[quote="Euphoric"]But, oh my god, what's that crappy Phi2 positive edge with a 6502A ???[/quote]

NMOS uses (I believe) active components to drive lines low, and passive to pull them high. The passive component is effectively a resistor, and it has to charge the capacitance on the bus. The active component has a much lower resistance, and can discharge the bus a lot faster.

If you're crazy enough to want to emulate that curve, you could put a resistor in series with the +5V supply. But if you want it to work... :-)

Ok, I'm not that crazy

Hey, I am now using 4 inverters before the AND gate, and IT WORKS GREAT !!
I mean, it works with all the memory speeds I have tried (70ns, 80ns, 150ns). With two inverters only, I could only use those 150ns rams.
(On the Orics, there is a ULA that does two memory cycles during the low level of Phi2, for video purposes. I guess the second cycle overflows on the beginning of the high level of Phi2, and thus conflicts with a cpu memory cycle).

Anyway, with those 4 inverters, the delay between Phi0 and Phi2 seems more compatible with the old nMos 6502. Do you agree to wire 4 inverters on your design, Daryl ?

Best regards,

Fabrice

Are they 74LS or 74HC? I'll use whatever the group decides is best. If possible, I'll make it optionable for both.

Daryl

I have used 74LS from the beginning but lately I've found that the Phi0 clock on my system is a rather neat signal compatible with CMOS levels...
So I could have gone with 74HC chips...

I don't know if we can assume all the systems have a clean clock...

Cheers,

Fabrice

> My question is about sourcing these pins. They have to be able to fit
> into a standard DIP socket. The IDC header pins cannot be used
> because they are square and would either not fit or damage the host's
> socket. Does anyone have any suggested sources for these pins?
> . . .
> What do we use for the pins???

I saw Winslow Adaptics at a trade show a couple of years ago and was very impressed. Unfortunately their website does not show much of what they're really capable of. This company makes all kinds of standard and custom adapters, some being very complex little circuit boards that plug into a DIP socket. You get just a minor idea from http://www.winslowadaptics.com/uk/custom.asp They're based in England but have offices in Florida. I wrote a note on the front of their catalog saying that NRE (non-recurring engineering charges) for custom parts is often $300. I don't know what the per-piece cost after that might be. Depending on how many you think people will buy, it might be worth looking into. If you do most of the layout for them, I expect that would make their job quicker and cheaper.

Another possibility would be to use a hybrid house, but I suspect their charges will be higher. Maybe they could at least suggest where to get the pins if you make the board.

After several months and many more revisions, I've completed laying out the traces for the converter. I still need to clean it up - straighten some traces, add 45 deg bends on the corners, check spacing around all pads and vias, etc.

I added to the schematic. There are two jumpers. J2 allows you to select the gate for the 74x245 and J3 selects 2 inverters or 4 inverters as inputs to the AND gate.

Check out the schematic, layout draft, and expressPCB files on my web site.

http://sbc.rictor.org/support/conv.html

Estimated cost for 10 boards would be $200. ($20 each)

What do you all think?

Daryl

The clean up is complete and the design is ready.

I am willing to place a bulk order for these boards. We'll need a minimum of 10 boards in order to get the price down to $20 each. More boards would bring the price down lower. I currently have requests for two. These are four layer standard boards produced by ExpressPCB.com.

Please note: I was asked to help design the board and used information from others to create the schematic. I have not tested it myself and cannot promise that it will work. Please look over the schematic and my layout before you decide to order. It can be found here:
http://sbc.rictor.org/support/conv.html

Thanks!

Daryl

As of March 29, I have a total of 4 boards requested. We'll need at least 10 to make a bulk purchase viable. You can contact me via email with any questions!

Daryl

Oh my god, I've been away of this forum for such a long time...

Count me in for two boards, and expect to have a few others from other Oric users: I'm going to post an announce in comp.sys.oric.
Shouldn't we do the same in all the 6502-based computers newsgroups ?
I'm sure we would get a lot of interested people in the Atari, Commodore and Apple communities...

Fabrice

Hallo allemaal,


> Check out the schematic, layout draft, and expressPCB files on my web > site.
>
> http://65c02.tripod.com/conv.html

Very strange, I checked the forum two weeks ago and missed this thread completely :( Even worse is that I maybe have to spoil the things more then a bit.

Why the 74HC245? I have used my module in various 6502 systems W/O any trouble.

If people want to use more the addresslines A16..23 and the original system uses the RDY-line, you maybe can come in trouble. I'm still doing research on it. The basic is that PHI1 is used to clock a 573 or equivalent to latch A16..23. But if RDY is (L), the databus will keep on outputting the data, also when PHI2 = (L).
My idea is to stretch PHI0 towards the 65816 until all is OK again. Cannot give you details at the moment as my notes on this subject are at home. FYI: you only need a 7402 and 7400 for the circuit

_________________

This is because some 6502 systems use the first half of the Phi2 clock (video display), so the module should not conflict with this use.

Best regards,

Fabrice

Hallo Fabrice


> This is because some 6502 systems use the first half of the Phi2 clock
> (video display), so the module should not conflict with this use.

I studied the SCH again and noticed that the 245 is tri-stated when PHI2 is (L). Hmmmm, I think I have it: the databus of the 6502 is always in input mode when PHI2 is (L). (IIRC, no datasheet to check it) I must remember this.
OK, I get it. But I then still have an objection: this design blocks the possibility to address above 64 KB. Unless you start to hack the PCB.

The above is only negative feedback but I hope to compensate that with some SCH's early next week.

_________________

Yes, you are correct. The purpose of this board is to provide a 65816 for use in an existing 6502 design, where 64k is the addressing limit.

However, if something could be incorporated onto the board (space permitting) to allow more addressing, then I would consider adding it.

Thanks!

Daryl

Thanks for the input. Before we can actually place an order, we still need to hash out something. We need a source for the pins that go into the DIP-40 pads. These need to be able to be soldered to the board and plug into an existing DIP socket on a target board. We need to find a source so that I can ensure the holes on the board are the proper size to accomodate them.

Anyone have any viable sources?

Thanks!

Daryl

> We need a source for the pins that go into the DIP-40 pads...
> Anyone have any viable sources?

Ok, if the Winslow AdaptICs idea above is a no-go, how about putting a row of pads (without holes) down each edge to solder a DIP header to? The board could be about .700" wide, protruding a little past the top of the DIP header pins for a good solder fillet. That would allow wider things to be put on the top instead of being forced to fit between the rows of pins. SMT parts can go on both sides. To solder the DIP header on, you would just tack two corners to hold it in place while you solder the other pins, and then come back if necessary to improve the joints at those two corners.