Congratulations -- it sure looks pretty! I'll bet you're excited.

Hmm, it might have been a little better to postpone soldering the memory and peripheral sockets til later. Without the latter you could still power up the board and do a preliminary test using a NOP generator.

What I'm thinking is it's conceivable that the NOP test might reveal a short or other problem -- and the fault might be rendered invisible and inaccessible if the memory and peripheral sockets have gotten soldered in. (Of course the odds of that happening are small. I just like to thwart Murphy's Law whenever I can!)

-- Jeff

_________________
In 1988 my 65C02 got six new registers and 44 new full-speed instructions!
https://laughtonelectronics.com/Arcana/ ... mmary.html

If temperature rise becomes a problem, you could get one of those cheap, stamped heat sinks (TO-220) and sandwich it between the regulator and the PCB. Your circuit isn't going to draw all that much juice, so I wouldn't expect the regulator to get very warm.


As Arlet noted, minor silkscreen imperfections are not uncommon. I see that every so often on 500 dollar server mainboards.

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

There's no real heat from it so all is good - I'm just irritated with myself for forgetting to put the pad in. Good idea re the heatsink though.

Here's the board built up and complete apart from a zener diode which I need to find. Current draw is 80mA while running NOPs from the EEPROM. Most of that is probably going to the GAL.

Board stats (for those interested):

Name: 65C02 GPD (General Purpose Design)
CPU: WDC 65C02S
CLOCK: Internal or external source selectable using jumper
RAM: 32kB, 55ns ($0000-$7FFF)
EEPROM: 30kB, 45ns ($8400-$FFFF)
I/O: 1x 65C22S VIA ($8100-$810F)
I/O: 1x 65C51N ACIA (using MAX238) ($8200-$8203) - for RS232C
Power In: 7-20V DC regulated
Power Out: 3x 5V DC out, 1x 7-20V out





Testing:

So far it runs at 1MHz fine (2MHz split down to 1MHz) with NOP instructions from EEPROM. Internal and external clock source selection works fine as does reading address and data bus from the bus connectors (the IDC style sockets on the board).
The 3.2mm holes are perfect for the motherboard standoffs found in PCs - thanks to all those who made suggestions in this area.

My only real issue is tryign to get the flippin' flux off of the PCB. Isoprop and a short bristle brush + lots of scrubbing. It's coming off slowly. Also tried using low fibre kitchen towels and scrubbing through that with Isoprop to try and lift the flux off.
Anyone got any tips?

Next steps: RAM tests, I/O tests (probably output to an LCD)

You can get special flux remover from electronics distributors.

Isopropanol can be used to remove flux, but in some cases, will dissolve the solder mask. Use with caution.

I use an aerosol flux remover but when I went onto Mouser's website posing as a UK buyer, I discovered that none of the aerosol defluxers that they sell could be purchased within the UK. I did find an aerosol defluxer at Digi-Key's site, but the £21.12 price seems kind of steep.

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

Just a small update: The board runs (so far) without any issues up to 16MHz (using a 32MHz oscillator can then divided by two by a flipflop). Tried a 40MHz / 2 for 20MHz but it then goes a bit mad and tries to read from random addresses.
I suspect that CPU, VIA or that the RAM might be the bottle neck given the following:
RAM: Lyontek LY62256PL @55ns
ROM: Winbond W27C512-45Z @45ns
CPU: WDC W65C02S @71ns (14MHz)
Glue logic: Lattice GAL 22V10B-15LP @ 15ns pin to pin
VIA: WDC W65C22S @71ns (14MHz)

Given that the WDC parts can normally go much higher, I think it's most likely the RAM. Not sure how to prove this other than find a faster DIP RAM chip of the same pinout - unlikely. I could try an SMD part through an adapter though... pin in the back side given that the adapters are wider than the RAM DIP socket .

I also need to find a way of suppressing the noise coming off of the board. I can hear static on my head phones if the wire gets within feet and I'm using the radio on my phone.

[edit]

Yeah, 16MHz = 60ns, 20MHz=50ns... so the RAM would work for 16 but 20 is outside the RAM spec.

Excellent!

FM radio ?

Yep FM radio in my Galaxy S2 mobile phone.

At 20MHz, the cycle time is 50ns; but that has to include setup time, hold time, glue-logic propagation delays, etc., leaving a lot less than 50ns for the memory. The memory has to be way faster than 50ns. It's only a small exaggeration to say the 6502 does a read or write operation in half a cycle.

When the data sheet says the processor is spec'ed for 14MHz, it does not mean it won't run any faster, but that it's not guaranteed to meet the specified timing margins at higher clock rates. The resulting speed requirement on the memory and supporting parts then becomes steeper and steeper until the curve goes vertical somewhere over 25MHz and not even infinitely fast parts will get you any more performance. See Jeff Laughton's to-scale animated timing diagrams at http://laughtonelectronics.com/Arcana/V ... iming.html and the related 65816 discussion at viewtopic.php?f=4&t=2438 .

_________________
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?

Yep I see what you mean with regard to the RAM speed. I've now tried a 12ns SRAM (Alliance AS7C1024B-12TJIN 1MBit) and that has the same issue .

However, at least my home made RAM bodge-adapter worked at 16MHz (with optional cat hair!) :-

What's the speed on that Winbond part?

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

45ns (22MHz) so this is might next port of call. I'll see if I can find a faster part. Alternatively, I can try copying the ROM to RAM as discussed in other posts, but this board isn't set up for that so would end up being a chunky add-on.

Might it be the EPROM that's holding you back? I don't quite understand how your test is conducted. You said the machine goes a bit mad and tries to read from random addresses. But what's the context -- what code is running, and from which memory? Can you be confident the initial bootup (from EPROM) was successful?

_________________
In 1988 my 65C02 got six new registers and 44 new full-speed instructions!
https://laughtonelectronics.com/Arcana/ ... mmary.html

I have a small piece of test code which sets port A of the VIA to $60 and then drops into an infinite loop. I then have an ISR which both the IRQ and NMI vectors point to. The ISR does a simple check at RAM location $0000 for the value $11. If it's not $11, then it sets it to $11 otherwise it says it to $22. It then sends the value ($11 or $22) to port A of the VIA.
I then connect my h/w bus monitor up to port A and also control lines for /ROM select, /RAM select and /VIA select.
Finally, I tap the NMI button on my board. Each time I do this it toggles VIA port A between $11 and $22 and I see the result on the bus monitor along with the signals.
This works fine and as expected up to and including 16MHz. At 20MHz the RAM and RAM select lines go on and off randomly and there is no response to the NMI button on VIA port A.
My test isn't the absolute best, but it does give me an indication of the when the system is running ok or not at a given speed.
It is possible that the GAL is the cause, but I suspect the Winbond W27C512 ROM more as it's the next slowest device and at 45ns (22MHz) is skirting the edge. As Garth says, the speed rating doesn't include set up times, etc., so it's very possible that this is the culprit.

My mistake, then; I thought you were blaming the RAM. The ROM does seem more likely to be problematic.

There's a very simple wait-state circuit here which will roughly triple the time available for the ROM to do every access. Using that you can crank things up and see what the rest of the system can do!

_________________
In 1988 my 65C02 got six new registers and 44 new full-speed instructions!
https://laughtonelectronics.com/Arcana/ ... mmary.html