Hi Guy, And with which conclusion? I once thought about that, too. Don't remember if I was inspired by some posting somewhere. It would be quite easy. A $2 AVR with a crystal oscillator would to the job as well. Then you would only need a switch for the number of sectors. One prerequesite for this "virtual sectoring" would be a good stability of the disk rotation speed. With a microcontroller you could also resynchronize the software PLL with the original index slots. That would make the thing even more accurate! So, new question: Would there be any interest in a little circuit that can be added to any RK05 drive? Hundreds and only four? I have about 15 16-sector packs. And less 12 sector packs. But as I wrote - there's 12-sector pack supply... Kind regards Philipp

Good morning, Ah, ok... I understand :-) As far as I remember the drive outputs a sector 0 pulse and index pulses. I think of a little (!!) board (2*3cm) that can be taped to the RK05's card cage. You solder 5 AWG30 wires to it. You'll need WireWrap tool and unwrapping tool to do the job in 3 minutes. - Two wires are wrapped to 5V supply and ground. - One wire is wrapped to the index pulse (the thing that tells about the double slot) - The connection between sector pulse output and the bus interface is unwrapped - One wire is wrapped to the now open sector pulse output of the RK05 electronics. - One wire is wrapped to the sector pulse line on the bus connection. So the only modification to the drive would be unwrapping of one backplane connection. I find this acceptable. The board itself will have a jumper where you can select the number of desired virtual sectors. The controller on the board will count the pack sectors. If it matches the desired count, it will just pass along the sector pulse with some few ns delay which won't be seen by anything in the system. If the desired sector count does not match, the virtual thing applies. The final use case would be: You jumper your drive according to the system you want to connect it to. And then use any (!!!) pack. Wouldn't that be nice? I will do the PCB design tomorrow. This will be a new product in my will-be-famous webshop! But the alignment pack are 12 sectors afaik :-P Haha! What a blasphemic thougt just to format them :-) I have seen very few bad packs yet. But some don't work. I seen packs that just keep crashing. And packs that look great, don't crash, but have errors. I also have moved a disk from a 12 sector pack to a 16 sector pack. Centering is difficult. If not centered properly the drive will rattle uncomfortably. Kind regards Philipp

has anyone joined their local maker space to gain access to tools to do this sorta thing? http://makezine.com/groups/ On Mon, Jun 10, 2013 at 11:19 PM, Ethan Dicks <ethan.dicks at gmail.com> wrote:

Hi Ethan, On 11.06.2013 07:59, Ethan Dicks wrote: If I get it to work, I'll build some and sell them. The hardware design is quite primitive: An AVR, crystal and some passive stuff. The trick will be the firmware. I'm quite sure you would be able to design the pulse regenerator! Kind regards Philipp -- Dipl.-Inf. (FH) Philipp Hachtmann Buchdruck, Bleisatz, Spezialit?ten Alemannstr. 21, D-30165 Hannover Tel. 0511/3522222, Mobil 0171/2632239 Fax. 0511/3500439 hachti at hachti.de www.tiegeldruck.de UStdID DE 202668329

Hi, I think of doing it interrupt driven in software :-) I already made some rough estimations on the timings involved. I will use a 20MHz ATmega-AVR ATmega48 or ATmega88. The reason is that I already have that part. It's available for less than EUR 3 and does the job. I know it's overkill. But I'm not designing for mass-production where every penny costs thousands. I want to get the job DONE and the thing RUNNING. So I use the MCU that I have on stock and where I already have made some schematic templates in Eagle. Everyone who wants it "more retro" is free to do so. But why has it not been done yet??? :-P Some prophylactic info: I will *NOT* use any special magic driver IC for the generated index pulse output. A regular AVR pin has to do the open collector job. I will *now* play a bit with a TU56 drive because there's a bit DECtaoe frustration. When I have finished that I will go back to my Eagle. And I will probably finish AND BUILD my little board prototype tonight. So it could be up and running in the next few days. If nothing keeps me from doing so... Kind regards Philipp :-) -- Dipl.-Inf. (FH) Philipp Hachtmann Buchdruck, Bleisatz, Spezialit?ten Alemannstr. 21, D-30165 Hannover Tel. 0511/3522222, Mobil 0171/2632239 Fax. 0511/3500439 hachti at hachti.de www.tiegeldruck.de UStdID DE 202668329

I prefer to modify my classic computers using devices/technology that is similar ot the rest of the machine, or that in the very least would have been aroudn when the machien was in active use. In this case, I guess using an early 8 bit microprocessor is contemporary with the machine (there were plenty of RK05s in use in 1975), but I would prefer to do it with TTL devices only. I shouldn't be too hard, given the quite good stability of the RK05 spindle speed. -tony

Sure... Writign the firmware takes work too... Remember this si a hobby (at least for me) so you do the sort of work you enjoy... I only want ti to do one thing -- resector the disk pack. Does it ahve to fit inside the drive? I am modering if you can make a device that fits ebtwee nthe drive and the cotroller, the advantage being that it then converts all deives on that inerface cable. The downside would be that you woudl ahve to reproduce the sector counter in the device, and that the secotro count would eb invalind if you change drives until the next index pulse. But is that a real problem? -tony

On 06/11/2013 12:41 PM, allison wrote: Pretty much what I was thinking. Microcontrollers aren't necessary. The PLL need only provide an output in the ratio of 3:4 of the actual vs. the synthetic pulses. In other words, every 3rd sector pulse from the RK05 should be coincident with every 4th synthesized pulse. --Chuck

I'm interested in the "little circuit that can be added to any RK05 drive" proposal, Philipp :->. On Mon, Jun 10, 2013 at 8:10 PM, Philipp Hachtmann <hachti at hachti.de> wrote:

A few monthes ago there was a discussion about modifieng 12 secor packs to make them 16. very handy but not something I have the equiopment to to, however I have been experimenting with softsectoring an RK05 and tonight got a susessfull 16 sectore format of a 12 sector pack. Im pulling out the raw secotor/index pulse train from the M7700, feeding this into a PIC18F8720 and using this to synthesize a 16 sector train which is fed back into the M7700, ok it does mean I have to modify the M7700 but its only 1 cut track and 2 wires out, also its not at all peroid correct ( sorry Tony ) but if it means I can have more than the 1 usable 16 sector pack I have right now then im a happy bunny. When I have done some more testing and tided it up a bit I will post more details, probably via the vintage-computer.com DEC forum as its quite active and I expect most DEC people here keep an eye on it DaveH

----- Original Message ----- From: "Tony Duell" <ard at p850ug1.demon.co.uk> To: <cctalk at classiccmp.org> Sent: Saturday, September 07, 2013 7:20 PM Subject: Re: Soft Sectoring a RK05 I dont know yet, so far al lI have done is half a dosen formats on one pack, all ok, yes read only check would give a better take , i will look at the format program, should be able to startit at the read check so missing out the format part, or I may just rty and do a os8 restore image and see how that goes. In general I agree and find fault finding and repairing ( and having to) down to basicly gate level gives me graet deal of stisfaction and to be honest is probably what " drives me", if it all worked all the time then it dowdent be half as interesting.and as you say this technology can still be repaired , later is much harder. This exercise has meant that I have had to delve into othe inner working of the RK05 drive so Iv leant a lot , this drive worked when I got it so iv not needed to look at it in detail before, the RK8E on the otther hand needed lots of work, this board set is a bit of a nightmare, at laest 2 versions and lots of engineering change patches, in fact I ended up changing the design a bit and adding an extra latch to get rid of a timing problem that was givving the CA register an extra incument on load. I am just wondering how impossible it This will be the next version, TTL only Thats going to be interesting , and lots of fun.. Dave

On Mon, Sep 9, 2013 at 12:49 PM, Tony Duell <ard at p850ug1.demon.co.uk> wrote: A DPLL seems like overkill, when a one-shot for index recovery, an analog PLL (e.g. 4066 or 74HC[T]4066) and a counter should be more than adequate to do the job, and would make the circuit much simpler to design and debug. If you want selectable sector count, a programmable down counter (e.g. CD4059, 74HC[T}4059, or MC14569B) could be used. Personally I'd be fine with using a PIC, since it's apparently already designed and working. Presumably the PIC circuit can be removed, so it's not permanently contaminating the system with non-period components.

I did concider trying a PLL, after all it is the obviouse solution, however I had the PIC dev kit lying unussed after a cancelled contract and this seemed a good reason to play with it, real application, not just led flashing. Also the last PLL i worked with, years ago now, could not be described as robust, locked to everything BUT the intended pulse. No its not rocket science, trap the index pulse, BTW its not midpoint but about 600uS after the 15sector pulse, time one rotation, divide by 16 and use in counter synked to the index, as Tony pointes out, on the RK05 the platter speed is not going to change quickly and the sector leadin at 100us is probably generous compared to more modern drives. You do have to watch the timing a bit though, as an error in the sector time is cumulative. The programming is in fact almost trivial, though there was still a small sense of achievment when I got the format compleat printout, espeshally as this code was unchanged from the final bench test ( pulse generator and scope) test, no patches, no timing fiddles. When I get the time, im sure most here know that problem, im going to do a TTL version, maybee PLL or maybee not, unless someone else does it first. Meanwiile im cleaning some 12 scotor packs so I can do further testing and do an os8 restore. Dave

Sure, and there are evenb official HP scheamtics for the 7900. there is, of course, the very minor problem that it's a twin platter drive (one fixed, one removable), with separate secor transducers, so the modification would have to be switched in only when the remvoeable pack was selected. That's trivial to do. Mod modifyign the pack is easy too. It's jsut a matter of remving the hub, putitng it in a dividing head, careully lining up a sutiable slitting saw with one of the existing slots, then rotatign it by 1/24th of a turn, cutting a an extra slot, rotating it by 1/12th of a turn cuttign another slot, and so on. And the resulting pack would then work i nan unmodifed drive, for exampel on somebody else's machine Actually, that is an issue with modifying the drive rather than the pack. A modified drive can;'t read 'original' packs, and the packs recorded on the modified drive can't be read on an unmodified drive. You cna't exchange packs with friends unless they also have the modification fitted. I could have sworn that it was a separate part, but it looks like _I_ was wrong. All the RK05 packs I've just looked at have the sectorign slots cut in the main part of the hub. Oh well. I am now tying to think where I saw a separate sectoring ring. Makign a new hub is not goign to be easy. The 2 faces (the one that lovates on the spindle flange and the one that carries the actual disk) have to be very parallel. I think the tolerance on the central conical hole it sight too, and that is too small to be able to use a boring tool ot make it. I think you'd need a special tapered reamer. Modifiyign the drive would thus appear to be a lot more practical for 16 sector machines. For HP 24 sector ones, cutting the extra notches seems quite possible. I must give it a go, if I can ever get my 7900 to work again. There seem ot be 4 (?) holes in the hub just inmside the centring ring. The edge of the hole i nthe platter is vising through them. I wonder if they were sed for some kind of alignemt fixture at the factory. For the final alignement, you put the disk/hub on a spidnle (an spare RK05 spindle is fine) and use a dial guage o nthe outer edge. Be warned its tedious to do right... Ah... My time is free. Or at least nobody will pay for it. -tony

[Cutting extra slots i nthe sectorign ring] Sure. As I mentioend, that is one that I do want to make. I have this mad idea of putting together a complete HP9880 system for my HP9830. For those not into old HP 'calculators', the HP9880 is the disk system for the HP9830. It consists of a number of units : HP11305 controller (a failry complex microcoded thing) HP11273 'cable' to link the HP9830 to the controller. This 'cable' ahs a PCB at each end with about 25 chips on each PCB. The sector buffer RAM is one one of the PCBs, for example. In fact uo can plug 4 such cabels into oen controlelr to chare it between 4 HP9830 machines. HP11273 ROM module. This fits in the HP9830 and contains the bootstrap, etc. HP11302 'cable' to link the controlelr to the disk drive. This has a PCB at only one end and contains mostly signal buffer ICs HP7900 drive. This is a twin-platter thing with oen fixed platter and oen remvoable pack which is similar to an RK05 pack. The drive has 4 heads on oen carrage and thus can onlt be used if there is a removable pack in place. There's a separate sectore sensor for each platter, the one for the fixed platter is set up for 24 sectors (25 slots total in the ring), the 11305 assuems this (although it takes sectors in pairs...). So the removabel pack ahd better be 24 sector too. Unlike the RK05, the comaprison of the cector counter is done in the drive -- the controller loads the wanted sector number into a register in the drive and there's a signal from the drive to the controller that is asserted when the sector counter (reset by index, ioncrements by sector) has the right value. A PSU (I forget the number) for the drive The removable pack A casette containing teh OS that is copied to the disk on initialisation. So far I have alsmot everything. THe only things I am missing are the correct 24 sector pack and the PSU for the drive. Alas my drive needs a lot of work to get it going again (I hope the heads are OK...) but I intend to give it a go. For tesitng the drive shoudl spin up with any pack in the drive. Since the fixed paltter is alsways 24 sectors I cna test with that. But a remvoeable pack would be nice too. YEs, I guess a mill and dividing head are not all that common.... Right... No, I didn't realsie it would autom,atically detect the original type of disk, I am pretty suree I have the machines to do it... Anyway, as I said last night, it might be possible to machine off the notched rign form a normal hub nad fit a new ring wit hthe right number of notches. That would mean you could make 16 sector disks. A small lathe/mill is something to consider. Be warned, thoguh that you will want ot buy/make a lot of add-ons, and that learning to use it well take a long time. And that it quickly becomes addictive (hmmm... Like computers, I guess). -tony