• Please review our updated Terms and Rules here

Vector 4 low-level disk format?

H

hjalfi

Experienced Member
Joined
Feb 11, 2017
Messages
265
Location
Zürich, Switzerland
Does anyone know anything, or know where I can find anything, about the Vector 4's low-level disk encoding? I have someone with a Micropolis drive who is sending me dumps of a Vector 4 disk, and they are extremely weird. I can see the disk structure easily enough, and it looks very like the Micropolis 1084 format, but the actual bit encoding doesn't look like anything I've ever seen before. It could be a drive fault, but I'd expect to see more noise than I am.

Attached is the interval histogram of one track. The intervals are only 1.5us apart, but the lowest peak is ~6.3us, which I would not expect.

Sadly the Vector Graphics manuals I've found on Bitsavers don't go into this much detail of the disk format.

Code: 
Clock detection histogram:
    (us)
 73 6.08    381 ██▉
 74 6.17   2051 ████████████████ 
 75 6.25   4714 ████████████████████████████████████▉
 76 6.33   5109 ████████████████████████████████████████ 
 77 6.42   2886 ██████████████████████▌
 78 6.50    940 ███████▎
 79 6.58     79 ▌
...
 92 7.67    134 █ 
 93 7.75    552 ████▎
 94 7.83   1371 ██████████▋
 95 7.92   1228 █████████▌
 96 8.00    576 ████▌
 97 8.08    124 ▉
...
 110 9.17     63 ▍
 111 9.25    324 ██▌
 112 9.33    725 █████▋
 113 9.42    777 ██████ 
 114 9.50    869 ██████▊
 115 9.58    716 █████▌
 116 9.67    209 █▋
...
 126 10.50     60 ▍
 127 10.58    237 █▊
 128 10.67    413 ███▏
 129 10.75    672 █████▎
 130 10.83    834 ██████▌
 131 10.92    767 ██████ 
 132 11.00    482 ███▊
 133 11.08    178 █▍
 
This is is definitely 100tpi, and the sector structure is about right for 16x128 byte sectors --- it's a hard sectored disk. It must be from one of the early 4s.
 
So the structure, if my guess is correct, should be about the same as the Vector 3 for the HS media. I did some recovery for that in the distant past; I could look around and check my code archives. Right now, I can't think of what I did, offhand.
 
I'm pretty sure the Vector 4 hard sector disks use the same format as the Micropolis controller and the Vector dual-mode controller used in Vector's S-100 machines. These are hard sector controllers using MFM for double density encoding. The drives are 100tpi. I'd expect pulse intervals to center around 4us, 6us, and 8us. Since the disk is not soft sector, there are no missing clock flags to generate 10us pulse intervals. Of course, other pulse intervals will appear in the histograms due to unwritten and previously written areas on the disk and also due to the variance in position that occurs every time a sector is written.

So yes, I agree your histogram results are not as expected. Are you trying to archive Vector 4 disks you have?

Mike
 
Okay, so this is pointing at a bad read somehow. The machine's not mine; I've got someone who's sending me flux files from their machine; they have a FluxEngine connected to a Micropolis 100tpi drive. I've got them to swap out the drive for a known working one to see if anything changes.
 
Thanks very much. Silly question: what format is that in? It looks easy enough to reverse engineer --- the tracks are concatenated together with a six-byte data containing track, side and size --- but it doesn't seem to be either the standard CatWeasel raw format or the CWF format you sent me previously.
 
Well, these were made about 10 years ago, so each track is prefixed with a six-byte header as you say--and otherwise follows CW data format; that is--high-order bit is set during an index hole, the other 7 bits are interval counts. As well as I can recall, either a full 128K of samples is recorded (size of the CW buffer) or two revs (I don't recall which). IIRC, the sample clock is either 14 or 28 MHz.

Here's a histogram of a sample track:

Code: 
Cylinder 2 Side 0

  0:       0      0      0      0      0      0      0      0      0      0
 10:       0      1      0      0      1      2      0      0      0      1
 20:       1      2      0      4      2     98   9948  17882   2185      1
 30:       1      0      1      0      0      0      0      0      0     92
 40:    3650   4706    689      3      0      0      0      1      3      0
 50:       0      0      1     75   2767   1657     94      0      0      0
 60:       0      0      0      0      0      0      0      0      0      0
 70:       0      0      0      0      0      0      0      0      0      0
 80:       0      0      0      0      0      0      0      0      0      0
 90:       0      0      0      0      0      0      0      0      0      0
100:       0      0      0      0      0      0      0      0      0      0
110:       0      0      0      0      0      0      0      0      0      0
120:       0      0      0      0      0      0      0      0
 
As I recall, these were not just hard sectored, they were 15 hard sectored.
I still have some punches for the 15 hard sectored.
Dwight
 
After analysis (which both I and the other party did simultaneously and indepently --- ah, the joys of crossed emails!) it turns out that:

  • the drive was faulty
  • the real disk format was a strange tweaked version of Micropolis 1084, with a 16-bit checksum and a mysterious extra four bytes after the checksum that we haven't been able to identify.
So, we have made it work. Thanks, everyone!

On an aside: anyone know where I can get one of these hard sector punches? I need a hard sectored 5.25" disk to test with, and have been completely unable to find any.
 
hjalfi said:
After analysis (which both I and the other party did simultaneously and indepently --- ah, the joys of crossed emails!) it turns out that:

  • the drive was faulty
  • the real disk format was a strange tweaked version of Micropolis 1084, with a 16-bit checksum and a mysterious extra four bytes after the checksum that we haven't been able to identify.
So, we have made it work. Thanks, everyone!

On an aside: anyone know where I can get one of these hard sector punches? I need a hard sectored 5.25" disk to test with, and have been completely unable to find any.
Click to expand...

The extra bytes on the end were added by Vector Graphic with the introduction of their Dual-Mode controller. These extra bytes form an error correcting code that can actually fix errors after the fact. The repair was not done in real time as the disk was in use, however, you could run a utility that would use the ECC bytes to fix errors on the disk. Here's a link to the Dual-Mode controller that mention the extra bytes: https://deramp.com/downloads/vector_graphic/hardware/HD-FD documentation.pdf

A couple of guys bought out Athana's floppy disk inventory a couple of years ago. These are brand new Athana disks that I've had good luck with. Here are some of the 16 hard sector disks: https://www.ebay.com/itm/16-Hard-Se...218007?hash=item4438975557:g:pmcAAOSwXlpcip7F

Mike
 
Or, he can buy one of my punches. The only problem people have it that the punch gets dull and leaves a poor cut. The punch is made with drill stock but I don't know if it has been heat treated or even if it is the right kind of drill stock that works with heat treating.
There is one type that you can easily machine and then just heat it and let it cool. It is then hardened.
If the edge, of the biscuit hole, is a little rough, I smooth it down with the back of my finger nail. It works well enough for 10 sectored but I'n not sure how it works with 16 sectored. I did sell some of them originally and not had anyone with different complaints.
I only ask $35 plus shipping. I'd like to sell them all as I have no need for them.
Dwight
 
You must log in or register to reply here.
Back
Top