Can a misaligned 12V kill /M1?

[dfzx]

The problem with some some Spectrums' Z80's /M1 lines not working is well known, but I recently noticed this blog post from Tynemouh Software:

http://blog.tynemouthsoftware.co.uk/201 ... gnals.html

He says:

"However, if you look at the ZX Spectrum edge connector, M1 is right next to the 12V AC line. I wonder how many misaligned joystick interfaces accidentally shorted the 12V AC into the M1 line and killed it?"

I've seen this post cited a couple of times in retro computing forums, where people agree that a misaligned interface is the likely cause of a dead /M1 line.

Is this credible? I struggle to see how it's possible to misalign an interface such that /M1 is damaged, but nothing else is. I'm also not sure whether such a mistake happens often enough to correlate with the number of failed /M1s out there. (I was recently chatting to someone who's made up one of my devices, and it turned out that 2 of his 3 Spectrums have /M1 failures.)

So is this a reasonable explanation for the observed /M1 failures? Does anyone have, or care to speculate on, what other credible explanations there might be?

[cj7hawk]

More than the M1 gets damaged by misaligned interfaces... And frequently it killed the entire spectrum.

I suppose the M1 fault is simple a fault that doesn't result in the machine's immediate failure as it wasn't really used for anything. So when this was the only resultant, it's likely no one realized the machine had been damaged.

Other failures would result in a dead machine and repairs, but even then the z80 M1 line wouldn't have been checked for operation on repair.

And yeah, it was pretty common. Once fully inserted, it's much harder to misalign the connector, but on insertion, it's fairly easy. Which is why accidentally forgetting to turn the computer off before inserting an external device was such a common cause of damage... You've probably seen the stickers that were on some older Spectrums warning people not to do this... And of course, a whole bunch of people would have bought spectrums back in the day, and just plugged stuff in and out while the power was on, feeling for the correct connection as things come together.

If it's the cause, the M1 will most likely be shorted to power or ground. If both, the chip would fail. You can check it with a multimeter.

[1024MAK]

The misalignment may occur if the locating pin has detached from the female edge-connector socket on an interface (such as a joystick interface).

Or if the user attempted to plug in, or remove the interface/expansion while the power is on, especially if it ends up being at an angle to the board (hence the contacts of the female edge-connector socket are at an angle and short between adjacent contacts).

And yes, edge-connector abuse can, and does sometimes kill other chips, or cause other damage. But as the /M1 line is not used internally, a fault with this output pin does not show up unless and until an interface/expansion is plugged in that needs it.

Note that some joystick interfaces do include the /M1 in the address decoding, normally only those that use a 74LS138 for the address decoding.

Edge-connectors come from industrial electronics applications, where they were normally used in rack mounted systems. Especially to form a backplane. Here the PCB cards are mounted vertically, and the rack has long guide rails/slots that keep the card perpendicular to the edge-connector socket (the socket is mounted at the back of the rack). Hence with some systems, inserting or removing cards was possible without causing damage.

The ZX Spectrum has absolutely no physical support or guides for expansions/interfaces, just like the ZX80 and ZX81. Which could be considered a design limitation.

Mark