ST536 STE EDITION

[exxos]

PS. my daughter is going through a period of not sleeping unless someone sits with her for an hour or more so I've had exactly no time the last couple of days for CPLD changing. :(

No worries, family first :)

I'm trying to figure out why its gone nuts for no reason. Part of the PLD is at 55c ! CPU is around 52C. But "stick on" heatsinks not great for thermal conduction IIRC. So the CPU core could be more like 70c for all I know.. Will put a fan on the whole board.. see if anything changes...

[exxos]

@PhilC I think I remember that CF card now.. its odd as the STE536 "last gen" won't boot up with it plugged in at all.. The current gen had the fix I did for TRUDIE to stop that issue... Though I wonder if theres something else "off" with that card as all my troubles started when I plugged that card in this morning :lol: :roll:

[PhilC]

@PhilC I think I remember that CF card now.. its odd as the STE536 "last gen" won't boot up with it plugged in at all.. The current gen had the fix I did for TRUDIE to stop that issue... Though I wonder if theres something else "off" with that card as all my troubles started when I plugged that card in this morning :lol: :roll:

Well it was a cheap Chinese one. BTW, I took the top off one of them the other day and it was branded SanDisk inside believe it or not.

[exxos]

Well it was a cheap Chinese one. BTW, I took the top off one of them the other day and it was branded SanDisk inside believe it or not.

Weird, is that one a bit crazy as well ?

[PhilC]

Well it was a cheap Chinese one. BTW, I took the top off one of them the other day and it was branded SanDisk inside believe it or not.

Weird, is that one a bit crazy as well ?

No, that one works a treat

[exxos]

I have noticed something else now :lol: :roll: I think there is something iffy with the IDE..

This does not always happen but if I slowed the address bus down slightly it becomes way more prominent...

IMG_3255.JPG

Currently I have latched the address bus to make it stable(er) but decode IDE again to disable the data buffers to the ST bus. I think the problem there is as its clocked by CLK100, its adding 10ns delay, plus the delay of the buffer (~5ns) I think its causing some conflict with the IDE interface. The problem also goes away when removing the IDE card.

On the current boards actually have a patch to turn off the IDE buffers automatically.. This is the fix, which if you do something daft like plug in a bad card or something, it can actually short out the bus preventing it from booting, which I had earlier ... But that doesn't seem to be enough to help with the current issue.

I think basically the IDE interface is just marginally running too fast (***) somehow, but I don't know how this ends up corrupting STram...

*** Or to slow ? Maybe buffers stay enabled to long corrupting the next bus cycle.. but it would seem unlikely considering how slow a 8MHz CPU is going to be reading data from RAM anyway.

BUT.. I could guess a glitching / transitioning address could inadvertently trigger the IDE interface while it's getting ready to access RAM.. Then by the time all the clock delays have gone through the IDE interface, its basically ended up with a bus-conflict I suppose..

I ran STram test for about 10 mins then it locked up. I ran TTram test and it locked up prettymuch right away. So I think the CPU is overheating. The board I have on at the moment doesn't even have a heatsink. But got my little fan on it now. So will see how it goes.

I was testing the voltages across the motherboard early and while the ground doesn't seem to have much drop, the 5v rail has more. Certainly not massive but is losing up to about 0.2v. So its a bit low for the 030 really. (4.90V) . I will investigate the voltage more later..

Ultimately the idea is to just have a gap filler pad between the heatsink and the top of the metalwork inside the case. The CPU sticks out a bit further than the edge of the case unfortunately but.. One problem at a time... Ultimately the sticky carbon sheet stuff can just be used on top of the CPU to the metalwork anyway to use as a heatsink!

[exxos]

Not seeing the corruption since doing the IDE fix.

I have confirmed that the voltage on my board was 4.96v which was causing Lockups after a few minutes. All this was documented and discovered previously on the DFB1X stuff. viewtopic.php?p=114456#p114456

I have upped the voltage so I get 5.10v on the board. STram Test got to loop 4 before I turned it off. I also did a hundred loops in the marpet RAM test just for the hell of it. It's on loop 2 of testing TTram now.

So some CPUs will run after slightly lower voltage than others, but I will have to say the same as the DFB1X, that the voltage on the power supply needs to be increased slightly for the CPUs to function properly.

I have just uploaded the current8 firmware. Along with TOSCOPY which copies ROM into TTram. It has the current fixes for IDE etc

The new firmware has the basecode (but not enabled) for a new function register. So various things can be turned off and on.... Oh... And I've used 80% of the PLD now. The compiler is grumbling at me now :lol:

[Steve]

@exxos amazing work Chris, you really are making some great achievements here. It's inspiring 💪

[exxos]

@exxos amazing work Chris, you really are making some great achievements here. It's inspiring 💪

:thanksyellow:

It's inspiring me as well.... to drink heavily. :lol:

[Badwolf]

Took the old CPLD off last night, but half-lifted a pad (144, I think -- the one to right right of pin one).

I think it's a ground. Is it connected 'outward' or 'inward', @exxos? It's the inward half that's lifted so may have to bodge wire it if the trace goes that way.

BW