On 04/01/2014 18:11, Robert Jarratt wrote: DEC were infamous for using "sniffer boots" during which the controller would use a number of tricks to identify drives by their idiosyncrasies. It does mean that apparently similar drives don't always work, depending on the controller firmware revision, most notably with RQDX1/2. Anyway, working out the number of heads is trivial. Every sector on any vaguely conventionally formatted MFM drive has a header with an ID mark which includes the track, sector and head numbers. So you just select each possible head in turn, from 0 up to 7 (most drives) or 15 for controllers that have a 4-bit head select, read a sector, and see if the returned head number is correct. The first time it's not, you've just exceeded the head count. -- Pete Peter Turnbull Network Manager University of York

on each that number So how does it do this when the disk is yet to be formatted? In other words, when it formats the disk how does it know what size the disk is? Regards Rob

On Sat, Jan 4, 2014 at 11:16 AM, emanuel stiebler <emu at e-bbes.com> wrote: How does that match with this information, if it is correct, that the MicroVAX 2000 disk recognizer works for the set of known disk drives even if the disk came from a different system and has no DEC specific information or formatting on it currently? http://home.iae.nl/users/pb0aia/vax/fmtbob.html "The standard DEC drives that are supported in the VS2000 and their industry equivalents are: RD54 - Maxtor XT-2190 RD53 - Micropolis 1325 or 1335 RD31 - Seagate ST-225 RD32 - Seagate ST-251-1 If you have any of these standard drives you can simply plug it into your 2000 and the ROM formatter will automatically recognize the drive types. I've never seen this fail, even if the drive has never seen a DEC machine before, or even if the drive was previously formatted in a PC."

Some (all?) ST412-interfaced drives report 'iilegal' head selections and seeks by asserting the WrtieFault/ line. The ST412 diagram I've jsut looked at suggests it will assert said line if either more than one head is selected (indicating a fault i nthe selection circuitry) or if no heads are selected (possibly due to the head selection lines tryign to select mroe heads than the drive has). There's a link on the ST412 PCB shich is removed if the board is used in an ST406 (I do not mena an ST506!) which has fewer heads, so as to cause this error if heads only present in the ST412 are selected The microcotnrolelr which cotnrosl the head posiitoner stepper motor i nthe ST412 can also assert WriteFault/ My guess (and I do not have the firmware source) is that it does this if you try to step in too far. If other drives behave in this way, its' farily easy to find out how many cylinders and heads they have. -tony

You don't need a very high sample rate to hadle the ST412 interface signasl, so jsu aobut any logic analysr would do. They are not that hard to find now, but if you buy a second-hand one, make sure it comes with the pods. Finding pods is not at all easy. Making them is perhaps a little easier, but still a lot of work. I am told you can now get a USB-interfaecd thing that turns a PC into a logic analyser. Since I want ot keep PCs ase far from my electroncis bench as possible, I've not looked into these, but I have heard good reports of at least some of them. -tony

On 01/05/2014 11:45 AM, Tony Duell wrote: Just remember that the read and write signals to an ST512 drive are differential, not single-ended. Not a major issue, but worth knowing. --Chuck

On 01/04/2014 12:19 PM, Tom Gardner wrote: First I keep a uVAX2000 as a DEC compatible formatter. I also have XXDP diags for my -11 based systems for RQDX2 and 3. It does one of two things. Read the disk for previously placed info and uses it. Failing that the user MANUALLY enters that. If you don't what the number are... Try google for the drive data all the the most wierd are there. Allison

On 01/05/2014 01:35 PM, Ethan Dicks wrote: The concept of BIOS is limited to two instances, PCs and the hardware abstraction layer of CP/M that was user customizable. CP/M came first. PCs made it a part of the hardware as resident firmware. The closest thing to that in DEC form was TEV-11 or The resident boots on 11/23, and later machines. Funny I was doing some file organization work to do some fun with CP/M and found a list from usenet of drives then current (likely around 1989) and it was a long list and some were quite odd. The way PCs did things were quite odd compared to the rest of the industry at the time. They would later be come the new-defacto standards by force of volume. The twist in the floppy cable for example was only on PCs and not seen elsewhere. The rest of the world had to know jumpers to set drive number but could have four rather than two. If you go into other than PC systems with a PC mindset, you are likely universally going to go down the wrong path of beliefs. Allison

On 01/05/2014 11:33 AM, allison wrote: The industry in general has suffered from a lack of forward-thinking. Shugart could have handled the motor issue by using two pins on the interface (one to turn the spindle motor on; the other to turn it off), conditioned on drive select--a trivial modification. I suspect Shugart was still locked in the 8" mindset (AC operated spindle motors running all the time) to do anything different. IBM's solution was pretty clever and also did away with the need to change drive-select jumpers. Hard drives were really a mess. When we started out with the very new 5.25" ST412-interface drives, we offered only two variations a Rodime 7MB and a 14MB model--and they were hard-coded. To our credit, communication with the driver was done by relative block, not CHS. After Rodime discontinued the 7 and 14MB models, and sold us 10 and 20MB models for the same price, we kept the same driver, so 1/3 of the drive went unused. This was *after* we'd offered the Shugard SA4000-series 14" drives up to, what, 40MB? Stupid, stupid, stupid. The same shortsightedness extended to display resolution, CPU speed assumptions and just about anything else you can name. I'll give Apple that one--they tend to think a bit ahead of everyone else. But we're still lacking in real visionaries. --Chuck

On 01/05/2014 05:14 PM, Peter Corlett wrote: Along that same thing, the NEC 765 controller as used in many PCs was originally an 8" beast. It has a rather novel feature of automatically polling the READY line of each drive and initiating an interrupt should the status of READY change. Of course, since the PC uses drives with no READY line, this had to be worked around. So the READY input on the 765 is permanently tied active and the unit-select outputs of the 765 go nowhere. IBM installed a separate motor-control and drive-select register to get around this. However, a drive not ready condition will "hang" the 765 until the drive comes ready (because of the permanently active READY input). So a software timer is used to reset the 765 should an operation take too long. Clumsy, but it gets the job done. Back in the 70s, when we first obtained 5.25" drives (Micropolis), we had a READY signal, but that hardly meant anything because we had a motor control register also. Our controller used the now-forgotten WD1781; not too different software-wise from the 1771. We had a problem with customers changing floppies with files on them still open for writing. We investigated door-lock solenoids, but dismissed them because of added expense and limited choices. So how do you tell if a drive is ready if the motor's not on? We figured that keeping the drive motor on so long as a file was open for writing, but that could be hours and hours. We eventually stumbled on sampling the write-protect signal which didn't depend on the motor being on. About every quarter-second, we sampled a drive's WP status if the drive had any files opened for writing--if it changed, we halted execution of the current application, put a message saying essentially "Now put that disk back!" and sat on the beeper, turned on the motor and waited for the drive to come ready again. We never had another complaint on that problem again. --Chuck