high density disks

for bbc micro/electron hardware, peripherals & programming issues (NOT emulators!)
frankoid
Posts: 53
Joined: Wed Jan 10, 2007 9:18 am
Location: London

Re: high density disks

Postby frankoid » Thu Oct 16, 2008 2:52 pm

Good stuff Martin, looking forward to your next post :)

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Thu Oct 16, 2008 11:34 pm

Thanks for the feedback everyone – I do enjoy this sort of thing but it’s only really satisfying if you have people to share it with so even if it’s just a good read for a few others then it’s worthwhile :)

Bit more reading, bit more research, lots more experimenting and finally on to the ‘mod’ I mentioned. For dramatic effect I’ll just start with a cryptic picture of what I did in “30 seconds” and then hopefully Mark will read to the end before he pokes me in the eye with a datasheet :lol:

[ Actually Mark, this is good timing (sic) since I think you have designed a 1772 replacement on one of those smart-ass chip thingy’s and before you roll the presses there’s a couple of extras you might like to embody.]

Ok, I take a Beeb with one (for simplicity) 3.5” PC drive, as we all use, and an Acorn 1772 DFS. This one happens to have my dual controller mod so there’s an 8271 sat there too but that’s not actually relevant. What is significant to my “30 second” trick is that there is a loose wire clipped to the Beeb motherboard via a hook probe. (See my other thread if you’re already confused.)

I put a HD floppy (no sticky tape) in the drive and type *FORM80 02 and, no prizes for guessing, the drive briefly whirrs before reporting a drive fault. So, all I do next is unclip the probe and move it to a different point on the motherboard. I again type *FORM80 02, the drive whirrs and then smoothly, quickly and quietly twice counts from 00 to 4F. I type *VERIFY 02 and the drive whips through the two counts again. I save test files, dump them, type them, delete them, cat them, load them etc. etc. I try new (HD) discs, old discs, full discs, blank discs, blue discs, red discs, black…..well, you get the idea and yes, the drive is operating in full HD high flux, high speed mode.

So, what’s going on? Well, I have been discussing finding a way to get the drive to use the correct flux for HD discs whilst keeping the DD (250Hz) interface. However, due to smart-ass chip thingies, I haven’t yet found a way of separating the functions in a drive.
However, in my brief excursion away from Acorn (shame on me), I formed an Atari collection of ST’s, Mega’s, STE’s etc.(which are all still in the loft) and I remembered that there were discussions about how one could do something to some of these machine to allow them to use HD drives and discs. I never followed this up until now and, with all due thanks to our contemporary Atari community, I now know what was behind the discussions and it’s all about the 1772 and clock speeds.
Apparently, in the Atari TT (and/or STE I think – need to check) there was a ‘super’ 1772, the 02-02 or ‘Ajax’, which was designed to accept a 2x faster clock at 16MHz rather than the original 8MHz. This allowed the chip to use the higher 500Khz interface and hence be compatible with HD drives. Now, and here’s where things become slightly anecdotal, it also seems that virtually all of the later ‘ordinary’ 1772 chips, although officially only rated upto 10MHz, would happily work at 16Mhz. As Mark alluded to in the DFDC thread, many of these older design chips were incredibly robust and over-engineered such that ‘overclocking’ was no big deal. It’s certainly not the case for the original 1770 (I’ve tried one and it couldn’t cope) but it does seem to be true of the 1772. In fact, somewhere deep in some internet folders (where I wasn’t meant to be!), I came across an informal technical memo from Western Digital which stated these things and even went on to say that some chips would even cope with 32MHz ED (Extended Density) but frustratingly I can’t remember where I found it :(

The Acorn 1770 upgrade kit uses a 1772 equivalent chip by VSLI called VL1772-02PC and I can’t specifically find a datasheet for it but I think it was still officially spec’d as a 10Mhz chip. So, as you’ve probably now guessed, when I moved my hook probe, I removed it from the 8MHz clock and connected it to the base 16MHz clock. This causes the 1772 to now run it’s interface at 500KHz and hence function with the HD drive and disc. The only extra ‘effect’ is that this also halves the Head Step time and so it’s important that the relevant keyboard links have not been changed from their open default. If we had previously set them to 3ms then we would now be trying to achieve 1.5ms which is too fast even for the best drives.

Of course, just hard changing the 1772 clock would commit the interface to HD only so in order to allow mixed drives and discs we would also need some simple clock switching logic which would ideally utilise things like Pin 2 ‘Selected Density’ output by the drives (now you know why I mentioned it) and ‘Drive Select’ etc. so that the 1772 could be automatically switched up and down according the drive and density in use. I am working on that but I won’t go into it now, I’ll wait to see what reaction I get to the basic idea.

So Mark, if nothing else, your ‘new’ 1772 based on a smart-ass chip could undoubtedly eat 16MHz for breakfast and with a little extra logic could make HD drives the media of choice. Yes?

Martin

User avatar
sorvad
Posts: 2172
Joined: Wed Aug 24, 2005 12:13 pm
Location: Back of beyond

Re: high density disks

Postby sorvad » Fri Oct 17, 2008 6:10 am

Gosh, you are so talented with hardware, seriously impressed, have you any big ideas for new hardware for the beeb besides a disk interface type thingy ? Bet there's loads of things mulling about in your head.

User avatar
retroclinic
Posts: 3016
Joined: Thu Jul 03, 2008 1:22 pm
Location: East Riding of Yorkshire
Contact:

Re: high density disks

Postby retroclinic » Fri Oct 17, 2008 4:20 pm

Hi.

Another good post there! When you've got the interface set up for 500KHz, presumeably the drive is not spinning the disc at double speed, so therefore when the system is in SD/HD mode it's writing all the data on only half the track each time? I'm presuming it doesn't read normally written 3.5" disks in that mode? If so, then that would be good, as a small hack to the 1770 DFS ROM to write twice the amount of sectors on a track would give us 400k per side instead of 200k? It also opens up the possibility of going double density and getting the full 1.6MB from ADFS in a similar mode to the arc, but then you're asking the Beeb to write data to the 1772 at more than double it's rate in normal ADFS, and I'm not sure the 6502 at 2MHz is quick enough.

The Interface I redesigned actually still uses a 1770 chip, I sourced some NOS VL1772-02PC from the states to use with it. I looked into reproducing it in an FPGA as the VHDL for it has already been written, but it worked out just too expensive to go that route. The difference from the Acorn board is it uses a PAL for the latch and control logic, so does allow some adjustment, but pin 18 which is the clock is just a straight feed from the 8271 socket, as there was no reason to change that - although it could be hacked with a flying lead I suppose like yours does.

I've not found the datasheet either for the 1770/2, just some register info. It would be good to have that!

Mark.

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Fri Oct 17, 2008 6:45 pm

Steve : On the one hand you’re right, my head’s full of these ideas but on the other hand you’re wrong because I think many of these ideas are going to outstrip my hardware talents!

Mark : Sorry, my mistake - I thought you had replaced the whole thing. I am actually familiar with these programmable logic devices but don’t have any personal experience so I wouldn’t pretend to know how complicated or expensive such a device would be :?

I did actually come across one of your VLSI 1772 purchases – it just popped up from one of my many searches. Made me smile that it was advertised as “a rare Phillips TV chip” :)

No, the drive speed is constant and I had wondered about the sector geometry but just assumed it would ‘sort itself out’. I’d have to sit down and do some serious thinking about how the data might look on the disc but I’m not too fussed at the moment. Not sure what you mean about normally written discs? Since there aren’t any genuine Beeb DFS HD discs until I start making some with this mod then there’s nothing to test. If you mean DOS/Windows 1.44Mb HD discs then I was going to try tweaking the disc map of my RAMagic DOS to DFS code and just give it whizz. I usually have more luck in practice if I spend less time on the theory! As you say, there may be timing issues but I haven’t sat down with my calculator yet.

At the end of the day, I was really only interested in opening up a supply of brand new drives and discs so that floppies can be used for a lot longer yet. I know that lots of people are moving on to solid state drives but I just like using floppies.

If you get chance, would you be able to give the fast clock tweak a try on another random 1772 for me? Just the basic single drive test as I did would be fine. It would be interesting to see if you get the same stable results. I’ve roughed out a simple automatic clock selector as I mentioned and if it all hangs together it might be useful to anyone else with a passion for discs.

More to come as usual :wink:

Martin

Edit Add : Forgot to answer your comments [Mark] about 5.25" drives - haven't got any HD floppies to hand but don't they have a second notch accross from the write protect notch which again triggers a HD/DD sensor in a HD drive? I could be mistaken but I think that's what I was assuming. I thought I could then just do exactly the same as the 3.5" drives, i.e. nothing :) , and that the drive would auto-select the density and the matching spindle speed. Anyone know for sure :?:

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Fri Oct 17, 2008 11:28 pm

Answering my own question - that's nonsense [-X . The notch I've seen must have been on 'flippable' discs meant to be used in ancient single-sided drives.

So, using a HD (but DD capable) drive the density would have to be selected with the 'DENSEL' input on pin 2. However, I don't see why there would ever be a requirement to use HD 5.25" discs. I can see that being able to use a PC HD drive would be beneficial because on average they're going to be several years younger but I think such a drive would only be needed in order to read existing 5.25" DFS discs. I even doubt that many users would want to write to a 5.25" especially if the drive pair includes a 3.5" unit in which we can now use new HD floppies.

So, for the 5.25" drive, I think I would tie DENSEL to the drive select line for this unit (conditioned as necessary) such that whenever it's active, its in DD mode. The same line would then signal back as a pseudo DENSEL to the clock switcher that 8MHz was required. Is there a seperate input for spindle speed or is everything linked to density input :?: If there is, I'd just tie it to whichever is needed for DD.

For the 3.5" drive, DENSEL would simply be an output back to the clock switcher (again gated with drive selected) to select 8 or 16 MHz depending on the density of the inserted disc.

Thus, the two drive selects would be used to enable one of the two DENSEL inputs which would in turn select the fast or slow clock. If a twin 3.5" setup was used, no logic change would be necessary but now either DENSEL line could be fast or slow.

Not much logic then, a data selector with maybe four enables selecting one of two clock inputs to a single clock output.

Martin

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Mon Oct 20, 2008 10:02 pm

Just an update to keep interested readers in the picture :wink:

I’ve tried three 1772 chips now, two VSLI and one WD and all three have worked fine at 16MHz. That’s not a guarantee that they will all work but on two 1770’s I’ve tried it’s immediately obvious when they don’t - They don’t break, they just can’t cope with the blistering speed!

I’ve also been looking at drives to see which output ‘Density’ on Pin 2 of the 34 way connector. They don’t all do it and there seems to be one of three configurations – ‘Doesn’t do it’, ‘Does it if you set the links correctly’ and ‘Does it’. Basically, this means you get either a logic ‘1’ or a logic ‘0’ depending on which type of disc is in the drive and hence which mode the drive is operating. Additionally, the signal will ideally be tri-state meaning it goes high impedance or ‘disappears’ when the drive is not selected so that a second drive on the cable can signal on the same line without a fight.

Digressing slightly for a moment but on the subject of links, I’ve found and attached an interesting document which details the available links and their options for a number of common drives. It gives a good general insight into how drives can be configured and in many cases you will see the references to density and Pin 2. Well worth a flick through if you use 3.5” drives.
All of these drives seem to be ones with the familiar pin jumpers but I’ve noticed from inspection that many drives, even when described as ‘jumperless’, do in fact have configurable links but not as mechanical jumpers. The Sony drives for example have pairs of solder wetted pads marked ‘JCnn’ which can be joined with an iron. These (as with other makes) are often factory joined with what look like tiny surface-mount resistors but are marked with ‘000’ which means ‘zero ohms’ or in other words, a link! You can often use these to change the default drive number of the unit where it is theoretically jumperless and fixed at 1 or 0.

Back to the project, I have found that Pin 2 of the 34 way disc interface is connected to the West-most pad of a three-way PCB pad, S4, but is otherwise isolated and not used. (Centre and East of S4 are connected by a PCB track.) Therefore, it is very easy to pick up Pin 2 and hence ‘Selected Density’ via a wire soldered to the West pad of S4. Drive select 0 and 1 can be picked up easily from chips near the connector (e.g. IC79) and the 16MHz clock is available from the East point of another PCB pad, S27. So, all the signals are readily available for use on a small one or two IC mod board. (The clock signal to the Acorn 1772 module must also be hijacked and this is described in my thread, 'Dual .. FDC', where the clock input line ends up on a flying lead. Note though that you don't have to do the whole DFDC mod.)

I think I will initially go for a 3.5” HD/DD + a 5.25” DD drive and I have found that the (only) 5.25” drive I have tested holds Pin 2 high (i.e. not tri-state) so this would need to be either disconnected or buffered through a tri-state using drive select as the enable. This is probably the simplest configuration because the clock switching circuit would simply default to 8MHz (DD) unless ‘Drive n’ (the 3.5”) is selected AND the ‘Density Selected’ line is indicating HD when we would switch to 16MHz.

This is actually quite a simple circuit and the end result would be totally transparent to the user except that both DD and HD floppies could be used freely in the 3.5” drive and both types would be read and written as designed – i.e. no fading data, brand new discs still in the shops :D

A twin 3.5” set-up wouldn’t be much different and again, totally transparent for both drives.

So that’s where I’m at, feel free anyone to run with it if you’ve got the general idea. I will eventually describe a practical mod in detail but as ever I’m multiplexing between many projects :roll:

Martin
Attachments
3dot5_fdd_jumpers.zip
(522.34 KiB) Downloaded 149 times

User avatar
retroclinic
Posts: 3016
Joined: Thu Jul 03, 2008 1:22 pm
Location: East Riding of Yorkshire
Contact:

Re: high density disks

Postby retroclinic » Tue Oct 21, 2008 10:04 am

Hi.

Keep the good work up! Just a note on 5.25". They hold it high, because it's an input. A 5.25" has no idea what type of disk is in it, so relies on an input on pin 2 to be told to either use SD/DD @ 300 RPM, or HD @ 360 RPM. Some later drives used only 360RPM and relied on intelligent controllers in the PC to adjust the data rate for DD. (Those georgeous single height dual 3.5" & 5.25" Canon drives do that....grrr....)

I'm not sure how this would co-exist when pin 2 is an output on a 3.5" with a 5.25" on the same bus?

The reason for trying to get 5.25" HD disks to work on a Beeb, is it instantly doubles the available blank media in existance. New HD disks tend to sell for peanuts as the only thing that can use them is PCs, on which people tend not to bother with floppies now anyway. Although DD are still out there, they tend to command a good price.

And by normally written disks, I mean a standard DD 3.5" disk, written as Acorn intended at 8MHz. I'm presuming that at 16MHz, it then won't read that as the data rate is doubled? If the spindle speed is the same, then it would mean it's only writing half the track - not an issue I suppose, but you do end up making a "propriatory" format that only machines with the mod could use.

Mark.

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Tue Oct 21, 2008 12:12 pm

Hi Mark.

Thanks for that, it’s good to have someone to bounce ideas off. Two heads… etc. :D

Hmm.., on the 5.25" HD drives then, I can feel a switch coming on :wink:. I didn't really want to get into modifying the DFS code because then things start to get a bit out of hand and the complexity just mushrooms. Having said that, it could be done in that the DFS could monitor for drive fault $0E, which seems to be the error code for situations such as mismatched densities, and in response switch Pin 2 to the 5.25” with the clock speed to see if the drive becomes responsive. This is something like the modern PC's do things I think.

However, it might be just as easy to add a switch on the drive itself to select HD/DD to the drive and then pass the selection back out on Pin 2 via a 10k or so (to avoid a chip) such that the other drive can still pull the line Hi or Lo as necessary. As I said, it would then be important that the 3.5" drive is tri-state on Pin 2 according to it being the currently selected drive. I guess this starts to come down to personal preference in terms of ease of use versus complexity of the mod. I think a switch would suit me as I mostly use 3.5" and I'd be happy to manually switch the 5.25" depending on the media type. We are already used to this idea to some extent with 40/80 switching?

Having thought about it, the sector geometry wouldn’t actually change would it? There would still be 10 sectors evenly around the track but the difference would be that each sector would only ever be half full @ 256 bytes instead of the PC’s 512 bytes. Whether or not this is then unique is a kind of self-fulfilling prophecy - since we are only going to be able to read and write HD discs with this mod then it doesn't matter if it's a unique layout. Hence, the ‘proprietary’ aspect is in itself the HD media.
More sums would suggest that the 2MHz Beeb isn't upto full spec. PC 1.44Mb so just leave it as standard DFS and the user will be non the wiser. The presence of the clock mod would be totally invisible apart from the (godsend!) ability to use DD and HD media and, as you say, made better still if that includes 5.25" HD floppies.

Keep talking Mark, there's decisions to be made..... :lol:

Martin

User avatar
retroclinic
Posts: 3016
Joined: Thu Jul 03, 2008 1:22 pm
Location: East Riding of Yorkshire
Contact:

Re: high density disks

Postby retroclinic » Tue Oct 21, 2008 1:20 pm

For 5.25" HDs, if we can get the drive to spin at 300 RPM, and just adjust the write current, then there's no need for a different data rate in the beeb - so we could as you suggest, just have like a disk density switch. The mod though would have to be done on each drive, so we'd need for startes a list like that one you posted for 3.5" (great BTW), for the 5.25" and the more common drives that the Beeb uses - I know most of the later ones are all HD compatible, and even some of the earlier ones are too, as they have 2 speed control trim pots on the motors for the two speeds.

I'll have a fiddle with a drive myself I think, you've got my interest going again on this....need to find a mech that has clearly adjustable motor speed control....

Mark.

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Tue Oct 21, 2008 1:56 pm

...you've got my interest going again on this...

I knew I'd break you down in the end - happy days :D

Martin

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Fri Oct 24, 2008 9:55 am

Progress report :)

Having been sidetracked onto adventuring again I haven't had chance to do much over the last few days. However, I did dig out two PC HD 5.25" drives and, without changing a thing, hung them onto my Beeb with the 1772 clocked at 16MHz and they both just worked. I used HD floppies and ran the usual suite of tests without a hitch. As it stands then, I have a Beeb with two nearly new (well, very low mileage) drives for which I have stacks of brand new HD floppies and all just by changing the clock line to the 1772.

If we were starting from scratch, i.e. nothing had gone before with Beebs and floppies, then this simple change would be just fine but the problem is that we have to look after the legacy discs which are of course all DD discs. The 3.5" drive will be easy, just need to make sure we use and configure a drive that outputs the media density on Pin 2 and we can use that to automatically switch the clock between 8MHz (DD) and 16MHz (HD).

As Mark and I have suggested, the tricky part looks as if it will be to get a 5.25" drive to switch between DD and HD as required. I think there will always be a need to play or at least copy those original 40 or 80 track floppies and so we can't just leave the 5.25" as a HD only drive. I've tried pulling and pushing Pin 2 on the 5.25" drives but haven't managed to get one to read a DD disc. Trouble is, these things have a dozen or so links inside, some of which are connected to/with this density selection configuration, and finding information on these links is not that easy.

Mark - since Pin 2 is accessible from inside the Beeb on S4, I think I will persue a switch on the back of the Beeb to select the 5.25" to HD or DD depending on the discs in use and the 3.5" will be automatic and transparent.

Ultimately it would be good if we can end up with a set of 'How to..." instructions for use by anyone interested in using HD drives and discs on a Beeb. (My motivation for these things is always to pretend that that 'anyone interested' is more than just me :wink: )

Martin

User avatar
regregex
Posts: 558
Joined: Sun Jan 02, 2005 9:51 pm
Location: London, UK
Contact:

Re: high density disks

Postby regregex » Fri Oct 24, 2008 10:34 pm

Thanks Martin for the detailed info. Nice to hear you got HD working.

Regarding 5.25" drives, I have a TEAC FD55GFR-140U that handles DD data on HD discs just fine. I've set the jumpers to make it spin at 300 rpm always. It would be a simple matter of giving an input to pin 2 to set the speed, but as pin 2 is currently unused this shows that selecting head current doesn't appear to be an issue.

I did try writing in HD at 300 rpm, this almost works but gets lots of errors, which is probably why they increased the speed to 360. The same controller has formatted 3.5" HD discs and written data to them so I know that's ok.

--Greg
Attachments
teac-fd55gfr-140u-jumpers-bbc.jpg
Jumper settings on TEAC FD55GFR 140-U. Made: D1, DC, E2, FG, I, LG. All others unmade. Serves DFS drives 1 and 3 (D1). Disconnect FG if the drive housing has its own earth connection.
(59.49 KiB) Downloaded 3095 times

User avatar
retroclinic
Posts: 3016
Joined: Thu Jul 03, 2008 1:22 pm
Location: East Riding of Yorkshire
Contact:

Re: high density disks

Postby retroclinic » Fri Oct 24, 2008 10:45 pm

Hi.

The issue with head current is that unlike the 3.5" disks, the coercivity difference on 5.25" is a factor of 2 - HD disks require a much higher write current in order to write the info correctly. The speed is so that the data rate can fill the track fully, and won't have an effect on it's own to the reliability of the data written.

I've found some early Mitsubishi drives, especially the ones with the push down clip, rather than the 90 deg turn, and some early Teac too, will sometimes appear to write HD correctly, whist others of exactly the same make and model number don't. I think it's a case of them just being on the "edge" of making the data readable again.

What we need to do is to be able to use pin 2 to select the write current, whist leaving the spindle speed at 300rpm, and the clock to the FDC at 8MHz. That will allow HDs to be used with DD data, and "should" allow them to be read back on non-modified DD drives, although on a read the head will output a higher voltage as there is more magnetism stored in the disk, so may distort if the input amp is not expecting it. Would need to be experimented on.

I had hoped to fiddle with this by now, but I've got some orders to complete so looks like my weekends taken up again!

Mark.

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Fri Oct 24, 2008 11:51 pm

Thanks for that Greg :D. I've got a couple of those drives somewhere (well, FD55... something or other, ex PC) so I'll have a play on those too.

Coincidentally, I've had a breakthrough tonight which probably mixes and matches something from what both you and Mark have suggested. I decided to focus on a Mitsubishi MF504C simply because it looks like new and it's got lots of the quick links to mess with :wink:. I'll include a photo in another post.

I basically just tried lots of permutations of 8MHz, 16MHz, 300rpm, 360rpm, DD discs, HD discs and of course Pin 2 Hi/Lo. In the end I didn't actually do all that much, in fact I only moved one link from SB to SS which, as ever, is confusing because one (the only) reference to this drive says SB is 300/360 and SS is 300 only. However, I don't think that is correct because during playing, and whilst contemplating building a strobe, I realised that I could audibly detect the difference between the two speeds and two speeds I'm getting!

Cutting to the chase, the end result is that with Pin 2 held low and the 1772 running at 8MHz (250KHz), the drive runs at 300rpm, is in DD mode and can happily read/write legacy DD discs. Did loads of tests as ever and never incurred a single error.

With Pin 2 pulled high (actually it's default state in the drive) and the 1772 running at 16MHz (500KHz), the drive runs at 360rpm, is in HD mode and can happily read/write HD discs I first formatted under DFS. Same bunch of tests, no errors.

Worth noting that if I jumpered such that the speed was 300 in HD or 360 in DD then I got drive or disc errors - not occasional errors, just 'not interested' terminal errors.

I will obviously bolster these results with a lot more testing but it certainly now looks as if the whole HD drives thing is possible with a simple 1772 frequency selector (probably 1 chip) and a manual HD/DD media switch on the Beeb. Just need to confirm that popular 3.5" drives do (or can via links) output 'Media Type' on Pin 2 and that it's tri-state according to Drive Select.

Then, in service, the 3.5" would automatically switch according to the inserted disc and the user would select the 5.25" media type on a switch which I don't think is any hassle given how the discs would likely be used.

Martin

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Tue Oct 28, 2008 1:10 pm

These flippin 3.5” drives are doing my head in - none of the ones I’ve tried so far is actually configured by default to output Density on Pin 2 :(. They’ve all got various links inside which doubtless can be set to do what’s needed but finding details is proving to be a nightmare :evil:. I think I’ll just buy one of the ones listed in the ‘Jumpers’ attachment I posted so that I can at least finish the project and publish some details. And I thought it was going to be the 5.25” that caused problems...... wrong!

On the plus side, I have found a datasheet for the 177x chips :D . As expected, it only specifies an 8MHz clock but I’ve still had no problems in hours and hours of use at 8/16 MHz.

Martin

Edit : And then I found an even better document which covers the 1770,1,2 and 3 in more detail and it's better quality too =D> I've replaced the attachment so whoever downloaded the first one - get this instead!
Attachments
wd1984storageProducts_01.zip
(2.67 MiB) Downloaded 148 times

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Wed Nov 05, 2008 12:29 am

Hello again - I'm back after some lengthy investigations and I've made some executive decisions on the way forward :D

The main delaying issue has proven to be the 3.5" drives - not their ability to handle both DD and HD media, they all do that perfectly, but because of a complete lack of consistency across the makes and models with respect to how, or if, they signal their density state back to the host computer - the Beeb for us. You may recall that Pin 2 on the 34 way connector is allegedly earmarked in the interface for either a 'density select' input or a 'selected density' output. Some drives claim to allow these functions but actually don't, some say they can but then don't provide user jumpers to configure the option, some just don't mention it, some do it perfectly, some..... etc. etc. :roll:

The problem is then, whilst I could undoubtedly find a suitable drive for myself and tinker with it until it's as required, this doesn't really allow me to offer a universal solution to others which would readily apply to any drive. The design I'm therefore proposing should allow those of you following this to embrace the core ideas and achieve the same pleasing end result as me. I do acknowledge that this won't be within reach of the complete technical novice but I've tried to keep it as straightforward as possible to include as many enthusiasts as I can. I'll have to do this in several stages as ever, and probably write some separate instructions as with my DFDC but for now, stage one, 'The Drives', isn't too complicated. The final full working system will only result in one cut track on the 1770 module, an extra IC on a bit of board and the addition of a few tacked on wires in the Beeb - no permanent modifications to the Beeb itself will be necessary. This will give a fully automatic system for 3.5" drives and a semi-automatic system for 5.25" drives which will have a DD/HD media switch.

Since there are few drives that signal back density on Pin 2 and even fewer that do it as a tri-state signal, I've decided to allocate Pin 2 exclusively to Drive 0/2 and bring Pin 4 into play for Drive 1/3. Pin 4 is unused on the (standard) Beeb interface but, as with Pin 2, it does emerge within the Beeb on a handy spare link. The configuration I've used for myself is a 3.5" as Drive 0/2 and a 5.25" as Drive 1/3 - any mix'n'match of drives is possible but you'll have to read-across the basic idea to implement whatever configuration you fancy. One feature seemingly common to all 3.5" drives is the HD/DD detect sensor which is a simple push-to-make switch that is closed when a DD disc, not having the 'hole', pushes the plunger. All the drives (about 12) that I have tested have 5v on one contact and 0v on the other such that the 5v contact is pulled down to 0v when a DD disc is inserted. So, what we will do is isolate Pin 2 of the drive from any existing connection and connect it to this 5v HD / 0v DD contact of the switch. This done, we will now have within the Beeb on the West pad of S4, a logic '1' when a HD disc is in Drive 0/2 and a logic '0' when a DD disc is present. We will then use this signal to switch the 1770 clock between 8 or 16 MHz as appropriate to the disc density in use. Easy as that :^o

Isolating Pin 2 of a drive simply means doing a visual inspection to see if it's connected to anything (mostly they aren't or just to blind pads) and if there is a track running away from the pin that you can't follow then cut it - trust me, it's not needed. Then, we simply solder a length of wire from Pin 2 to the 'active' side of the density switch. The active side can be found with a meter, no need for a scope - with just power applied to the drive, identify the switch contact that is at 5v with a HD disc inserted (or no disc) and is at 0v with a DD disc inserted.

Note that if you are considering using two 3.5" drives then one will have to be configured as described above and the second drive will have to have both Pin 2 and Pin 4 isolated (cutting as necessary) and the switch wire will be connected to Pin 4.

The photo below shows a typical drive with the new wire in place. On some fully enclosed drives it is necessary to unscrew and/or unclip one side cover to access the switch and interface connector pins.
35P2wire.JPG
(77.67 KiB) Downloaded 2333 times

On the next instalment I'll show the change to the 5.25" drive. Hoorah!

Martin

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Wed Nov 05, 2008 11:06 pm

An interim thought - if you are planning on adding this HD capability then you need to first check that your particular 1770 or 1772 can run at the higher 16MHz clock speed. I haven't yet found a 1772 that doesn't (various makes) but the one 1770 I tried didn't like sprinting [-X

[ As an aside here, I recently picked up a scrap Atari Mega ST and that has a WD1772 02-02 fitted which is the dog's doo-dahs of 1772's, potentially 32MHz, and could be scavenged into a Beeb module.]

For the check, you don't need to do anything to your drives (assuming you have at least one 3.5") but you will need to do the 1770 module track cut I used for the DFDC mod to be able to test the higher speed. This isn't an irreversible cut and it can easily be linked back together if required. I use the genuine Acorn modules and the included photo is of that item but the clock change can be done on any 177x module. (Ignore the piggy-backed 8271, you don't have to do my full DFDC mod :wink: ) The clock is fed to the 177x on Pin 18 so to allow us to connect a faster clock on a different module you would need to do a similar track cut somewhere appropriate such that Pin 18 of the 177x chip (not module pin 18) is isolated from the original 8MHz clock input. We then attach a wire to the 177x side of the cut (i.e. connected to Pin 18) and it is this wire that we then move around for the test. It’s best to make the wire long enough to reach IC6, the video chip, because here is where we easily can pick up the 8MHz and 16MHz clocks on pins 7 & 8 respectively.
clockcut_s.JPG
(52.85 KiB) Downloaded 2359 times

So, using an unmodified PC 3.5” drive, we start by testing our cut & wire tack at the standard 8MHz clock speed. Connect the loose wire to pin 7 of IC6 by fitting a hook probe or just ‘wedging’ it in the pin socket ensuring there are no adjacent pin shorts. Switch on the Beeb noting the Acorn 1770 DFS message (good first indication) and test the disc system with a genuine DD (720k) disc. Once happy, switch off and move the wire to pin 8 of IC6 and switch back on. Again, note the 1770 message and then put a genuine HD floppy in the drive. It can be a pre-used PC disc, no need for bulk-erasing anymore, and type *FORM80 02 (or 13 if it’s your second drive). You should see that the disc happily formats with no errors. Try some verifies, some random saves (e.g. *S.TEST 1900 C000), some *TYPE’s etc. etc. If it all looks ok then your 177x is good to go at 16MHz and you can carry on with this exciting project :D

For the time being, re-make the cut track or make the wire connection to IC6 Pin 7 (standard 8MHz) secure for use until we get round to adding the auto-switch.

Martin

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Sat Nov 08, 2008 12:47 am

The next bit of tinkering is on the 5.25” drive. The problem with these is that they don’t have a media density sensor so we can’t fully automate these drives unless we were to undertake some serious re-moding of the DFS – maybe I’ll look at that later if HD floppies catch on :wink:.
So, I’ve decided that the best solution for 525’s is to add a density select switch to the drive itself where the switch is set to DD or HD according to the media in use. Contrary to how that might sound, it’s actually no hassle in use and since the whole HD objective is to allow us to gradually discard our ageing DD discs then apart from using the occasional original non-copyable floppy, I expect the switch to fairly quickly start collecting dust in the HD position.

If you read some of the earlier posts by myself and others on 525 drives, you’ll know that there are as ever lots of different makes and models and hence lots of different jumper options. Consequently, unlike the 35’s, there isn’t going to be a one-size-fits-all solution that I can offer but understanding the final objective should make things quite straightforward.

What we are aiming for is a 525 configured such that we have a switch on the drive which sets it to DD or HD mode (remember - head current, interface speed, disc rpm, blah, blah…) and in addition, it signals it’s mode back to the Beeb on interface Pin 4 where hi (+5v) = HD and lo (0v) = DD. Simple enough then :?

Now, having played with several different PC drives, they all do seem to allow their density mode to be selected via Pin 2. They aren’t necessarily set-up that way by default but there is more than likely a link or two which allows this mode of operation. However, due to a lack of link information, I figured out how to get my randomly chosen drive (a Mitsubishi 504C-368U) working by connecting it to the Beeb as the only drive (to keep Pin 2 & wire 2 clear for the experiments) and then repeatedly trying HD discs with the 177x clock at 16MHz and DD discs with the clock at 8MHz. Before trying the latter, I each time pulled Pin 2 of the 525 drive low by shorting the West side of S4 to ground (logic 0) and I moved a drive link which I thought might be ‘the one’. (See photo for exact location of S4 which is slightly under the keyboard near the disc interface connector.) You will of course have to start in HD 16MHz mode and format a floppy (actually the easy part) so that you have a HD reference disc to use and also have a ‘normal’ Beeb 525 to hand for the DD check. When I had both modes working, I could audibly hear that in DD mode, the drive was running slower (actually 300 rpm) than in HD mode (360 rpm).

That might all sound a bit fraught but it only took me about 20 minutes to discover that I only needed to move one link from the PC default to get what was needed. I moved a link from SB to SS and then I could control the density from S4 where grounding the West pad gave DD and letting it float gave HD (pulled high by the drive). I’m now guessing SS means ‘selectable speed’ or such so if you see something similar, start there!

Having got the drive doing what I wanted, the next job was to transfer the density control/status to Pin 4 (since I have used Pin 2 for my 35 drive 0) and that just involved a bit of track cutting and linking which is probably best explained by the schematic and photo below :
525schematic.JPG
(20.84 KiB) Downloaded 2264 times

525mods_s.JPG
(27.66 KiB) Downloaded 2273 times
Blue = cut to isolate Pin 2, Red = cut to isolate Pin 4, Yellow = link to connect previous Pin 2 line to Pad 4 and Green = switch wires

The final job was to fit a switch to pull the drive density select line low for DD and let it float high (drive default) for HD. This drive happened to have a further link, SD, originally connected to Pin 2 and paired with ground which would have allowed the drive to be forced to DD only so I simply wired a switch across these link pins and fitted the switch at the back of the drive housing.

I fitted both drives back in the dual case, buttoned it all up and the end result is that I now have an auto density selecting 3.5” whose density is available on the West pad of S4 and a manual density select 5.25” whose density is available on the East pad of S10. Thus, all the necessary signals are now available within the Beeb for the final bit of gadgetry which will automatically set the 177x clock speed to 8 (DD) or 16 (HD) MHz. =D>
links_s.JPG
(34.17 KiB) Downloaded 2274 times
Note that you don't have to have a 525 HD capability if it all seems a bit too much work, I will permutate the options for various combinations of drives.

Final instalment to follow – any questions, comments, thoughts, criticisms, cash and cheques are all welcome :D

Martin

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Mon Nov 10, 2008 11:46 pm

=D> =D> =D> Yay! =D> =D> =D>

Just a quickie for now because I couldn’t contain my excitement :D

Here’s a pic of the one-chip add-on to auto-control the DD/HD switching and it works a treat! Needs eight wires tacking on here and there but there’s still only the one track cut on the 177x board, no messing with the Beeb as such. The only extra tweak I had to do was make keyboard Link 3 to reduce the drive step rate to 12ms which is then immediately halved again in HD mode – I found that some drives were a bit ‘on the edge’ at 3ms but all were happy at 6ms.

I can have any combination of DD and HD in the drives and the auto clock switching just handles it seamlessly. I can copy and backup between formats and I've not seen a single error even with crappy old pre-used PC discs. The great thing though, I’ve already copied some of my valuable discs onto brand-new & shiny HD discs – safe for thirty years :wink:

Works fine with my 8271 DFDC too!

I’ll write up the mod and post the details in the next day or so – if I can tear myself away from playing with the thing! Floppy heaven :D

clock switch.JPG
(80.32 KiB) Downloaded 2191 times
Martin

User avatar
sorvad
Posts: 2172
Joined: Wed Aug 24, 2005 12:13 pm
Location: Back of beyond

Re: high density disks

Postby sorvad » Tue Nov 11, 2008 11:15 am

=D> =D> Well Done =D> =D>

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Tue Nov 11, 2008 4:18 pm

Thanks Steve. I'll get it all tidied up and finish the details and then I'll see if a similar trick is possible on the Elk. It has a 16MHz clock but of course there are a few different disc interfaces and I only have an ACP APx so if it can be done it'll just be for that interface. Might not be possible due to the slower CPU speed but we shall see :wink:

Nick (corelki) actually started this HD floppy thread with a question about his Elk so I owe it to him to at least have a try!

Cheers, Martin

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Wed Nov 12, 2008 11:52 am

So, finally on to the details of the clock switch and if you’re planning on adding this HD capability then you’re probably comfortable with electronics so I won’t bore you with too much detail :^o

I should say that I came across the idea of a clock switch whilst trawling the Atari forums as I mentioned previously so I want to credit (at least) Martin Graiter for inspiring me to produce this BBC implementation =D>

The clock switch needs to identify when a floppy drive is being accessed and also determine whether that drive contains a DD or HD floppy. In response to these triggers the circuit must then clock the 177x at 8MHz for DD discs or at 16MHz for HD. We are using the automatic DD/HD media density sensor in 35 drives and have added a switch for 525’s and these two signals are, due to our drive mods, available within the Beeb together with all the other signals we need.

The chip which will perform the clock switching for us is a 74x151 which is a 1 of 8 Data Selector or Multiplexer. It essentially takes in 8 input lines and, according to the state of 3 control lines, will pass one of the 8 input signals to a single true or complement output line. (We will only use the true output since we have no need to invert the clock.) For our 3 control lines we will use our two ‘HD present’ lines from Drive 0/2 (interface Pin 2) and Drive 1/3 (interface Pin 4) together with the Beeb’s own Drive Select 0. We don’t need to use Drive Select 1 since we can safely assume it is always the opposite of Drive Select 0.
Now, we don’t have 8 inputs to choose from, only two – the 8MHz clock and the 16MHz clock - so what we will do is connect 4 of the input lines to 8MHz and 4 to 16MHz and then devise a truth table for the three control lines such that for all eight states of ‘Drive Select n’ and ‘HD present n’ we will get the appropriate clock output to the 177x. Our Drive Select lines are active low and our HD present lines are active high. Thus, the truth table for our 3 control lines is as follows :

Code: Select all

DS0(C) HD0(B) HD1(A) Data & Clock Selected
  0      0      0      0      8
  0      0      1      1      8
  0      1      0      2     16
  0      1      1      3     16
  1      0      0      4      8
  1      0      1      5     16
  1      1      0      6      8
  1      1      1      7     16
The significant pins on the 151 are 9-11 (control lines C, B, A), 1-4 (input D3-D0) & 12-15 (D7-D4) with the true output on Pin 5. Given this, our chip circuit becomes :
clk_sw_schematic.JPG
(18.17 KiB) Downloaded 2061 times
For the chip family, I settled on a 74HC151 which is hi-speed CMOS and hence low power but more particularly because it presents virtually no load to the 8 and 16 MHz clocks it is connected to. The clocks in the Beeb are not the best in the world and we don’t want to add any unnecessary additional load, even that of an extra LS chip input. In some applications the HC range doesn’t always sit well with existing LS logic but in this case it is just fine. The two pull-down resistors are there to drop us into 8MHz (DD) by default if we aren’t using signalling drives and the slightly high 10k is because we won’t know for sure what pull-ups we are facing when we are using signalling drives. It might be wise to add a 0.1uf decoupler close to the chip (across +ve and gnd) but if the power leads to the chip are fairly short then it’s not essential. I picked up +5v and 0v from a couple of empty IC pads in the Econet area of the motherboard but anywhere handy will do. The longest flying lead is the 16MHz clock feed and as such it’s better not to try and wind it in a convoluted path round the board, just let it lie slack across the Beeb and all will be well.

Here are some pics of the various connections to the new chip. Note that this is the quick and easy prototype version (which is fine) but when I do another Beeb I’ll probably run the connections under motherboard for neatness.
clock_sw_1.JPG
The DS0, HDn and 8MHz connections
(97.65 KiB) Downloaded 2061 times
clock_sw_2.JPG
The +5v and 0v connections
(82.72 KiB) Downloaded 2060 times
clock_sw_3.JPG
The 16Mhz connection to S27 East
(100.93 KiB) Downloaded 2062 times
clock_sw_4.JPG
Clock output connected to the 177x module
(86.63 KiB) Downloaded 2061 times
Don’t forget to set the drive stepping speed to 12ms by making link 3 on the keyboard configuration links so that the minimum speed we get at 16MHz is 6ms. This does mean that you will then get 12ms at 8MHz but I doubt you’ll care given the benefits of having a HD capability :D

And in summary....

Although this has been a long and tortuous thread :roll:, I don’t think that the end product is actually too involved. As a minimum, we do need to build and install the switching circuit but, in the case of the drives, the amount of work needed will depend on the nature of the drives themselves and on the degree of flexibility required by the end user. For example, if you have a Drive 0 that is a 35 and already reports density on it’s Pin 2 then you don’t need to do anything at all apart from adding the 151 chip. If you have a second drive and want that to remain as DD only then simply tie the HD1 input to the 151 to ground rather than connecting it to S10. If you have twin 35’s and want them both to be DD/HD capable then simply connect the density sense wire (covered previously) to Pin 2 on Drive 0 and to Pin 4 on Drive 1 making sure you isolate these pins first. Basically, if you've followed the basic design concept, the permutations are limitless and down to your own ingenuity. Of course, the whole project does hinge on your 177x being able to run at 16MHz but all of the 1772’s I have tested thus far seem perfectly happy.

I guess that's about it then - Enjoy :D

Martin

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Fri Nov 14, 2008 12:31 am

I've had an initial look at the Electron and my initial thoughts and findings are these :

1. The FDC in my AP4 is a 1770 so I would expect to have to swap that for a 1772. The good news though is that it's in a socket so will be an easy switch. They can still be bought (admittedly from sporadic sources) but I will probably free up a Beeb one by salvaging the Atari 02-02 I mentioned.

2. The 16MHz clock does pass throuh the expansion connector and, better still, appears on the Plus 1 cartridge port edge connectors. In fact, I presume the AP4 divides it by two to get the standard 8MHz so the clock access side of things will be a breeze.

3. The Drive Select 0 will be on the 34-way interface exactly as per the Beeb and our newly allocated density tell-tales on Pins 2 & 4 must be there & spare, at least on the AP4 IDC connector.

So, hardware-wise, much better than I was expecting :D

The remaining two potential issues then are One - whether the slower Elk can hoover up the HD data now flying off at 500KHz (might require a tweak to an NMI routine or similar in the DFS :?: ) and Two - how we can set the disc step rate to a slower value (as I did on the Beeb via the keyboard links) which again might need a change in the ACP DFS.

So, although the only (possible) issues appear to be software in the ACP, that's still a bit tricky given that without listings it will require some serious disassembly to find the relevant bits of code :(

However, I am always optimistic and maybe it will just work..... [-o<

Martin

[ Although I haven't got one, there should be easy read-across to the Plus 3 since it can't be much different to an AP4 in relation to the 177x, the clocks and the 34-way interface. ]

User avatar
Elk Towers
Posts: 489
Joined: Sun Apr 23, 2006 2:10 am
Location: Kettering, Northants
Contact:

Re: high density disks

Postby Elk Towers » Fri Nov 14, 2008 7:18 pm

Thank you Martin for thinking of me, have read the articles with great interest. =D>

If I had known you were going to try this on the elk I would have sent you my plus 3 to have a play with instead of selling it( on here). I do however have two AP34s, you can turn your ap4 into one by adding an acp ap3 adfs rom (version 1.15) :D if you so wish.

If you cant do eproms and want one I can blow an eprom and post it to you, just pm your address and it will be done.


Nick.

ps I am having fun trying to sort out a scsi hdd i/f etc , just need to get a couple of pcbs etched and drilled.
Nick

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Fri Nov 14, 2008 8:22 pm

Hey Nick, no worries - that’s the great thing about a forum like this, you get involved in stuff because someone has a thought and shares it. If you hadn’t started the thread I probably would never have got into it and in truth it’s been a blast =D>

Thanks for the offer but I’m fine for eproms – I’ve actually got more programmers than I can shake a stick at :roll:. The main one I use a is large Stag PP39 hooked up to an old DOS only laptop.

I don’t bother with adfs :shock:. I don’t have a problem with it but I like the simplicity of dfs and it just does what I need quickly and reliably. My AP4 came as an AP34 but I promptly removed the original ACP adfs eprom and fitted ADT.

I knew you were on with a hard drive and I joked somewhere about getting the Watford 32k Shadow working on an Elk before you got that HDD going but due to a lack of technical info I think that’s on the back burner for now :(

However, I will now see if I can get an Elk to run HD floppies before you can get one to run scsi..... :lol:

Martin

User avatar
regregex
Posts: 558
Joined: Sun Jan 02, 2005 9:51 pm
Location: London, UK
Contact:

Re: high density disks

Postby regregex » Sat Nov 15, 2008 1:25 pm

Props to ya Martin, you've done it again :)

MartinB wrote:The remaining two potential issues then are One - whether the slower Elk can hoover up the HD data now flying off at 500KHz (might require a tweak to an NMI routine or similar in the DFS :?: )
If you've not modified the DFS then it will still be recording in FM -- ie 250 kbit/s data rate, so the same NMI routines designed for double density will work. Accessing PC floppies OTOH, can do it on the BBC with an extremely minimal NMI routine, on the Elk it's touch and go. In theory it can keep up -- just --, as long as the destination addresses are all at 2 MHz. By minimal I mean not even saving registers or counting the bytes :o

--Greg

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Sun Nov 16, 2008 12:16 am

Many thanks for the support Greg, much appreciated :wink:

When you say PC discs, I presume you mean Windows/DOS format? If so, I haven’t gone that far yet since doing the HD upgrade. I am familiar with that side of things though since I wrote my own 720k PC disc reader for RAMagic’s DOS-to-DFS transfer feature and this has a fast NMI routine which I will (in time) try a re-write of. If I just push up the sector buffer size to 512 from 256 then timing shouldn’t be an issue. No, my worry bead on the Elk was whether it could pull the data from the 500KHz stream before the 1772 times out given that it (the Elk) has a variable speed system clock depending on Ram access. The ACP has its own 8k static ram chip which is how PAGE is kept at $E00 but there’s no DMA or anything fancy like that so it’s still down to the (slower) 6502 to store the data.

Anyway, after a bit of time in the workshop, the speculation is over..... :)

I extracted the 1772PH-02-02 from the Atari ( %#$*!@ soldered in) and fitted it to the ACP in place of the original 1770. I confirmed that the ACP uses the 16MHz clock divided by two and then just bypassed the divider. For the initial test this is all I did, effectively hardwiring the ACP to HD, because if this doesn’t work there’s no point in fitting the clock switch.

[ I had a momentary troublesome thought about doing this because the Acorn Elk expansion info suggests that clocks generated from this 16MHz must be properly in synch with it’s own 2MHz clock but I decided this wasn’t relevant here since the FDC 250/500 KHz data stream is effectively asynchronous already. ]

Mod done, I refitted the ACP into the Plus1 and I’ll swear that the Elk, now sporting the Rolls Royce of 1772’s, smiled at me as I switched it on :lol:. In went a couple of HD DFS discs pre-prepared on it’s newly HD capable big brother and, lo and behold, it works! I used the 35/525 pack I have been using on the Beeb and basically everything seems fine. I did the usual bunch of tests and even the head step seems good so I’m presuming ACP erred on the safe side in the first place and set a slow rate. I’ll obviously do more extensive testing with the 151 clock switch to try mixing densities on the fly but I’m pretty confident that all will be well. Looks like a result then =D>

Here’s a couple of pics showing the flashy 1772 but note that any 1772 happy at 16MHz will be fine, just wanted to show that I haven’t dreamt all this :D

Martin
elk1772ph0202a.JPG
An ACP temporarily without it's case
(64.85 KiB) Downloaded 1819 times
elk1772ph0202b.JPG
The brown wire is the temporary clock jumper feeding the Bad-Boy 1772
(56.98 KiB) Downloaded 1817 times

User avatar
Elk Towers
Posts: 489
Joined: Sun Apr 23, 2006 2:10 am
Location: Kettering, Northants
Contact:

Re: high density disks

Postby Elk Towers » Sun Nov 16, 2008 11:32 am

I take it then that the 1772 and 1770 are pin compatible?
Nick

User avatar
MartinB
Posts: 4555
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity

Re: high density disks

Postby MartinB » Sun Nov 16, 2008 11:39 am

Yeah, absolutely identical. The 1772 revision was principally driven by the need to provide faster track-to-track head step rates and at the same time they improved the performance of the whole chip taking advantage of the rapidly evolving chip manufacture techniques.

If you download the WD177x attachment I posted further up you can see for yourself :)

Martin


Return to “hardware”

Who is online

Users browsing this forum: BigEd, danielj, jregel and 8 guests