@GTV reviews the Cosmic Fantasy 1-2 Switch collection by Edia, provides examples of the poor English editing/localization work. It's much worse for CF1. Rated "D" for disappointment, finding that TurboGrafx CF2 is better & while CF1's the real draw, Edia screwed it up...
Main Menu

PCE help diagnose video problem

Started by Chrom_, 06/09/2015, 03:00 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Chrom_

Hi all. This is my first message in the board, and I'm very glad to be part of this community. I hope I'll be able to give help, more than I'll need.
   
On my JPN White PCE (RGB modded) I'm getting corrupted sprites and backgrounds. The characters look fine instead. I'm planning to check the continuity of the traces this weekend because I have a vague suspect for a cut trace between the 6270 and the flat cable. There a bit of solder mask and possibly traces that came off under L102. Could it also be bad RAM?

Any idea or your past experience in similar problems is a very welcome help. All relevant images posted here

IMG
IMG
IMG
IMG

Here's an high res pic of the board. The place where I suspect a cut trace is in the lower left part of the image, between L102 and the area where there was the RF unit (removed for RGB mod)
IMG

Thanks in adavance for any sugestion

Keith Courage

#1
It's either bad ram or bad traces between the ram chips or Hu6270 chip.

The memory/ram chips are the top two in your pic that are identical.

Chrom_

I'll check all the traces as soon as possible. In the case the traces are OK, is there any trusted selled where RAM chips can be sourced?

Chrom_

Today I've checked for continuity ALL the traces going from and to the 3 main chips and the 2 RAM banks, no broken traces found. I've also checked the continuity to and from the HuCard slot, with no issues found. After that I've checked the voltage regulator, and it outputs exactly 5 volts. After that I have checked ALL the chips and RAM pins with a logic probe, and I see pulse to all the RAMbus and DATAbus. I'm really out of clues now. I have tried to bypass the RGB amp, but I'm not sure I managed to bypass it completely, so I can't say if the test was relevant. could it just be a broken HuCard? I don't have another one at the moment for testing. Any other idea to help me?

pnauts

I have a weird PC genjin 2 which make some glitches with incorrects backgrounds like you.
As you don't have another Huc, I think that it could really be the source of your issue.
Get some Huc to confirm.
IMG

esteban

Yes, I have a PC Genjin 2 with glitches as well.
IMGIMG IMG  |  IMG  |  IMG IMG

Chrom_

I'm sourcing a few hu cards, so that I'm able to test. If it's not that, my last hope is to bypass the RGB amp completely and see what happens. If that fails too, I may try replacing the RAM, but I have no idea where to get some spares from :(

mickcris

you can try to hook up composite video to the expansion port.  not sure if your tv supports NTSC composite video, but this would be bypassing the RGB amp.

http://www.cgquarterly.com/articles/tg16_cable.htm

Chrom_

Today I've completely bypassed the RGB amp by tapping svideo from the EXT slot. Same output.

At this point the chances are:
- broken Hu card
- some strange fault at the RAM level that is not evident with the logic probe

I still have one doubt about the small IC that is located on the Hu Card reader doughter board. Pin 1 seems connected to +5v or GND (I get continuity), but it is the only pin that gives me no reading with the logic probe. Any idea about this?

Also any reference for RAM spare parts?

Thank you all

pnauts

Do you try to reflow this pin ?
IMG

Chrom_

I'll get a couple of new Hu Cards next week. I'll test the unit with the new cards first, then I'll try reflowing. I'll keep you posted guys. Thank you all

Chrom_

I have tried with 2 other Hu Cards, and unfortunately I have the same videou output problems. I have noticed that to make the hucard work I had to plug and unplug it a few times. I have also noticed that the 6270 runs considerably hotter than all the other chips. Is that normal?

thesteve

if you have a chip running hot odds are its a bad chip (unless its got pins bridged)

Chrom_

Is there any good method of finding if there's a bridged pin? Would checking continuity between pins work even if most of them refer to the same databus or IObus?

Chrom_

Visually I'm not able to see any bridge... Check the attached image

Capture.JPG

thesteve

several pins look like they could be shorted
take a meter and check for adjacent pin shorts
or flux the pins and touch them with hot iron
the flux causes the solder surface tension to increase beading up clearing most shorts

thesteve

49 to 50
37 to 38
54 to 55
see the lines between the pins, they look like solder splash to me, but may or may not be causing a short

Chrom_

I'll check again very carefully for any shorts in the weekend. Thanks a lot for the tip. I'll keep you all posted

Chrom_

#18
Though I must add that trying different hucards, I always have the same problem: bad palette on sprites and disappearing sprites (meaning possibly that the sprites are there but covered in trasnparent colour, so again a palette problem).

According to the following documentation (http://www.emulatronia.com/doctec/consolas/pce/vdcdox.txt), it is the HuC6260, not the HuC6270 that supplies the colour palette to the Video System. Excluding a problem with the RAM (that would result in far worst effects and also a variety of glitches depending on the loaded game), and despite the fact that the HuC6270 runs hotter than the others, at this point I'm suspecting the HuC6260.

It's also true that HuC6270 does Background character processing and Sprite processing, so it is indeed possible that something is wrong with it...

Any wrong assumptions in what I wrote above? Other ideas?

thesteve

Only thing is the tile/sprite info is supplied to the 6260 from the 6270
I have seen palette and scale issues on the 6260 but such failures have been rare

Sent from my VS920 4G using Tapatalk


pnauts

picture you provided is too bad quality to see if there some shorts.
IMG

Chrom_

No shorts detected on 6270. To be honest the chips look in perfect cosmetic conditions, and no bad soldering is evident at all. Would it just be a fried 6270? Does it fail like that?

thesteve

it does fail like that
it also fails white screen
it could still be another chip as the issue is missing/corrupted data but the 6270 is a good bet
have seen 6260 fail that way only once, 6280 failures tend to be more severe

Chrom_

Any known option to buy a replacement 6270?

thesteve


thesteve

just fixed a duo with palette/grafix issues
was solder bridges on the 6270

Chrom_

I need to source some flux then I'll surely try refluxing the chips. The problem is that in my case the contacts look perfect to the eye and no short was found with the multimeter. Was the short evident in your case?

thesteve

had 2 shorts
the first was causing grafix corruption (i saw it)
after fixing that one the colors were wrong
reflowed the 6270 and it fixed the colors

Chrom_

In the next few days I'll try reflowing anyway, just in case. I'll keep you posted. Thanks! :)

Chrom_

today I have refluxed all chips, including RAM and the one on the Hucard doughterboard but the problem is still there. At this point I'm almost sure 6270 is failing :(

With all those custom chips it's almost impossible to cheap repair the Pc Engine...

thesteve

because the PCE is the cheapest unit to buy as a doner, its not often worth replacing the customs (3 chips)
everything else in the system is off the shelf common parts