Is anyone interested with the Z80?

Did you know that the Z80 has different memory access timings for instruction fetches and standard memory accesses?

I worked at Mostek maintaining the Fairchild Sentry testers used to test the Z80s when they first began to manufacturer them. The second computer I every

built (parts, pcb, solder) was the SDB-80 which was the Mostek Software Development Board for the Z80 with 16K bytes (8 - MK4116 16K bit) dynamic RAMs

which Mostek also made. The first wire wrap board I ever made was a 48K byte dynamic RAM board also using MK4116s that worked with the SDB-80. I can

tell you there were a bunch of happy technicians when we were told that Mostek would be second sourcing the Z-80 and even happier when we were told we

could buy SDB-80 kits and build them ourselves! Those were really fun days. I still have a Zilog Z80 in a ceramic package that says it was manufactured in

Dallas even though Mostek was actually in Carrollton a suburb in Dallas County just north of north-west Dallas.

Are you still in the Dallas area. So am I.

Clem

Wow! that's amazing that you got such a large DRAM board working with Wire-wrap.

Do you still have any photos or schematics of it?

It was quite rare to come across Mostek parts in Australia. We usually got Zilog, SGS and Sharp.

And from memory the SGS parts didn't support the undocumented instructions.

Yep, lived here all my life (more years than I want to admit to).

No photos but the schematics were from a Mostek Z80 application note which I may still have. I used a tool from Vector called the Slit and Wrap which used

small spool of wire coated with a soft insulator kind of like magnet wire that was slit by the wrap bit just before it was wrapped on the post. It was a great tool

for speeding up wire wrapping with one fatal flaw. If two wires pressed against each other hard enough, they could displace the soft insulator and short.

When ever I powered it up and it did not work I would have to look for two wires pressed against each other in a rats nest of wires (actually I was more

systematic than that but there were always problem areas) and separate them. When the offending wires were separated it would usually make a distinctive

"plink" sound so I knew when I found it.

I used the same tool and a Z80 to wrap up a 256 x 240 x 4 bit/pixel graphics display controller. The schematic was from Byte Magazine and I modified it from

the original 1 bit/pixel configuration and to use static RAMs that were fast enough to allow the Z80 full speed access to the video RAM by interleaving the

accesses with video timing generator. The Z80 assembly language software had to be entered in by hand to be assembled and of course modified to handle

the multi-bit/pixel mod and to use a 8 bit parallel interface to the SDB-80. I believe it only had commands for pixel read/write, line and circle drawing.

Unfortunately after quite a few sessions of find the Slit and Wrap short I gave up and disassembled it for the parts and never used my Slit and Wrap again.

Like I said, fun (and sometimes frustrating) days.

Ahhh!

The days of triple power supplies (+5, +12 and -12V).

Slit-N-Wrap from Vector.

It's still available today but not on the e14 catalogue.

I used a competing product called JUST WRAP and I think I've still got the blighter!

Yes, three voltage power supplies ruled then. And you can add +24V for the 8" floppy drives but they were usually in their own case with a separate

+5V, +24V power supply. The MK4116, and all other dynamic RAMs of the time used +5V, +12V, -5V which was usually derived from the -12V which

came in handy for  RS-232RS-232 drivers

I thought it was Slit-N-Wrap but I couldn't remember so I went for the long version. I thought Vector was still around, I wonder if they fixed the shorting

problem ?

I remember the Just Wrap too and I know I was very interested but for some reason, lost in the mists of time, I never bought one which is made even

more curious after all the Slit-N Wrap problems. I think I just went to pre-cut wire wrap wire (which I still cut to length) with an electric wire wrap tool but ...

I still have a motorized Slit-N-Wrap and I still use it once in a while. It used to be really fast for wiring low-speed backplanes - much faster than manually soldering connectors on PCBs. I built a circuit that would make it automatically stop wrapping after 7 turns.

I never had problems with shorts. They have 2 types of wire, one with a Tefzel (Teflon) jacket that is thicker than normal Kynar, the other is  a thin green enamel coating (a bit like magnet wire). I never had problems with either type although I was always careful not to let the green stuff rub on square pins.

The Slit-N-Wrap wire is 28AWG and normally 30AWG is used for Wire Wrapping.

I usually use 24, 26 and 28AWG for prototyping because I find that 30AWG is a bit too flimsy.

Dobs of glue are also useful for holding wires in place.

COMPACT  wrote:

 

It was quite rare to come across Mostek parts in Australia. We usually got Zilog, SGS and Sharp.

And from memory the SGS parts didn't support the undocumented instructions.

The latest post in this thread was recent enough for me to dare 'reply' to this 2016 entry - hope that's OK with you ?

During the '80s I've been professionally involved in a number of commercial projects, built around the PC-1500A pocket computer from Sharp.

This was often referred to then as a 'BASIC programmable calculator', though its functionality, capabilities, and the actual applications surpassed

that status by far.

With only superficial knowledge of microprocessors, I've always wondered about Sharp's own LH-58xx family. As far as I know, it has only been

utilized in the PC-1500 and its successor, the PC-1600. It wasn't until recently that I learned that Sharp had (and still has) its own line of OEM

4 and 8 bit microprocessors. The LH-5080 e.g. is offered as a member of the Sharp  Z80 family, LH-0080.

For me this raises several questions:

- Is the LH-5801 really a Z80 derivative, as most sources over time have stated ? Doesn't the model number, 5800, seem odd, even peculiar,

   when one realizes the importance of the Motorola 6800 in those days ?! In other words: LH-0080 makes sense, better still: is clear - LH-5800

   is confusing. The only explanation I could think of, would be, Sharp meaning to discern her OEM products from her microprocessors intended

   only for use in her own apparati - i.c. her line of pocket computers. Btw Sharp herself makes no mention of the Z80 (nor any other microprocessor

   for that matter) in her Technical Reference Manual. (http://www.pc-1500.info/Data/Service_Manuals/PC-1500_Technical_Reference_Manual.pdf )

- Anybody has a clue as to what the letters, LH, stand for, and what the 5 might refer to ?

  (Sharp being proud of her early chip design and production capabilities, called one of her seventies calculator lines: Elsi Mate, where Elsi stands

   for ELSI - Extra Large Scale Integration)

- The CMOS LH-5801 had an impressive number of on-chip facilities/peripherals - probably more than the (standard ?) Z80.

   I have not been able to find a (useful*) comparison with the Z80 itself, derivatives, or other microprocessors from the same era. Anybody ?

    *(cpu-world hasn't brought me any further).

Thanks or answering !

Ray

Hi Ray,

I've not looked at Z80 for an extremely long time, so I could be wrong, but looking at this PC1500 machine code article (PDF) the instructions do not look like Z80.

For instance, on the last page of the PDF, there is this example program:

Taking an example, the red instruction, the Z80 does not have an RTS instruction according to here: Z80 / R800 instruction set

There's a RET instruction for the Z80, but it has a different opcode too (0xC9, not 0x9A).

Also, according to here: Frank C. Odds : Probe! a disassembler for the Sharp PC-1500 - PC-1500.info

the PC-1500's CPU only has three registers.

So, I believe that while it may have similarities to CPUs such as the Z80 or indeed any other CPU from the time, it's not a Z80 derivative (or not a very close one), nor compatible with z80 at mnemonic or opcode level.

Hi Ray,

I've not looked at Z80 for an extremely long time, so I could be wrong, but looking at this PC1500 machine code article (PDF) the instructions do not look like Z80.

For instance, on the last page of the PDF, there is this example program:

Taking an example, the red instruction, the Z80 does not have an RTS instruction according to here: Z80 / R800 instruction set

There's a RET instruction for the Z80, but it has a different opcode too (0xC9, not 0x9A).

Also, according to here: Frank C. Odds : Probe! a disassembler for the Sharp PC-1500 - PC-1500.info

the PC-1500's CPU only has three registers.

So, I believe that while it may have similarities to CPUs such as the Z80 or indeed any other CPU from the time, it's not a Z80 derivative (or not a very close one), nor compatible with z80 at mnemonic or opcode level.

One also has to remember that R650x and Z80 had altered instruction sets to evade intellectual property claims.

Sharp did make Z80 products and to me the PC-1500 looks like a Z80 inspired CPU with the index registers missing (like the gameboy CPU) and altered object code and a disguised instruction set.

e.g. BC = X pair and DE= Y pair.

The Texas Instruments TI-84 Plus CE Calculator that does have a eZ80 (a Z80 derivative) in it.  With it one can play Gameboy and other games with it.

The processor itself is running at 48MHz but is using wait states to interface its slower memory and i/o.

If you do have one, don't upgrade to the latest firmware because it places unwanted scholastic restrictions upon it.

Edit media

Dimensions x Small Medium Large Custom

Subject (required) Brief Description Tags (separated by comma) Video visibility in search results Visible Hidden

Parent content

RE: Is anyone interested with the Z80?

Poster

Upload Preview

Ok

Cancel

I concur those don't look like Z80 comands.

As an aside, when playing with my ZxSpectrum I found an interesting scenario when I was stepping through some code. It ran fine at full speed but errored in the debugger. There turned out to be a couple of commands that the chip knew about but the debugger (and documentation) did not.

Turns out that the undocmented commands are now documented

http://www.worldofspectrum.org/z88forever/dn327/z80undoc.htm

The CPU used for the PC-1500 is the LH5801. The native mnemonics have been made to look like 6800/6500 style ones.

Here is a quick partial cross reference between its mnemonics and Z80 equivalents.

http://www.pc-1500.info/Data/LH5801-ASMCodes_Sharp.htm

That's only a partial list of undocumented instructions.  There are various sets posted around the Internet including Z80 Undocumented Instructions .

Also be aware that some instructions are used for alternate purposes of later generation Z80s and derivatives such as the Rabbit 2000, Z180, ez80 etc.

For example the relatively useless LD L,L (0x6D) is not supported by the eZ80.

COMPACT  wrote:

 

The native mnemonics have been made to look like 6800/6500 style ones.

 

Interesting. For what reason would they do that? So it can look like a 6800/6500 derivative too? Or some other reason?

Hi Shabaz,

Thank you for your reply. This was actually the first time ever (I'm 73) I participated in any professional community discussion on the Net.

So please don't mind my many newbie mistakes - one of them not being a prompt replier.

For everyone else here, interested in the Sharp PC-1500(A), your link to the pc-1500.info site is spot on, as this site offers the best,

i.c. most comprehensive info there is.

The LH-5801 microprocessor turned out to be a very powerful tool indeed in many applications of this pocket computer.

One of the projects I've been involved in was at a Nestlé plant for the production of powdered milk. The Swiss manufacturer

of the 5 parallel operating in-stream weighing-doseing units to be installed, required 5 full-fledged PCs as controllers - the chosen

brand being Commodore. The engineer in charge however judged this brand to be below industrial reliability standards, and

proposed the PC-1500A instead, i.e. 1 pocket computer instead of 5 PCs ! (a printer - not its 4-color plotter mate, CE-150 - was added).

Later system SW for real multi-tasking was developped by a third party - we utilized a MUX.

Before the mid-eighties there was also assembler/disassembler, hexmonitor, etc. software available; furthermore a C-compiler was offered,

along with Forth and several extended BASICs.

From your and Compacts reply it would seem the LH-5801 had only

(continued in post nr. 106)

Hi Ray,

That's interesting that it was used to control part of a Nestle production line. I do recall an episode on TV of 'How It's Made', which showed a production line in a factory which made piano music rolls, and part of it was controlled by an ancient Apple IIe machine. Since it worked, I guess the factory felt no need to port the code to anything more modern. Better to trust what is working : )

Into modern times, and a lot of retail stores in my area run very suboptimal hardware - consumer PCs! - always clogging up with dust from the Manager's office, running slow, and on the floor getting kicked around.

Hi Shabaz,

Let me try to recapture the lost part of my post with the Nestlé story first.

I meant to argue that your and COMPACT's reply suggested the LH-5801 had less hardware facilities than the Z80:

- "the PC-1500's CPU only has three registers" (you)

- "a Z80 inspired CPU with the index registers missing " (C)

For me this makes it a bit hard to understand, how a pocket computer with this CPU has been able to achieve such extraordinary results

in numerous high-level applications. Or is this just a matter of clever trickery by the application programmers ?

First of all (as far as I know), other than Z80, LH-5801 had 8-bit (full-width) registers - esp. in the accumulator.

Program counter and stack pointer were even 16 bit.

Secondly some items from the feature list of the Sharp CPU:

- (3) Use of a 6-byte general purpose register. in addition to the accumulator. allows to comprise three pairs of 2-byte date pointers. (misspelling Sharp)

-  (12) Reducing program steps by means of 28-kind single step vector subroutine jurnp

If only there were a feature-by-feature or architecture comparison of the two microprocessors ...

Judging from the applications side (remember I lack sufficient knowledge of micoprocessors), I would say, LH-5801 was - maybe inspired by some

standard Z80, 6502, or even 6800 - yet an advanced proprietary design with a remarkable number of on-chip peripheral facilities. Your and COMPACT's

observations concerning family traits, in one way or another, are very interesting and seem to confirm my assumption all along, that LH-5801 was just not

your run-off-the-mill Z80 clone.

As stated before, the Nestlé project utilized a MUX. Later someone else devised real multi-tasking software for the PC-1500A. I suppose that requires

a quite sophisticated interrupt scheme/structure (esp. at a clock rate of only little more than 1 MHz), or can one do this by clever programming via a

high-level language ?

Last but not least: thank you all for a friendly welcome into this community !

Cheers

Ray

Hi Ray,

The Nestlé project sounds fascinating.

You'd expect the LH-5801 to be superior to the Z80 for its designated applications. The designers could pick, choose, optimise and add the features that they wanted.

It is was unlikely that a LH-5801 device needed to have the DRAM memory and 8080 legacy support because it was for low power consumption devices such as battery powered calculators.

Itsinbuilt timer is very handy.  The Z80 required the addition of a CTC (Counter Timer Counter) or similar to do this.

There's nothing wrong about creating a CPU that's lacking the Z80's index registers. If fact these index registers can be a pain to use. The Game Boy version of the Z80 is also missing these index registers and look at the quality of the software for that.

If you look at most Z80 code, the IX and IY index registers if used, were seemingly used as a last resort. This is because the Index Register op codes need an 8 bit offset to be specified within the instruction.

It's also interesting to note that Cristophe Gottheimer's assembler for the LH-5801 uses the Z80 JR (jump relative) and mnemonics instead official Bcc (Branch) mnemonics. A clue to the LH-5801's origins perhaps?

One thing people always seem to overlook is that the Z80 limitations were set by the memory device performance limitations of the day. The designers at Zilog even incorporated inbuilt wait state support to handle slower memories and DRAM refresh support.

This this in mind the Zilog designers worked out that they could use a 4-bit accumulator and run it twice for 8 bit results in the time taken for a single memory fetch.

At the time common memory devices were small with 1K * 1 Static and 4K * 1 Dynamic RAMs being the most affordable. The larger and faster ones came years later.

Cheers and welcome to the community.

Very!

Hi Ray,

At around that time, a lot of the handheld devices were getting powerful, so the ultimate decision to use that instead of any other device may have been simply because it was what someone was experienced in.

From the very little production/shop floor experience I have, it was interesting (at the places I worked) to see that the tools used in production were different to what the rest of the organisation were using. In other words, different CAD tools, different procedures, and different hardware. I'm not saying it is ropey, but a lot of stuff is designed to be one-off or very low quantity production.

Regarding the suggestion of Z80 origins made to look like a non-Z80 device, I've no idea. Certainly it won't execute Z80 software without modification and assembly with a different assembler.

Many processors at the time were similar, from a birds-eye view. Even the different processor designs that soon emerged like RISC and SPARC still have instruction sets and registers, and buses.

There was historically plenty of anti-Japanese tech sentiment in the 1980's, when many people thought that Japanese firms were copying or stealing ideas. That idea still exists today. It's unfounded, and likely no more than western firms stealing western ideas from each other. Perhaps that sentiment inspired the suggestion that it looks similar to the Z80 I cannot tell, since I don't know where and when that suggestion originated. The Japanese government encouraged firms to buy rights to patented tech, that was a major policy at the time.

Depending how keen you are to research down that road, there are some books worth reading, they mention names of people who would know better than speculation, of which unfortunately there is a lot of on the Internet. The first is We Were Burning, and the second is Industrial Innovation in Japan. The first is an easier book to read. According to that, Sharp was extremely quick off the block developing calculators, and amongst the first to productize (not invent) CMOS, so it's quite natural they would either license a Z80 design, or design their own chip. Neither of which would be good reasons to fake instruction names to look like a different processor.

Hi COMPACT,

Sorry for my late response.

Thanks for your friendly reply, and for sharing your broad knowledge to fill-in some of the white spots in mine.

Further to the Nestlé project (and risking being thrown out by the moderator on account of wandering too much off topic ...): the 'poor'

PC-1500A had to communicate with 68000 processors in the Swiss units. When the info required to tackel this problem didn't come in time

to meet the project deadline, I decided to try and overcome the speed differences ... in BASIC (call me crazy). The only solution to get around

a formal handshake was applying probability calculation. It worked fine and reliable.

Reading the history of the Z80, it seems safe tot conclude that Sharp must have been (one of) the very first (1980/81) to produce a CMOS

derivative (?) with many on-chip extras. Just out of curiosity I've once tested the PC-1500A for power consumption (keep in mind that there

was also an I/O controller on board, the LH-5811). Battery life was specified as 300 hrs, i.e. with the usual unrealistic operational limitations.

So after a fortnight of continously running an endlessly repeated program with heavy duty calculations, that made as much use as possible

of display functions, and most of all of the BEEP option, I expected to find my PC-1500A 'dying'. But after more than 3 weeks only the LCD

contrast showed some signs of a nearing 'low battery' condition. Unfortunately I had to abort the experiment, simply because I needed the

pocket computer for my programming work.

The first assembler, as far as I know, was made by Rasso von Schlichtegroll, fairly soon after the PC-1500's introduction; later he developped

a successful line of system programs and tools under the RvS brand, eventually commercialized by the Holtkötter publishing firm.

Jumping to the present, recently produced emulators may be the best way to keep the PC-1500(A) alive (and leave curious young users

in awe). As to the limited storing capacity of RAM chips in those days: the LH-5811 has a direct adressing capability of 64 kB - nothing fancy;

but because of the PU and PV flipflop registers, the full addressing range was 512 kB. And would you believe it: in the early eighties a small

third party company, AMS in Germany on the Danish border, managed to pack this huge amount of storage inside the pocket computer !

I'm not sure if/how his maybe conflicted with an eventual RAM module in the bay for external memory ... Later they had their Gold Card

produced in Taiwan for the top models in Sharp's organizer line: 256 kB, while Sharp was still fiddling with 32/64 kB.

Didn't know the TI-84 was Z80-based; always wanted to have one, never got around to it ... The origin of my interest stems from the late

seventies, and was thus calculator-oriented. I could go now into the old AOS vs. RPN debate, but this would be a) off-topic, and b) maybe

not of interest to anybody in this group. Suffice it to say that It is useless to try and grasp Sharp's design concept for its pocket computers

without taking her EL-5100 scientific calculator into account.

Cheers

Ray