*** CP/M-COMPATIBLE COMPUTING ALIVE AND HEALTHY IN 1988!! *** P.S. -- For us Morrow users I've added a couple of notes from Scott Moore from Ladera Z-Node in L.A. at the end noting that he's been able to put in the Zedux DIRECT Z280 board (the computer is no longer using the Z80 chip at all, not even as a coprocessor!) into the Morrow MD3 we've lent him! Once again Morrow is at the top of the pack! All (?) we need to do is rewrite the Morrow BIOS (remember the 64180 project?) - rc ------- Message #13005 ** General ** Posted: 11/23/87 at 8:19 am From: Dave Vanhorn To: Ron Bardarson [read 01/01/78 at 12:59 am] About: z-280 (4 lines) I have a chip. My folks at Zilog tell me to expect the 50Mhz versions in the next 90 days. I have not found any trap problems as previously alluded. The traps are only active in user mode and the chip powers up in system mode. I have a Z-280 to Z-80 plugboard running also. ------- Message #13028 ** Z-System ** Posted: 12/05/87 at 12:55 pm From: Ron Bardarson To: Ben Grey [read 12/07/87 at 6:21 am] About: Z-280 (12 lines) Hi Ben, actually I was hoping to keep this Z-280 discussion public. I'm a very big fan of Z-80's and Zilog and was playing devil's avocate to stir up the msgs here at ZNC. I have also seen a Z-280 demo organized by Rich Charnes. Unfortunately whoever gave the demo didn't know a thing about Z-System, they were running their own private OS. Their Z-280 impletmention wasn't going to allow interrupts or poking system bytes, as an owner of 3 BigBoard II's ( Mode-2 interrupts) I wasn't impressed. As a hacker, I would have been handcuffed. They were also going to trap I/O calls, another poor idea. As for Kaypros, I don't know what could have failed to follow spec, I have mine running at 8 MHz without problems, but then I've tinkered with it a bit. ------- Message #13030 ** Z-System ** Posted: 12/05/87 at 1:07 pm From: Ben Grey To: Ron Bardarson [read 12/05/87 at 1:08 pm] About: Z-280 (17 lines) Ron: The implementation of the Z280 OS/Z by Zedux is not limited in the manner you suggest. I suspect it is because you don't understand what it takes to "simulate" CP/M (ZRDOS or whatever) and in the environment that they were working in. I've spent many hours working on some of the problems that one has to deal with to put a Z280 in a Z80 system. It is not simple! The fact that your demo didn't come off too well, cuz of the broken boards, should not have been a factor. Scott Moore is a Unix/VM-370 type and he has implemented all the various functions we normally have under Z_system in his new OS. The fact that they have to trap I/O and don't allow poking is all very well and good. But what you might not realize is that any program which uses self modifying code will NOT run in a Z280... There simply is no way to address the bytes in the program (data segments, yes, code segments, no). FYI, Greg Trice, Toronto, has a working S-100 version and is quite a good Hardware type. You and he probably ought to chat when you can... Ben. ------- Message #13032 ** Z-System ** [reply to #13030] Posted: 12/05/87 at 1:37 pm From: Ron Bardarson To: Ben Grey [read 12/07/87 at 6:23 am] About: Z-280 (20 lines) If you can't poke system bytes then : the poke and go techniques are lost! Zex is lost Memory Editors are reduced in usefullness Error handling and message communications are lost! .... All of these programs and many more use parts of the operating system to communicate or gain services. They REQUIRE poking system bytes to work. If the Z-280 doesn't allow this, then it is NOT an upgrade to Z-System. If you can't use I/O outside of the operating system then : Kiss your modem programs goodbye Forget about BYE and RAS's ( BYE would need to be re-written anyway) No more special printer drivers No more custom I/O drivers for controller purposes ( SCSI, etc.. .... Most of these programs don't use the operating system for I/O. Again if the Z-280 doesn't allow this, then it's a step backwards. The fact that Scott's demo didn't work had very little to do with my opinions, they were generated as a direct result of Scott's talk. ------- Message #13041 ** Z-System ** [reply to #13032] Posted: 12/07/87 at 6:30 am From: BEN GREY To: Ron Bardarson [read 12/07/87 at 7:25 pm] About: Z-280 (19 lines) Well Ron, It's apparent you do not understand how the memory management of the Z280 works. Code segments simply cannot modify themselves. There is no need for a BYE or RAS or any of the usual selfmodifying programs in a system that uses the Z280 (at least not the OS/Z). There is nothing preventing loading of "system" segments that take care of printers and terminals, hard disks, and so on. The idea that the Z280 can't do I/O is patently absurd. It can, but in general it cannot run directly in a Z80 environment. The co-processor can -- but it has the old Z80 to do it's I/O and transfer info into and outof the Z280's memory. I suggest you contact Zedux on your own and get their literature before you rail at me for explaining what most every system designer who has looked at the Z280 already knows. For $10 I'll send you the newest data book. As soon as I get my 1MB DRAM, I'll but running the Z-node with a Z80-280 combination. We'll see what I can/can't use. Programs that worked before will work just as they always did, just the data areas will be in data memory and not in code memory. There are very good reasons for not permitting self-modifying code. We surely don't want lowly CP/M programs crashing our OS... ------- Message #13045 ** Z-System ** [reply to #13041] Posted: 12/07/87 at 7:32 pm From: Ron Bardarson To: BEN GREY About: Z-280 (10 lines) Ben ah take a closer look at my replies, nowhere do I say we can't use 280's. ALL THAT I HAVE SAID IS THAT THEY WILL NOT BE Z-SYSTEM COMPATIBLE If you would like to regenerate the code, fine. As for what you'll be able to run or not, thats the very important test that I'm waiting for before I invest a penny, since I know that the 280 is not Z-System. No doubt that it's a better processor, but so are several others. The only thing that's attractive is if it is almost fully Z-System compatible, thus allowing all of us to hardware upgrade without software re-investment. If it takes that ( both hardware and software) then I'll get a 68030 system. ------- Message #13053 ** General ** Posted: 12/08/87 at 7:41 pm From: Dave Vanhorn To: Ron Bardarson About: z-280 (20 lines) FIFTY MEGAHERTZ. 50 CPU clock then = 25 mhz, and system clock is (on the buss) 12.5 mhz. (Assuming you can dig up some Z-80(Q?) I/o chips that will go that fast!) Waitstates up the wazoo I think! Unless some bright person used a Z-80 as an I/O processor and blasted the I/O information out to the Z-80's ram using the DMA and common memory areas... (Who would do a thing like that?) I'm making a shift in my 280 experiments. The CP/M machines that exist now are too diverse in their hardware for a Z-280 fit-in to be anything but a cripple, so i'm looking at putting a Z-280 into a PC and using the 80XXX as an I/O manager. It will be fully Z-System compatible, running Z-80 code unmodified. The trickiness comes in when the bios calls are done, and they actually pop the data in and out of the pc's ram for handling by the 80xxx on the pc iron. I'll probably put a SCSI port on the 280 board, and might consider a floppy port as well to make r/w of existing CPM formats easier. The use of the PC iron gives us CHEAP boxes, and bits,, and should go a long way twoard standardizin g the software as well. (Maybe the best of both worlds?) Imagine: One board turns a PC into a real computer! What a concept! ------- Message #13054 ** General ** Posted: 12/08/87 at 7:46 pm From: Dave Vanhorn To: Ron Bardarson About: z-280 (11 lines) I forgot to mention: Caching WORKS! I have made measurements on code executing in rom-based systems that PROVE the cache does work. It's real spooky to see it suddenly stop doing bus transactions (xcpt for refresh) when it drops into a tight loop!) As for CP/M programs not running, well.... My ampro dosen't seem to notice that it has a Z-280 card plugged into it's Z-80 port... Z-system, MEX-PC, and all my little wierd programs are running just fine. It strikes me that someone is piping lots of anti-280 rumours out into the world. Could it be that the cloneheads are getting nervous? ------- **** Messages re Z280 captured from Z-Node Central 12/14/87 ------- Message #13069 ** General ** Posted: 12/10/87 at 11:21 pm From: TONY PARKER To: DAVE VANHORN [read 12/12/87 at 2:43 pm] About: Z280 on a Kaypro??? (3 lines) Dave; Is there a real live "non-vaporware" ready to buy board or series of parts That I could get up and running Z on my Kaypro!!!!TP ------- Message #13078 ** General ** [reply to #13069] Posted: 12/12/87 at 2:44 pm From: Dave Vanhorn To: TONY PARKER About: Z280 on a Kaypro??? (2 lines) Only one I know of is the ZeDux machine, and I'm not sure of it's status as vapor/hard ware. (I've not seen one...) ------- Message #13082 ** General ** [reply to #13053] Posted: 12/12/87 at 3:14 pm From: Ron Bardarson To: Dave Vanhorn [read 12/14/87 at 12:06 am] About: z-280 (5 lines) I'd heard of the idea of using PC chassis as a home for 280's, seems like a very good idea since the cost of clones is so low and there are expansion abilities built right in ( along with the cards to fit). In short, are you looking for a poor soul to test if all that CP/M software can run on it? I know one or two..... ------- Message #13101 ** General ** [reply to #13082] Posted: 12/14/87 at 12:08 am From: Dave Vanhorn To: Ron Bardarson [read 12/14/87 at 5:46 pm] About: z-280 (4 lines) I checked with Z-world, and they aren't planning to do a board like this. I put in a message also to Rick over at Ampro, but I rather doubt they are planning anything. I may be forced to drag out the old wire wrapping gun again! ------- Message #13105 ** General ** [reply to #13101] Posted: 12/14/87 at 5:50 pm From: Ron Bardarson To: Dave Vanhorn [read 12/14/87 at 6:31 pm] About: z-280 (9 lines) Do you have a reply to Grey's cache statement? Never pay much attention to people who tell me that it would take too long to explain myself. Old rule of thumb for physicists, if it doesn't fit on one sheet of paper and you can't explain it to everybody, then you don't understand it yourself. Have you tried your board in any other machines? I may be travelling down your direction in the next few months, perhaps I could get a peek? ------- Message #13106 ** General ** [reply to #13105] Posted: 12/14/87 at 6:35 pm From: Dave Vanhorn To: Ron Bardarson About: z-280 (6 lines) I measured the caching by looking at the buss while executing code which all fits within the 16 byte limit of the cache. I saw caching take effect when the 280 stopped doing the memory reads required to fetch the data off the eprom. Apparently in some other cases it acts differently. Remember, it's not a BIG cache, just 16 bytes. Still, it's better than nothing. ------- Message #13083 ** General ** [reply to #13054] Posted: 12/12/87 at 3:22 pm From: Ron Bardarson To: Dave Vanhorn [read 12/14/87 at 12:08 am] About: z-280 (11 lines) Well I heard that the cache didn't work from someone trying to sell Z-280 boards, but rumors are only as reliable as their source. Glad to hear that your weird programs run, as I have a number myself that sounded like they wouldn't work. I'm also an amateur radio operator and have some custom code running my FT-980 transceiver, for which I need several I/O ports and interrupts. I'm working toward a frequency agile CW robot one of these days..... As for anti-280 rumors, I doubt if the cloneheads understand hardware enough to realize what the excitment is about, glad to hear you have a running system. ------- Message #13085 ** General ** Posted: 12/12/87 at 3:41 pm From: Ron Bardarson To: TONY PARKER About: Z280 on a Kaypro? (1 lines) Check with Ben Grey, you can leave a msg here or on his Z-Node. ------- Message #13090 ** General ** Posted: 12/13/87 at 10:29 am From: BEN GREY To: TONY PARKER About: Z280 on a Kaypro!!!??? (1 lines) There is no Z280 for *ANY* Kaypro machines at this time. ------- Message #13091 ** General ** Posted: 12/13/87 at 10:35 am From: BEN GREY To: Ron Bardarson [read 12/13/87 at 8:30 pm] About: Z280 (10 lines) I guess the problem with even the Z80 users and those that are not familiar with the way the hardware works, it is impossible to explain what the difficulties with the Z280-Z80 situation is. As for running Z-System on a Z280 is concerned, there is absolutely no problem with do so. The fact that you don't have a BIOS or BDOS plays absolutely no role in the scenario whatsoever. It is the function of the OS to provide the "system functions" we normally associate with the BDOS and BIOS. If they don't exist, then the problem becomes on of how do CP/M programs of any sort expect to run on a Z280? The fact is, they can and do run, without difficulties. \ ------- Message #13092 ** General ** Posted: 12/13/87 at 10:39 am From: BEN GREY To: Dave Vanhorn [read 12/14/87 at 12:09 am] About: Z280 Cache (9 lines) The Z280 cache problem is NOT fixed!! Unless you have received a chip in the past two weeks (unlikely) you are still running with the last batch issued in July. The problem is not easily explained, but it can be demonstrated with programs that set the stack to the address below the CCP and programs which gobble memory up to that point. The cache controller simply gets confused and mixes bit/bytes/words in such a way it nearly impossible to track. The cache does work in some cases where the program is *very well behaved*. No word from Zilog on when the "cache insect" will be removed. I suspect in Jan 88. ------- Message #13102 ** General ** [reply to #13092] Posted: 12/14/87 at 12:12 am From: Dave Vanhorn To: BEN GREY About: Z280 Cache (4 lines) Hmm.. The program I set up to test cacheing is probably a 'very well behaved' program. I've not monitored the buss with normal programs running since the TEK 1240 has this problem in not recognizing opcode fetches without the M1 signal. Have faith. ------- Message #13094 ** Z-System ** [reply to #13045] Posted: 12/13/87 at 11:00 am From: BEN GREY To: Ron Bardarson [read 12/13/87 at 8:33 pm] About: Z-280 (25 lines) I guess i need to iterate again that the difference between Z-System and Z280 is like the difference between dumplings and road apples. Hardware does not an operating system make. It is the OS that makes the computer in the long run and it is the OS function to provide the user with the necessary hooks to accomplish that which he want to create. It is too unfortunate that folks like to compare hardware and software in the same sentence. They are distinctly different. The problem with putting a Z280 in an existing Z80 machine would not be nearly so complex if each manufacturer had followed the letter of the Zilog Z80 specifications and designed for the worst case. The did not, and played games with the bus timings and various other signal timings making a "universal" replacement exceedingly difficult. The classical case is the Kaypro machines. If you really care to follow along with Z280 info, you should be checking in on Al Hawley's system from time to time (weekly if you can) as there is much discussion about it and the "new breed" of OS we are likely to see in the very near future. I have no problem playing devils advocate with this issue, since I've got a Z280 here too, and am busy (what with all my other projects) designing a Z280 system that has a full potential. The fact that I've unloaded the Z280 from it's mundane chores and left it for the work it can do best is academic. Who wants to bog such a machine down with slooow I/O or such? Oh well, such is the life of a hardware/software engineer. ------- Z280 Messages captured from Z-Node Central 12/19/87 ------- Message #13109 ** General ** [reply to #13106] Posted: 12/14/87 at 11:20 pm From: Ron Bardarson To: Dave Vanhorn [read 12/16/87 at 10:57 pm] About: z-280 (3 lines) Thanks for the info on the cache test, did your wire wrap comment indicate that you were planning on building more Z-280 boards? Have you tried it in any machines other than an Ampro? ------- Message #13135 ** General ** [reply to #13109] Posted: 12/16/87 at 11:00 pm From: Dave Vanhorn To: Ron Bardarson [read 12/17/87 at 1:00 am] About: z-280 (5 lines) I've tried it in the ampro, and in several of our Z-80 based credit card terminals. I am looking at doing a Z-280 in a PC card. (Similar to the Z-world (Formerly Decmation) blue lightning/thunder cards, but with Z-system as the main DOS with PC operation as an auxilliary) Mainly for a fixed environment with cheap hardware. ------- Message #13146 ** General ** [reply to #13135] Posted: 12/17/87 at 1:07 am From: Ron Bardarson To: Dave Vanhorn About: z-280 (4 lines) Hey you could use the extra memory clone buyers get ( 1 Meg) only to find out they can access 640K....seriously clone chassis make sense, as I said earlier. Ampro schematic looks like most I've seen, wonder why other people are having problems. ------- Message #13114 ** General ** Posted: 12/15/87 at 5:35 pm From: Frank Gaude' To: ALL USERS About: PZDOS (5 lines) Just a quick note to state that Echelon is actively working out an arrangement with Cameron Cotrill, et al, to handle PZDOS, the Z80 DOS we intend to be the DOS that upgrades ZRDOS beyond v1.9. More as the agreement and product near finalization and release, respectively. Echelon, Inc.--from high in the Sierras... ------- Message #13147 ** General ** Posted: 12/17/87 at 1:23 am From: Ron Bardarson To: TONY PARKER About: Z280 GRIPE ## (13 lines) Scott Moore has just applied for user status, perhaps your comments would be better directed to him. As for msg content, the rule of illegal activities applies ( foul language, credit/telephone card info, etc..). I believe in everybodys right to put their foot in their mouth, but at the same time, think that a point of view should be presented without emotion. I have not deleted any of the Z-280 discussion and will not, I believe that the majority of sales of the 280 will be to Z-System folkes and wish to allow all interested parties their say. As for your comments, if you already feel that they are too strong, then perhaps you should tone them down until you feel satisfied ( but you're the judge ). when entering a msg, remember that's the only medium via which a reader may form an opinion of you, so present your msg accordingly. ------- Message #13151 ** General ** Posted: 12/17/87 at 2:58 am From: Scott Moore To: Dave Vanhorn About: Z280 (28 lines) Very good to hear that someone out there is playing around with the Z280. Noted your message about the cache problem (or lack of). Turning the cache off on this system, and looking about for programs that will not run under cache, I too was hard pressed to find one. One I did come across was turbo pascal 2.0, which here just sticks. The cache problem gets activated as follows: the jump instruction that CP/M places at 5 (bdos entry), is also used as the stack location by many programs: ; | Bdos | ------------ -> jump at 5 points here | stack | ; what happens is that a program which makes a call to $5, with the stack empty, will first stack the address under the bdos entry, then jump to the bdos. In Z280 cache terms, the program MAY both do a data access (the stack push) and an instuction access in the same cache "line" of 16 bytes. When the bdos returns, the fetch of the address to perform the return messes up as follows: it reads only the low part of the address, then copies the low byte to both halves of the PC, which then crashes. The problem is dependent on where the location of bdos is, and how much data is already on the stack. I wasn't saying the cache bug appeared frequently, but even such a once in a blue moon bug justifyes truning off the cache in my opionion. Also a word to the wise (unofficially). dont run the processor at a speed greater than 16 mhz (the xtal speed to the part). Keep us apprised ! sam ------- Message #13152 ** General ** Posted: 12/17/87 at 3:10 am From: Scott Moore To: ALL USERS About: ZEDUX (11 lines) I think the community does deseverve an apology from me concerning the difficulty of getting through to Zedux for answers. I dropped answering my phones and used an answering machine instead for the reason I was not getting much done. This is no longer true, and getting me between the hours of 10 am to 6pm or perhaps later is usually possible. The number at my desk is 818-787-0113. Be happy to discuss any issues, and would especially like to hear from those who just want to discuss the Z280 in general. I have been running CP/M on the Z280 since JAN of this year, and have had extensive experience with the part. sam ------- Message #13153 ** General ** [reply to #13053] Posted: 12/17/87 at 10:45 am From: Greg Trice To: Dave Vanhorn About: z-280 (6 lines) Interesting. I have had the concept of replacing the 80286 in an AT with a Z280 and a PAL or two, plus new system roms, which would produce a low cost Z280 machine, and all the plug in peripherals available for the At would then be available. I may actually get round to manufacturing such a machine (with a purpose designed motherboard). Using low cost clone parts would keep the cost down. ------- Message #13154 ** Z-System ** Posted: 12/17/87 at 10:45 pm From: Ron Bardarson To: Scott Moore About: Z-280 (2 lines) Why don't you upload the latest version of your file that explains how to install your product in a Z-80 system? ------- Message #13161 ** General ** Posted: 12/18/87 at 6:16 am From: BEN GREY To: Dave Vanhorn About: Z280 Cache (3 lines) I assume you understand that the Z280 has no need for an M1 since it is a pipelined machine. Fetching an instruction has no meaning to the bus either, since it will NOT be the next instruction executed. Ben. ------- Message #13182 ** Z-System ** Posted: 12/19/87 at 1:59 pm From: BEN GREY To: Ron Bardarson [read 12/19/87 at 3:41 pm] About: Z280 M1 (2 lines) The Z280 does not issue any M1, impossible, since it does not know when it is fetching an instruction or data. ------- Message #13185 ** Z-System ** [reply to #13182] Posted: 12/19/87 at 3:46 pm From: Ron Bardarson To: BEN GREY About: Z280 M1 (6 lines) So.... you need MREQ* and M1* to determine if it's an opcode fetch or a memory read. Can't imagine anyone decoding the difference, so I don't see the importance of the Z-280 not providing M1* during an opcode fetch. It certainly provides M1* during int ack. Still wondering what the problem is with the ZIP board in Kaypro's...maybe Scott will upload his info file. ------- ***** THESE ARE FROM LADERA Z-NODE IN L.A... ***** - rc Msg #2473 posted 12/13/87 at 12:28 am by SCOTT MOORE To: Tony Parker About: Z280 (30 lines) I really can't imagine where you looked. Zedux's phone is either (818) 787-0113 or (213) 301-1935 A general update for interested parties: Had a frank conversation with Hi-Tech, was told (and yes, I asked if I could pass the information on) that due to problems, they were postponing all work on the "ultra-board" until april or may 1988. Our basic or "direct" version of our board has been shipping for 6 months now. This has been described ad - nauseum (z280.lbr). Our only shipments to date have been to people who are updating (industrial) Z80 controllers to Z280 in the field (assume that any new product they would just do a redesign without our help). As far as CP/M / OS et. all, these have fallen into these two general cases: 1. The board is completely explained to the purchaser, including the fact that software/hardware updates may be required. Without fail, the boards sent out to such customers have come back. Each and every one of them was contacted via phone. Basically they were called on the bluff of knowing a DAMM THING about the hardware/software involved. We found out FAST that people will tell us ANYTHING, with the actual intent to get "a free trial". This has gotten rapidly silly, the last customer we virtually GRILLED about his real knowledge about hardware/software. The MD-3: After having an MD-3 graciously loaned to us for a month, on a whim decided to plug the direct board into it. It came up with no modification. However, A software mod to the BIOS needs to be done to avoid the dread Z280 "cache bug" in the current chip. Noone stepped forward with the nessary skills to do this (in fact, I have not seen one person with the genuine ability to handle or even know what is in the MD-3 BIOS). [* ahem! * - rc] I will do the conversion myself, here, hopefully as a demontration of how easy such a conversion can be for people with the appropriate backround. (cont) Msg #2474 posted 12/13/87 at 12:44 am by SCOTT MOORE To: Tony Parker About: cont (42 lines) The thing that must be realized here is that NO ONE COMPANY, PERSON OR DIETY has the power to singlehandedly update every CP/M / Z80 system ever made. THE "UNIVERSAL" board: The coprocessor version of the board has been put together and tested in prototype form. The problem with that unit is that it is about 10 times more complex than the present board, and requires about 10kb of software, and a VERY high density PAL. I have postponed the final version of that board till at least january (when we have a standing promise to ship to a paid customer). THIS BOARD IS NOT MAGIC. It obtains better compatibility by using (read "wasting") Z280 power. The basic rules of compatibility are: **** READ THIS ***** ; 1. Unless you use the new instructions of the Z280, You will see no speed improvement from the more powerfull instruction set. ; 2. Since the old Z80 instructions execute basically the same way on the Z280 (the same number of bytes fetched, etc). The only way to speed up the REAL EXECUTION SPEED is to make the cycles faster. This is the EXACT SAME PROBLEM as moving up to a higher speed version of the Z80, which you all know is impossible without faster memory, etc. Z80's are avalible up to 12 mhz NOW. ; 3. The only possible improvement offered by the Z280 in speed is via the onboard cache, and that is BUSTED on the currently shipping chip. Despite rumors, Zilog's shipping date for the "fixed" version is UNKNOWN, and NOT SOON. ; End of lecture. Anyways, I have been taking the time to update the Z280 operating system OS/Z280 to version 3.0. In january, that system will be 2 years old (yes, you read correctly). We remain the only users of the system, as the current Z280 adapter users aren't OS users (as explained). I expect to show a MD-3 version in january. [Wow! - rc] Time will tell. If you have any questions GIVE US A CALL. Forgive the frankness, but I am bloody sick of hearing 2nd, 3rd and nth party rumors and trash. sam