My SNES troubleshooting workflow for “black screen of death” systems

SPOILER ALERT: They ALL have dead CPUs.

SPOILER ALERT: They ALL have dead CPUs.

I have dozens of these black screen systems. I have a basic flowchart for troubleshooting all SNESes in my head. It’s simple and progresses from the easiest fixes to the most difficult. But I’ll spoil it and reveal ahead of time that they almost all have dead CPUs that need to be replaced. It’s never the easy stuff.

The first steps are disassembly and cleaning the housing. Of course, the latter doesn’t fix anything, but time is money, and by cleaning the housing first I can leave it to dry while I work on the motherboard itself.

Step one is to clean the cartridge slot with a toothbrush and 91% alcohol. 99% is better but it’s more expensive and harder to find. 91% is good enough. Most folks recommend wrapping a credit card in a cloth, dipping it in alcohol, and then inserting and removing it repeatedly to clean the cartridge slot, but I’ve never understood this method at all. I don’t understand how that could possibly be very effective. A toothbrush seems like the obvious answer to me. For one thing, with the credit card and cloth method you’re only cleaning the removable top part of the connector, which is completely pointless if you don’t also clean the pins beneath it that are soldered directly to the board. For RGB, APU, and 1CHIP models I suppose the credit card method may be all right, since you can’t just lift the connector off on those. But a toothbrush just seems like a far better method to me. You need to use serious elbow grease when cleaning these things, and there’s no way to put the kind of force necessary behind your scrubbing if you’re just using a credit card wrapped in cloth.

Of course, rather than clean it, it’s faster to just grab a known working cartridge slot connector to test the system quickly. Keeping one handy saves time.

It's pointless only cleaning the removable top part of the cartridge connector if you don't also clean beneath. How would you possibly clean a system like this with the credit card method?

It’s pointless only cleaning the removable top part of the cartridge connector if you don’t also clean beneath. How on earth could you possibly clean a system like this with just the credit card method?

Cleaning beneath the connector is very important. If someone spilled something on the console long ago, you have to consider where gravity would have taken it. Soda spilled on top of the console wouldn’t have just sat on the top removable piece of the connector for all those years, so it’s rare to find much corrosion on the removable part. Any liquid spilled on top of the system would have run down through the connector and settled on the pins beneath. That’s why you tend to find rust and corrosion on these pins. Sometimes they look all green, like the Statue of Liberty. I scrub these with a brass brush to remove any corrosion, then clean them off with a different toothbrush and some cotton swabs. Deoxit is also good to use here.

Knowing how gravity works allows you to predict this before you even fully disassemble the console.

Knowing how gravity works means this kind of result is predictable.

At this point you can test the system. If it works, great. You’re done. But the premise here is that these simple things usually don’t work. Most consoles don’t have so much corrosion, so, while cleaning is always a good idea for sanitary reasons, it rarely actually fixes anything. Ordinary dust and dirt won’t stop the console from reading games. So let’s move on to the next step in my flowchart.

If you hold the reset button on a working console while powering it on with a game inserted, you get a black screen until you release the button. The idea here is that if the reset button is very dirty (again, think spilled soda) then it can be stuck in the activated position, causing the same symptoms as if it were actually being held down. I’ve never actually seen this personally, but it’s an explanation that makes a lot of sense to me, so it has a place in my flowchart and I always first try cleaning the reset button with a toothbrush and a bit of alcohol. If it seems sticky I temporarily desolder and remove it just for testing. It never turns out to be the culprit, but it’s pretty quick to remove so it’s not much of a waste of time.

C62 and the reset button.

C62 and the reset button.

C62 is a small 2.2µF capacitor right above the CIC chip, near the reset button. I’m told that if this is bad it has basically the same effect as holding the reset button down, resulting in a black screen on all games. Again, I’ve never actually come across a system where this had happened, but it’s an easy thing to replace, so I sometimes try replacing it if it looks funny. It has never fixed anything for me though.

I’ve seen this next thing once and only once. Since it happened once though, it does have a place in my flowchart, since I suppose I could come across it again.

I flip the board upside down and do a visual and tactile check of the solder joints on the bottom of the cartridge slot. When I say “tactile” what I really mean is that I press on them one at a time with my fingers to see if they’re cracked. It doesn’t hurt to quickly reflow them all. It almost never fixes anything, but it can help you eliminate the cartridge slot completely as a possibility. If you are an insane person you can get a multimeter, connect the top removable connector, and then check each pin on the bottom of the board for continuity with the corresponding pin inside the removable connector. This is a colossal waste of time though and you should only do this if you are a masochist. Even with a third hand tool, you’ll nudge the board constantly, your hand will slip, you’ll drop one of the probes, you’ll lose count of which pin you were at and have to start over and you will want to off yourself in no time. Besides, you really don’t need to test. It’s never the cartridge slot that causes these issues. When it comes to the cartridge slot, if everything looks good, it is good. The one time I had a system where this sort of thing was an issue, the solder joint was so badly cracked on the underside of the board I could wiggle it with my finger. If something is wrong it will be obvious.

Next is to check for broken traces. There are no shortcuts here. You just need a jeweler’s loupe and a lot of time on your hands. What I’ve learned from experience though is that you shouldn’t waste your time on this step unless you have good reason to suspect there will, in fact, be some broken traces. Basically that means if you have a system that had liquid damage or was in a very damp, wet, humid, or dirty environment you may want to spend some time looking at it carefully under magnification. So if you open it up and find lots of rust or dead bugs, it may actually have some broken traces. But if you open it and it’s nice and clean, don’t waste your time. Unless, of course, someone else worked on it prior to you. If that’s the case, you should absolutely check for broken traces, scratches, lifted solder pads, and that sort of thing, since you never know what the last guy may have done to it.

If the system is an SHVC model, yes, you can try swapping out the sound module. Some games will give a black screen if the sound module is disconnected or bad. But many games actually load to the first screen and freeze when the sound module is bad or disconnected, so if you’re using a game like that and you get a black screen, don’t waste your time, since it’s not the sound module. An Everdrive will load and display the contents of your SD card even with a bad/disconnected sound module. If you try to run a ROM, it’ll freeze.

An Everdrive, by the way, is something that can be very helpful when you’re not quite sure of the extent of the problem. Some systems may give a black screen on most games, but display garbled graphics on another. Still others give a black screen on 9 out of 10 games but might play one specific game just fine. Those consoles may have hope. To help understand the extent of the problem a bit better I see if it’ll read an Everdrive. If it reads the Everdrive and loads the burn-in test rom, I run it and see what it says. These black screen systems may not read any retail games, but sometimes they do read the Everdrive, though it doesn’t always actually load up fully. It often crashes when trying to display the contents of the SD card. But if it does load and I can get the burn-in test rom to run, it usually is very straightforward and simply says, “CPU —— FAIL”. That’s about as clear-cut an answer as it gets. Almost all the failures are CPU-related, but occasionally you do see some VRAM problems. Those are nice since you can easily grab the VRAM from another console. There’s also plenty of space between the pins so soldering them in is easy. But I’ve only seen bad VRAM two or three times and those had all been worked on previously. I don’t think the VRAM is typically prone to failure. Normally it’s a CPU problem.

At this point the only thing left is to replace the CPU. That’s not as hard as it sounds if you have hot air rework equipment. It’s very easy to remove the old CPU, but you do need to be moderately good at soldering to put in the new one. I usually end up with a few solder bridges at the end that need to be fixed. The hardest part though is locating a good CPU. The reason I have dozens of dead black screen SNES boards is because I have no good CPUs to put in them. Most, I’m sure, would work fine with a new CPU, but the trouble is that there’s nowhere to get them. I found a few IC dealers online that claim to have a small quantity in stock, but they are asking such high prices it would actually be cheaper to buy working SNES consoles for the CPUs than to buy from those bloodsuckers. When I get really badly water-damaged or otherwise screwed-up boards I take the CPUs. Sometimes they’re bad, too. But occasionally they’re good and I can revive one dead system from my stack. It always feels good. Plus I end up producing some unique SNES consoles like 1990 SHVC boards equipped with the later (and much more resilient) “S-CPU B”, which was normally only found in the GPM-02, RGB, and APU motherboard revisions.

Don't worry. I didn't cannibalize a working RGB board. It had been eaten up by roaches and was totally beyond repair with broken traces all over and the solder mask peeling up on the back.

Don’t worry. I didn’t cannibalize a working RGB board. It had been eaten up by roaches and was totally beyond repair with lots of broken traces and the solder mask peeling up all over the place.

Maggots? In my SNES?

It's more likely than you think.

It’s more likely than you think.

No, don’t worry, this isn’t my personal console. I got this one on eBay. Mine, of course, is immaculate, as you would expect. Incidentally, buy only from me or else this is the kind of thing that’ll show up at your doorstep after shopping online.

But wouldn’t you know it? This disgusting pile of filth and disease worked fine after I spent several hours washing it. I had to replace the capacitors, but otherwise it was fine. It’s aggravating that horrific-looking systems like this survive despite such neglect and abuse and yet so many well-cared for pristine-looking systems stored in safe and clean conditions by responsible, civilized people just die for no good reason.

DSCF2842

DSCF2843

DSCF2851

DSCF2848

DSCF2847

DSCF2846

DSCF2845

I really do wonder if I’m blacklisted by the USPS

Today I had the 5th consecutive parcel I’ve attempted to have delivered go straight to the post office instead of to my home address. None of them have had signature confirmation. They’ve all been either priority mail or parcel post. Most had DC. There was one that didn’t. Either the USPS has blacklisted my address or they’ve put a new carrier on my route who is consistently doing his job terribly wrong.

It really adds insult to injury when, in addition to having to travel every single day to pick up junk mail and letters, you also have to travel to the post office to pick up parcels. There’s no reason to pay extra for signature confirmation if the recipient has to travel to the post office in person and show I.D. in order to get a parcel that doesn’t even have DC on it.

Maybe I’ll stop defending the USPS now. In the past 2 months they’ve destroyed one parcel, sent 3 to the wrong city, and decided to send these 5 parcels to the post office for pickup when they should have been delivered by a carrier.

Stupid USPS keeps wasting my time

Normally I’m the first person to defend the USPS. I’ve generally had good experiences with them. I’ve only had a few packages permanently lost in my entire life. This time, however, I must complain.

When I send and receive parcels I nearly always use delivery confirmation. The way this works is that the carrier scans the package using a handheld scanner when it’s delivered. This way the seller can confirm that the item arrived at the destination. This is useful in situations where the buyer claims the item was never received and asks for a refund or a re-shipment. With DC you can just point to the confirmation of delivery on the USPS website and the would-be scam artist gets shut up.

Signature confirmation is like a much more extreme version of delivery confirmation for more paranoid sellers. It’s more or less the same as what happens with any UPS delivery. The carrier must give the parcel to a human being at the address and get the signature of the human being. If there’s no human being to accept and sign for the package, it doesn’t get delivered. With DC it’s possible that someone other than the intended recipient will steal the package (eg. it’s left by the mailbox and the neighbor steals it before the addressee arrives home). Signature confirmation is a pain in the neck if you’re the recipient for obvious reasons: you must physically be at the address and answer the doorbell when the mail carrier arrives. When you’re like me and your mail carrier arrives anywhere between 1:30 pm and 5:30 pm depending on the day of the week, it’s a terrible inconvenience to wait around all day.

Registered mail is about equally inconvenient as non-registered signature confirmation parcels if you’re the recipient. It’s better for the seller though because supposedly registered mail is handled in a more secure manner than non-registered mail while in transit. I don’t know the details, but registered mail automatically requires a signature so for the recipient it’s more or less the same hassle as non-registered mail with signature confirmation, regardless of class.

Anyway, I always try to use only ordinary delivery confirmation. Regardless of whether we’re talking about media mail, parcel post, first-class mail or priority mail, I always try for just delivery confirmation when ordering packages. Ordinarily my mail carrier just leaves the DC parcels outside the mail box and I pick them up when I get home. The last three parcels I’ve ordered with DC, however, have all gone straight to the post office and I’ve had to go pick them up. I find that very annoying. DC is DC. There’s no requirement, as far as I know, for a human being to physically accept the parcel if it’s just got DC. That’s the point of the handheld scanner, if I’m not mistaken. That’s also why signature confirmation (which ensures a human being accepts the parcel) is more expensive. If DC required a human being to accept the package, there wouldn’t be much use for signature confirmation since it wouldn’t add any protection that DC didn’t already provide other than the name of the specific human being who accepted the package.

Anyway, it’s just annoying that the USPS would suddenly start requiring a human to accept parcels at the exact time that my apartment was destroyed. It’s enough of a pain in the neck having to travel to the remnants of what was once your home to pick up the mail without also having to travel to the post office, wait in line, and show an I.D. every time a piece of mail with delivery confirmation is sent to you. I’m starting to wonder if maybe the USPS started blacklisting addresses of destroyed homes.

In defense of USPS, I do understand how some people might see this as a an upgrade to the DC service, since requiring the addressee to travel to the post office and show identification makes it much more likely that the intended recipient gets the package, rather than some neighbor or even somebody else (such as a family member) who lives at the same address. Of course, it doesn’t stop an impostor from picking the package up using a fake or stolen I.D. card, but it’s still much more secure than leaving the package at the mailbox.

I don’t see it as an upgrade though. That’s because I’m not ordering any sensitive materials. If I were, I’d use registered mail with signature confirmation or maybe FedEx.

Anyway, hopefully this is just bad luck I’ve had on my last three parcels and not an indication of an actual change to the way delivery confirmation works.

I wasted my entire stupid day troubleshooting this stupid Arris modem

So I’m living elsewhere temporarily while my apartment is being demolished. I’ll be here until it’s rebuilt. Who knows when that’ll be. But that’s not the point. The point is that I have Time Warner Cable here and I’ve had the chance to use Road Runner for an extended period of time now. I’ve used Road Runner at friends’ homes several times in the past and had opportunities to do speed tests and the like, but I’ve never had the chance to use it on my own terms, with my own computers and home network equipment. I’m very upset at somebody — I know not whom — about a grievous oversight in the instruction manual for the DOCSIS 3.0 modem/router TWC gave me. The trouble is that I don’t know who to blame. It’s an Arris TG862G. Frankly, I had never even heard of Arris before they gave me this thing. The installer, who mentioned, by the way, that he has FiOS at home, said that the device is both a modem and a router, but that TWC doesn’t let customers change the SSID or the passphrase on the network, so if a customer wants to change that information, he or she must use his or her own router. I didn’t really care, since I do have my own router and I figured I’d just have to live with a suboptimal home network split on two different subnets (192.168.0.xxx on the Arris router and 192.168.1.xxx on my router). I didn’t think I was able to have TWC put the Arris in bridge mode since the TWC connection is on the account of the owner of the place in which I’m living (though nobody else will be using the connection). I figured I’d either do the 2 subnet thing or I’d simply use my own router as a switch and actually use the Arris router as a router. I figured I’d decide once I logged into the web configuration pages on the Arris and saw what features it had. If they compared favorably with my own router I’d just go ahead and use the Arris and use my own router as a switch (I have a real 24-port switch, but it, along with most of my stuff, is in storage until the apartment is fixed)

Anyway, the reason I’m angry is because either Arris wrote bad instructions or TWC made a slightly modified firmware for the Arris modem with an annoying feature. The manual says you can access the web GUI at 192.168.0.1, which of course is similar to most routers. I connected my computer directly to the Arris via Ethernet cable with nothing else connected to the modem but the coax cable and the AC adapter. I set my computer to get an IP address automatically to rule out the possibility that some pre-existing configuration on my computer was causing problems. I typed in the address, waited… and it timed out. After trying all sorts of other combinations (10.0.0.1, 192.168.1.1, 192.168.1.100, 192.168.0.100, etc…) I finally figured out today (2 days of web searching later) that you can only access the web configuration GUI if you unplug the coaxial cable from the Arris modem first. If you have the coaxial cable plugged in and you try to access 192.168.0.1, it’ll just time out. The fact that this isn’t mentioned in the manual is a major oversight which caused me a huge pain in the neck. All I wanted to do was access the port forwarding settings page, which should be the simplest thing in the world but because I lacked this simple bit of information I had to go on a wild goose chase of searching through forums and support pages, none of which actually mentioned this. Hopefully posting this information will save somebody a bit of time configuring his or her Arris cable modem/router in the future. I just wish I knew whether it’s Arris that made this feature or if it was an adjustment that TWC does to the units they send to customers. I know that Comcast uses this same modem for some of their customers, so I’d be interested in learning if they also suffer from this “feature”.

P.S. The connection tests about 35Mbps/5Mbps to test sites in the NY/NJ area.

I’ll be offline for some weeks

This is the only water-related picture I have handy at the moment.

my apartment is flooded by Hurricane Sandy. Most of my hard drives are undamaged because I managed to move them to high shelves, but a few were submerged. Anyway, the damage means the place is not presently safe to live in, though it’s just as well since all the furniture was ruined. Once the place is fixed and I get some furniture I’ll be back online.

These SRO-R updates have utterly ruined the game for both botters and legitimate players alike

My sosun-wielding goldbots have no gold to pick

Others have said it before me, but I feel the need to reiterate how poorly thought out an update this was for the iSRO economy. Servers like Xian and my old home of Babel have players (or at least abandoned accounts) from 2006, filled with items. Those items retained their value for these six years. This is one of the reasons why the supposed database leak was of interest to so many people who didn’t even play; they hoped to be able to harvest items from old, unused accounts and sell them. Equipment and other items are not like pastries; they don’t lose value as they age. At least they shouldn’t. In fact, some, like GDFs and other items that were available only for a limited time, actually increase significantly in value with the passage of time. But now the majority of the items people have been saving in reserve, ready to sell in the event of a rainy day, have been rendered worthless in one fell swoop, like all my SoX 8th degree weapons. No other update has caused a catastrophe of comparable proportions.

Say what you will about Joymax, but none of their updates to the game thus far, not even the much-maligned abolition of the triangular trade conflict system that was the main selling point of the game (it’s called Silkroad Online for fuck’s sake) have done as much damage to as many players as the phase-in of SRO-R features in iSRO. All items below 9th degree have lost their value. Even SOSun 8th degree weapons are being sold for 100 to 200 million gold. You can’t even give Spell Paper away for 30m when it ought to fly off the shelves at 70m on my server. There is no point to the Forgotten World prior to 9th degree. What this update essentially does is transform the entire player experience from level 1 through 70 into an agonizingly long and boring tutorial before they are allowed to play the “real game”. Levels 20 through 70 used to be part of that “real game”. There was competition with other players. There was envy when someone came about with a golden glowing +7 weapon. Not to deny the repetitiveness of the game, but at least back then there was suspense, since you could get a drop at any moment. That was one of the incentives to play legitimately, without the use of bots. Now there’s really no reason at all not to bot and go afk all day.

I made a goldbot party recently in celebration of learning to use the clientless feature in iBot. Now my goldbot party is useless. I have to raise them to level 75 or so and send them into the Donwhang Cave or put them at Niyas in Taklamakan for them to be of any use. I had been counting on getting a good spot fighting Hun Archers or something and letting them all stay there as long as possible with a 9 level gap. Now I can’t do that. All this update will serve to do is drive the last remaining legitimate players to use bots, and force goldbot companies and players with goldbots such as myself to level them up more quickly, leading to armies of SoSun-clad underfarmed lightning nukers who will ks you at every turn and spawn party giants and envies to kill your wizard while you’re afk or playing Skyrim on the other monitor.

pffft…

I’ve dumped Comodo

Haha, I just love this kid.

Comodo firewall is great if you like to micromanage every single process on your computer. Nothing slips past Comodo, seemingly even if you set the firewall to “disabled”. On the one hand, it’s a bit idiot-proof to the extent that it’s hard to accidentally disable the whole program by some mistake. On the other hand, I’m rarely able to get it to do exactly what I want. For example, just to get a simple P2P program working required making 4 non-intuitive application rules (one for port 80 for tracker announces, one for TCP in/out, one for UDP in/out, one to deny everything else) and a global rule. I also had to make these 5 new rules top priority or else they didn’t take effect. There’s a section for managing programs on an application-by-application basis so you’d think I could just point the firewall to P2P_program.exe and tell it something along the lines of “trust everything P2P_program tries to do; it’s safe” but that doesn’t work. Even the “allow all” setting doesn’t really allow all. You have to dig deep in the forums to figure out how to allow a single program access through the firewall on a single port or port range, as though this were some sort of uncommon task that only superusers need to know how to do. It’s not though. Whenever you install a program that needs relatively unfettered access to your Internet connection you need to let it through your firewall but there’s so much prerequisite knowledge needed to be able to do that in Comodo firewall that it can be extremely frustrating and time consuming. I should be able to choose a program, specify some ports and a direction (ie. IN/OUT) and be done with it. I shouldn’t have to mess with global rules, stealth port settings, and application-specific rules or be familiar with some arcane settings buried in an obscure “advanced” menu.

Comodo is not a bad firewall. Of course, I know nothing really about computer security, so even if it were a bad firewall I wouldn’t be qualified to judge. What I can say though is that Comodo makes it so difficult to actually allow a program to run normally that I may as well unplug my Ethernet cable for 100% security. The program is so secure that it makes it a real challenge to weaken it even slightly. A more user friendly approach is what I’m looking for, even if it may put me in a less secure environment than Comodo. I’d rather be put in a bit of danger by human error resulting from my own foolishness than do things safely but painstakingly with Comodo.

My other problem with Comodo though is cfp.exe, which unexpectedly jumps to 25% CPU utilization occasionally. I’ve read that this can be caused by conflict with another firewall or AV program but I don’t have another firewall and I’ve tried uninstalling my AV program Avast, with attention to meticulous detail including using CCleaner and the official Avast manual removal tool. I’ve also specified the entire “COMODO” directory to be excluded when Avast does virus scans. In the Defense+ settings of Comodo I’ve likewise specified all the Avast folders to be excluded. Neither of those ideas helped. The high CPU usage is not a reproducible reaction. It’s totally unpredictable and has no upside, which is what makes it so aggravating, unlike the strictness of the rules system which at least has the benefit of preventing me from doing something stupid. Sometimes I’ll simply be browsing the web without any P2P program open at all and cfp.exe will cripple my computer, sometimes for just a few minutes, sometimes for as much as a half an hour.

I didn’t want to use a commercial firewall as a replacement. I did a google search for “lightweight firewall” and came across PrivateFirewall. I’ve only been using it for about 30 minutes now but I already know how to make global rules, new application rules, how to trust an application completely, and how to make application rules that allow a program through the firewall only under certain conditions such as only via specific ports or only in one direction or the other. These are all things that were not satisfactorily covered in the Comodo help menus which meant I had to read countless threads on the Comodo forums before I could figure out how to do them and even then, my rules did not always work properly because I had missed some other minor setting that would render them ineffective.

I also like the PrivateFirewall taskbar icon of a friendly police officer.