Hidden MSX hardware

Page 9/14
2 | 3 | 4 | 5 | 6 | 7 | 8 | | 10 | 11 | 12 | 13 | 14

By Piter Punk

Master (224)

Piter Punk's picture

11-11-2016, 03:03

l_oliveira wrote:

That's a Hitachi HD64180, the same one seen on the high end JVC/Victor MSX2, as turbo mode CPU, no? (edit: seen it mentioned on the article)

It's the original chip which gave origin to the Z180 Zilog sells.

The Zilog from eighties and nineties seems to not have much skill to make 16-bits Z80 processors. The Z180 is licensed from Hitachi and you can find on internet some old news talking about Zilog making a deal with ASCII, to license the R800.

l_oliveira wrote:

Edit2: Modem chipset manufacturer is Sierra Semiconductor which left the modem business in 1996. Later it changed name to "PMC-Sierra".

https://en.wikipedia.org/wiki/PMC-Sierra

Interesting. Maybe Zeal have a pile of modem chips, or find an equivalent somewhere to keep selling E/PASO until 1998 (or maybe this is the reason it stops selling it?)

PAC wrote:

It sounds like a boosted MSX2+ :) .

Yes, a boosted MSX2+ is a good definition. To me there are two downsides on this machine:

  1. Too little RAM and
  2. No expansion capabilities

If you ignore these two items, it's an amazing MSX2+, very compact, with an embedded modem and is the *first* MSX based appliance that I found which doesn't have a superimposer!!!

PAC wrote:

The most notable aspect of this is how the MSX architecture was the ideal (and versatile) platform in which several companies decided to base their devices.

The MSX architecture was planned to be versatile and "embeddable", there are articles talking about to include a MSX-on-a-chip inside TVs or VCRs. In one of those articles there is a picture with Nishi showing an MSX-Engine from Toshiba (probably the T7937). I believe that some of those ideas got implemented, the hard part is to prove that.

The process to search from MSX based appliances is a bit paradoxal. For each hidden machine that I found, more I think about where are the missing ones. I don't feel closer to an end.

PAC wrote:

Again thank you for these interesting articles.

You're welcome! It's been a very exciting journey and is nice to have fellow travellers.

By maxis

Champion (512)

maxis's picture

11-11-2016, 22:50

Piter Punk wrote:
l_oliveira wrote:

That's a Hitachi HD64180, the same one seen on the high end JVC/Victor MSX2, as turbo mode CPU, no? (edit: seen it mentioned on the article)

It's the original chip which gave origin to the Z180 Zilog sells.

The Zilog from eighties and nineties seems to not have much skill to make 16-bits Z80 processors. The Z180 is licensed from Hitachi and you can find on internet some old news talking about Zilog making a deal with ASCII, to license the R800.

Pardon me, but Zilog by all the means had an extraordinary skills in building chips of ANY complexity by those days.

Talking about Z80 compatible line, Zilog never considered Z80 as an important and viable business in mid 80s. They were into the CPU wars with their Z8000 and Z80000 plans gradually loosing the edge. They were into high performance computing and industrial automation with their Z8 controllers.

But, hey, I still do have a board with Z280, which by all the means is an extraodrinary 16 bit processor and yes, it can run natively Z80 code! And it has a cache memory and it's high performance.

HD64180 or even R800 are not even get close when comparing their complexity to Z280 or Z380.

BTW when Intel acquired NVidia they didn't kill their x86 Tegra line. Having Tegra within Intel doens't mean that Intel doesn't have a good skill to build x86 processors Cool

But apart from all this, thanks for publishing the findings in your blog! Excellent post! Very interesting hardware. S3527 is used up to its spec -> MSX1 engine, no mapper.

By Lord_Zett

Paladin (807)

Lord_Zett's picture

11-11-2016, 21:44

that Zeal E/PASO is a great find. where did you buy it. i love to look into that one to. good work!

By Piter Punk

Master (224)

Piter Punk's picture

12-11-2016, 03:02

maxis wrote:

Pardon me, but Zilog by all the means had an extraordinary skills in building chips of ANY complexity by those days.

Talking about Z80 compatible line, Zilog never considered Z80 as an important and viable business in mid 80s. They were into the CPU wars with their Z8000 and Z80000 plans gradually loosing the edge. They were into high performance computing and industrial automation with their Z8 controllers.

Ok, so Zilog had the skill, but not the will to create a 16 bits processor compatible with Z80. So they licensed it from other sources.

maxis wrote:

But apart from all this, thanks for publishing the findings in your blog! Excellent post! Very interesting hardware. S3527 is used up to its spec -> MSX1 engine, no mapper.

AFAIK there is no MSX machine with native turbo that uses the S1985. The Panasonics uses the T9769, the Victors uses discrete components, the video titlers uses different custom gate-arrays and this Zeal machine uses the S3527. There is something wrong with S1985?

Lord_Zett wrote:

that Zeal E/PASO is a great find. where did you buy it. i love to look into that one to. good work!

There is still time to take a plane and go to MSX Jau. I'll be there Saturday and Sunday with this machine and the NIA-2001. So you can look the machines live.

I found and bought the E/PASO from Yahoo Auctions. It was the only one that I saw in one year interval. If you see one for sale, keep in mind that this same case was used by other equipment from Zeal, as I said in the blog post.

The process to choose an item to buy is a bit hard and too much error prone. The E/PASO was the result of a quest for communication devices based in MSX.

This quest, I need to say, was not very sucessfull, have a lot of "Not-MSX" to post in the blog: three CAPTAIN terminals from NTT and one Homemaster from Fujitsu, to connect to JRA-VAN.

By Lord_Zett

Paladin (807)

Lord_Zett's picture

12-11-2016, 09:36

i cant go to the msx fair. i wish i could

By maxis

Champion (512)

maxis's picture

12-11-2016, 19:12

Piter Punk wrote:
maxis wrote:

Pardon me, but Zilog by all the means had an extraordinary skills in building chips of ANY complexity by those days.

Talking about Z80 compatible line, Zilog never considered Z80 as an important and viable business in mid 80s. They were into the CPU wars with their Z8000 and Z80000 plans gradually loosing the edge. They were into high performance computing and industrial automation with their Z8 controllers.

Ok, so Zilog had the skill, but not the will to create a 16 bits processor compatible with Z80. So they licensed it from other sources.

HD64180 by any means isn't a 16 bit processor, neither Z80. Actually Zilog at some point realized, that Z80 can act as the communitaion controller and they jump started the product line by cross-licensing Hitachi. For example Z80182 was very widely used in HAYES modems as the main controller till late 90s. And these cores are purely 8 bit without any simple magic, including the pipeline, prefetch, etc.

However, Z280 is a 16 bit microprocessor.

Piter Punk wrote:
maxis wrote:

But apart from all this, thanks for publishing the findings in your blog! Excellent post! Very interesting hardware. S3527 is used up to its spec -> MSX1 engine, no mapper.

AFAIK there is no MSX machine with native turbo that uses the S1985. The Panasonics uses the T9769, the Victors uses discrete components, the video titlers uses different custom gate-arrays and this Zeal machine uses the S3527. There is something wrong with S1985?

IMHO, S3527, is fast, and likely was freely available at the time of the design. Built-in mapper in S1985 introduces the delay into the address bus. Maybe the designers decided not to do any risk taking (pure speculation).
S3527 based machine can go up to 10 MHz w/o any noticable degradation of reliability.

Have you noticed somewhere an additional circuit with the waitstate generator to sync up the CPU to VDP?

By yzi

Champion (444)

yzi's picture

12-11-2016, 21:26

Piter Punk wrote:

Ok, so Zilog had the skill, but not the will to create a 16 bits processor compatible with Z80.

I don't think it was about skills. It was just not possible to create a 16-bit processor that's binary compatible with Z80. Here's the story as told by Zilog themselves, "Oral History Panel on the Development and Promotion of the Zilog Z8000 Microprocessor"
http://archive.computerhistory.org/resources/text/Oral_Histo...

Quote:

Peuto: I’m sorry, pins. But for cost reasons it was plastic versus ceramic, so all of those things were put as constraints on me. I spent the first three months analyzing the Z80 and proving to myself through a lot of the techniques that I’d done in my paper on architecture, proving to myself that the Z8000 would have better code density, faster speed from an architectural standpoint, and would allow you to have every instruction of the Z80 basically either executable in one instruction, or translatable. So I did a lot of work in the first three months to come with an instruction set that was basically Z80 supportive but the minute you had to decide that you had more than 16-bit for an address you could not be binary compatible anymore.

And it continues:

Quote:

It’s interesting that every vendor made the same decision

Then the interesting question, why didn't Z8000 win the whole market? Why didn't the Motorola 68000 win the whole market? Why was Intel's 8086 chosen for the IBM PC? (or actually the cut-down version 8088)

Zilog's Z8000 was out of the question, because it was owned by Exxon, which IBM considered a competitor:

Quote:

Faggin: ...
But a major reason why IBM chose the Intel was that we were perceived by IBM as competitors and I was told many, many years later by an IBM insider that Cary [Frank Cary IBM CEO] had a memo, internal memo forbidding IBM to use Zilog products because of the affiliation. At that point we were an affiliate of Exxon Enterprises.

Motorola's 68000 was out of the question, because they refused to make a version with an 8-bit external bus. More quotes from "Oral History Panel on the Development and Promotion of the Motorola 68000"
http://archive.computerhistory.org/resources/text/Oral_Histo...

Quote:

House: From my experience at Intel I know that the 68000 was our primary competition and it started out that we were looking at the Z80 and decided we had to have a 16–bit machine. And ZiLog was kind of dismissed because of Exxon's role in office products and their stated objective to be strong in office products and stuff. I'd be interested in hearing what the experiences were. I think probably we were able to get in because of our success with the Display Writer, to get into that design group and we maybe had better access to them. It was really done at the design engineer level. You just couldn't get to talk to those people.

Gunter: No. You could probably answer a million questions for Jack and I on that. We were constantly told that the real decision at IBM was they wanted to make sure that we were going to have a competitive product from a cost perspective; and the fact that you had the 8088, an 8–bit implementation. And we didn’t. That gave you the option to build—IBM at that time didn’t have the PC

House: Oh yes.

Gunter: We refused to do the 8–bit version of the 68000. In retrospect, it would take one person less than a month to do the conversion. We basically didn't do that until after the fact. That would even have fit in a 48–pin package.

By maxis

Champion (512)

maxis's picture

13-11-2016, 17:52

yzi wrote:

I don't think it was about skills. It was just not possible to create a 16-bit processor that's binary compatible with Z80. Here's the story as told by Zilog themselves, "Oral History Panel on the Development and Promotion of the Zilog Z8000 Microprocessor"

This is a very interesting piece of the information.

Reading the WEB sources, Z800 (NMOS design prototype) and Z280 its CMOS commercial counterpart went to the market before the deal with HITACHI to cross license HD64180.
And Z280 is a high performance processor with MMU and capable enough to run UNIX System V. I used the dev. system based on Z280 in the end of 90s (looked more like the industrial controller).
Quite likely the price tag of Z280 was exorbital to integrate it into a 8 bit home computer.

Talking about Motorola SPS, sometimes in a history a company or an influential individual have a chance. And then it works out pretty well giving the competitive edge. It worked well for 68000. But it didn't later. I remember the optimistic PPC presentations in Mot SPS about the merger with IBM and the future of the ColdFire and now we see that all that is the history.
On one hand stopping MC68060 in favor of the PPC finished Apple high performance computing, disrupted Quadra AV lines, Commodore Amiga and Atari 68K products among with many others. On the other hand killing MC88000 RISC created the Nexgen team of the pissed off ex-Motorolans, which was acquired lately by AMD. So, having a chance when defining the company techno roadmap is important for the company future. Why this happened? Why 68K came to an end? Likely because the floating point performance of 68060 were not as good as Pentium. And also process wise the yield was relatively low at the beginning. If Motorola would continue pushing their 68060 with the same dedication quite likely they would be still afloat.

By msxholder

Champion (393)

msxholder's picture

13-11-2016, 22:21

Looking at the Vram of the Zeal E/PASO it's d ram is 512 kb !

By maxis

Champion (512)

maxis's picture

13-11-2016, 22:37

msxholder wrote:

Looking at the Vram of the Zeal E/PASO it's d ram is 512 kb !

The ICs are 256Kx4, but you see these 4 green wires, connecting the A8 pin of the each VRAM to GND. Hence, A8 is not used.
It's the standard 128K configuration, IMHO.

Page 9/14
2 | 3 | 4 | 5 | 6 | 7 | 8 | | 10 | 11 | 12 | 13 | 14