STE video chip

[exxos]

Sign of the times trying to order a few inverters farnell want £10 handling plus postage , so it ends up over £20 just for a couple of chips evilbay were like £5-£10 each anyway not like i'd trust chips from there anyway. Long story short, I ended up ordering 100 from RS as there was no where else selling them

[stephen_usher]

Farnell are awful for handling charges.

[sporniket]

yup, same with mouser and the 20€ charge for orders below 50€. (plus the fact that as I have to not forget the VAT in my final computation)

[exxos]

Tried the inverter and oddly the signals come out the same. So I ramped the capacitor value to increase the delay and things got worse with the video. So as luck would have it I guess, the way the board turned out is the best fix.

If anyone is doing a order and has a 1712 shifter with the video faults as described, and is willing to actually test the board, then I will include the board for free.

So audio noise and video corruption - case closed!

Of course if anyone tries the fixes and it doesn't solve the problem then "at some point" I can take a look at your board. Though the fixes work on the earlier external blitter boards, and integrated blitter boards. So don't really expect there can be much other variation out there.

[JezC]

So don't really expect there can be much other variation out there.

That's a very brave statement for someone with so much experience of how unpredictable the original machines are!


Let's hope you are proved right though - at some point I will need to check my STE's (when they're fixed...) to see if I have the suspect video chip and the video or audio problems (adds yet another task to his massive retro to-do list...)

[ijor]

I had a quick look into this issue. It looks like the 16Mhz clock is to blame for the issues.
...
Falling edges on the C300588 look a lot better synchronised than the C301712 .

That's very interesting. Has this been confirmed that the C301712 exhibits a more significant clock skew than the C300588?

[ijor]

I don't know if it was ever reported if the shifter has any wake-up modes or not ? but something is clearly changing when the machine is powered down and powered up again. It seems very hit and miss if the video is basically perfect or totally trashed. It is not thermal because the video was basically fine for the past few hours. But I have noticed it repeatedly that turning machine off and on seems to have different effects on the video.

As Troed said, the STE does have Shifter wakeups. But these wakeup modes normally affect very special software only. I'm not sure how this could affect the video in such a way as you describe.

Don't know, conceivable, if noise on the bus is affecting the video, the wakeup mode might be relevant. But I guess you wouldn't expect the video to be completely trashed by noise.

I assume you are using a standard Shifter video output, right? I'm asking because if some custom video hardware is attempting to sync to the 16 MHz clock, this certainly might be affected by the different wakeups.

[Steve]

A quick "bodge it in test".


IMG_0369.JPG

Hey @exxos I'm wanting to test the board in an STe which is getting video corruption in some demos, I can't quite see how it is hooked up here, could you advise? Cheers

[Steve]

Does anyone else have an idea? Thank you

[Badwolf]

Hi Steve,

AIUI (and I don't have one and haven't tried this, so YMMV), this delay board sits in line between pin 52 of the shifter (the 16MHz out line) and the trace.

So you'd have to break the trace out of pin 52, 'IN' would be the chip side of the cut and one of the two OUT lines to the far side. Presumably you try both consequtively and see which is better.

As to the how? Sorry. Dunno. Perhaps insulation tape the chip leg for a non-destructive test?

BW