RAM speed issue

[msknight]

I'm doing my best to try and understand what's happening, and I'm wondering if TR1 is starting to fail... or somehow not switching fast enough?

But I'm now back down to guessing.

[1024MAK]

It's extremely unlikely that TR1 is causing an issue. Transistors in circuits like this are not under any stress whatsoever, and transistors are fairly hardy and reliable devices.

I'm sorry that I can't help you much. But with edge cases especially when it comes to timing issues, they can be extremely frustrating and difficult to track down.

Mark

[msknight]

Thanks Mark,

I know you've got your hands full with life right now, but if you can remember cases in the past and what the resolutions were, then I can give them a try and see if anything works.

But I know you've got a lot higher priorities in your life, and rightfully so.

Michelle.

[dr beep]

I'm not dissing the game, or your fine work in any way, shape or manner.

I don't take it in that way.

I can take a look what value from which address must be read/shown if that can help?

Hardware itself I can't help you with.

[dr beep]

Closer look at the picture

First thought:
The text is ok, but some bits are
somehow set.

The best thing you could have is a simple
display with just 1 (adjustable) value and see how that display appears.

[dr beep]

I made a simple testprogram out of a game.

HR-test.p

(542 Bytes) Downloaded 95 times

After NEWLINE you will see 16 hires lines (well more, but also lines with zero)

First line 1 and 1
spaceline
Second 2 and 2
spaceline
Third 4 and 4
etc....

So you can see which bit are set by your computer.

[msknight]

Thanks very much for this.

I eventually socketed the whole board and I was using two Ferranti ULAs. After swapping pretty much everything, I reluctantly pulled a uLA81 from another motherboard (because the new stock still isn't in yet) and tried that...

It worked.

So I've got two partially failed ULAs ... or potentially to be generous, two ULA's that can't perform to the standard where WRX hires graphics will work, even if I was using the original 1k RAM.

That currently places me at a 50% ULA failure rate... and I'm wondering how long it will be, before the other 50% will eventually fail as well.

[dr beep]

So I've got two partially failed ULAs ... or potentially to be generous, two ULA's that can't perform to the standard where WRX hires graphics will work, even if I was using the original 1k RAM.

I think you have bunked why "some" ZX81s did not work with my hiresgames.
We wanted to release a tape with all games but due to this issue we never did.

[msknight]

I think you have bunked why "some" ZX81s did not work with my hiresgames.
We wanted to release a tape with all games but due to this issue we never did.

Cheers. That explains it.

I've noted this down on the ULAs and will put them one side, noting that other than the HiRes side of it, they are still functioning fine. In future I will also test zeddies for this issue before I do the ram upgrade.

Michelle.

[1024MAK]

It is possible that not all original Ferranti ULA chips would have worked correctly with WRX hires graphics software.

Remember, the hardware was not designed for this. With the wide variety of components used (different makes/models/manufacturers of processors, RAM, ROM, the different ULA versions and different batches of ULAs) there are bound to be some machines that work normally in the standard SLOW and FAST modes, but which don't work correctly with WRX hires routines.

Are we rediscovering something that was already known about (but has been forgotten), or...

Mark