PDA

View Full Version : LOGOS-I RAM Board



glitch
March 31st, 2011, 03:10 PM
Since I'm in S-100 mode, I decided to check out two problematic S-100 RAM boards. One of them is a SRAM board populated with many 21L02s...an 8K board by the number of SRAMs. The only markings on it are battery connection, MEM PROT, an area labeled with 0-5, and LOGOS-I, which I assume to be the manufacturer or the system the board was used in.

Anyone ever heard of these and/or have a schematic? There's another 21L02 in a random 16-pin socket on the board...pretty sure that's not supposed to be there!

scorch
April 4th, 2011, 07:58 AM
Hi Glitch, I vaguely remember that board, but can't help you specifically. I worked for Cromemco from about 1976 to 1980, and built probably half a dozen S-100 kits from various mfr's or cobbled together out of components. There was a plethora of memory boards around back then from 4k to 64k as any old issues of Byte or Kilobaud would show ads for.

glitch
April 4th, 2011, 08:06 AM
Yeah, it seems that unless the board was wildly common or you got the original manual with it, there's not a lot of hope for finding documentation! So many S-100 kit boards out there...oh well! I ordered a new logic probe, which will detect pulses as well as showing static logic states, so I should be able to leave a program loop accessing memory and try to figure out what the decode logic on the board is doing.

Interestingly enough, this RAM board uses the same dual TO-220 heatsink that Cromemco used with their boards. What's more, it's being used with a ZPU and 4FDC! (I'm getting a Cromemco Z-2D back up and running)

scorch
April 4th, 2011, 10:24 AM
Those 2102's were pretty straight forward 1k x 1 bit devices, so were generally arranged in 8 chip rows. This being the case, look for simple 74138 (1 of 8 ), or 74139 (1 of 4) TTL decoder chips, to directly translate the address bus lines A9 .. A11 to the chip enable mapping. A single 74138 would allow an 8K board to be designed without a lot of other signaling required.

glitch
April 4th, 2011, 10:52 AM
Those 2102's were pretty straight forward 1k x 1 bit devices, so were generally arranged in 8 chip rows. This being the case, look for simple 74138 (1 of 8 ), or 74139 (1 of 4) TTL decoder chips, to directly translate the address bus lines A9 .. A11 to the chip enable mapping. A single 74138 would allow an 8K board to be designed without a lot of other signaling required.

Yup, I'm guessing it's two 74LS138s doing the decoding -- there's an 8-pin socket that had another 2102 stuffed in it, which I doubt was correct. I haven't fully followed the traces yet, but I suspect that I will find it's supposed to be a '138.

scorch
April 4th, 2011, 04:15 PM
Actually, after building quite a few boards back in the 70's, I don't remember if the 2102 tristated its outputs to the data bus? If not, then that extra chip could be a tristate buffer. I also stand corrected, the address lines to the decoder would have been A10..A12, because A0..A9 went straight to the 2102 chip pins for the 1k span. Of course you realize that today a single 28 pin 62256 static RAM available for $2 or $3 replaces an entire board with 32 rows X 8 columns of 2102's. Lets see, at 8 chips/row that's 256 chips - LOL - and the 62256 probably uses 1% of the power.

glitch
April 4th, 2011, 05:13 PM
Of course you realize that today a single 28 pin 62256 static RAM available for $2 or $3 replaces an entire board with 32 rows X 8 columns of 2102's. Lets see, at 8 chips/row that's 256 chips - LOL - and the 62256 probably uses 1% of the power.

Heh, yes, I actually wire-wrapped an 8K board with a 6264 and 4 other ICs: two 74LS245s, a 74LS00 and a 74LS138! I recently picked up a bunch (100+) of 62256 SRAMs for around $0.20 each, so there may be a proper etched board that comes of that...

scorch
April 5th, 2011, 09:10 AM
Hey great buy on the 62256's. Those chips come in handy for all sorts of projects from single board computers to full bus boards. I think I paid around $2.50 apiece at Futurlec.

glitch
April 5th, 2011, 09:21 AM
Hey great buy on the 62256's. Those chips come in handy for all sorts of projects from single board computers to full bus boards. I think I paid around $2.50 apiece at Futurlec.

I picked them up mainly for the second version of my 8085 SBC:

http://www.glitchwrks.com/8085projects

I'll be adding 32K RAM in the upper half of the address space, with a /MASK line to allow other offboard devices to overlay RAM/ROM as needed. The new board will also be able to write its own EEPROMs -- very handy as I plan to build a serial interface board that will provide a console monitor!

MikeS
April 5th, 2011, 09:25 AM
Hey great buy on the 62256's. Those chips come in handy for all sorts of projects from single board computers to full bus boards. I think I paid around $2.50 apiece at Futurlec.Hey, nice to meetcha! I was a Cromemco sub-dealer starting around '82, a little after your time there. Had about 20 installed systems, the last one of which was actually in use for 18 years until 1999 (with upgrades along the way of course); still have a few taking up space in my basement and on my porch.

And yeah, those higher capacity SRAMs sure make many projects more fun and less drudgery... ;-)

mike

scorch
April 5th, 2011, 11:22 AM
Hi, Nice meeting you too MikeS. Yeah, I left Cromemco and moved on to Apple in 1980 joining the Lisa engineering team as a test engineer. Both very cool jobs for computer geeks though. :-)