ewen@actrix.co.nz (Ewen McNeill) (11/21/90)
In article <1990Nov19.090518.3479@simasd.uucp> donm@pnet07.cts.com (Don Maslin) writes: > ewen@actrix.co.nz (Ewen McNeill) writes: > >In article <5543@uafhp.uark.edu> bbs00167@uafcseg.uucp (Keefe Jackson) writes: > >> > >> Does anyone know anything about the Amstrad "128k CP/M" machine? > >> Is there a hard disk available for it? Is it still availble? > > > >I know a reasonable amount about the Amstrad CPC6128 (which I presume > >[...] > >There is not a hard disc available for the 6128, to my knowledge. > > Assuming that it is a Z-80 based machine (display of ignorance here!), it > seems that it might be possible to get a HD kit from Emerald Microware which > uses a small daughter board that plugs into the Z-80 socket plus the WD1002-05 > hard disk controller. The one problem that might crop up would be the > software modifications to the BIOS. I know Emerald has it for 2.2, but I'm > not sure about 3.0. A telephone call should answer that though. > > Ewen, would you care to comment? Certainly, I will comment :-) The machine is definately Z80 based, so your solution might well work. Software modifications of the BIOS could be difficult, as could be keeping up with the speed of a HD. The machine runs at 4Mhz with one wait state (effective speed is said to be 3.3Mhz). The machine does not support DMA, or Interupt driven Disk IO (the floppy IO is done by polling the controller as is the serial IO -- both _cannot_ be done at the same time :-) I think the answer is that technically it might be possible, but the only real way to approach the problem (from my point of view) would be to build a seperate (Z80 controlled) Disk drive manager (supporting Floppies and HDs). That would allow an implementation of DMA - and with big enough cache, reasonable amounts to be read at one time. Again, BIOS would be a problem. I don't have source to it (though, I may well reverse-engineer it one day :-) and the CPM 2.2 implementation on the Amstrad only allows for 42K TPA. (The screen on this machine is a graphics only one -- and it takes 16K!). Amstrads are definately cut price machines. One possible solution (with expanded memory (of which I have 512K!)) might be to run CP/M in a different bank from the screen. Still, someone would have to do a lot of work. One final problem -- the HD would still cost more than the machine :-) -- Ewen McNeill. Email: ewen@actrix.gen.nz
donm@pnet07.cts.com (Don Maslin) (11/22/90)
Touche'!
- don
Keeper of the CP/M System Disk | UUCP: {nosc ucsd crash ncr-sd}!pnet07!donm
Archives for the Dino(saur)SIG | ARPA: simasd!pnet07!donm@nosc.mil
- San Diego Computer Society - | INET: donm@pnet07.cts.comslsw2@cc.usu.edu (11/24/90)
In article <1990Nov20.212904.14578@actrix.co.nz>, ewen@actrix.co.nz (Ewen McNeill) writes: > In article <1990Nov19.090518.3479@simasd.uucp> donm@pnet07.cts.com (Don Maslin) writes: >> >> Assuming that it is a Z-80 based machine (display of ignorance here!), it >> seems that it might be possible to get a HD kit from Emerald Microware which >> uses a small daughter board that plugs into the Z-80 socket plus the WD1002-05 >> hard disk controller... > > The machine is definately Z80 based, so your solution might well > work. Software modifications of the BIOS could be difficult, as > could be keeping up with the speed of a HD. The machine runs at > 4Mhz with one wait state (effective speed is said to be 3.3Mhz). > The machine does not support DMA, or Interupt driven Disk IO (the > floppy IO is done by polling the controller as is the serial IO -- > both _cannot_ be done at the same time :-) Actually, the WD1002-05 has a sector buffer on board. The data is transferred from the disk into this sector buffer, after which the host processor can pull the data from the sector buffer at its leisure. DMA is not required, nor are interrupts. -- =============================================================================== Roger Ivie 35 S 300 W Logan, Ut. 84321 (801) 752-8633 ===============================================================================
ewen@actrix.gen.nz (Ewen McNeill) (11/25/90)
In article <1990Nov23.114001.43015@cc.usu.edu> slsw2@cc.usu.edu writes: > >> > >> [...] HD kit from Emerald Microware which > >> uses a small daughter board that plugs into the Z-80 socket plus the > >> WD1002-05 hard disk controller... > > [...] > > could be keeping up with the speed of a HD. The machine runs at > > 4Mhz with one wait state (effective speed is said to be 3.3Mhz). > > The machine does not support DMA, or Interupt driven Disk IO > > Actually, the WD1002-05 has a sector buffer on board. The data is transferred > from the disk into this sector buffer, after which the host processor can > pull the data from the sector buffer at its leisure. DMA is not required, nor > are interrupts. This begins to sound very interesting. Can someone mail me the address of Emerald Microware, and also the cost of this small daughter board? The design of my machine (Amstrad CPC6128) is such that it should go in quite easily. I am also interested in any comments that people have about the reliability of this controller, and the sort of through-put that you get. Of course, then I have to find a HD... but that shouldn't be difficult. There seem to be a lot of oldish 10MB drives around NZ, going cheaply. Anyone know of a reasonable access speed for this type of setup (I would be inclined to think that 65ms would be more than fast enough). Thank you all for your help. -- Ewen McNeill. Email: ewen@actrix.gen.nz
donm@pnet07.cts.com (Don Maslin) (11/25/90)
ewen@actrix.gen.nz (Ewen McNeill) writes: > >This begins to sound very interesting. Can someone mail me the >address of Emerald Microware, and also the cost of this small >daughter board? The design of my machine (Amstrad CPC6128) is such >that it should go in quite easily. > Emerald Microware P.O. Box 1726 Beaverton OR 97075 USA 503/641-8088 HDS Host Board w/ software $79.95 HDS Board, WD-1002-05 & s/w $245.00 WD-1002-05 HDC only $185.00 >I am also interested in any comments that people have about the >reliability of this controller, and the sort of through-put that you >get. Should be roughly comparable to a Kaypro K-10. Its virtually the same HDC board (floppy controller added to the -05). > >Of course, then I have to find a HD... but that shouldn't be >difficult. There seem to be a lot of oldish 10MB drives around NZ, >going cheaply. Anyone know of a reasonable access speed for this >type of setup (I would be inclined to think that 65ms would be more >than fast enough). > Probably 65-85 would work just fine. Have fun! - don Keeper of the CP/M System Disk | UUCP: {nosc ucsd crash ncr-sd}!pnet07!donm Archives for the Dino(saur)SIG | ARPA: simasd!pnet07!donm@nosc.mil - San Diego Computer Society - | INET: donm@pnet07.cts.com
mlinar@eve.usc.edu (Mitch Mlinar) (11/26/90)
In article <1990Nov25.062505.26432@simasd.uucp> donm@pnet07.cts.com (Don Maslin) writes: > >>I am also interested in any comments that people have about the >>reliability of this [WD1002] controller, and the sort of through-put that you >>get. > >Should be roughly comparable to a Kaypro K-10. Its virtually the same HDC >board (floppy controller added to the -05). Actually, the throughput of HDS is slightly (10%) higher since the original Kaypro s/w was pretty inefficient, despite the fact it was only targeted at a one drive, 10M hard disk. With Winchester Connection (the s/w which is attached to HDS as well as part of all KayPLUS ROMs), you can handle any hard drive from 5 to 64M (up to 2 hard drives). Note that the throughput of the KayPLUS ROMs is 30% greater on hard disk than either MicroCornucopia or TurboROM, and 25% faster on video. Just different coding. >>type of setup (I would be inclined to think that 65ms would be more >>than fast enough). Yep. 65 is good enough for the WD1002; they are not all that fast anyway. -Mitch