Current 6502 Kits

[BigDumbDinosaur]

BTW, I did a search for “Argon Lite” and couldn’t find a computer by that name. I did find references to welding shielding gas.

[cbmeeks]

BigDumbDinosaur wrote:

BTW, I did a search for “Argon Lite” and couldn’t find a computer by that name. I did find references to welding shielding gas.

My spelling was incorrect. It's "Agon Light". I was WAY off. lol

Anyway, I am talking about this guy:

https://www.youtube.com/watch?v=R2dYZmlu1D4

[sburrow]

BigDumbDinosaur wrote:

Probably not a good requirement. Many of those chargers have regulation problems with any significant load. They are designed to charge a battery, which lessens the requirement for good regulation and noise-free output.
Also, the maximum output per USB spec would be 500 mA, which sounds like a lot until you add more to the basic system. Running any power supply near its maximum output rating is never a good idea. Also, running your system through the thin wiring of a USB connector inserts some resistance into a circuit that should have as close to zero resistance as possible.

So, I actually have a barrel jack footprint built underneath my USB port footprint, in case someone wants to use that instead of the USB. Likewise I have a 2-pin header nearby for VCC and GND, in case someone would rather just use a direct connection to the board (or using the board's power for other things).

My computer runs off a external phone battery for MANY hours, as discussed on a different topic here. If a phone battery can power the thing without any problem for hours and hours, I'm ok with plain ol' USB doing it's thing too. I'm not looking for efficiency, just flexibility.

cbmeeks wrote:

Can you imagine a modern 6502 running at the same 18MHz the Argon runs at?

That's one thing that's not part of my list of 'requirements', a super-fast computer, or something with a ton of RAM. I know that many of y'all here do speed-tests and all kinds of crazy things with a 6502 that just boggles my mind. Cool, but my goal isn't trying to out-perform, but to be simple and usable. I'm quite ok with 3.14 MHz and 64KB of RAM.

And maybe that's why I know many of y'all are not my target audience (if my board does eventually become a kit). Got a question, in general: Who here has bought the Dodo 6502 kit? Who plans on buying it soon? I would like to because it is a very neat design and something I've actually wanted for a while: A simple yet efficient 6502 handheld! But, at the point where I am at now, I would just build my own instead. Nothing against the Dodo, I just think it's fun to design that stuff myself now!

I do not know much about the Z80 front, but it's neat to see retro-style computers prosper at all, 6502 or not.

Thanks everyone, good discussion.

Chad

[cbmeeks]

@ sburrow

For the record, I think building 6502 kits is a great idea. I started a couple myself (one based on a 6502 driven calculator and one as a hybrid KIM-1/LCD "trainer" computer).

Best of luck!

[plasmo]

BigDumbDinosaur wrote:

BTW, I did a search for “Argon Lite” and couldn’t find a computer by that name. I did find references to welding shielding gas.

It is Argon Light based on eZ80 but only 18MHz, so it is a far cry from Min-eZ, same processor running at 50MHz. eZ80 is pretty impressive device, clock for clock, it is 2-3 times faster than Z80, which means it is about same speed as 6502, clock for clock. This also means the faster 6502/65816 designs in 6502.org can outrun Argon Light.

OT, I thought I can out-Google BDD by searching for "Argon Lite homebrew". It turned up many topics about using Argon/CO2 to pressurized beer. Since I do weld and make wine/beer, those were rather interesting reading. I readily admit burping Argon gas is safer than burping hydrogen gas...ya, I'm truly off topic now...
Bill
Edit I see cbmeeks corrected the spelling while I was enjoying myself reading about Argon Lite homebrew...

[Paganini]

I will mention again the RC6502. The backplane version is basically an RC2014 slightly modified to use a 6502. There is also a 6502 CPU card for the RC2014. When I was first thinking about building something the RC2014 was really appealing. It looks cool, is well tested, and is available! However, it is pretty expensive!

Looking at Peanutbutter-1 sitting here next to me, most of the components cost less than $1. The RAM, ROM, MPU, VIA, and can oscillator all cost more than $1, but not much. I would estimate the whole thing cost around $30 in parts. If you go much higher than that in price point, potential customers will have to answer the question "why build this (expensive) kit, when I could just buy an already built SXB board from WDC?"

The folks that cbmeeks mentioned who "are interested in buying a simple computer and just programming it" have the SXB board available for $65.02. Here are some ways I can think of to make a kit more attractive than the SXB board:

* Offer some standard I/O connections (PS/2 or USB keyboard input, VGA output)
* Make the kit customizable
* Make the kit inexpensive

[kernelthread]

plasmo wrote:

It is Argon Light based on eZ80 but only 18MHz, so it is a far cry from Min-eZ, same processor running at 50MHz. eZ80 is pretty impressive device, clock for clock, it is 2-3 times faster than Z80, which means it is about same speed as 6502, clock for clock. This also means the faster 6502/65816 designs in 6502.org can outrun Argon Light.

According to the data sheet (http://www.zilog.com/docs/um0077.pdf) eZ80 approaches 4 times as fast as a Z80 at the same clock frequency. It does real pipelining, so it can do some instructions in 1 clock cycle (e.g. bog standard 8 bit data processing instructions on register operands). It does block move (LDIR) in 2 cycles per byte - i.e. maximum possible rate - compared to 7 cycles per byte for 65816 MVN/MVP, or the best I managed was 6.4 cycles/byte using dynamically generated code. This post https://www.retrobrewcomputers.org/foru ... &#msg_7905 has some benchmarks for an eZ80 at 50MHz - the equivalent Z80 clock frequency would be 170 to 190MHz. A 6502 may be twice as fast as a Z80 at the same clock frequency, so it seems the eZ80 will be almost twice as fast as the 6502.

[BigDumbDinosaur]

plasmo wrote:

OT, I thought I can out-Google BDD by searching for "Argon Lite homebrew". It turned up many topics about using Argon/CO2 to pressurized beer. Since I do weld and make wine/beer, those were rather interesting reading. I readily admit burping Argon gas is safer than burping hydrogen gas...ya, I'm truly off topic now...

I mostly use 75/25 with my MIG welder and don’t inhale the argon. I do, however, exhale CO2.

As for burping hydrogen, I'll pass. No need for any fires and explosions at the dinosaur pen.

[plasmo]

kernelthread wrote:

This post https://www.retrobrewcomputers.org/foru ... &#msg_7905 has some benchmarks for an eZ80 at 50MHz - the equivalent Z80 clock frequency would be 170 to 190MHz. A 6502 may be twice as fast as a Z80 at the same clock frequency, so it seems the eZ80 will be almost twice as fast as the 6502.

eZ80 is an impressive 8-bit uP, no doubt. I want find out how Argon Light's claim of fastest computer at 18MHz matches up to several 6502 forum's 6502/65816 running in mid 20Mhz (my own CRC65 is 29.5MHz).

I have WSM's 50MHz Min-eZ for a couple years now. It is so darn small, I can't find it now to do more benchmarks. As I remembered, it was certainly very fast, especially with computation-bound benchmarks and it was still fast with mixture of disk and serial port operations. I want to see how it does compiling large C programs.
Bill

[barnacle]

I think the points mentioned earlier around newly available parts are valid; I try to design my own stuff that way though the board I'm currently playing with has an ebay 68B05 for serial comms and an ebay 28C256 ('new' - HAH, explain the solder on the legs, the data in the chip, and the fact that it was software write locked! Oh, here's some discount. OK, it seems to work )

One thing which I *never* see mentioned regarding kits or simple designs: EMI regulations. Without very expensive testing (I get to see that fun at work) it's unlikely that many of these kits are anywhere near meeting the regulations, though best practice PCB design may help: the fast edges that modern parts can generate have harmonics waaay up there. Hmm, I note that some SoC parts and ARM processors are available with spread spectrum jitter built into the clock circuits...

Neil

[GARTHWILSON]

barnacle wrote:

One thing which I *never* see mentioned regarding kits or simple designs: EMI regulations. Without very expensive testing (I get to see that fun at work) it's unlikely that many of these kits are anywhere near meeting the regulations, though best practice PCB design may help: the fast edges that modern parts can generate have harmonics waaay up there. Hmm, I note that some SoC parts and ARM processors are available with spread spectrum jitter built into the clock circuits...

When the company I work for first incorporated a computer into a high-end aircraft intercom, we looked into this, and I visited a local lab that does the testing for it too.  The answer was that we didn't need the testing and certification for mobile use.  Obviously it could be a problem if it interfered with the aircraft radios; but I did a good design and there was no such trouble in the 13 years we sold that product.

The only time we had that kind of trouble was with a switching regulator on another model that didn't have a computer.  Initially it was fine; but the manufacturer of the controller IC, Maxim, changed the process perhaps a year after we started production, and suddenly this thing was putting a lot of RF noise into the radios, because of the faster switching edges (although the switching frequency was unchanged).  We had a lot of boards already assembled, and it would have been very expensive to scrap them and hold up production and sales until I could re-design the board and get new ones made; so what we did instead was some manual re-work, soldering capacitors directly to IC pins to shorten the connections, and adding ferrite beads to wires.  That worked.  I applied what I learned from the experience to subsequent designs.

I think I read, but it was long ago so I couldn't tell you where, and I would have to do a search, that kits also don't have to meet the same standards.  I also wonder if the FCC is not paying as much attention anymore since very few people watch TV from signals from the air (as opposed to cable or optical fiber), and very few people are listening to AM and FM radio anymore, especially when not in their cars.

[plasmo]

plasmo wrote:

Here is a 6502 kit idea: You can buy kits on eBay with collection of IC. Here is $10 + $6 shipping for "Barebones/minimal 6 chip 6502 computer IC kit". I just ordered one, it says "118 sold". Yes, I understand 6502 from eBay is particularly problematic but this one is Rockwell R65C02 which tends to be less problematic. The rest of the kit should be OK.

Received the "barebones/minimal 6 chip 6502 computer ic kit" yesterday and the result is abysmal. Picture shows the received package, it is well packaged in antistatic foam. There are 7 chips in the package, the 3 TTL tested OK with TL866II programmer; the HM62256 failed TL866II test and also failed to run in a working system; AT28C256 reads OK, but can not be programmed successfully with TL866II; MC68B50 works; and Rockwell R65C02 draws more current than normal CMOS 6502 and does not work even at 1.84MHz.

I'm quite disappointed with the vendor who had provided good Z80 kits over multiple purchases. I will write to the vendor and complaint about the defective parts. My good experience with Z80 didn't translate to 6502. I'm really quite surprised by the defective RAM and EEPROM, this illustrates the hazard of making kits from untested eBay parts.
Bill
PS, this does not stop me from purchasing on eBay; eBay has good return policy so I'm ready to test parts immediately upon reception. The few defective shipments I had were fully refunded without exception.

[barnacle]

Which is why I'm unhappy with designs that can't be fulfilled with parts from a mainstream supplier. Nothing to stop you using ebay parts of course -- my own very similar design uses an AT26C256 of which I bought five 'new' which miraculously had bent pins, solder sticking to them (and in one case a short between two pins), contained data, and were software locked... though all five work apparently correctly. I also bought half a dozen 68B50 (from a couple of sellers), all of which appear to work correctly and are perhaps NOS - everything is at 1.8MHz.

But I try do design stuff with parts which are available from the likes of Digikey, Mouser, Arrow, Farnell... Mouser are the European suppliers for the WDC 65C02 so there's no reason not to (though it's a bit galling to realise that STM can supply a Nucleo development board for about the same money )

The problems are always going to be (a) programming the eeprom on board - I'm thinking about a mechanism to automatically copy from a serial flash into ram, but haven't thought it fully through yet! - and (b) serial comms. If you don't require the 6522 it's not beyond the wit of man to implement simple latched parallel ports in both directions, and of course that's also available from Mouser (and as expensive too!)

I think that SPI interface is the way to go for both bulk memory (e.g. 128Mb/16MB flash is only a couple of quid) and for serial comms. I don't mind using an external microcontroller (likely a Nucleo) to get the initial code onto the machine but I don't want a second processor alongside the 6502 except perhaps for video.

A supplier I do use a lot is LCSC in Shenzen; some of the locally made glue logic is remarkably inexpensive and I've never had a failure. Takes a week or three for the bits to get here (German post have been intermittently on strike this month) but the PCBs are coming from next door at JLC anyway.

Neil

p.s. I need to restrict myself to no more than three or four projects at a time!

[BigDumbDinosaur]

plasmo wrote:

plasmo wrote:

Yes, I understand 6502 from eBay is particularly problematic but this one is Rockwell R65C02 which tends to be less problematic. The rest of the kit should be OK.

...Rockwell R65C02 draws more current than normal CMOS 6502 and does not work even at 1.84MHz.

That “65C02” looks to have been re-marked. Also, the 0044 date is suspect. Rockwell discontinued 65C02 production in 2000.

Quote:

I'm quite disappointed with the vendor who had provided good Z80 kits over multiple purchases. I will write to the vendor and complaint about the defective parts.

I’d give him/her a “neutral” or “negative” rating for shipping you junk parts. Sending parts that were re-marked or outright defective demonstrates either intention or lack of due diligence on the part of the vendor. He/she deserves to be dinged for that.

Quote:

PS, this does not stop me from purchasing on eBay; eBay has good return policy so I'm ready to test parts immediately upon reception. The few defective shipments I had were fully refunded without exception.

Yes, eBay’s buyer protection is pretty good. However, it won't compensate you for the waste of your time, nor will it be of any help should a defective part inflict damage on other parts in your system.

[Paganini]

plasmo wrote:

Received the "barebones/minimal 6 chip 6502 computer ic kit" yesterday and the result is abysmal.

Alas! I had high hopes. I have a better result to report: I just received 4 ICs from eBay seller joma0818. They shipped from CA, came in a reasonable amount of time (about 3 days, I think). Although listed in eBay's "new" category, I am 100% sure they're pulls, as I expected they would be. However, they are in good shape, and to the best of my ability to test them they seem to be the real thing. Which is to say, both the VIAs work at 4 MHZ as advertised. I don't have any way to test the ACIAs yet.

On this topic, has anyone had any experience ordering from UTSOURCE? I have run across references to them a few times in the context of "where can I get parts when Mouser/DigiKey can't help me in this time of supply chain shortage?" I took a risk and placed an order with them. I expect it will take another week or so for the order to arrive, since the parts ship from China. However, I think they might be legit, as they sent me an e-mail cancelling and refunding me for one part that was out of stock. I don't think the guys peddling fake chips ever run out.