Sally replacement by standard 6502 ;)

[Sikor]

O.K. Mq from atariarea made adapter to replace 6502C by stabndard 6502: http://www.atari.org.pl/forum/viewtopic.php?id=15241(Polish). For now it was tested with non-ECI 65XE (not working property - tests in progress) and with 4-bits RAM chip ECI board - working perfectly (test in progres).

I think, that it is good idea to use normal 6502, idea was borned by me and few other peaples, adapter was created by reverse-enginering of Atari 400/800 computers, where was use standard 6502. Adapter is simply and little.

If anyone interested, haw it was made - please visit atariarea and ask Mq (eventually i can write to him with link to this thread, if anyone interested it).

 

[+David_P]

Can you also use a 65816 with the adapter?

[Sikor]

I dont try use it - so, I don;t know...

[JamesD]

Can you also use a 65816 with the adapter?

The 65816 has a different pinout, so it would probably need more work.

The 65C02 should work though.

[Heaven/TQA]

What is the benefit?

[Rybags]

Standard 6502 probably somewhat easier to obtain than the Sally model.

 

65C02 has the "advantage" of extra instructions and probably lower power consumption, though IMO it's not worth specifically doing 65C02-only coding

 

Disadvantage of 65C02 is that commonly used illegal instructions won't work and some other subtle differences that can create incompatibilities.

[Sikor]

If i know - 65C02 working fine (I'm see picture), and Yes, it hasn't illegal opcodes

Mq wrote, that probably write here some more info - he has account on atariage

[guus.assmann]

Hello,

 

Have done this many years ago and developed a PCB. It has also been published in the Dutch Atari magazine. (SAG)

Yes, a 65C816 will also work, just a few minor changes to some pins.

I do still have a PCB layout, in PDF format.

 

BR/

Guus Assmann

65C02.zip

[peteym5]

I always wondered how someone have it use the "halt signal" when the Antic/GTIA needs to access the bus?

[Sikor]

Nice, but i see - big PCB I attach photo made by Wq with actual PCB (monted).

[guus.assmann]

Hello Sikor,

 

New PCB is better, I agree.

But my desgn was from 1987.

And the PCB is single sided.

It also contains an extra 6520 to accomodate a printer port and RTC.

 

BR/

Guus Assmann

[+CharlieChaplin]

Hello,

 

Have done this many years ago and developed a PCB. It has also been published in the Dutch Atari magazine. (SAG)

Yes, a 65C816 will also work, just a few minor changes to some pins.

I do still have a PCB layout, in PDF format.

 

BR/

Guus Assmann

 

And if you want that in german or english, here you go:

http://www.b-pahl.de/atari8bit/65c02/65c02e.html

http://www.b-pahl.de/atari8bit/65c02/65c02.html

[Mq.]

Hello.

I'm the designer of that new PCB. The main purpose of this project was to build it for fun, and to use 6502 as a replacement of our beloved Sally (for fun too:-)).

Some people on our atari.area forum found cheap UMC6502I (Sally) in China on eBay I believe, and they bought some of them. Unfortunately the seller sels them 6502/6502A - not Sally.

So we found that we can build adapter using the solution from Atari 800 schematic for generating HALT signal.

I designed PCB, built one board, and succesfully launched it. But...

It works perfectly on Atari 65XE/130XE motherboards with 4-bit RAM chips (tested on two such boards without any issues).

There are issues on Atari 65XE motherboards with 1-bit RAM chips, and without ECI. The issues was the same on two such motherboards. Generally everything is ok, but there are some problems with GTIA. In some games P/M graphics disapear and there are no players sprites. Motherboards and GTIA's works perfectly with original Sally.

Any ideas what is the diffrence between that motherboards (65XE with/without ECI) which causes that problems?

I attach a schematic and pcb design of the adapter.

(sorry for my english, I hope You understand what I wrote and what I mean:-))

Edited April 8, 2018 by Mq.

[Sikor]

Hello Sikor,

 

New PCB is better, I agree.

But my desgn was from 1987.

And the PCB is single sided.

It also contains an extra 6520 to accomodate a printer port and RTC.

 

BR/

Guus Assmann

Your PCB is super for year of made it. New PCB was made little, because idea was close it in standard cover with all elements. And Yes, i don't know about your projest.

As Mq wrote - it is made for fun... And for keep ours Ataris live - even, where Sally was death...

Now, I'm wait for POKEY, GTIA and AANTIC replacement

[Level42]

Yeah, easy enough to replicate the 400/800 extra hardware of course.

 

Nice little project ! I have some Russian/Romanian 6502 clones that would be cool to try with one of these PCBs

 

Since the goal was a small PCB, why not also use SMD chips for 74HCT541's ? You could have put chips on both sides of PCB and it would be about the same size as just the chip....

No criticism....just wondering

[Mq.]

74HCT541 in SMD are in DW package, which don't fits between 6502 and Sally socket pins at all. Paradoxically in this case 541 in DIL package is better choice and takes up less space than SMD chip. That's why I purposely used SMD for 74HCT02 and 74HCT74, but DIL for buffers chips.

[Level42]

Aha I see

[dmsc]

Hi!

 

74HCT541 in SMD are in DW package, which don't fits between 6502 and Sally socket pins at all. Paradoxically in this case 541 in DIL package is better choice and takes up less space than SMD chip. That's why I purposely used SMD for 74HCT02 and 74HCT74, but DIL for buffers chips.

Other solution could be using TSSOP-20 package, like this at mouser. I had soldered those a few times, requieres good eyesight (or a magnifier glass) but it's not difficult.

 

Sometimes I wonder if soon it would be cheaper/easier to simply use a CPLD like the ATF1500AL for this kind of adapter.

Edited April 9, 2018 by dmsc

[+tf_hh]

 

It works perfectly on Atari 65XE/130XE motherboards with 4-bit RAM chips (tested on two such boards without any issues).

There are issues on Atari 65XE motherboards with 1-bit RAM chips, and without ECI. The issues was the same on two such motherboards. Generally everything is ok, but there are some problems with GTIA. In some games P/M graphics disapear and there are no players sprites. Motherboards and GTIA's works perfectly with original Sally.

Any ideas what is the diffrence between that motherboards (65XE with/without ECI) which causes that problems?

 

I´m wondering that you didn´t have problems in the XE generally. I´m curious, how many systems you test? What is the Freddie´s marking in the XE computer where you have no issues?

 

I´ve made the same thing >5 years ago, but this one works only fine in all XL except the "Rose" mainboard (with Freddie) - and also not in all XE / XEGS computers. When analyzing the issues I found out, that Freddie causes the problem.

 

Using the 400/800 way to halt the CPU when ANTIC claims the bus causes PHI2 remaining high during a very long time sometimes (in relation to the standard PHI2 phase). Freddie itself uses PHI2 to synchronize DRAM access. When you connect a logic analyzer, you will see, that CAS sometimes remains high, even when the ANTIC/GTIA want to fetch data from RAM. In my version also PMG failures occur, also disturbation in high-res (Graphics 8 mode). So I dropped the whole thing, it was only a proof of concept.

[Mq.]

As I wrote, I tested the adapter on two 65XE motherboards without ECI (with eight 1-bit RAM chips). There were the same PMG issues on the both of them.

Then my friend built the adapter and test it on the 65XE/130XE board with ECI and 4-bit RAM chips, and he told it works fine. So I found the same motherboard, and my adapter works on it fine too.

I attach the picture with issues on no-ECI motherboard, and second picture without issues on ECI motherboard.

I will check the numbers on Freddie's chips later, or maybe tomorrow...

Edited April 9, 2018 by Mq.

[Wilheim]

Imho, it looks like the antic dma access is not working. Maybe you need to synchronize the clock signals.

Edited April 9, 2018 by Wilheim

[guus.assmann]

Hello Mq. ,

 

Digging into memory now.... The 7474 needed to be an LS type in my design.

And also the 74LS02 was needed.

I did test with S and F types, but that was no good.

So the total timing is quite sensitive.

The LS02 is most critical.

For me, it was good that I had the IC's in DIL. That means using sockets.

 

About 65C816, see the link to Berhard Pahl's site.

There you can read what to do to get the 816 working.

And yes, I tested it a long time ago.

 

BR/

Guus

Edited April 9, 2018 by guus.assmann

[Mq.]

@guus.assmann: Thanks for the hint. I know in this circuit it is important to have proper propagation delay times to get proper phi1 and phi2 signals. I think it is the reason to F or S series may not work, because they has to short propagation delay time. The LS series has the same propagation delay as the HCT series, so it's equivalent in this case. My friend used LS series in his adapter board, but that adapter was tested only with 65XE/130XE motherboard with ECI and 4-bit RAM's - and works fine.

And You are right: if You have DIL chips in the sockets, then You can try few different IC's. Anyway, I think if there are so sensitive timing, then maybe the circuit needs to be fine tuned to every specified machine.

Adapter boards I designed and ordered, I shipped to few people, and they will build few adapters more. I think they will test them on some different motherboards in the near future, so we will get some more information about that.

[Sikor]

Test with rapidus card on grawitacja game jam:

[Spotted Lady]

I think I know a workaround to get this to work with more machines, if Freddie is the problem. You can put 64k SRAM on the board (or use the universal memory upgrade/repair kit with the SRAM), remove any memory boards, and also run a wire to the CAS inhibit pin. If it still doesn't run, then use a Freddie replacement that removes the refresh while keeping any other tasks it does.

 

I was going to suggest using Potato SMD chips for the SMDs you used and 74F for the others, but I see that won't work, since you tried similar, since I guess it it better to come in a cycle late than it is to come in late in the same cycle.