Strange Pegasus DFS fault

for bbc micro/electron hardware, peripherals & programming issues (NOT emulators!)
User avatar
jms2
Posts: 2089
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Strange Pegasus DFS fault

Post by jms2 » Thu Dec 06, 2018 10:36 pm

I've recently been using my Pegasus interface in ADFS mode, because I was checking out the AP6 format problem which was being discussed in a separate thread. Although the Pegasus card has a rom slot for ADFS, mine is empty because the ADFS rom lives on an AP5 (where it needs to be to support BeebSCSI). Anyway, when using ADFS the card was working perfectly (as usual).

So today I needed to revert to DFS mode so I can use UPURS. And without any warning, DFS doesn't work. This is what happens:

- The DFS ROM responds to *HELP etc without problems.
- In DFS mode (irrespective of whether there is an ADFS rom in the system or not), I get one of two things (at random) in response to attempting to catalogue a disc. Either the disc grinds a bit and returns nothing at all, or I do get a catalogue but its garbage.

At any time I can plop the AP5 back in, switch to ADFS mode and read and write discs without any problems.

For a while I was flummoxed as to how ADFS could still be working while DFS was completely crippled. Then I realised that there is one crucial difference: DFS uses the onboard RAM. Inspecting the RAM chip revealed a remarkable amount of whitish corrosion on the chip legs. So I very carefully scraped it off, but to my surprise it hasn't made any difference. The DFS rom was affected similarly (though not quite as badly), but the 1770 and glue logic seem fine. I have reseated all the chips.

The only thing I can think of is that the RAM chip must have died for some reason (although I can't think why that might have happened). Does anyone have any thoughts?

User avatar
1024MAK
Posts: 8107
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Strange Pegasus DFS fault

Post by 1024MAK » Thu Dec 06, 2018 10:43 pm

What is the state of the socket contacts?

Mark

Prime
Posts: 2601
Joined: Sun May 31, 2009 11:52 pm
Contact:

Re: Strange Pegasus DFS fault

Post by Prime » Thu Dec 06, 2018 10:45 pm

How are the jumpers set on the Pegasus?

I currently have DFS in leftmost ROM socket, ADFS in rightmost and jumpers are set as follows :

Left->Right : down-up-down-down.

The jumpers control which bits $8000-$C000 area are ROM and which are RAM.

Cheers.

Phill.

User avatar
jms2
Posts: 2089
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Pegasus DFS fault

Post by jms2 » Fri Dec 07, 2018 7:12 am

1024MAK wrote:
Thu Dec 06, 2018 10:43 pm
What is the state of the socket contacts?
They are turned pin sockets, and gold plated (well, gold colour at least!). They look fine as far as I can see.
Prime wrote:
Thu Dec 06, 2018 10:45 pm
How are the jumpers set on the Pegasus?
I need to check, but the thing is - I have certainly not changed them. Could be corroded though I suppose.

User avatar
jms2
Posts: 2089
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Pegasus DFS fault

Post by jms2 » Fri Dec 07, 2018 12:04 pm

I checked the links, and found that the third one was noticeably loose (no signs of any corrosion though). I replaced it with a new one, which fits a lot better, but it has not helped. Having checked the manual I now understand why:

The links are arranged as two pairs, the leftmost pair for the left hand ROM and the right hand pair for the right hand ROM. For each pair the meaning of the links is:

up,down = All rom, no overlay
down,up = 14k rom, 2k ram
up, up = 12k rom, 4k ram.

As my loose link was in the right hand group, fixing it clearly makes no difference as I don't have a right hand rom installed. I need to double check the links in the left group.

What I do notice is that your arrangement of the right hand group (down,down) is meaningless according to the manual. Do you have a right hand rom installed?

User avatar
CMcDougall
Posts: 6356
Joined: Wed Feb 02, 2005 3:13 pm
Location: Shadow in a Valley of Scotland
Contact:

Re: Strange Pegasus DFS fault

Post by CMcDougall » Fri Dec 07, 2018 5:11 pm

Never had any problems with mine, but it never sees ADFS as it's utter sh!te :lol:
Attachments
29572546_10155694284722730_5564328416104188096_n.jpeg
ImageImageImage

User avatar
daveejhitchins
Posts: 4532
Joined: Wed Jun 13, 2012 5:23 pm
Location: Newton Aycliffe, County Durham
Contact:

Re: Strange Pegasus DFS fault

Post by daveejhitchins » Fri Dec 07, 2018 5:15 pm

If anyone needs a printed copy of the Slogger DFS User Guide for the Pegasus Interface (?) I now have them. Reformatted from the download that kicking around.

I also have the SEDFS ROM and can supply copies with a genuine Slogger label or just the lebel, if required.

Dave H :D
Parts: UM6502CE, GAL22V10D, GAL16V8D, AS6C62256A, TC514400AZ, WD1772, R6522, TMS27C512, AT28C256
Products: ARA II, ARA III, ABR, ATI, AP6, MGC, AP5 . . .
For a price list, contact me at: Retro Hardware AT dave ej hitchins DOT plus DOT com

User avatar
MartinB
Posts: 5025
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity
Contact:

Re: Strange Pegasus DFS fault

Post by MartinB » Fri Dec 07, 2018 5:27 pm

Col wrote:Never had any problems with mine, but it never sees ADFS as it's utter sh!te :lol
I recognise that Peg, right down to it’s little blue foot! :wink:

You and I appear to agree Col - ADFS = Absolutely Dreadful Filing System :lol:

User avatar
CMcDougall
Posts: 6356
Joined: Wed Feb 02, 2005 3:13 pm
Location: Shadow in a Valley of Scotland
Contact:

Re: Strange Pegasus DFS fault

Post by CMcDougall » Fri Dec 07, 2018 6:42 pm

MartinB wrote:
Fri Dec 07, 2018 5:27 pm
recognise that Peg, right down to it’s little blue foot
yeah I wonder who made a good job of that without solder jumping across to neighbouring close un protected track's =D> :wink:

don't know why people bother with Absolute Dreadful FS on a electron, should rename the hardware sub forum to Broken beebs & elk ADFS nonsense! :roll:
ImageImageImage

User avatar
jms2
Posts: 2089
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Pegasus DFS fault

Post by jms2 » Fri Dec 07, 2018 11:02 pm

There's something else going on too: this is what I get in response to *verìfy 0 (still in DFS mode):
20181207_230653.jpg
:shock:

User avatar
jms2
Posts: 2089
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Pegasus DFS fault

Post by jms2 » Fri Dec 07, 2018 11:32 pm

... and now the Elk fails to recognise the interface altogether. Something on the board has clearly failed. :(

User avatar
MartinB
Posts: 5025
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity
Contact:

Re: Strange Pegasus DFS fault

Post by MartinB » Fri Dec 07, 2018 11:33 pm

I assume you have UPURS on eprom and on an EUP or an AP5? If so, you can use UPXROM to image the DFS direct to the PC where you can confirm that the rom isn’t corrupt through an image binary compare or checksum. Also, if you put the 6264 ram chip into a writable rom socket (in a cart or whatever), you could use UPLOAD together with UPXROM to load and re-image the ram chip (with say test patterns first of AAs and then of 55s) and hence integrity check the ram device.

I’m suggesting these methods since those aspects of UPURS don’t require any storage media whatsoever, you only need a PC.

User avatar
daveejhitchins
Posts: 4532
Joined: Wed Jun 13, 2012 5:23 pm
Location: Newton Aycliffe, County Durham
Contact:

Re: Strange Pegasus DFS fault

Post by daveejhitchins » Sat Dec 08, 2018 7:53 am

Just to throw a spanner in the works, at this point:

I've been building the rest of the Pegasus PCB stock. Plus I have some later issue PCBs that look as if they were production or development repairs - the thing that stands out with both sets of PCBs, not counting the layout and shorts for the one with 'curly' tracks, is the poor PCB production in the area of vias. Even on what I'd say were the 'good' ones - if you inspect the vias there are a lot with incomplete via plating. As IanS may remember, from the last ABUG, I was trying to get one of the "repair' boards working - after my tenth via link I gave up - saving that one and the others for a rainy day!

So, what I'm trying to say is: get the multimeter out and with Prim's schematic start checking continuity. I suspect any form of corrosion may have just been the Camel's straw!

Dave H :D
Parts: UM6502CE, GAL22V10D, GAL16V8D, AS6C62256A, TC514400AZ, WD1772, R6522, TMS27C512, AT28C256
Products: ARA II, ARA III, ABR, ATI, AP6, MGC, AP5 . . .
For a price list, contact me at: Retro Hardware AT dave ej hitchins DOT plus DOT com

User avatar
jms2
Posts: 2089
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Pegasus DFS fault

Post by jms2 » Sat Dec 08, 2018 9:02 am

MartinB wrote:
Fri Dec 07, 2018 11:33 pm
I assume you have UPURS on eprom and on an EUP or an AP5?
Actually I'm not sure I do - but its not a problem because I do have it on the BeebSCSI and can load it from there.

I think testing the DFS rom is not going to reveal very much, but the ram may well be defective as you say.

What I can't figure out is how any kind of disk system fault would lead to spurious data getting plastered across the screen though.

User avatar
MartinB
Posts: 5025
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity
Contact:

Re: Strange Pegasus DFS fault

Post by MartinB » Sat Dec 08, 2018 9:35 am

John wrote:What I can't figure out is how any kind of disk system fault would lead to spurious data getting plastered across the screen though.
Any rom or software given authority to write to memory can get it hopelessly wrong if things go awry, even to the point of targeting the screen ram - I only suggested checking the DFS rom and 6264 early doors because you're probably going to end up ticking boxes for elimination so the easy things might as well be first.

I do agree with Dave though - if you remember, I had to repair (perhaps more 'finish') many of the original boards by separating tracks that were simply too close and had shorted during etching so as Dave says, it wouldn't take much 'fur' to cause problems. It might get tricky if that were the case though because a few of the closest tracks are under IC's and sockets but easy stages..... :)

User avatar
1024MAK
Posts: 8107
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Strange Pegasus DFS fault

Post by 1024MAK » Sat Dec 08, 2018 10:39 am

It is worthwhile removing all the socketed chips. Then visually inspecting as much of the board as possible.
Certainly do as Martin suggests with regards to testing the EPROM and SRAM chips. There is nothing special about the SRAM chip, so if you have other devices that also use the same size SRAM chip, it can be tested in those.

Once you get past the above actions, you then have a number of choices. Use a multimeter to test the PCB tracks, both continuity between all points for each track. And testing on resistance to ensure no unexpected short circuits.

Are the glue logic chips socketed? If yes, test by substitution.

If the glue logic is not socketed, or none of the above revealed the fault, and if you have a spare edge-connector, a 5V bench PSU (that has current limiting), and a bread board, a bank of switches or similar, you can build a temporary test rig. The purpose is to supply signals to the glue logic chips to test it’s operation. You can then chase the logic levels around the board with a logic probe or multimeter at your leisure :wink:

I suggest this method, as in system testing is a bit tricky as the edge connector part of the board gets hidden by the cartridge slot. Of course if you have a ‘scope and remove the cover of the Plus 1 or ROM Box, in which case in system testing can be done.

Mark
Last edited by 1024MAK on Sat Dec 08, 2018 10:41 am, edited 1 time in total.

User avatar
MartinB
Posts: 5025
Joined: Mon Mar 31, 2008 9:04 pm
Location: Obscurity
Contact:

Re: Strange Pegasus DFS fault

Post by MartinB » Sat Dec 08, 2018 10:58 am

Did you have one of the original boards from me John? (I can't remember!)

Anyway, if so, below is a reminder of where the worst of the tacking problems were.....

Peg track faults (1).jpg

Peg tracking gap issues.jpg

User avatar
jms2
Posts: 2089
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Pegasus DFS fault

Post by jms2 » Sat Dec 08, 2018 12:15 pm

MartinB wrote:
Sat Dec 08, 2018 10:58 am
Did you have one of the original boards from me John? (I can't remember!)
Anyway, if so, below is a reminder of where the worst of the tacking problems were.....
I think I probably did. I remember checking those areas carefully, but didn't find any problems at the time. As you say, it is possible that a tiny piece of conductive material has bridged one of these very small gaps - although the corrosion that I scraped off was more like a thin layer of paint than big lumpy crystals.

It's interesting, and possibly a clue, that these problems only started after I unplugged the board from the Plus 1 (and took out the DFS rom). Prior to that, it had been intact and in-situ for quite a while. This must have disturbed something I imagine.

User avatar
daveejhitchins
Posts: 4532
Joined: Wed Jun 13, 2012 5:23 pm
Location: Newton Aycliffe, County Durham
Contact:

Re: Strange Pegasus DFS fault

Post by daveejhitchins » Sat Dec 08, 2018 1:15 pm

jms2 wrote:
Sat Dec 08, 2018 12:15 pm
It's interesting, and possibly a clue, that these problems only started after I unplugged the board from the Plus 1 (and took out the DFS rom). Prior to that, it had been intact and in-situ for quite a while. This must have disturbed something I imagine.
That's a similar problem to the one I've just had on the Cumana interface here. The stress of plugging in and out was just too much for some of the partly corroded tracks.

Dave H :D
Parts: UM6502CE, GAL22V10D, GAL16V8D, AS6C62256A, TC514400AZ, WD1772, R6522, TMS27C512, AT28C256
Products: ARA II, ARA III, ABR, ATI, AP6, MGC, AP5 . . .
For a price list, contact me at: Retro Hardware AT dave ej hitchins DOT plus DOT com

User avatar
jms2
Posts: 2089
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Pegasus DFS fault

Post by jms2 » Sun Dec 09, 2018 10:42 pm

I found what seemed to be a dry joint on the /OE pin of the DFS rom, and fixed it. But this didn't solve the problem, so maybe I just wasn't firm enough pressing the multimeter probe onto the joint.

But I then realised I could move the DFS rom to the spare socket. It still didn't show up when put there. It also doesn't respond to *ROMS when inserted in a rom socket cartridge. So... it looks like the rom is dead.

I could try UV erasing it and re-writingk it, but is that likely to reset whatever is wrong? Or is it more likely that the whole chip has died?

EDIT: it's readable in the Eprom programmer, but is clearly corrupt. The only bit of the header which is untouched is "logge" out of "Slogger".
Last edited by jms2 on Sun Dec 09, 2018 11:54 pm, edited 1 time in total.

User avatar
1024MAK
Posts: 8107
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Strange Pegasus DFS fault

Post by 1024MAK » Mon Dec 10, 2018 12:03 pm

Is it dead, probably not.

Internally EPROMs use the stored charge memory cell system, when a voltage (charge) level on a small ‘capacitor’ is used to record the state. When a read is performed, voltage comparators decide if the analogue voltage level is above or below a threshold, and hence is it a logic one or a logic zero.

Depending on how well and how long ago the EPROM was programmed, it’s very likely that the voltage levels on the cells has degraded. When NMOS EPROMs were made, no one expected any to be in use after ten years...

So it is worthwhile reprogramming it if the corrupt bits are where logic zeros have become logic ones. If this is the case, you don’t need to erase if first if you are using the exact same data. Otherwise, yes, erase it first. A ‘blank erased EPROM is actually filled with 0xFF, when programmed, the relevant bits are changed from logic one to logic zero.

The normal failure mode for EPROMs is when they have been through too many erase/programming cycles and will no longer erase completely or fail to program properly. They can also be destroyed if inserted the wrong way round and then power is applied...

Mark

Prime
Posts: 2601
Joined: Sun May 31, 2009 11:52 pm
Contact:

Re: Strange Pegasus DFS fault

Post by Prime » Mon Dec 10, 2018 2:21 pm

1024MAK wrote:
Mon Dec 10, 2018 12:03 pm
They can also be destroyed if inserted the wrong way round and then power is applied...
Especially if you attempt to program them like that, and even more especially the higher the programming voltage.......

Cheers.

Phill.

User avatar
jms2
Posts: 2089
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Pegasus DFS fault

Post by jms2 » Mon Dec 10, 2018 10:18 pm

1024MAK wrote:
Mon Dec 10, 2018 12:03 pm
They can also be destroyed if inserted the wrong way round and then power is applied...
:lol: I appreciate your diplomacy skills in not assuming that I actually did just that! I don't think I did... but I suppose it is the most likely explanation.
So it is worthwhile reprogramming it if the corrupt bits are where logic zeros have become logic ones.
I have compared the original rom against the corrupt version. A good guide is what has happened to the word "PEGASUS". It has become "VEGEWUW". This means that b1 and b2 sometimes (but not always) change from 0 to 1.

So I tried reprogramming the eprom, but my (Willem) programmer was not happy - even when I erased it to FF throughout. It was able to write the first byte, correctly, to 00 but failed immediately saying that the byte was '04' rather than the required '00'. This makes no sense to me at all and suggests that the eprom is broken I think...

I only have a few manky old 27128s left, and whilst all of them can be wiped successfully,the programmer doesn't seem to be happy writing any of them. This isn't all that surprising, because these are ones I have struggled to program in the past. I probably should have chucked them in the bin, rather than put them back in the box!

User avatar
CMcDougall
Posts: 6356
Joined: Wed Feb 02, 2005 3:13 pm
Location: Shadow in a Valley of Scotland
Contact:

Re: Strange Pegasus DFS fault

Post by CMcDougall » Mon Dec 10, 2018 11:15 pm

^ round wrong way 100% does do that to them :(
or a 12.5v eprom being programmed at 21v (not so much 20%, they usually work for me :? )

i've done the upside down eprom before #-o , but I was drunk as a skunk doing four things at once! [-X

got about 50+ spare, so I dont care :D
ImageImageImage

User avatar
daveejhitchins
Posts: 4532
Joined: Wed Jun 13, 2012 5:23 pm
Location: Newton Aycliffe, County Durham
Contact:

Re: Strange Pegasus DFS fault

Post by daveejhitchins » Tue Dec 11, 2018 8:29 am

John . . . PM sent.

Dave H :D
Parts: UM6502CE, GAL22V10D, GAL16V8D, AS6C62256A, TC514400AZ, WD1772, R6522, TMS27C512, AT28C256
Products: ARA II, ARA III, ABR, ATI, AP6, MGC, AP5 . . .
For a price list, contact me at: Retro Hardware AT dave ej hitchins DOT plus DOT com

User avatar
1024MAK
Posts: 8107
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Strange Pegasus DFS fault

Post by 1024MAK » Tue Dec 11, 2018 4:48 pm

How are you powering your programmer? A lot of the cheap designs made in China don’t have a very good DC/DC step up converter, they struggle to get and maintain the required 21V programming voltage at the current required by older NMOS EPROM chips.

So if you are using a 5V supply or USB lead, and it has an external power input that can take a 12V DC input (or thereabouts), and you have a suitable PSU, try that.

Mark

User avatar
jms2
Posts: 2089
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Pegasus DFS fault

Post by jms2 » Thu Dec 13, 2018 11:16 pm

Good thought about the programmer Mark, but as Dave H has kindly supplied me with a Pegasus rom its not something I need to look at straight away.

With the new rom in, the interface is now recognised once again. But it still doesn't catalogue DFS disks properly... at least not the one I tried anyway. It is possible that this disc has become corrupted by whatever problem existed previously, so I need to try one or two more.

The fault looks different on screen though. Rather than stepping the disc head, now I get a pause followed by a kind of proper error message that reports the drive, track and sector number followed by something like OSCLIOSCLIOSCLI etc.

User avatar
1024MAK
Posts: 8107
Joined: Mon Apr 18, 2011 4:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Strange Pegasus DFS fault

Post by 1024MAK » Fri Dec 14, 2018 3:59 am

Well, at least that is some progress...

Mark

User avatar
jms2
Posts: 2089
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Pegasus DFS fault

Post by jms2 » Fri Dec 14, 2018 6:55 am

Yes. I have also checked out all the rom and ram control lines, and not found any issues. I'll check the ram next.

EDIT: Just tried putting the ram into a Care Electronics cartridge, to see if it is recognised by the AP6 *ROMS command. It wasn't. I know its only an 8k ram, but it should still be recognised shouldn't it?
Last edited by jms2 on Fri Dec 14, 2018 8:18 am, edited 1 time in total.

User avatar
danielj
Posts: 6693
Joined: Thu Oct 02, 2008 4:51 pm
Location: Manchester
Contact:

Re: Strange Pegasus DFS fault

Post by danielj » Fri Dec 14, 2018 11:14 am

Is the care cartridge wired up properly for R/W on the elk? I have it in my head it might not be...

Post Reply