Apple I power-on memory state?

3 replies [Last post]
Offline
Joined: May 11 2005
Posts: 4

Can anyone tell me what the _actual_ state of memory locations are on a _real_ (i.e. not emulated, not clone) Apple I at power-on? (I'm trying to establish whether the DRAMs come up with 0s in all locations on poweron or if the values are random, and if the latter, whether the firmware itself clears them on initial startup.)

Rodney

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
Offline
Joined: Jun 5 2008
Posts: 381
Re: Apple I power-on memory state?

With my Mimeo 1 that has Mostek 4096 DRAM - date code twenty ninth week of '77, memory comes up >95% ones, with a few zeros scattered here and there. The monitor doesn't do any clear memory function at startup. The monitor source can be found online in the Apple 1 Operations manual.

video memory looks like the image in the Mimeo 1 build manual (every other character a flashing cursor and the rest are underscores).

There is no reason to think that the "real" thing will behave any differently, though someone else with an operating original unit may be able to confirm.

Regards,
Mike Willegal

Offline
Joined: May 11 2005
Posts: 4
Re: Apple I power-on memory state?

Thanks, Mike!

My main reason for asking is a desire for more accurate Apple I emulation...namely, the power-on memory state should accurately reflect reality (however ugly).

Unfortunately, emulating this particular 'quirk' would be, shall we say, interesting, because there's no pragmatic way of emulation at the gate level. The best that could be done here is simulation, but without a predictive model (i.e. should I just fill the memory space with 95% of the bits randomly set?), it doesn't really improve accuracy (particularly if different DRAM timings affect the result in real life).

I'll take a look at the Mimeo 1 build manual about video memory.

My comment about wanting to know about the behavior of real hardware was more a result of knowing that some clones use modified firmware that could potentially affect the result (and to be very honest I haven't looked at the original or current disassemblies to verify that, but you confirmed what I suspected - heck, there's barely enough ROM there to bring up a monitor prompt! Wink. I'd say the 1977 Mosteks are a more than adequate bench case. =)

Rodney

Offline
Joined: Jun 5 2008
Posts: 381
Re: Apple I power-on memory state?

This becomes real ugly if you really want to emulate not just power on, but a quick power cycle. Flipping the power switch briefly off and then back on, will result in different results than power off for a while (say five minutes), then power back on. Some memory cells apparently retain their charges longer than others. Smile

Regards,
Mike W.