Other reasons for White or solid colour screen

Discuss hardware-related topics, such as development cartridges, CopyNES, PowerPak, EPROMs, or whatever.

Moderators: B00daW, Moderators

VajskiDs
Posts: 16
Joined: Mon Oct 12, 2020 6:07 am

Other reasons for White or solid colour screen

Post by VajskiDs » Mon Oct 12, 2020 6:30 am

Hi guys,

Recently decided to use opentendo boards to then put old scrap boards components on to use and turn them into complete original Nintendos again.

First board - all brand new parts - the boards based off the CPU-11 revision.
My 15pf caps hadn't arrived (theres ONE on the board) so I used an 18pf instead as there's one 18pf on the board as well.
I used 2 x 56pf caps instead of 2 x 51pf because I couldn't get/ find 51's

Then i basically replaced everything with new components bar the PPU/ CPU/ WRAM and VRAM.. maybe the 2021 transistors as well.

Added the RGB board - Rejuvenated the 72pin connector so it worked around 75%-90% of the time first go instead of 1% of the time then re-assembled and all good. Systems great - flawless. I've left it on overnight to finish difficult games and it's been fine even using only the one 7805 regulator for the entire system + RGB board. (PAL components - I'm in Australia - Mattel)


cool! I'll do another one - pulled apart a trashed Famicom - CPU/PPU/Crystal will give me an NTSC board! yay - this will be full screen. no overscan on my PVM required! I always prefer 60hz!

Does not work! I've spent now the most part of 2 full days trying to sort it out and I give up.
Everywhere i read, EVERYWHERE says its the 72 pin connector.. all the forums! the final say is always "oh it WAS the 72pin connector after all"
I've poked it with meters and pocket scopes and done comparisons to working non modded non opentendo systems...

I've tried CPU/PPU revisions off the Famicom
CPU/PPU revisions off my MINT condition boxed AV Famicom (I'm waiting for an FC everdrive, takes months.. haven't even used it!)
CPU/ PPU off a donour board purchased on ebay from the US

The famicom PPU mainly gives a green screen - sometimes dark blue/ sometimes light blue - switching on and off rapidly its orange/yellow
The famicom AV PPU is sometimes grey - but often green as well - or yellow - I think it's irrelevant

The US side loader revisions PPU only gives a white screen - from what I see on US videos on NES fixes this is completely normal.

But its not the 72 pin connector! I was lucky enough to get one that had being recently refurbished and hardly used - not sure where she got a decent aftermarket 72 pin connector from but this thing loads games first go - every go without even pushing the game down on the NES it came from.

Things i've noticed: CPU doesn't need to be installed for these "bad 72 pin connector" symptoms to show. WRAM doesn't need to be installed either. I still get the same symptoms. initially i figured it was the CPU but I've tried 3 NTSC ones now. I've tried 3 NTSC PPU's. I've tried swapping WRAM 3 times. I've tried swapping out the transistors responsible for CPU clock out of desperation and even though I get a "60hz 480i" output to my LCD (interpreted as 480i) i still swapped the crystal too! I'm ready to throw the board in the bin!

As for the CIC chip. I initially tried with out one installed, then I tried a spare PAL one with leg 4 lifted
Then i tried the "2 wire" mods with the PAL CIC installed to bypass it
Then I tried a few other "2 wire" mods with the CIC uninstalled

Then the NULL CIC hack as per the OpenTendo creators guide - (if you see this you have a diode array installed backwards on your photo)

Then i tried the bypass suggested here: http://forums.nesdev.com/viewtopic.php? ... 0&start=60
With the 100k resistor - this should turn it into basically a top loader.

Then I even tried removing the CIC chip in a worn out game I have - which only prevents it from working on my RGB opentendo system as I didn't want to mess with the CIC in this and leave it stock.

So that's it. I've even gone through *nearly* every 72 pin connection with continuity on my meter to make sure JLCPCB didn't send me a dud board.

Any suggestions?!
Last edited by VajskiDs on Mon Oct 12, 2020 6:57 pm, edited 1 time in total.

lidnariq
Posts: 9777
Joined: Sun Apr 13, 2008 11:12 am
Location: Seattle

Re: Other reasons for White or solid colour screen

Post by lidnariq » Mon Oct 12, 2020 10:16 am

VajskiDs wrote:
Mon Oct 12, 2020 6:30 am
The famicom PPU mainly gives a green screen - sometimes dark blue/ sometimes light blue - switching on and off rapidly its orange/yellow
The famicom AV PPU is sometimes grey - but often green as well - or yellow - I think it's irrelevant

The US side loader revisions PPU only gives a white screen - from what I see on US videos on NES fixes this is completely normal.
[...]
initially i figured it was the CPU but I've tried 3 NTSC ones now. I've tried 3 NTSC PPU's. I've tried swapping WRAM 3 times.
This symptom almost always means that the CPU never gets as far as initializing the PPU.

Left to its own devices, just given power and the correct clock, the PPU will generate a valid 240p (or 288p) signal that's a solid color. No RAM, latch, or cart needed. You can put it in a breadboard (although you'll need a suitable amplifier for the video output to be capable of driving a TV).

So my advice is to make sure the CPU is working. This means:
1- Is M2 (pin 31) oscillating?
2- Is A0 (pin 4) oscillating?

A logic tester is adequate, but better testing equipment may be easier.

VajskiDs
Posts: 16
Joined: Mon Oct 12, 2020 6:07 am

Re: Other reasons for White or solid colour screen

Post by VajskiDs » Mon Oct 12, 2020 5:50 pm

My little pocket oscilloscope reads all legs alive. I’m a learner with these but I get a nice clean waveform on all legs apparently 50hz (well yeah matches the power pack) - and all the CPUs I’ve used warm up a little over time. But still a dead NES board

lidnariq
Posts: 9777
Joined: Sun Apr 13, 2008 11:12 am
Location: Seattle

Re: Other reasons for White or solid colour screen

Post by lidnariq » Mon Oct 12, 2020 7:10 pm

Nope, 50Hz won't do. You should see 1.8MHz on M2, and something slower - between 300 and 900kHz - on A0.

50Hz means your oscilloscope isn't correctly referenced to ground on the NES mainboard.

VajskiDs
Posts: 16
Joined: Mon Oct 12, 2020 6:07 am

Re: Other reasons for White or solid colour screen

Post by VajskiDs » Tue Oct 13, 2020 12:00 am

It gets confusing for me with the settings it has
You are correct I wasn’t grounding at the time

I only purchased it to read VPP levels for video circuits I was prototyping which are pretty well finalised now

Would you know what settings to use so it’s not going haywire? I can change to MHz no worries, and the time frame of reading updates but all the other settings get a bit confusing I will have another check

VajskiDs
Posts: 16
Joined: Mon Oct 12, 2020 6:07 am

Re: Other reasons for White or solid colour screen

Post by VajskiDs » Tue Oct 13, 2020 12:08 am

I had a play with settings

Around 6 to 7khz on pin 4 but this seems to drop to 4 to 5khz after a minute of being on

Tried to read 31
This o scope has a mind of its own it switched itself of and now I’m getting nothing on either pin gimme some time.
Ok so pin 4 has dropped to 2.5khz after a reset

It seems to stop showing any activity once fully warmed up until a reset

A full power off and on its back to 7 to 8khz
Pin 31 is dead

I guess I can open the board in kicad and follow to the components of the pins you referenced. Sockets definitely in fine and traced continuity from cart pins to most CPU legs

VajskiDs
Posts: 16
Joined: Mon Oct 12, 2020 6:07 am

Re: Other reasons for White or solid colour screen

Post by VajskiDs » Tue Oct 13, 2020 12:21 am

M2 goes to pin 14 of the 74HC139 and other than that just the CPU pin 31 you referenced and to the cart slot

I actually changed this to a brand new IC as originally used an old one

lidnariq
Posts: 9777
Joined: Sun Apr 13, 2008 11:12 am
Location: Seattle

Re: Other reasons for White or solid colour screen

Post by lidnariq » Tue Oct 13, 2020 10:14 am

M2 / pin 31 should always be in the range of 1.6 to 1.8MHz. If it's not, either its /RESET pin is not being pulled high / allowed to float high, or the CPU is not receiving 21.5MHz from the clock, or the CPU is broken.

What oscilloscope do you have? I might be able to give you better instructions if I know what.

VajskiDs
Posts: 16
Joined: Mon Oct 12, 2020 6:07 am

Re: Other reasons for White or solid colour screen

Post by VajskiDs » Tue Oct 13, 2020 4:09 pm

Thanks I’ll take another look. It’s a DSO112A - I’ll get something more substantial down the track but it’s served it’s main purpose well

lidnariq
Posts: 9777
Joined: Sun Apr 13, 2008 11:12 am
Location: Seattle

Re: Other reasons for White or solid colour screen

Post by lidnariq » Tue Oct 13, 2020 4:32 pm

Ok, so DSO112A has an analog bandwidth of 2MHz and runs at 5MSa/s. That ought to be plenty for this purpose.

The 1.8MHz signal on pin 31 and the 300-900kHz signal on pin 4 should be around 4Vpp; set "Select sensitivity and couple" to 1V and DC.
set "Select time base" to 1µs.

Pin 31 should be high and low roughly twice per division.
Pin 4 should high or low every 0.5-to-2.5 divisions.

VajskiDs
Posts: 16
Joined: Mon Oct 12, 2020 6:07 am

Re: Other reasons for White or solid colour screen

Post by VajskiDs » Tue Oct 13, 2020 5:46 pm

Thanks so much
So it looks like I replaced the component above and below a misplaced cap- but this wasn’t the issue

I’ve put a 10nf (the most common value on the board) in place of the 100nf in the clock circuit. This has been rectified but hasn’t fixed the issue. I did more comparisons to my working pal board and I definitely have no clock at the CPU.

I then stumbled across this after googling no clock stuff

viewtopic.php?f=9&t=20406

This got me wondering which 150k resistor he was referring to as there’s none on the OpenTendo board. I have a couple of these CPU-11 boards and found they are indeed populated with a 150k resistor in the clock circuit. Well OpenTendo for some reason has 1.2k here- not sure what’s going on here ! Doesn’t affect my working pal one done prior.

Just to be exact here his schematic shows 2x pull downs, one either side of the 220pf cap.

He shows 150k+1.2k
OpenTendo is 1.2k and a 510r
Original board is 150k and a 510r (also on 08/10 revisions I also have handy)

OpenTendo should be based off the 11 revision as that’s what on the schematic
Last edited by VajskiDs on Tue Oct 13, 2020 6:15 pm, edited 4 times in total.

VajskiDs
Posts: 16
Joined: Mon Oct 12, 2020 6:07 am

Re: Other reasons for White or solid colour screen

Post by VajskiDs » Tue Oct 13, 2020 5:48 pm

I will check those settings out later after it’s charged a bit- but at least comparing those pins you mentioned has shown a missing clock at CPU. Maybe the 2021 transistor on top ? Only replaced the bottom one...

VajskiDs
Posts: 16
Joined: Mon Oct 12, 2020 6:07 am

Re: Other reasons for White or solid colour screen

Post by VajskiDs » Tue Oct 13, 2020 6:24 pm

Ok with those settings pin 4 is dead until a reset then seems to correspond with what you mentioned... only after a reset though

lidnariq
Posts: 9777
Joined: Sun Apr 13, 2008 11:12 am
Location: Seattle

Re: Other reasons for White or solid colour screen

Post by lidnariq » Tue Oct 13, 2020 6:59 pm

Pin 4 should keep on changing as long as the CPU is executing code. If the CPU ever fetches one of the "bad" instructions that cause it to stop, it'll stay low or high at that point until reset.

Are these tests with a cart in?

VajskiDs
Posts: 16
Joined: Mon Oct 12, 2020 6:07 am

Re: Other reasons for White or solid colour screen

Post by VajskiDs » Tue Oct 13, 2020 7:58 pm

Yes, I made some notes that I figured were irrelevant due to the settings I used before you informed me. It was just differences on the A0 line between the working and non working machines and with and without a cart. It’s different when a cart is not plugged in.

Still probably all irrelevant as I found a setting that Gave a massive and clear reading on the pin 29 clock leg on my working machine and it’s not there on my OpenTendo board. The CPU/s seem to be floating ducks on the mainboard- all 3!

Basically I’ve given up but I’m left so confused

My pal machine worked straight away. No messing about.

I mean fair enough I put a 10nf in a 100nf footprint on this one, found a large discrepancy in pull down resistor between OpenTendo and the original board

and originally I didn’t think the 6.8k resistor bus had an orientation- but it does - so i put a new one in.

But that’s not system breaking stuff!
I’ve tried 3 CPU’s and ones out of a tested AV famicom (at least tested before I purchased it, sold as working and its mint condition)

The lockout circuit bypass has being tested and proven as working as well and this wouldn’t kill the CPU.

Basically I’m baffled and left scratching my head wondering what the hell is wrong with this system...

It’s no big deal in the sense of not having a working NTSC machine, it’s just that I’ve tried 3 CPUs, I’m using a known perfect cart connector and I’ve replaced so many components and checked continuity between so many points with my meter. JLCPCB is not at fault. I feel I’m not at fault and I feel there’s no way I have 3 dud CPUs.

Post Reply