Last updated July 10 2012. Edited by Herb Johnson, (c) Herb Johnson, except for content written by Lee Hart and others. Contact Herb at www.retrotechnology.com, an email address is on that page..
Here's the Rev D Membership Card support page. There are links there to Rev C support info also. Refer to the Membership Card home page for the current status of the kit and more information.. - Herb Johnson
Here is what I changed:
1. dual CPU board option, for ROM and RAM & more I/O. Support for stacking two Membership CPU Cards; to share the CPU chip with a RAM card and a ROM card. Here's how I did that:
a. Added jumpers (under the 1802) to pick an I/O port address.
- With no cuts/patches, it has the same OUT4/INP4 addresses as the previous board.
- By cutting a trace and adding a patch wire, you can get three non-conflicting addresses.
- This allows two (or 3) Membership CPU Cards to stack with different I/O port addresses.
b. Allow memory to be addressed from 0-32k (8000H), or 32-64k. ONe can be RAM, the other ROM. I added a pad so transistor Q1 can be installed 2 ways:
- The first way is just like the present board. Memory is from 0-32k, and only active during TPB. - The second way puts memory from 32-64k. But it is active during the entire bus cycle, so power consumption will be higher (like the Rev.A board).See this tech note about the changes from Rev A to B, for an explanation.
[Here's some details about Q1.] There are 3 pads where the leads of Q1 go (gate, source, drain). I would add a 4th pad. Let me number these 4 pads 1-2-3-4.
1 = /CE of memory (drain of Q1, and pullup resistor)
2 = TPA (from 1802)
3 = A15 (from address latch)
4 = VSS (ground)
0000: Q1 goes in 1-2-3. gate = TPA, source = A15.
/CE goes low when A15=0 and TPA=1.
8000: Q1 goes in 1-3-4. gate = A15, source = ground.
Q1 inverts A15, which directly chip selects memory.
There is so little room that I can't label these pads on the board; it will depend on the manual to get the right holes. I don't like it, but don't see a choice.
c. Stack two CPU boards through the CPU's 40-pin socket. To use two CPU boards, install a 40-pin wirewrap socket on the top board, and plug the 1802 into this socket. Cut off conflicting pins, such as pin 1 (CLOCK) of the wirewrap socket so the clock on the lower board doesn't conflict with the clock on the upper board. Plug the bottom ends of the wirewrap socket into the lower MC card (which has no 1802, of course). Jumper the I/O to be at different addresses for the two boards.
You can use the Molex sockets I supply for the 30-pin connector, though they would have the effect of offsetting the upper board by about 0.1". The wirewrap pins can be bent .1" to compensate. Or, use non-offset Molex sockets.
2. 8-bit I/O for OUT and IN. Support to use the 25-pin D-connector (J2) for general purpose I/O that is not PC printer-port compatible. Here's how I do that:
a. Added a jumper to enable/disable the 74HC257 multiplexer.
- Shorted, it works like the previous boards. 4 pins in J2 are OUT Q0-Q3 when IN is low, OUT Q4-Q7 when IN is high. (IN is the input pushbutton or J2 STROBE pin 1).
- Open, it leaves Q4-Q7 on these 4 pins all the time, regardless of the state of IN.
b. Eliminated grounds on J2 pins 21-25. Added jumpers to
connect them to OUT Q0-Q3 instead.
- Jumpers open leaves J2 pins 21-25 floating. They were were grounded on previous boards. Leaving them open shouldn't matter, as there are 8 ground wires.
- Jumpers Q0 Q1 Q2 Q3 installed connects OUT Q0-Q3 to J2 pins 21-25 so you have all 8 parallel output bits on the D-connector.
- it's *NOT* PC compatible if the jumpers are installed, as the PC would short the outputs to ground!
3. Added +5V in on DB-25 pin 18, to power the card.
Added diode D12; anode to DB25 connector J2 pin 18, cathode to VDD. Pin 18 was grounded, again; we don't need 8 ground pins; I think the remaining 3 are enough.
- PC compatible; Used with a PC printer card, the diode does nothing. The PC card grounds the anode of the diode.
- When J2 is used for general purpose I/O, you can use pin 18 as a +POWER pin. The diode has a few functions:
a. "Idiot" diode; blocks if power is hooked up backwards. In the previous design, backwards power made zener diode D11 conduct, shorting the external supply.
b. Allows 5v +/10% external supply without overvoltaging zener D11 or supercapacitor C5.
c. It will probably work even with an AC supply, or an RS-232 data line! :-)
d. It serves as the blocking diode for solar panel, so it won't discharge the supercapacitor if the panel is dark.
4. Added jumpers between the twin rows of pads for the Front Panel 30-pin connector. This was at the request of a customer, who wanted to put the female connector on the MC card, and the male pins on the FP card. There were holes for the pins on the FP card, but no pads on them and they weren't connected to the pads for the female connector.
I didn't have room to put both male and female holes on the MC card. But there are sockets for 0.025" square pins whose solder pads are right in line with the male pins. He'd have to use this type of socket on the MC card.
a. Minor tweaks to increase spacing (less chance of solder shorts) and to improve silkscreen legibility.
b. Took the soldermask off the Cover Card's foil side. As it was, you had to scrape it off to solder it to the Altoids tin.
c. Option for lower-power blue LEDs. If you look in the picture in my latest manual, you'll see that my Membership Card has clear (instead of red) LEDs. Those are blue LEDs. I also changed the LED's 1k series resistor network to 3.3k, as they are *way* too bright at 1k. If people are willing to pay more [for blue LED's], it would be worth adding it as an option.
d. Reduced the hole size for the toggle switches to allow a different model/brand switch than in Rev C.
- Lee Hart
Here's a discussion from Jan 2012 about ways to modify Rev C CPU boards to support stacking two cards. The point of using two CPU cards, is to use one to support ROM, the other to support RAM. Also, it provides an additional IN and OUT port. To do those things without conflicts of signals, some cuts and wiring is needed. Rev D, among other things, makes that easier to do.
Lee Hart also provided these related and similar comments, when announcing the Rev D kits. - Herb Johnson
It would be a challenge to write up a comprehensive manual that was good enough to allow a novice to cut-n-patch a rev.C board into a rev.D. While the changes are pretty simple, actually *making* them is hard due to the close spacings and small traces.
But I can outline the basics. All changes are made to the Front Panel card.
1. Cut the trace to U9 pin 1. 2. Add a jumper wire from U9 pin 1 to VDD. 3. Cut the foil trace on the bottom of the board that ground J2 pins 18, 22, 23, 24, and 25. 4. Add a new diode (D12, 1N4148). Anode to J2 pin 18, cathode to VDD. Pin 1 of any of the resistor networks (R11, R12, R13) is a convenient place to find VDD. 5. Add a jumper wire from the anode of LED D3 (side toward the toggle switches) to J2 pin 22. 6. Add a jumper wire from the anode of LED D2 (side toward the toggle switches) to J2 pin 23. 7. Add a jumper wire from the anode of LED D1 (side toward the toggle switches) to J2 pin 24. 8. Add a jumper wire from the anode of LED D0 (side toward the toggle switches) to J2 pin 25. 9. Cut the trace to J2 pin 1. Add a 5.6k resistor between J2 pin 1 and R13 pin 2.
That's basically it! There were other changes, but they are cosmetic or physical, and not worth making to a rev.C board. For example, I added jumpers to allow changing the I/O port address, and to remap memory from 0-32k to 32-64k, and reduced the hole size for the toggle switches to allow a different switch (some people had trouble with the ones I sent with the rev.C boards, so I'm trying a different switch vendor).
- Lee Hart, July 2012
This page and edited content is copyright Herb Johnson (c) 2012. Contact Herb at www.retrotechnology.com, an email address is available on that page..