• Please review our updated Terms and Rules here

Memory Compatibility Z80 vs 8080

H

Hugo Holden

Veteran Member
Joined
Dec 23, 2015
Messages
4,773
Location
Australia
I read in a manual for a S-100 Vector Graphics memory card that it was for use with a Z80 and wouldn't work with the 8080. But they did not explain why. Since the card was designed for the S-100 bus (and should work with standard bus signals ?) what would it be about it that would stop it from working in an 8080 computer ?
 
I expect it is relying on the Z80 to run the DRAM refresh cycles. That is hinted at in the restriction of the longest DMA burst latency in the specifications section.
 
Note section 2.1.1. The RFRSH signal needs to be on pin 66 of the S100 bus. Pin 66 on the standard S100 bus isn't defined--and the 8080 doesn't produce a RFRSH signal, nor does it have a refresh counter or register. Could a refresh circuit be designed for an 8080? Sure. But 8080/8085 usually employ something like an 8202 DRAM Controller, which makes the whole affair much simpler.
 
Chuck(G) said:
Note section 2.1.1. The RFRSH signal needs to be on pin 66 of the S100 bus. Pin 66 on the standard S100 bus isn't defined--and the 8080 doesn't produce a RFRSH signal, nor does it have a refresh counter or register. Could a refresh circuit be designed for an 8080? Sure. But 8080/8085 usually employ something like an 8202 DRAM Controller, which makes the whole affair much simpler.
Click to expand...

Thanks, so its a hardware issue.

I have tried a Seatle Computer Products (SCP) 16k Ram card in the Sol-20 and it works fine. But I'm wanting a higher capacity single card and looking around at options. Obviously the PT 48k ram would work, but I can't get my hands on one. So I have been looking at other brands of 64k cards.

I wonder if the SCP 64k Ram board would work in the Sol ? since they made their 16k card compatible.

Recently I tried the YANG 32k card in the SOL, didn't work, but it could have hardware fault.

It there a known list of 48k to 64k cards that are known to be compatible with the Sol-20 ?
 
I like the 64K static RAM boards which use the 2K x 8 static RAM chips. These are low power and very reliable. Most can disable 2K blocks to make memory space holes as required for the Sol-20 and North Star FDC. Be careful though, some can only disable a single 2K block or a full 16K block. With most, you can also throw a 2716 or 2732 (half used) EPROM into one of the RAM sockets. I’ve had the best luck with the CompuPro RAM17 as it is easy to disable multiple 2K blocks and is compatible with front panel machines like the Altair.

For a DRAM board, the Central Data Corp 64K Board is uniquely configurable for a 64K board. It seems to be designed with early systems in mind that need various holes in RAM for PROM, video, FDC, etc.

Mike
 
deramp5113 said:
I’ve had the best luck with the CompuPro RAM17 as it is easy to disable multiple 2K blocks and is compatible with front panel machines like the Altair.

Mike
Click to expand...

Interesting, these boards carry the same HM6116P IC's as the Yang board I have been trying to get to work (but the Yang simply has half the number as its 32k) The photo of the Yang board is on this thread:

http://www.vcfed.org/forum/showthread.php?71036-YANG-single-board-computer-amp-32k-Memory

They also mention that 2k blocks can be easily disabled and 2716 Eproms can be used. Maybe the Yang board has a fault or needs the refresh signal of the Z80 system, I will investigate.

Can I deduce from what has been said by Chuck(G) in post #4 that if I come across a memory board that makes no connection to pin 66, that likely it will be 8080/Sol-20 compatible ?
 
Somewhere I have an old SD Systems 64K DRAM board that runs with an 8080 CPU, so DRAM boards for 8080/8085 CPUs do exist.

I don't know if anyone ever made an S100 SRAM board using the InMOS 16K SRAMs. Those were really fast (and probably expensive).
 
Last edited:
Hugo Holden said:
Can I deduce from what has been said by Chuck(G) in post #4 that if I come across a memory board that makes no connection to pin 66, that likely it will be 8080/Sol-20 compatible ?
Click to expand...

Almost certainly,yes--unless there's a weird one out there that uses the RFSH signal on a different S100 pin.
 
Chuck(G) said:
Almost certainly,yes--unless there's a weird one out there that uses the RFSH signal on a different S100 pin.
Click to expand...

Thanks.

I just checked on the Yang board, pin 66 not used. Si I think one of the IC's has failed probably in the address decoder as it has an LED that is supposed to light hen the board is active.

It has a lot of DIP switches as it can be set for extended addressing. It looks like address lines A 14 & A15 are used to select the A and B memory IC banks and later combined with A12 and A13 to create the Chip select signals, with A0 to A11 being the address lines for the 6116 IC's.

A16 thru A21 appear to be used for an optional extended address scheme.

If I set the A bank to start at say 4000H, I need to send an out instruction to see if I can activate the address decoder for a test...any suggestions ?
 
deramp5113 said:
If the Yang board uses the 2Kx8 6116 RAMs then it is a static RAM board which does not use any sort of refresh.

Mike
Click to expand...

Thanks Mike. Of course.

So now there are two ways to spot a likely compatible RAM board for the SOL - 20:

If its DRAM it must not use pin 66.
If its SRAM its ok.

I'll report what was wrong if I can repair the Yang board.
 
I recall making a small board with a TTL gate that I added to my Poly88 CPU card to handle a SD DRAM board. It was an 8080 processor. It just needed to tell the SD card when to do refresh. Some boards expect there to be an external refresh address counter, like the one you were looking at. This was a build in feature of the Z80 as it would automatically generate the address for 128 cycle refresh( but not 256 cycle as some 64K DRAMs required a full DRAM controller ).
Your static board should work fine with a 8080 processor.
At the VCFNW, I bought a 64K static board for $5 :). I removed the SD card I had in my Poly88 and used it at the LAST Maker Faire :(
Dwight
 
Do note that some banked cards need a write to some port address to be enabled. This should be easy to defeat once identified.
The S100 buss was designed to have one set of 8 bit lines for writing from the processor as data and another set of 8 bit lines for data returning to the processor. Some RAM cards were intended to be used with a 16 bit system and used both 8 bit busses for a bidirectional 16 bit bus. Look at how the bus buffers are connected.
Dwight
 
I have repaired the Yang board. And it works in the SOL-20. There were two issues. It has jumpers for normal S-100 mode addressing and another mode used with the YES-5 computer. But the silk screen labels on the board for the settings are incorrect for four of the jumpers, so it was not responding to the correct addresses. I figured this out from the schematic of the address decoder where it is correct and the manual is not wonderfully helpful otherwise. Once it started working then I discovered that one of the 16 HM6116P IC's was defective.

Are the HM6116P's thought to be a reliable IC or do they often require replacing or is it rare for a failure of these types?
 
There is a little oddity on this Yang board. There is an indicator LED which it driven on when any of the 16 memory chip select lines are active, the odd thing is that the designer didn't put a current limiting resistor in series with the LED and just relied on the gate's ability to sink current being limited, not a wonderful idea I would have thought. Has anyone seen LED's driven like this from TTL circuits before, with no resistor ? Circuit fragment attached, there is no resistor on the actual board either.

Also, looking at the jumpers on pin 5 of U9 it is possible if two were there, to short out the +5V rail. This probably explains why there were separate soldered jumpers , even though the silk screening on the board said DIP switch.It probably wouldn't have been a bad idea to have put a small series resistor on the connection to +5v.
 

Attachments

  • yes5.jpg
    yes5.jpg
    94.5 KB · Views: 1
Last edited:
That is strange--you're depending on the resistance of the pulldown transistor in the output totem-pole datasheet. Maybe it's enough to avoid toasting the LED. Oddly, you'd be better off reversing the LED and connecting the other side to ground, as the high-side output transistor has a 120 ohm resistor in series. Can't say that it would work well with a 74HC04, however.
 
Chuck(G) said:
That is strange--you're depending on the resistance of the pulldown transistor in the output totem-pole datasheet. Maybe it's enough to avoid toasting the LED. Oddly, you'd be better off reversing the LED and connecting the other side to ground, as the high-side output transistor has a 120 ohm resistor in series. Can't say that it would work well with a 74HC04, however.
Click to expand...

Double checked the device actually on the pcb its an LS04. I probably should put a 100R resistor in there at least, but I would have to modify the board, cut a track etc... hmmm.... maybe just leave it as the curio that it is.
 
If you hack in and measure current through the LED, you’ll find it’s less than you think - probably in the 20-40ma range.

Mike
 
You must log in or register to reply here.
Back
Top