Cart Wierdness

[GadgetUK]

Hi,

I've got a really really strange problem where I get some artifacts on the display with one of those dodgy multicarts. And yes, I know they are crap quality and dont last etc etc. But that said I've got a very strange problem that I am trying to understand.

The artifacts disappear completely when I 'touch' the main board silkscreen in one particular spot, near what I think is the video RAM. Now the obvious suggestions are:-

1) Board fault, proven its not the MVS by testing 2 other working boards - exact same behaviour.

2) PSU, used 2 completely different PSU's. One PC PSU and one dedicated 5v and both the same.

3) Cart slots - as per 1, i've used 3 different MVS boards all with their own cart slots, all cleaned up just the same.

4) The cart - i've cleaned the cart edges very throughly and its making a perfect connection. I've also gone as far as resoldering all the NVROMs, the CPLD and the CPU - no effect at all.

What is going on here, is this some form of capacitive reactance from me hovering my finger over that particular part of the board?

And the next question, if it is, why the hell would that be?

 

[xsq]

woha, your cart is haunted. Could you make a short vid?

 

[GadgetUK]

Yeah, i've got one vid up at the moment (see signature) - that was filmed last night, sorry its dark... That video was filmed before my endevours at lunch where I discovered after futher investigation that its not lifting it that is making the artifacts stop and start - it seems to be some sort of capacitive reactance or something and that i've now narrowed it down (2nd video uploaded later) to a 1/2 centimetre spot on the board, next to the palette RAM chips I think (left side, an inch or something down from the cart connector and LO ROM, in the middle of the 4 chips positioned there. I can move finger over board and everything is perfect, move finger away and artifacts come back, and occasionally it will reset or even get stuck in the black screen click watchdog, move my finger back and it then boots if in watchdog and there are no problems. I can litterally keep my finger hovering over that part of the board and everything with that cart is perfect until I remove my finger.

I am wondering if I could substitute my finger for something else lol - anyone got any dead fingers they dont need?

 

[GadgetUK]

By the way, it would be easy to just bin this cart but I feel that there's something we can all learn from this. I've heard a lot in the past about people suffering watchdog resets etc and I believe there's a fundemental design issue affecting some of these carts that we might be able fix. ie. carts that dont work at all for people might be salvageable. I think ultimately that a decoupling cap in the right place could resolve this, but I am just not sure where just yet.

I've discovered today (obvious to many but I didnt know until now) that the CPLDs, CPU and NVPROM chips on these boards work off 3.3v. 4 diodes are used to drop the voltage, 2 on top board, 2 on the bottom board. Each obviously passes the 5v through reducing it by 0.7v per diode. ~3.6v provided to the chips.

Now this could be part of the problem in that I dont know how these CPLDs work, I am guessing the logic level isnt the same as say a 27C800 or 27C160 etc, maybe 2.5 to 3v logic level, which could make it susceptible to a reactance type problem that I am experiencing. That said, the 161 in 1 has almost identical parts and isnt suffering from the problem the 120 in 1 is.

EDIT: It's also worth pointing out it's not just related to Top Hunter shown in that first video. It's doing it on ALL games on the cart. And if left without a finger it will go into a watchdog black screen loop with ticking noise etc, leading you to believe the cart is knackered completely. One finger later and everthing is working again lol.

I would have loved it if the problem wasnt there on another MVS board, because at least I could then focus on a fix for my MVS, but it doesnt look like theres anything wrong with the MVS as it's doing the exact same thing on 3 different boards with different PSU's, different video cables etc.

 

[Asure]

Ah... Electrons. So :

- Is the power supply connected to a grounded plug?
- If not, did you try grounding the MVS yet? (or try on a grounded mvs?)

 

[GadgetUK]

Good point... You know I am not sure if it is. I will check it. It's been a long time but am I right in thinking it should be safe to connect the - connection of the PSU to earth?

 

[SNKorSWM]

Put it in something like a metal container or mesh and see if there's anything nearby that could be causing interference.

 

[pulstar]

Have you tried replacing the plastic cover thingy (not sure what else to call it) and the feet and trying the cartridge like that? It seems to me to be some kind of ground issue, probably because of the manufacturing process of that particular cart or maybe just a defect in that model or something.

 

[shadowkn55]

Put it in something like a metal container or mesh and see if there's anything nearby that could be causing interference.

I'd go with the mesh. It's supposed to act like a Faraday cage and block EMI.

 

[Asure]

Also, those multicarts are probably not FCC approved.

But, it may also be another device causing this by radiating into the back of your mvs.

 

[ne7]

It could also be that something has been left on the silkscreen from mfg process - have you tried giving the board surface a good cleaning? Oh and another thing to check is the diodes if they are dropping the voltage... best o'luck \o_

 

[GadgetUK]

Thanks for all the suggestions! I've earthed it (makes cart watchdog reset), lifted off floor on different surfaces etc, nothing changes except that when -ve is earthed it watchdogs with that cart (on both of my MVS boards), so knowing that my original carts are fine, and the 161 in 1 is fine I've got to assume that the CPLDs or NVPROM chips are causing this. I've ordered another 120 in 1 and will experiment swapping the prog and char boards around to isolate which board it is, I will then measure from ground to each pin and compare board to board to see if I can work which IC might be causing it.

When I first got an MVS and discovered the multicarts I though they were great... I am firmly of the opinion now that they suck and are over priced pieces of junk.

If they made them quality they could net £200 per card...

Edit: I might get my scope onto the cart edge to see if I can see any noise etc.

 

[GadgetUK]

Thanks for the suggestions / advice guys!

I 'think' i've sorted this... I am not 100% sure yet, so don't take this comment as a case closed but there were 2 things i've just found / done.

1) There was a beige coloured particle of 'something' inbetween 2 pins on one of the NVPROM chips (C chips), which I removed.
2) At manufacture there was a scratch across 5 or more traces, which I had previously inspected and looked OK. I ran my nail over this scratch today and it did feel like something came off it - possibly the interaction there has joined a broken trace - they are so very small I cannot even see them through magnifying glass.

Following that it has been working for a good hour and been taken in and out of the MVS several times, and re-assembled and put back in and its still working...

Time will tell!

What else was interesting is that lifting it off the floor whilst holding it made it work but as soon as I put it on ANY surface it stopped working again. Earthing the -ve connection made it mostly watchdog reset - so it was actually worse earthed, which ties in with why in the 2nd MVS using a grounded PC PSU it would mostly watchdog in that system.

 

[GadgetUK]

It could also be that something has been left on the silkscreen from mfg process - have you tried giving the board surface a good cleaning? Oh and another thing to check is the diodes if they are dropping the voltage... best o'luck \o_

Thanks, you might be right - certainly regards the scratched trace I mention in my post above.

Regards the crappy diodes, I am going to have a go at fitting some SMT 3.3v regulators, they are rated at 1 amp which is more than sufficient for the cart. It doesnt need doing as it looks like i've resolved the problem for now but I dont like the idea of the only thing standing between the 5v line and a correct 3.3v source as 2 crappy diodes. They arent even providing what the chips on the board really need - ie. 3.3v, it ends up 3.6v ish which is a little high in my mind.