ST536 STE EDITION

[Cyprian]

great progress

[exxos]

Has anyone got a ST536 built up and working they can loan me for a few days ? (UK)

There is all sorts of weird issues on my STE, and I don't know if they are hardware or firmware related at this point. Though as I have a PLCC adapter for the ST536 to PLCC socket, I can plug in a ST536 to see if it works better or not than the STE536.. The firmware is basically the same at this point..

[JezC]

I've got the one I bought from @PhilC that you can borrow

[exxos]

What may be the problem, is the PLD likely doesn't like a 3.3v clock and actually needs 5v to function properly.Something is very borderline...

I've been through all this osc crap with the CT60... some oscs would work others not at all... It was one reason I developed the adjustable clock modules... But I had a right time with those... I ended up with a bias network to get them to work with the CT60 at all. I never figured out the problem back then.

With a 100R pullup on the clock, its doing similar, shifting the DC bias up a tiny bit... Increasing the logic high level fractionally and now things are way more stable .

I first noticed the problem as if I ran the CPU in the slower clock domain, with the regenarated 8mhz clock, it was unstable. Lucky to even get to desktop. Running from the raw 8mhz clock, it could run all day. The only difference was the regenerated 8mhz was clocked by the 100mhz oscillator. I thought it was clock sync or timing related at first but no change to the firmware made any real difference. So that's the summery of about 50hours work there!

So next I have 2 options.. Either buffer the osc with a 5v buffer, or run the osc on maybe it's own 3.8volt regulator..

[exxos]

The only regulator I could find easily was a 3.6V one, so I have added one of those in which will power the OSC now. I've also found a different brand of oscillators so there will be fitted in the second set of pads.. There is a jumper to select between two oscillators anyway. I also added in a termination resistor which can be switched between ground and 3.3V. This way I can "force" some "tolerances" to make sure its more stable. I also added in some special high-frequency bypass ceramics right next to both oscillators should help a bit I suppose.

Is all a bit cluttered but it fits :)

Capture.PNG

[JezC]

@exxos Do you want to borrow the ST536?

If so, I might be able to post it on Monday morning, just let me know

[exxos]

@exxos Do you want to borrow the ST536?

Ah sorry I didn't reply yet... Probably won't need it.... I think, as mentioned in the above post, it seems there are a few problems around the ocsilator... It's going to take a few weeks to respin the board and test again...

[exxos]

Clock patch bodge boards now on order....

Capture.PNG

[exxos]

Patch boards *might* be here tomorrow :)

BTW - This patch board will have to be verified and tested before any ST536 boards can be produced.. While TF536 / ST536 seem to work fine (***), I am reluctant to produce the ST/STE536 until the clock problem has been solved. While it doesn't (seemingly) seem to be a problem with the TF/ST536, At this point in time, I believe timings are very borderline. This is what seems to have been so far have been witnessed with the STE536.

(***) Problem is due to parts shortages, I moved to a different brand of OSC on later ST536 and STE536 builds. So while past ST536 might have been fine.. the later ones probably won't be..

[Steve]

Cool! Hoping they work well :)