Thanks again guys.

kitsch wrote:

i'm not sure if you've tried this, but put something in the second banks, and see if it helps.  switching to them, when there is nothing to load, may be part of the issue in why things are seemingly becoming stuck after having accessed the empty banks.

Unfortunately my troubleshooting options are limited with respect to reformatting etc.; being an OSX user without Parallels or Windows (yet) I had to use a loaner PC to install the file management utility, cartridge drivers, etc., and I don't have access to that machine any longer. :\  Will probably have access to other PCs in the future so as soon as I'm able I'll try to start over.

kitsch wrote:

on the back, bottom-right, of the cartridge (melted into the plastic) is a date code.  something like 1005, 1006A, 0912, whatever...

what is the date code?  the latest PCB revision (which is what the date codes partially reflect, the letter is for more than one batch in the same revision) is 1006A, which are functioning beautifully.  it must be a model EMS is comfortable with as well, because they've been using it the last 200 we got, and usually they cycle to a new one quicker than this...

the date code may reveal a lot.

Unfortunately the one I've got on-hand is marked "1006A." Both carts actually belong to my roommate and he's got the other one with him at the moment. I'll check it when he's back.

In the "Does This Mean I Fucked Something Up" category -- one anomalous thing I did notice during the formatting and flashing process (which did, at the time, involve several iterations) was that when using the formatted & ROM-flashed cartridge, the descending Nintendo logo was mangled. Worryingly, it was mangled consistently (i.e. the manner of glitching was consistent every time, not variable the way you'd expect if it were just a bad contact connection etc.). So maybe something I did during formatting punched these carts in the brain. Possible?

In my experience that mangled logo is a result of having no software on a given page.  Mine does the same thing when unloaded.

Unfortunately my troubleshooting options are limited with respect to reformatting etc.; being an OSX user without Parallels or Windows (yet) I had to use a loaner PC to install the file management utility, cartridge drivers, etc., and I don't have access to that machine any longer. :\  Will probably have access to other PCs in the future so as soon as I'm able I'll try to start over.

Unfortunately the one I've got on-hand is marked "1006A." Both carts actually belong to my roommate and he's got the other one with him at the moment. I'll check it when he's back.

In the "Does This Mean I Fucked Something Up" category -- one anomalous thing I did notice during the formatting and flashing process (which did, at the time, involve several iterations) was that when using the formatted & ROM-flashed cartridge, the descending Nintendo logo was mangled. Worryingly, it was mangled consistently (i.e. the manner of glitching was consistent every time, not variable the way you'd expect if it were just a bad contact connection etc.). So maybe something I did during formatting punched these carts in the brain. Possible?

i have an idea!!!!

a big thing i noticed about the 1006A batch of carts is that C3 (the big orangish one, just to be technical) is being left off the board.  I contacted Melody about this, and she said indeed it was intentional.  i was assuming it was a decoupling cap, and my worry as it was mistakenly left out of assembly.

removing C3 shouldn't be hard if you've got a soldering iron handy...

with that being said, i never actually asked what C3 was for, and haven't bothered to follow the traces to ask.  1006A works, C3 is gone, i guess that cap was an issue in the design.

and, since we are on the issue of caps...

possibly

its worth a shot.

nanovoice demo, muddyGB, GB electric drum...  all good apps to try which won't mess with your sav at all...

also, can you check something on the cartridge for me, as it might help...

on the back, bottom-right, of the cartridge (melted into the plastic) is a date code.  something like 1005, 1006A, 0912, whatever...

what is the date code?  the latest PCB revision (which is what the date codes partially reflect, the letter is for more than one batch in the same revision) is 1006A, which are functioning beautifully.  it must be a model EMS is comfortable with as well, because they've been using it the last 200 we got, and usually they cycle to a new one quicker than this...

the date code may reveal a lot. 

as well as putting a ROM in bank 2 of both carts. 

then again, they may just be faulty.  which would be very odd if they are 1006A, not totally weird for some of the 2009 dates, but still odd you'd get two with the same issue (apparently).

you should contact the person from whence they come, and as long as they are in good business terms with EMS, they should be able to work something out (EMS products are guaranteed by EMS, and the buyer (your cartridge seller) can work out an exchange even if your warranty with them personally may be up.  if they are willing, that is.  EMS, after all, wants to keep their wholesale purchasers happy and the person you bought it from can make a personal repair request on your behalf with no issue).

@ nitro: i'll send you a cart to fool around with, my pleasure.  i'll be sending you a little something soon anyways (surprise!).

EDIT: it occurs to me that a faulty capacitor would cause both kinds of behaviour in 100% of cases. The cap holds a charge that, when seen by the bank switching circuit, send it into bank 2. This cap is not supposed to hold a charge longer than a very short while, allowing for switching back to bank 1 when not flicking the power on and off. But a capacitor which holds too much charge will not allow going back into bank 1, while a cap which won't hold a charge will not allow switching into bank 2. My guess is that they used the cheapest possible electrolytic caps, allowing for manufacturing variance which would explain this kind of behavior.

My guess is that one could just bypass the bank-switching circuit, which uses a capacitor based circuit to determine time since last power up. with a toggle switch, thus keeping both banks but preventing power-time cycle.

That sounds like a good option.

Turn on, bank 1. Turn off, then quickly back on, bank 2. Turn off, wait for, say, 10 seconds or so, turn back on, bank 1. Reset by pressing start-select-a-b, and it should only reset the selected game, not return to the menu.

Seriously, put LSDJ or something in both banks. Empty uninitialized banks would totally bork something, I should know.

On further inspection it seems like the cart, if left sitting out for around 15 minutes or more, will then successfully revert to bank 1 on the next power-up. But power-cycle again, and then the Trapped In Bank 2 problem starts all over again.

A second cartridge seems to have the opposite problem -- will boot into bank 1 every time (no light on the back of the cart), and will never boot into bank 2. Pragmatically speaking this is less of a problem for me because, really, fuck a bank 2 -- but the basic fact that the unit's not working the way it's theoretically supposed to is still worrying.

So -- I'm not crazy, right? These two carts are not behaving the way they're supposed to? In other words, there's not some arcane procedural omission on my part (forgot to recite the incantation, etc.)?