Drive reads & writes Fine but won't format...

[DavidMil]

I've discovered that one of my 1050's will read and write fine but won't format any disk. Even disks that have just been formatted by another drive.

When you put the disk in with DOS (2.5) on the screen and try to format a disk the drive light comes on for a few seconds, the head moves once and it just

spins for a second, stops, then says "bad media try another disk". No other error message or code. I checked the drive out with the diagnostic disk,

And it passed all the tests multiple times. I was concerned about the track zero, speed, and head settle tests, so I ran them for about 2 minutes each

but it passes all tests without any problems. I'm open to suggestions at this point.

 

DavidMil

[Joey Z]

I'm guessing bad index sensor, or otherwise misaligned. I think the 1050 doesn't use it (nor do most drives for that matter) when reading and writing, but they *do* use it for formatting.

[_The Doctor__]

The 1050 does NOT contain any index hole led or detector. It is faked. The XF does however work the way Joey describes. There is a specific mention of this in either SAMs or tech serv. There is a description of the method also. I am certain it's on Atari age or someone has it. I no longer do, or if I do it is buried, I can't do anything for it at this time... look to the index pulse generator on the 1050 pcb.

Edited May 6, 2017 by _The Doctor__

[cwc]

I had the same issues and got lots of help here back in 2015. It turned out to be U21 (3086).

 

http://atariage.com/forums/topic/239200-another-strange-1050-fault/

[_The Doctor__]

and cr10 so it would seem

[DavidMil]

Thank you all and especially cwc for pointing me to that good thread! I wonder why my search didn't find it? Because I'm lazy, I'll try swapping IC's first and I'll replace

CR10 also. I will get back to you and let you know what happens.

David

 

PS. Something I forgot to mention in my original post was the transport swap with a good drive. I knew the problem was on the board when the problem didn't move with

the transport. Sorry for forgetting to mention that, but the pizza guy arrived at the door and the ball game was getting ready to start; so with food and sports as a distraction

I ended the post quickly.

 

DM

[_The Doctor__]

make sure to use quality part and that it installs with correct orientation keep the old parts just in case... replace and try one part at a time. start with chips end with cr10 and other supporting circuitry. Can't wait to here.. It's working It's working!

Edited May 6, 2017 by _The Doctor__

[DavidMil]

Good call Folks! I got out my Dremel and cut out the RCA LM3086N chip, removed the pins from the board, used solder wick with addition flux added to clean

the holes, and installed a new 14 pin socket for U21. IT WORKS AGAIN! My sincere thanks to all, and I'm going to add this fix into my 1050 folder. My wife

says I'm nuts for keeping several hundred various IC's, transistors, diodes, caps, and resistors in almost 300 plastic draws, but I wouldn't have had the 3086

chip if I didn't keep all that stuff.

 

Thanks Again!

 

David Milsop

[ijor]

Very interesting. And I'm not sure I can see a reasonable explanation as why a defect on that chip affects format but not writes.

 

It might be surprising for some people, but writing and formatting is the same operation at that level. The drive mechanism, the head and the analog logic doesn't distinguish between write and format. Format just writes a whole track, while write does it for an individual sector.

[gargoyle]

but I wouldn't have had the 3086

chip if I didn't keep all that stuff.

 

They're still around it seems, just in case anybody else needs it:https://tinyurl.com/n7recgs

[cwc]

Really pleased you got it fixed, David!

[Rybags]

So what's the resolution and reason it could write but not format?

[_The Doctor__]

@rybags LM3086N chip u21 was bad and he put in a socket and replaced it....

 

Congratulations! again.

[Rybags]

Why did writes work but not format?

There was some mention of index pulse generation - is that used by the firmware to simulate the index sensor for formatting and RPM measuring?

[+kheller2]

My completely unscientific view of the schematics hints at the fact the Erase head runs through U21 on different pins that the R/W head.

Its also a different voltage.

 

But then again, there are common signals on the E and R/W heads that confuse that theory.

[Joey Z]

Why did writes work but not format?

There was some mention of index pulse generation - is that used by the firmware to simulate the index sensor for formatting and RPM measuring?

well actually there's no index sensor on the 1050 as it turns out, so I think I was wrong when I mentioned anything about that. My guess is that the 1050 doesn't bother with the index, and probably uses timing to do any track skew.

[_The Doctor__]

as was mentioned in similar threads the the disk was NOT erasing during formats.... the data was still there.....

the 'index' pulse as mentioned previously is simulated... it is not from a standard emitter detector pair.... as I and others have mentioned in more detail previously

Edited May 7, 2017 by _The Doctor__

[ijor]

as was mentioned in similar threads the the disk was NOT erasing during formats.... the data was still there.....

 

But again, writes are format are the thing as far as the head and analog logic concerns. There is no such thing as writing without erasing in these drives. If erasing wouldn't work, then you couldn't write either.

 

Write vs format is a high level concept implemented at the FDC level. From the FDC to the drive, there is no difference. There is a single write signal, and a write gate signal. There is no such thing as an erase, or erase enable signal.

[DavidMil]

You guys are too deep for me. All I can tell you is that the drive now formats after changing out the LM3086 chip!

 

David

[ijor]

 

But again, writes are format are the thing as far as the head and analog logic concerns.

 

Ouch. I wrote that? .. I meant:

But again, writes and format are exactly the same thing as far as the head and analog logic is concerned.

Edited May 8, 2017 by ijor

[1050]

Not entirely the same ijor. At format, the track is laid down in

track at once mode which must have intergap data and MUST erase

everything but especially old intergap data and write zeroes for

the sector data field.

 

During sector writes, only the sector data field is altered after

the proper sector number has been found via the intergap data

info which is not overlaid onto the old intergap data. If we have

a weak erase/write drive here it can still change some data and give

the appearance of working while it's really a crap shoot as to

how well it is working. A grey area for sure. Erase/write heads

(plural) are married for life, the power supplied to those heads can

be weak enough to appear to write sectors while not being able to

pull off an error free and proper track at once format is my

understanding.

 

I prefer to think of the heads as erase and write/read in two different

locations with erase preceding the write/read combo. Only an opinion

there, I have no proof of any of the above. It just seems to fit

what I've experienced in my drive trouble shooting episodes.

 

These '3086 fixed it' drives all seem to have a partial write

capability while failing to format error free in a timely

manner. I highly suspect that what writing they are doing is at

best only marginal enough to give the appearance of a quality write.

I would not trust one any further than I can throw it.

 

DavidMil,

That would be the front 3086 for the fix, there is a rear 3086 too.

Congrats, that's another one for the books. My count is three so

far (one by PM) but that might be a biased accounting. It does seem

nebulous sort of like the symptoms, but changing out the front 3086

is something to try when the 1050 is not formatting.

[DavidMil]

DavidMil,

That would be the front 3086 for the fix, there is a rear 3086 too.

Congrats, that's another one for the books. My count is three so

far (one by PM) but that might be a biased accounting. It does seem

nebulous sort of like the symptoms, but changing out the front 3086

is something to try when the 1050 is not formatting.

 

I'm confused, Why do you mention U1? I did replace a U1 about 30 some-odd years ago when a lighting strike a few houses

down the street vaporized the Atari 400 that the drive was attached to. It was easy to spot the bad 3086 chip because there was

a nice round hole in the top of it (along with a discolored resistor).

 

David

[1050]

Didn't mean to do that. Mainly just for the next guy or in

case you didn't know that there were two of those birds in

there. Don't want the next guy changing the wrong one either.

 

Wow, impressive story. That chip would be a collector's

item here, took some juice to do that.

 

It wasn't repairable after taking a hit like that, right?

[DavidMil]

I fooled myself into thinking that I did fix it (several different times). But every time I thought I had it fixed, I found another dead chip,

or a badly swollen cap, or another blown diode. The 1050 is a very well built drive, but I guess the engineers didn't plan on 10,000

or so volts down a 5 volt command line. Go figure... Eventually I pulled the guts out, saved what I could from the transport, and

pitched the rest in the garbage. It was the kids 400, so then I had to let them play on my really cool 800.

 

David