It's been several years but UTSource supplied 'good' parts compared to some of the scratched-up 'used', 'pulls', and 'refurbs' you might see from other vendors back then.

BTW, speaking of parts... I was pleasantly surprised to receive five (5) W65C22S6TPG-14 chips from China last week ($10.99 total, including shipping) which looked very nice and tested good at 8-MHz on my new SBC + Programmer prototype (see below). It was very exciting (for me) to be able to run Nick Gammon's G-Pascal package at 8-MHz with bit-banged 38,400 bps serial after replacing the Rockwell chips on the board with WDC chips.

I used to recommend UTSource but after a few recent setbacks and price hike after ordering, I don't recommend it anymore, but I still order for myself. UTSource is still a cut above eBay and their refund policy is excellent. Don't buy W65C02 from them because they are re-labelled 6502, some of them are not even CMOS. They may run to 4MHz, but not 14MHz.
Bill

Hi all,

Apologies for the delay in getting back to this post,
So many responses where to start,

I'll start with Chads reply,

If one wants to make the lowest cost project / kit, whether for sale or personal use, then old parts are probably the way to go....with some risk of receiving dud parts. New, WDC, parts from Mouser are not cheap. One will require new parts to run at higher speeds.

I managed to get away with "strandard" parts like the 6264 (8kbyte) / 62256 (32kbye) (55ns) and Microchip SST39SF010/020/040. I saw no point in allowing for things like 27Cxxx EPROM's and am not a fan of the 28 series EEPROM's. I did not see this as a requirement.

I find the whole Windows vs Linux thing interesting, but that a topic for another day. I'm a Windows user....
I think CPLD's, specifically the (MAX7000)EPM7128 / Microchip AT1508, are the way to go.
Forget Wincupl, Use the Altera (Intel) tools and draw schematics and simulate. I can't program HDL/VHDL....on my leaning list.
The CPLD above is actually quite limited. Just enough for basic VGA, complex address decoding with latches for high address etc.
Much more efficient than breadboarding and fewer mistakes in PCB layout. EPM7128 is obsolete, but low cost and still available. Microchip AT1508 is more expensive and still in production. If anybody wants to know more just ask.

PS2 keyboards are becoming hard to buy in stores in Australia.

Agree, need to have video either VGA, TMS9918 or similar.

Your logic is sound as everyone has 20 of these lying around their house, but personally I'm not a fan. Other replies have made valid comments on this.
Personally I would prefer to regulate down to the required voltage on a PCB rather than feed it in from an external source.
My design will end up requiring around 18V as I want 15V to be able to power some A-to-D/D-to-A parts (on plugin cards) and audio amplifier requires 10-22V.

I was not planning solder it your self as I have used SMT for the resistors / capacitors and power supply.

The price of a board is tricky. Has anybody calculated the cost of WDC parts? (Let's use Mouser 1 piece price)
W65C22N6TPG-14 USD10.5
W65C02S6TPG-14 USD10.75
AS6C6264-55PCN USD4.03 (8kbyte)
AS6C62256-55PCN USD4.10 (32kbyte)
AS6C4008-55PCN USD6.38 (512kbyte)
SST39SF010A-70-4C-PHE USD2.92 (128kbyte)
SST39SF040-70-4C-PHE USD3.91 (512kbyte)

Depending on your design you are already at USD30.
In my design I have 1 x 62256, 3 x 6264, 2 x SST39SF040A and 1 x AS6C4008-55PCN, but I could drop one SST39SF040A and the AS6C4008-55PCN leaving it to the buyer to buy later.
I thought of having the PCB assembled in China (I already have several projects in production in China.) and shipping it without the major IC's.
I would fit tested CPLD's and AY-3-8910. Why ship WDC part from USA to Australia and back again. Make it an option, with or without major parts.
All parts socketed of course.
(Another topic we could spend hours on is how do the Chinese make certain things so cheap. I often see items on Ebay / Aliexpress where the parts from Mouser cost more than the item.)

I guess like many of you here I wanted to build my own. That was part of the challenge. I've never bought kits even when it came to 8051, AVR, ARM etc. I would just study the datasheet and make a PCB.
As an electronics engineer getting started was easier. If I were a programmer or someone with no real hardware/software background then I would probably have started with a kit and progressed from there.
The choice of kit would come down to what I could afford.
I would probably buy an assembled SBC if I had no assembly equipment / skills.
I would want some type of screen (VGA/TMS9918 etc.) and keyboard otherwise your programing options are limited.

I like these thoughts but agree that there are problems when it comes to “design-by-committee”.
Not having a common platform is something that holds everyone back.
If I were to ask, how many of you are either busy writing or have intension to write your own DOS, BASIC, graphics driver, KERNAL, etc. in the near future.
I would have to raise my hand to most of the above...as time allows.
We're all writing these things in isolation and from scratch and even though code is shared via GitHub or similar, how often do you re-use other peoples code?
Probably not that often because it's not compatible with your hardware or has features that you don't want/need.

Lots to think about.
Andre

_________________
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?

It might be worth noting that the RC2014 computer has been successful enough that the inventor and supplier was able to give up their day job. It's a backplane design, which makes it very flexible, and there are even 6502 CPU cards for it.

But as noted, it is one person's vision and implementation. They may well have been in discussions and learnt from them, but it wasn't a committee.

One thing to note about the various NOS parts available online from various sellers: there is uncertainty, but if you buy in volume and do your own testing, you can be a trusted resupplier (if that's your aim.) There's a world of difference between parts claiming to be something, and parts which have passed a test.

I thought I would check out Bill's observations so I ordered ten (10) W65C02S8P-14 CPU's ($13.01) and five (5) W65C22S6TPG-14 VIA's ($9.46) from one vendor and ten (1) 39SF010A 128-KiB Flash ROM's ($9.36) from another vendor. Results were good and bad;

W65C02S8P-14 CPU

All ten CPU's were nicely marked and tested good at 1-MHz and 2-MHz but failed at 4-MHz. So bad news is they're not WDC chips but good news is I seem to have ten 'free' mismarked 65C02P2 chips after receiving a full refund. Possible "give-aways" with a PCB purchase.

W65C22S6TPG-14 VIA

The five VIA chips were nicely marked and tested good but the pins had been cut short and the chips would just barely seat into a machined pin socket. I requested and received a full refund for the VIA chips.

39SF010A 128K Flash ROM

The SST 39SF010A chips were nicely marked and all tested good with manufacturer ID returning the correct BF B5 code. I would normally buy these from Mouser but they appear to have increased in price from $1.70 (last year) to $2.92 (now) and you still have to add shipping. This is still much better than the $12+ price plus shipping for AT28C256 EEPROM's.

Interesting experience. Have fun guys. Mike

Nice experiment! It's crucial that you ask for a refund - that makes it much lower risk.

Hi all,

I've been running some numbers on my 6502 project.
There are some photos on previous posts if you have not seen it.

Would you pay USD200 (or 199 if you prefer) for a main board with the following specification.
Bare in mind that there will be a significant DHL/Fedex shipping cost from Australia.

65C02 Computer Specification

WDC 65C02 (5MHz)
WDC 65C22N
WDC 65C51 (5V CMOS serial port…connect to 5V FTDI USB-UART cable not supplied)

512K BOOT FLASH (8K banked)
32K SRAM

VGA
Monochrome
Resolution: 256 x 240
7680 bytes Video RAM Read/Write
7680 bytes Video RAM (page write only)

PS2 Keyboard
(Memory mapped AVR uController)

Game Port
2 x SNES game ports memory mapped / decoded by CPLD

Audio
2 x AY-3-8910 (controlled by 65C22)
LM380 audio amplifier
Audio jack connector

Power Supply (on PCB)
Input 18Vdc (1A max)
5V 3A
15V 1A (limited by heat dissipation of TO-263 package and 1A total input current limitation)

I/O
5 x 50 pin edge connector for expansion
2 x 20 pin headers with 32 I/O pins from AY-3-8910’s

Optional Extras not included at this price (sockets fitted on PCB)

8K SRAM (I think most people would fit this)
512K SRAM (8K banked) (I think most people would fit this)
512K FLASH (8K banked) (maybe)
Real Time Clock (DS1511Y+) (maybe)


Some notes,

It will not be available in kit form due to the use of SMT parts.
All IC's socketed.
All parts new/modern/WDC/in production except CPLD's (MAX7000 series) and AY-3-8910. (CPLD's and AY-3-8910 will be tested...if it plays a tune it works...) CPLD could be modern/in production at some additional cost.

An option that may or may not reduce cost is to sell the assembled board with only CPLD's and AY-3-8910 fitted. It's up to the buyer to source the rest, but old/slow parts probably won't work at 5MHz so this option may be more hassle than it is worth.

Let's keep replies simple. I'm sure we could pull apart the specification and find plenty of reasons why it does not suit a particular use case or personal preference. The specification is the specification. If there is something that you would have liked to add, turn it into add-on card. I already have a few like SD card, TMS9918A and prototype card. It's already a sizeable PCB at 170x220mm, 4-layer.

You either like it or you don't. You would buy it or you would not.
If you have questions please ask.

This specification is a slight upgrade from my current board.
The 5V power supply will go from 2A to 3A.
The 15V regulator is new. (for future ADC / DAC / OPAMP circuits on plugin cards)
The SNES game ports are new.
I will probably make this version for my own use anyway as a last update to this project.
Future work will focus on plugin cards and software as time and money allow.

One more thing, though I've done a fair bit of software in C and assembler using CC65/CA65, it's not in particularly usable state.
Bottom line is, if you're not comfortable writing and debugging at a very low level then this is probably is not for you.
I don't think this applies to most of the readers of this post. It will take some time and probably the efforts of initial users to create things like emulators, kernels, file systems etc.
The hardware on this board is substantial and fairly complex with 2 CPLD's and a small uController. I have a glitch on power up that, occasionally, requires a manual reset to restart the system. I have not had the time to fix it. It's not a big deal to me. Fixing it may be simple or may not possible at all without a redesign. To date, after hundreds if not thousands of hours of software development, this is the only issue that I am aware of. The CPLD files and uController source would be available if people want to tinker. (CPLD designs are schematics not HDL/VHDL so anybody with a reasonable understanding of digital logic should be able to make changes.)

There's a lot to digest. Let's keep replies simple and on topic.
Let me know what you think.
Andre

I really like the idea, particularly as your kit has video, sound, keyboard and game controllers.

The short answer to would I pay USD200 (closer to 300 after shipping) is probably not. That works out to about R6000 in South African money and that's a significant chunk of cash.

On the other hand this is, I guess, not a typical answer as the exchange rate is horribly against me. I mean I'd really like something 65xx that meets all the criteria I mentioned above and that also doesn't see me designing it from scratch. My only generally usable 65 device at the moment is a W65C816SXB and frankly it's kinda boring.

What I've been thinking about while looking at this forum - and slowly plodding along with my own design - is to use someone's existing SBC schematic, PCB it up with the components that I do have try and do fun stuff with their already existing software. Of course what's actually happened every time I've thought this is that I've been missing a crucial piece. And then stopped myself from making changes that would make it incompatible with the original authors spec / idea.

I'll probably have to come and clean this message up later once the power is back (sent from my phone) but hopefully this ramble offers some perspective.

Andrew!

I've been thinking for a long time about making my Ruby 6502 (actually, probably the '816) board available as a kit (it's all through hole).

But there seems so much against me - cost is just one factor, finding time to buy the parts, get the PCBs store the parts then source suitable boxes to ship it out in ....

Then there's the hideous reality of "Brexit" which is doing no-one any favours - that means that for anyone outside the UK shipping costs and import/customs dutys are a real thing. But then again no different from me if I ever wanted to buy a kit, etc. from the US, so that makes it fair, right? (Not really).

Then what level of "kit"? I need to pre-program GALs and the ATmega - test them on system - test the memory chip, etc. so making up a test platform with ZIF sockets would be a thing too.

And then .... The actual kit? I have a couple of versions of my system - one the main board plus plug-in key keypad and LCD, can also use a serial connection to a dumb or "smart" terminal, but the smart terminal is Linux only right now. Then there's the "backplane" version with "slots" for "stuff" like video card (replaces the keypad/LCD) and other peripherals (SPI, I2C, GPIO) so the variants soon add up.

So it starts to become somewhat more a challenge than a fun thing to do. Or maybe not? Can't quite say right now as I've not tried it. Maybe I (and we) ought to just "suck it and see?" ...

Going back to "would I buy ... " Probably not, but mostly because I have my own and also because I've been using the 6502 since 78 and made a few little systems in the early 80s (doesn't stop me desiring a KIM/SYM/Aim-65 though) ...

Hmm..

-Gordon

_________________
--
Gordon Henderson.
See my Ruby 6502 and 65816 SBC projects here: https://projects.drogon.net/ruby/

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

Not a chance!

_________________
"The key is not to let the hardware sense any fear." - Radical Brad

I think the people here will not be typical purchasers of such a kit - and that really doesn't mean that there's no market. The people here, especially those who spend a lot of time here, have probably already built what they'd want to, or are in the middle of designing and building. The people to reach are the people interested in something new, perhaps interested in their first SBC or their first 6502 system.

So, you'll see some people saying they won't buy, but that's not really giving any information.

I'll add one more thing here - I don't think "we" are your target audience for this - in the same way "we" were not the target for Ben Eaters project, but that didn't stop him selling 100s if not 1000s of his kits ...

So never say never ... (is my motto here!)

Cheers,

-Gordon

_________________
--
Gordon Henderson.
See my Ruby 6502 and 65816 SBC projects here: https://projects.drogon.net/ruby/

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