Hi all,

Here, and searches on the net reveals tons of 6502 stuff, but I was wondering if anyone had any designs or small implementations of the 65c816 cpu - specifically when it comes to dealing with the shared data/address bus?

I've searched a bit, but haven't found much I can work with (either no circuit info, info from apple with regards to the IIgs, or in another language ) - does anyone have any good references they could point me to so I can get started with some basic designs with this chip?

Thanks in advance!!

-T

Tancor,

Check out Figure 5-1 (page 46) from the WDC '816 datasheet. It gives a simple circuit to split the adress and data info from the bus.

I have not found any actual schematics but there are a few builders on the 6502.org site who have built a '816 system. Have you tried to contact them for help?

I just got my '816 from the bulk order and plan to incorportate it into my next design.

Good luck!!

Daryl

For lower speeds, all you need is the 74AC573 transparent latch that latches the data bus's logic levels onto the high address byte when phase 2 goes up. Very simple. As Daryl said, the WDC data sheet shows this circuit. This will work fine at 4MHz, but the timings are violated badly if you try to do the same thing at 14MHz. I talked to Bill Mensch (the 65816's designer) about this a couple of years ago, and he acknowledged that the circuit shown was not suitable for the higher speeds. I haven't done it yet, but I think you'll need variable delay lines to move the latching edge forward and back from the rising phase-2 edge to find where you can operate the fastest.

Remember however that if you don't specifically need more that 64K address space, you don't have to latch or use the high address byte at all. You can totally ignore it and still get tons of other benefits from the '816. In fact, you can make a board that uses a few option-select jumpers to make it able to use either an '02 or an '816 in the processor socket.

Garth

8bit:
There was only one project that I saw that the linked was good for in terms of the 816 in the links page - the creator said on his site that he might put up schematics and such when he had time - the sites been up for a while, and as it hasn't changed in the about year that I've been following it, I'm guessing he's just too busy, so, I really didn't want to harrass him.

Garth:
Thanks for the info, didn't realize it would be that easy. I still have a lot of reading to do on 6502 and 65816 interfacing - some of it is still new to me. In terms of address space, I'm not sure how much I will ultimately need, but I like knowing what I can get before I get involved with something.

Also - did that guy make any suggestions as to how to create the circuit so that it would work at the higher speeds?

and, have you ever tried the 6502 at a higher speed? the data sheet talks about 8MHz and 14MHz, and I was wondering if you had ever tried it?

> Also - did that guy make any suggestions as to how to create the circuit
> so that it would work at the higher speeds?

No, he just said to go ahead with the delay-line idea.


> and, have you ever tried the 6502 at a higher speed? the data sheet
> talks about 8MHz and 14MHz, and I was wondering if you had ever tried
> it?

My current workbench computer starts having problems just over 7MHz, so I run it at 5MHz. It has 70ns ROM, 55ns SRAM, a 6MHz 65c802, and everything else (3 65c22's and 3 65c51's) is 4MHz. I have PLCC parts to go much faster, but the ROM speed will get in the way. So my next workbench computer won't have any ROM. Instead, it will load the BIOS/OS/whatever from a tiny serial EEPROM into fast RAM before boot-up.

Garth

This looks like a nasty limitation preventing 65C816 use at high speeds. I'm not sure I feel too comfortable with some delay-line solution either. I see the PLCC package doesn't have the address-multiplexing issue the 40-pin DIP does, so now I understand another reason why the PLCC version is so much better at higher clock rates.

My chip collection includes Rockwell 65C02s at 10MHz ratings, but I don't see 65C22s around rated higher than 6MHz and none on the market above 4MHz. Looks like we'd need to go with WDC parts here too even for use with faster Rockwell CPUs.

I have no doubt that it'll work just fine, since others have done it. Bill Mensch said one of their customers is using the off-the-shelf 65816's at 20MHz, but he didn't know how they were taking care of the timing of the decoding of the high 8 address bits.

I think the only nasty limitation is that the manufacturer-supplied data is not as complete as we would like it. Remember that their main business is licensing IP, and they might have quit the game a long time ago if it were only a matter of selling hardware. Timing info is primarily tied to hardware and specific geometries and manufacturing processes, not to the HDL information that is licensed to the companies that are still putting out about two hundred million 65c02 cores this year alone in their dedicated microcontrollers. In about 1987 when we showed interest in the 68000 processors for something at work, Motorola sent us not only data books, but a pile of ap notes about an inch and a half high! I guess we're slowly developing the 6500 ap. notes here on 6502.org.

The PLCC with 44 pins still multiplexes the 8 high address bits (bank address) the same as the 40-pin DIP, but it has three extra ground pins and one extra power supply pin. This, in addition to shorter internal connections with the smaller package, reduce the most critical inductances to allow faster edge rates and higher-speed operation.

I think California Micro Devices' 65c22's were rated to 6MHz, but they just discontinued hundreds of parts last year, including the 5 or 6 that were in the 6500 family.

Garth

Well this just leads me to another topic I haven't found much on. Once again, I apologize if this has already been discussed here and I've simply missed it.

Are there any microcontrollers using 6502/816 cores generic enough to be of hobbyist or general product-development interest? Something besides extremely narrowly targeted items? Or does everyone using them construct one-off designs?

I'm not thinking of general-pupose computer CPUs, but things more like PICs.

Rockwell and CMD have discontinued their 65-based general-purpose microcontrollers, and Mitsubishi didn't seem to be interested in selling in the US anymore when we were designing a product ten years ago where theirs would have been nice. WDC's 65134 (based on the '02) and 65265 (based on the '816) are really nice-- especially the latter. The '134 has 68 leads, and the '265 has 84.

First the negatives:
-> rather expensive, at somewhere around $12 and $25 each
-> The hobbyist would have to have his program in a separate (E)EPROM unless you want to pay for mask programming
-> no onboard A/D converters (although the '265 does have D/A to use with the signal generators)
-> not available with lots of options

'134 has
-> 20 interrupt vectors, plus BRK and RST.
-> 192 bytes of RAM (you can connect more externally)
-> 4KB of ROM containing ready-made software to run various part of the I/O, if I understand it right
-> 16MB segmented memory space
-> UART
-> serial interface bus (SIB)
-> 4 16-bit counter/timers
-> watchdog timer
-> 8 decoded chip select outputs
-> real-time clock features
-> connecting up to 64KB external memory still leaves 4 parallel ports free


The '265, without losing any of the above features except the SIB, offers more:
-> 8KB ROM and 576 bytes of RAM onboard
-> 2 tone generators
-> 25 interrupt vectors, plus RST, BRK, ABORT, and COProcessor interrupts
-> 4 UARTs
-> parallel interface bus (PIB) (but no SIB)
-> 8 16-bit counter/timers
-> I/O is mostly in the 00DF00-00DFFF address range.

This is what I understand from my WDC data book. I'll welcome any corrections or clarifications from anyone who has actual experience with these microcontrollers. (Wally??)

I've been using Mitsubishi cored chips. I'v even done a version of EhBASIC that allows you direct access to the onboard I/O and timers.

If the site is up (thank you b!**£y Lycos!) you can see it here ..

http://members.lycos.co.uk/leeedavison/6502/suprchips/

There's not a huge ammount of detail yet but I'm working on it.

Lee.

Nope, not up yet

Looking forward to reading it

Hi Everyone,

The ST7 family from ST Semi (formerly SGS Thomson) has
a programmers model that is very
similar to the 6502. You might
leverage your 6502 experience with
it. Note, I haven't used the ST7
family myself so I can't make any
claims either-way about their usefulness.

http://www.stmcu.com/modules.php?name=Downloads

Cheers,

Paul

It's back up as of this morning and I've updated the version of EhBASIC for the Mitsubishi 3806.

Somewhat OT. There is an unconventional picture of me up as well.

There's a rather obscure link to it from the home page.

It's not for the faint-hearted.

Lee.

Could you not use a higher speed xtal and divide it to produce a system clock and an out of phase trigger for the address latch?

> Could you not use a higher speed xtal and divide it to produce a system
> clock and an out of phase trigger for the address latch?

I apologize for the delay in responding. This computer has been down and in and out of the shop for the last two weeks. If I had known what it was all going to turn into, I would have just bought another computer.

I suppose you could do what you're saying, but that higher speed would have to be really, really high, since getting the maximum performance from your '816 system would require being able to put that latching edge forward or back in steps no bigger than a few nanoseconds each until you find that optimum point. Once that point is found, you could calculate the initial clock frequency (which could be a few hundred MHz) and then implement the appropriate high-speed counters.

Garth