Escom/AT A1200, floppy led always on (but floppy is working)

[Turrican_3]

Hey there,

I'm having a slightly annoying (but apparently harmless) issue with my Escom/AT A1200: when powered on, after an extremely brief pause, the floppy LED turns on and keeps staying this way regardless of any disk activity.

It's doesn't seem to be a floppy issue (tried an A600 drive, exactly same results) so I wonder what might be causing this.

Thanks to AmigaPCBexplorer I see - unless I'm mistaken - this signal is driven by Gayle via a small SMD transistor near the LED panel connector. Is this correct and, arguably more important, what checks could be made?

EDIT - please note the built-in floppy LED seems to be working fine, it's the PCB one (i.e. the one wired to the motherboard together with Power and HDD leds) that stays always on

[solarmon]

The FLOPPY LED is based on the _MTR0 signal. This goes to a circuit that drives the LED.

Either the _MTR0 is always held low (by Gayle [U5] pin 67) , or the circuit that drives the LED is keep the LED on.

Use the schematics and amigapcb.org to check this signal and drive circuit (R633, R634 and Q632).

[Turrican_3]

Thanks for your reply!

If it was the _MTR0 signal being stuck, would that mean the drive would always spin?

Sorry if the question feels a bit dumb, but I'm not very familiar yet with the Amiga inner workings.

[Turrican_3]

Ok, luckily there's a via next to Gayle and I was able to easily verify the signal: it seems the _MTR0 is roughly 5V when the A1200 is turned on, then after roughly a second it goes around 0V and stays that way straight from there.

Dead Gayle or what? Any further checks I could do?

Right now there doesn't seem to be any other noticeable issue, but perhaps _MTR0 being kept to 0V IS actually a serious issue, i.e. what might go wrong during regular use of the A1200 ?

PS: please note there are no visible cap leaks, and the system has been recapped a few years ago (2019 IIRC)

[solarmon]

Use Amiga Test Kit and see if you turn the floppy disk motor on and off.

Also, use Amiga Test Kit to do other health checks.

[Turrican_3]

I can confirm (via ATK) the motor always spins as soon as I insert a disk / start the test.

To avoid a further element of uncertainty I used an A600 floppy drive instead of the Escom supplied PC Panasonic floppy, which has a small logic PCB between the flat and the motherboard connector (both appear to be working fine though except for this always-on spin)

Specifically, ATK sees a floppy ID of 00000000

Register CIAAPRA is:

0x e0 (no disk inserted)
0x c4 (with both motor on and off, but shows 0x 84 as long as I keep pressed the "Motor On/Off" button in ATK)

Everything else seems to be ok: IDE & PCMCIA under Workbench, RAM, floppy disk (except the motor of course), CIA timing tests, etc. under ATK

[Turrican_3]

I've had a quick look with my untrained eyes at the A1200 rev2 schematics and I noticed Gayle has a _MTR *input* that comes from CIA (U8)

Should toggling DF0: (or possibly any floppy motor?) motor on and off make this signal change? Because all I currently see is a 0V voltage.

Perhaps - assuming the two are actually related, that is - it's this CIA that isn't properly working, not Gayle?

[Turrican_3]

Another quick update.

Checked the _MTR signal going from Even CIA to Gary on an A500 and it seems to be related indeed, as I can see it going from 0 to roughly 5V each time I toggle the motor with ATK, while - as I said - it stays stuck to 0V on the AT A1200.

So I guess it's very likely I'm having a faulty CIA on the A1200 then?

[Jope]

You have deduced it correctly, it does go via Gayle.

The motor signal originates from the CIA.. The Gary/Gayle generates the internal motor signal from this based on the motor + sel0 signals. MTRX is for the external floppy drives.

It could be a CIA problem, it could be a Gayle problem. CIAs are more common to fail though, so I would start from there. Too bad it is so difficult to isolate signals in the A1200. In a through hole Amiga you could just bend out the motor signal pin and see if that helps.

[Turrican_3]

Yup, plus had it been a through hole Amiga I could have very easily tried a different chip, even with a simple swap on the very same machine. But on SMD Amigas it's a completely different story. :-\

I think I'll try to desolder one from an A600 and see what happens... perhaps in the meantime I can source a few spares.

[solarmon]

Originally Posted by Turrican_3

I've had a quick look with my untrained eyes at the A1200 rev2 schematics and I noticed Gayle has a _MTR *input* that comes from CIA (U8)

Should toggling DF0: (or possibly any floppy motor?) motor on and off make this signal change? Because all I currently see is a 0V voltage.

Perhaps - assuming the two are actually related, that is - it's this CIA that isn't properly working, not Gayle?

Yes, there is a 'flip flop' in Gayle that takes the _MTR (71) and _SEL0 (72) signals as input. If _SEL0 is low it will output the state of _MTR as _MTR0 (67).

I have drawn up a logic diagram to try to show this and the other floppy drive signals, in a different way:

https://sites.google.com/one-n.co.uk...ve-connections

[Turrican_3]

Thanks, I am aware of your site, very informative and well made if you ask me! :-)

Here's hoping it's actually a CIA issue...

[Turrican_3]

Successfully removed two 8520PL CIAs from a donor A600 and the supposedly faulty one from the A1200.

Though looking again at the motherboard layout, I noticed the _MTR signal from the CIA is also connected to a 74LS86 IC.

So I'm wondering whether a faulty 7486 might be keeping the signal stuck to 0V or not.
I suppose that could be worth further investigations should the replaced CIA show the same behaviour (which in turn would likely mean the original one is not faulty)

On a side note, I am also wondering whether it makes sense to put PLCC sockets in place for the CIAs (that is, assuming I can find them easily), as I'm a bit unsure they would fit in the A1200 motherboard. Available space seems a bit too tight, hmm...

[Jope]

PLCC sockets will fit.

The LS86 might be your next suspect if a CIA swap doesn't fix this.

[Turrican_3]

Great to know!

Fingers crossed then...

[Turrican_3]

_MTR seems to be shorted to ground.

I have completely removed CIA (U8) and the 74LS86 (U26) and the short is still there, so I guess it's actually a fried Gayle? :-(

[Turrican_3]

Sorry for the quick bump but I need help to decide what to do.

I'm basically back to where I started: assuming the _MTR signal stuck to 0V is the only issue this A1200 has, how is software expected to behave?

Because I can definitely live with that if (and that's a big if!) this brings little or no compatibility issues.

[Jope]

The software probably won't care, but if you have a disk in the drive, it might always be spinning, depending on the mechanism.

[Turrican_3]

Thanks.

It does keep spinning indeed. But this system is mostly a WHDload one nowadays so I don't care that much if no other issues are present.

[solarmon]

Do you have any floppy write issues on this A1200?