one that works!!)
32KB Alliance RAM/70ns
32KB Atmel EEPROM/150ns
74HC00, 74HC30, 74HC138
10.00MHz half-size can osc. for 65C02
3.6864MHz half-size can osc. for 65C51 (38.4K baud)
2- Red LEDS for power indication
- 65C51 links to a FTDI DB9-USB-5V module

The next plan is to replace the 3- 74HC logic chips with a single Atmel AT750CL and give up two I/O selects.

_________________
Regards, KM
https://github.com/floobydust

Most of the remaining 31mA is undoubtedly going to the oscillator cans. My first 'c02 design took 2mA for the entire computer, including the LCD, but it just used an RC hung directly on the processor for generating the clock, at 170kHz.

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?