Old 7406 driver at the floppy interface

[ijor]

I recently found out that the TTL chip that drives the floppy interface in the ST is a plain 7406, and not a 74LS06 chip. I have seen the schematics numerous times and also the actual chip itself, but I admit I never paid attention before, until Pera mentioned this in other forum.

I am wondering why Atari decided to use a 74S chip. The first idea that comes to mind is because the older TTL family has higher driving power than the LS one. That would make a lot of sense for the floppy interface. Typically, 74S chips can drive at least twice the current than 74LS ones. But not in this case, the 74x06 is already a high current driver in either family. According to TI datasheets, both the 7406 and the 74LS06 are rated for the same 40ma maximum sinking current.

So why use a 7406 and not a 74LS06? May be the 74LS06 was rather new at the time? That's possible, I guess. But note that all the schematics, ST, STE, MSTE, and even the Falcon, still put a plain old 7406. May be they just didn't bother to change that and kept using the old original design without any specific reason?

[czietz]

I would say: the LS06 simply came too late. Browsing through the data books from common TTL logic manufacturers from the late 1980s one can see that no LS-variant existed back then. Three examples, courtesy of Bitsavers:

TI, 1988:

ls06-ti-1988.PNG

Motorola, 1986:

ls06-motorola-1986.PNG

National Semiconductor, 1987:

ls06-natsemi-1987.PNG

I did not research when exactly the 74LS06 was introduced by the respective manufacturers. Maybe it would have been soon enough for the Falcon; but Atari did not care, anymore, then.

[ijor]

I would say: the LS06 simply came too late.
...
I did not research when exactly the 74LS06 was introduced by the respective manufacturers. Maybe it would have been soon enough for the Falcon; but Atari did not care, anymore, then.

Thank Christian. I suspected that the 74LS06 wasn't available at the time. But surely it was when the Falcon was released, and even probably also when the MSTE was released. But yes, I guess it was already too late for Atari anyway.

[czietz]

It's only speculation, sure, but I can imagine they simply kept that part of the circuit (with the 7406) because it was working. One would only change a part if either one knew it's not functioning properly, or it isn't available anymore, or the price of the replacement is significantly lower.

[Icky]

It's only speculation, sure, but I can imagine they simply kept that part of the circuit (with the 7406) because it was working. One would only change a part if either one knew it's not functioning properly, or it isn't available anymore, or the price of the replacement is significantly lower.

I was speculating similarly thinking that if they bought a large stock of the 7406 and as you mention @czietz if it works don’t fix it then they continued with the 7406

[ijor]

It's only speculation, sure, but I can imagine they simply kept that part of the circuit (with the 7406) because it was working. One would only change a part if either one knew it's not functioning properly, or it isn't available anymore, or the price of the replacement is significantly lower.

Well, that might be possible. But according to the BOM there are other 7406 chips used in the motherboard. So may be it is like Icky is saying, they just had many in stock.

[Steve]

@ijor Just wondering - what would the LS varient offer over the older version? How does its electrical characteristics differ?

[ijor]

@ijor Just wondering - what would the LS varient offer over the older version? How does its electrical characteristics differ?

Well. 74LS chips use much less power and also they are slightly faster. That's why they replaced the old TTL family (almost) completely. But it is true that in this particular case is not that important. The main power consumption here is determined by the strong pullups at the motherboard, and not by the chip itself. There would probably still be a difference when the floppy is not used and the chip is idle with all signals at the high level.

As said, it is also slightly faster, but a few nsecs at the slow FDC interface is not really significant.