Pair of Twiggys

[Fred Cisin]

On Tue, 14 Mar 2017, Brad H wrote:
> I'm assuming anything can be interfaced to old tech.  But if I had 
> Twiggys I do have a Lisa they could go into.  Or I'd just sell them and 
> buy something a lot more useful. :)
> What'd be cool if replicas could be made somehow.  I don't know what all 
> goes into a disk drive but I imagine it's in the realm of possibility at 
> least.

It would be possible, but a bit of work to make them.
A lot depends on whether you want to make replicas, or compatible.

Jobs wanted a closed system with significantly increased difficulty for 
after-market supply of common items.  The bizarre Twiggy drive system did 
not provide improvement over 96tpi double density ("quad") standard 
systems.


A long time ago, Eric Smith explained to  me the details about them (he 
could probably build one):
1) track to track spacing is 62.5? tracks per inch, as opposed to the 
"standard" 48tpi or 96tpi (or less common 100tpi (Micropois))  That should 
not be very hard to do, but not trivial.

2) They used variable rotation speed, from 200RPM? to 350RPM?, to make 
less variation in the FCPI (flus changes per inch)  Relatively trivial, for a 
drive to be used on something other than the original Lisa, as the 
rotational speed could be left constant, and vary the data transfer rate.

3) Instead of a "normal" pair of facing heads, they wanted to keep single 
sided heads, with their felt pressure pads, so they went with that strange 
double slot, providing unmatched opportunities for always leaving a 
thumbprint on the media.  That could be easily worked around, by simply 
using conventional double sided head, and maybe, in some cases, losing 
half a revolution waiting for the desired sector to come around.  Very few 
pieces of software (copy protection) rely on the relative timing of the 
two sides.   Since the Twiggy heads are connected to the same mechanism, 
but on opposite sides of the spindle, when one head is hubward, the other 
one is rimward, so for most purposes, switching to a conventional head 
system would provide improvements in speed.
The optimum speed method for reading a Lisa Twiggy would be to stick with 
one side, reading all tracks, and then when at the last track, the other 
head would be in position to read the first track of that side.  That is 
not the same as what we have become accustomed to, where after reading one 
track, switch to the other head, and read that track of the other side, 
and then step to get to the next two tracks.

4) GCR.    I don't know the specific GCR pattern used, but that 
information is out there, or can be empirically determined.

5a) There is an extra hole, for latching the disk in place? or for later 
plans for a disk changer?

5b) Notch out of one corner to avoid ibverted insertion.

5c) Write protect/enable notch is in a different position.  Trivial.
(I was amazed that on the "Computer Bowl" quiz show, nobody on Bill Gates' 
team could remember where the write protect notch on an 8" disk was!)


Total capacity was 850K?

Media appears to be 600 Oersted  (same as 1.2M)

You will not be able to use a Western Digital/IBM style of disk 
controller, due to the GCR, non-WD/IBM sector headers, and maybe the 
variable speed.

BUT, if you manage the track spacing, then it should be possible 
to read them with a flux transition board, such as Kryoflux.
Or, design and build a suitable logic board.

There are a couple of ways to manage the track spacing, ranging from 
analog positioner (Amlyn), extremely crude milling a new "record" for an 
SA390/SA400, different diameter winding hub for split band positioner, 
gearing interposed between stepper and positioner, different stepper 
motor, etc.  I know none of the details for doing those, but it does not 
seem insurmountable.


Judging by the eBay response, it looks like a replica (or counterfeit?) 
would be far more valuable than a usable substitute.