It has to be close to the 1000 because of thee chipset used. However, the
chipset deals with the drive-side of the controller, and not at all with the
host interface. Consequently, the DAVONG folks, whose documentation for
this baby is lying, even as I type, in my lap, accompanied by the four 360K!
diskettes, did not see fit to describe this product in much detail in their
user manual or installation guide.
They didn't even tell you much about the hardware with which you presumably
bought this baby. The usual assumptions were made, i.e. that you bought the
system interface (which uses a SCSI-type 50-contact connector, hence the
conclusion that it was SCSI, which it isn't) so they didn't include the
signal definitions for the cables. The odd thing about the box is that
there are numerous connectors. There's a power connector, a DA-15, an
external drive connector on a DB-25, a drive control cable connector which
is a DC-37, and one of those common 50-pin SCSI-1 connectors seldom seen on
SCSI equipment except in pairs. There's no ID switch for the controller.
Why they'd provide power from one powered box to another isn't clear.
That's how it is though.
It's likely from a functional standpoint that the controller works more or
less like a WD-1000-series controller of the earliest type, i.e. without the
WD1010 chip. These had a different arrangement of the registers, probably
just inverted, as I recall, from that of the WD1010-chip type. They use a
varactor-tuned VCO, just like the WD1000's and again unlike the smaller
board with the WD1010 chip. The fact they use the Signetics microcontroller
does suggest that these would be similar to the WD1000, but the WD1000 used
an 8X300, which maxxed out at 4 MHz while the newer and "improved" 8X305 was
called that because it had a few more instructions and features, and
operated at 5 MHz.
For now, that's all I've got about this DAVONG controller.
Dick
-----Original Message-----
From: Tony Duell <ard_at_p850ug1.demon.co.uk>
To: Discussion re-collecting of classic computers
<classiccmp_at_u.washington.edu>
Date: Saturday, April 17, 1999 7:41 PM
Subject: Re: Ancient disk controllers
>> Connector J2 and J1 are together on one side of the long end (J1 is
34-pins
>> with half of them grounds, J2 is 20 pins with pins 2, 4, 6, 8, 12, 16 and
20
>> grounded). J3, J4 and J5 are in line on a short end, 20 pins each,
similar
>> ground pattern to J2. J2 through J5 appear to have connections to a
Motorola
>> AM26LS32 and a TI AM26LS31 which I take to be some sort of analog chip.
>
>Those are almost certainly ST506-like data connectors (for up to 4
>drives). The 26LS31 and 26LS32 are differential line drivers/receivers
>which can be used for these signals.
>
>> The final connector, J6 is 50 pins.
>>
>> J1 appears to be the control cable for an ST506 drive, J2-J5 appear to be
>> data cables for talking to four drives. The interesting chips on the
board
>> include a N8X305N processor, some N82S181N ROMs, an N8X371N with leads
going
>> right to the 50-pin connector, and five socketed WDC parts copyrighted in
1980:
>
>Which pins are used on the 50 pin connector? Could it be pinned out as a
>SCSI port. Or is it possibly some custom host interface? For example I
>have here the data sheet for the WD1001 controller. It uses essentially
>the chipset you mentioned. It has a 50 pin host connector, but it sure
>ain't SCSI.
>
>> WD1100V-03, WD1100V-01, WD1100V-04, WD1100V-05 and WD1100V-12. There is
a
>
>I have data sheets for those chips...
>
>-03 : Adress mark detector
>-01 : Serial/Parallel converter
>-04 : CRC generator/checker
>-05 : Parallel-serial converter
>-12 : Improved MFM generator.
>
>> crystal at 20Mhz in the analog section of the board and an 8Mhz crystal
by
>
>20MHz/4 = 5MHz = standard ST506 data rate.
>8MHz is a common enough clock for the 8x300 series of CPUs as well.
>
>> the processor. In the middle of the board are three vias that are
labelled
>> as if they are configuration pads, in an inverted-L, labelled "1", "2"
and
>> "3", with a "W" above them,
>
>What bothers me, if this _is_ a SCSI controller is that there seems to be
>no way of setting the device address.
>
>My guess is that it's a sort-of WD1001 clone.
>
>-tony
>
Received on Sat Apr 17 1999 - 23:56:02 BST