| Click here to select a new forum. |
| SE/30 Pixel 'Snow' Artifacts After 10 Minutes or More |
Posted by: hfrazier on 2018-03-13 07:22:08
That's what I am thinking as well, because in my experiments once you add a PowerCache to the mix the artifacts are gone.
The PowerCache has buffers which most probably deliver more power than the bare CPU driving the bus on the stock logicboard. @Bolle But would that explain why the artifacts are not present after a ~20 minute cool-down period??
|
Posted by: Bolle on 2018-03-13 07:38:04
Maybe. Once the CPU heats up the maximum current it can drive will get lower. This is only a tiny little bit but as it obviously is operating at the edge of what is possible that tiny bit might be enough.
|
Posted by: hfrazier on 2018-03-13 08:15:28
Maybe. Once the CPU heats up the maximum current it can drive will get lower. This is only a tiny little bit but as it obviously is operating at the edge of what is possible that tiny bit might be enough. @Bolle That makes sense. Guess I need to install some liquid-cooling 😛
I'll freeze the CPU first in case we can root it out. The only other thing to consider is that -12V rail being at ~-9.8V.. Not sure what that powers or if it matters. I'd figure the 5V lines are all we need to consider.
|
Posted by: IlikeTech on 2018-03-13 11:14:24
That sounds way, way low.
|
Posted by: hfrazier on 2018-03-16 10:13:38
Hey everyone, I'm back with updates!
It's UE8, I'm pretty sure.
-With the logic board removed and the cables extended, I ran it for close to two hours and the artifacting was extremely light, like in the first picture I posted earlier. (Points toward the heat issue)
-Placed a piece of paper on top of the board and ran it longer. That helped trap heat and it produced a bit more artifacts. Funny though, after I removed the paper, they got less pretty quickly.
-Froze UE8, and the artifacts went away.
I'm wondering if I should just place a little adhesive heat sink on UE8..
|
Posted by: hfrazier on 2018-03-16 10:48:05
Yeah. It's the 8-Bit shift register UE8. Check this out!! @Bolle @techknight @apm
|
Posted by: Bolle on 2018-03-16 10:50:42
Sweet.
|
Posted by: MOS8_030 on 2018-03-16 11:34:09
Your troubleshooting technique is... novel to say the least.. 😀
|
Posted by: apm on 2018-03-16 13:56:01
Interesting! The question then is whether UE8 is faulty, or if it's just sitting on the very edge of spec because of the other load on the PDS slot. I wouldn't bother heatsinking UE8: try replacing it and see if the problems still persist.
|
Posted by: techknight on 2018-03-16 14:15:48
Yea your case is unique just like mine. I have replaced UE8 and all the other ICs with no change with the board I have. So its sitting in the junk pile until I can get another look at it.
|
Posted by: joethezombie on 2018-03-16 21:33:49
I have to say, putting flame right to the logic is one of the coolest videos I've seen. Bravo.
|
Posted by: techknight on 2018-03-17 10:13:09
Ive done it alot of times with a handheld torch for heatshrink, But I prefer the heatgun method. a flame could overtemp the chip and actually damage it if your not careful.
|
Posted by: james_w on 2018-03-18 13:44:34
Haha is this an approved technique? 😉
|
Posted by: techknight on 2018-03-18 14:29:31
Eh. Not really, but if you need pinpoint accuracy theres no other way.
|
Posted by: hfrazier on 2018-03-19 11:11:59
Interesting! The question then is whether UE8 is faulty, or if it's just sitting on the very edge of spec because of the other load on the PDS slot. I wouldn't bother heatsinking UE8: try replacing it and see if the problems still persist. That is what I'm going to do... The heat sink didn't work for squat! Got one on order from digi-key. This one will work.. right?
I'm also just hoping that the chip is simply faulty. ... I would think that it is a faulty chip, since even in the lighter test one or two columns look like they are still functioning properly... Hmmm. :/
@techknight Yeah I have used a lighter many times for heat shrink so I felt fairly confident that I wouldn't mess it up. 🙂
@joethezombie Thanks! It definitely gave me a definitive answer!
|
Posted by: hfrazier on 2018-03-19 16:03:06
Update - I was able to replace it with the new one that came in, but unfortunately the problem still persists, although much lighter. Dang! It must be juuuust on the edge, just like @apm was suggesting.
Of course, removing the pivot makes it go away completely. Is there anything else I should try or just forget about making it perfect? I'm afraid it is probably just a hardware limitation (or at least until I find an accelerator card to put in with Bolle's ProtoCache.)
|
Posted by: techknight on 2018-03-20 14:58:22
You need a riser card with internal buffers IMHO. I know they make bi-directional buffers without the need for the select or R/W line. Maybe thats the route to go.
|
Posted by: apm on 2018-03-29 17:28:38
Here's a weird idea, maybe not a good one but probably not harmful (@techknight?). Try shorting out R29, a 22-ohm SMD resistor on the bottom side of the board, just under UI6 (I think).
R29 is in series with the CLK16M signal (i.e. the main system clock), which is generated by UI6. If there's enough capacitance on the bus, for example through too many cards, then the rise time of the signal could be just barely too slow. But shorting that resistor will make the drive just a little bit stronger, which might be enough. I imagine it is there to meet emissions specs and that shorting it probably wouldn't damage anything, though I suppose there's a chance that it could create a stability issue.
The even stronger version of this would be to place a 5V high-speed buffer in place of R29.
However I still don't have a good explanation for why a laggy clock signal would produce this pattern on the display, so maybe it won't make any difference.
|
| < 2 |