nesdev.org

I decided instead of lifting the pin just to desolder the entire 5b and cut the trace underneath where it sits. My experience with lifting pins is that they tend to break when being lifted so this seemed like the better option to me. I used a heat gun to safely desolder the 5b in record time. Here's how the pcb looks without the 5b there:



I disconnected pin 2 from vcc, the other pins needed for audio (3 and 27) are floating so those ones are all good. Here's the 5b resoldered with some pins soldered onto the audio pins so I can solder wires easily when the time comes:

I probed everything, all the connections are good and pin 2 is no longer connected to vcc. Gremlins 2 runs fine with pin 2 disconnected from vcc....thanks a lot sunsoft for making this 100x harder by needlessly connecting pin 2 to vcc underneath the chip.

*edit*

I wired up a fourth pin to the prg a17 spot of the 5b since that's been left floating. I'll probably need that pin wired up for the 256k prg eprom of gimmick!



It still runs gremlins 2 fine.

So first off, thanks a lot for that picture of the PCB with the SS5B removed; it lets me finish the trace with 100% confidence.

Drakon wrote:thanks a lot sunsoft for making this 100x harder by needlessly connecting pin 2

It's unfortunately necessary, amplifiers do bad things when their inputs aren't connected.

Drakon wrote:to vcc

Actually it's ground. (it was connected to pin 6 also)

Drakon wrote:underneath the chip.

Yeah, that's just obnoxious.

Of course, now that pin 2 is known to not be PRG A18 on the 5B, I have to wonder if it's actually present on the 5A or FME-7...

lidnariq wrote:So first off, thanks a lot for that picture of the PCB with the SS5B removed; it lets me finish the trace with 100% confidence.

Drakon wrote:thanks a lot sunsoft for making this 100x harder by needlessly connecting pin 2

It's unfortunately necessary, amplifiers do bad things when their inputs aren't connected.

Drakon wrote:to vcc

Actually it's ground. (it was connected to pin 6 also)

Drakon wrote:underneath the chip.

Yeah, that's just obnoxious.

Of course, now that pin 2 is known to not be PRG A18 on the 5B, I have to wonder if it's actually present on the 5A or FME-7...

Oh, I should ground pin 2 until I wire up the audio circuit otherwise something bad will happen?

Also since you appreciate depopulated pcb pics:

http://imgboot.com/images/Drakon/fme7desoldered.jpg

Batman return of the joker US cart with the maskroms and fme7 removed. The fme7 I didn't intend to remove but when I put the pcb down after removing the maskroms I placed it upside down and the fme7 fell out (I guess the entire pcb was still hot enough oops). I resoldered the fme7 back on and socketed the maskroms the cart works fine.

Drakon wrote:Oh, I should ground pin 2 until I wire up the audio circuit otherwise something bad will happen?

Eh, since you've already powered it up and it hasn't done anything, I wouldn't worry about it since you're not planning on leaving it that way. The canonical problems with leaving it disconnected are excessive power draw and causing interference.

Drakon wrote:Also since you appreciate depopulated pcb pics

oooooo.

But would you believe that no-one has a picture of the back side of NES-BTR? It's hard to figure out what pin 2 is connected to here without it

Drakon wrote: Wait a minute...I need to lift my 5b pins?

Here's another random famicom cart that usually has a fme7 but possibly can come with a 5b:

http://rockn69.blog54.fc2.com/blog-entry-565.html

Honoo no Tokyuji Dodge Danpei

With possibility 10-15%. Discussion about Dodge Danpei was on FW while ago, but soon shifted to IRC.
Potencially ANY cart manufactured around Gimmick release date and based on FME7/5x can have 5B.

80sFREAK wrote:With possibility 10-15%. Discussion about Dodge Danpei was on FW while ago, but soon shifted to IRC.
Potencially ANY cart manufactured around Gimmick release date and based on FME7/5x can have 5B.

Oh okay, is there a manufacture date written on the case somewhere? Also is this exclusive to famicom carts or has anyone ever found an american sunsoft pcb with a 5b in it?

lidnariq wrote:But would you believe that no-one has a picture of the back side of NES-BTR? It's hard to figure out what pin 2 is connected to here without it

Well I re-attached pin 2 to ground for now it runs fine with it that way. I can believe what you say there's a lot of undocumented stuff out there. I'd really love to know what the av famicom audio circuit is. Also my hvc-cpu-gpm-02 famicom has almost the same audio circuit as the av famicom but with a slightly bigger low pass filter causing a little more muffle and a little more bass. Again I'd love to know the audio circuit of that famicom but the pcb is masked.

nes-btr solder side:

http://imgboot.com/images/Drakon/nesbtrsolderside.jpg

Pin 2 of the fme7 goes to a18 of the prg maskrom. Normally that would worry me but if you look at the solder side of the gimmick! famicom pcb:



a18 of the prg maskrom is connected to vcc, so all's good for using that pin of the 5b for the audio circuit when the cart is converted to gimmick!

Looking at the parts side of the gimmick! pcb:



a17 of the prg maskrom indeed is connected to pin 32 of the 5b. On my gremlins pcb pin 32 of the 5b is left floating, so I wired up this pin along with the audio pins:



This thing should now be ready to wire up when my eproms / audio circuit parts show up.

Drakon wrote:Pin 2 of the fme7 goes to a18 of the prg maskrom. Normally that would worry me but if you look at the solder side of the gimmick! famicom pcb:

a18 of the prg maskrom is connected to vcc, so all's good for using that pin of the 5b for the audio circuit when the cart is converted to gimmick!

The thing that troubles me about this is that they've switched one pin, and only one pin, from an output to an input, and subsequently exchange which IC any given PCB is using seemingly willy-nilly.

Specifically, these two games (Gremlins 2 and Dodge Danpei) both use GRM-E301. Because of your rework, we know that pin 2 is tied to ground on that PCB. But Gremlins 2 (here) is using a 5B (where that's an input), and Dodge Danpei is using an FME-7 (where that's theoretically an output)

BBW-E301, DD2-E301, and PYK-E301 have solder pads to tie PRG A18 to either FME-7 pin 2 or Vcc, but they're all pre-jumpered for Vcc.
FC-DBT, NES-JSROM, NES-BTR, and NES-JLROM all tie FME-7 pin 2 to PRG A18.
FC-GMK ties PRG A18 to Vcc (obviously)

I mean, my problem is mostly that this is damage-causingly bad electrical engineering; either the pin is externally shorted and the game could destroy the output driver for that pin, or the pin is a floating input to an inverter. (Or ... whatever is true for the 5A)

There are a few kinds of input pins that can be either grounded or left floating without causing a problem.

An open collector output can be either Z (high impedance) or 0 (ground), never 1 (Vcc). /IRQ pins are often like this, where the pin is pulled up to Vcc through a resistor but any one of several devices on the bus can ground it. If one version of a chip uses it as an input but another uses it as an output, the other version's output can be made open collector to keep compatibility with old boards. The opposite is an open emitter, which is either Z or 1, relying on a pull-down elsewhere.

There are also input pins with a weak internal pull-up or pull-down. A well-known example of these in the NES is the CIC, where pin 4 (lock/key) has a pull-down that enables key behavior if pulled down. This is the origin of cutting pin 4 inside the console so that sync failure in a key-key system with no lock wouldn't cause resets.

There were capacitance problems with open collector outputs at high speeds. This led to widespread use of the three-state output, which can force 0, force 1, or get off the bus (Z), such as in National Semiconductor's Tri-state product line. ROMs, for example, are three-state devices: they output only when enabled.

Some pins may be bidirectional, where some data direction mechanism controls whether a particular pin is input or output. The non-RGB NES PPU has a data direction register controlling whether it reads the EXT pins for backdrop color or writes the current pixel's color index to the EXT pins. And of course, every RAM has a data direction input that controls whether D0-D7 is an input (/WE) or output (/OE). If one pin A is set to treat another pin B as always input, then B can be grounded without problem. An IC might even detect which board it's in by pulsing pull-up 0 and 1 voltages on the bidirectional pin. If it's grounded, perform one behavior; if it's tied to Vcc, perform another; otherwise, it's an output.

If only Visual 6502 weren't so backlogged.

Drakon wrote:

80sFREAK wrote:With possibility 10-15%. Discussion about Dodge Danpei was on FW while ago, but soon shifted to IRC.
Potencially ANY cart manufactured around Gimmick release date and based on FME7/5x can have 5B.

Oh okay, is there a manufacture date written on the case somewhere? Also is this exclusive to famicom carts or has anyone ever found an american sunsoft pcb with a 5b in it?

No, there is NO MANUFACTURE DATE on the carts, but there is some unproven(yet) differences, related to date of manufacture.

pin2-pin3 is built-in amplifier or(which is most likely) CMOS invertor gate, used as amplifier. CMOS, so, if not used, input MUST be grounded.

Hint: look at release date(keep in mind, bootgod's database is good, but not 100% accurate) of Famicom carts and US releases.

80sFREAK wrote:pin2-pin3 is built-in amplifier or(which is most likely) CMOS invertor gate, used as amplifier. CMOS, so, if not used, input MUST be grounded.

Hint: look at release date(keep in mind, bootgod's database is good, but not 100% accurate) of Famicom carts and US releases.

Okay, so what happens if you don't ground pin 2 when it's not in use?

Drakon wrote:Okay, so what happens if you don't ground pin 2 when it's not in use?

Oddly, I can't coax google into giving me any good visualization of it. So, here's what it won't do:
It won't catch on fire
It's exceptionally unlikely to cause permanent damage

The problem is "not connected to anything" isn't actually true; there's actually a little bit of capacitance between everything. And that little tiny bit of capacitance (roughly 1pF), when combined with the intrinsic inductance of the leads of the IC (about 3nH) can cause oscillations which drain huge amounts of power and cause massive interference. Sometimes either can be big enough to cause the game to stop working. Sometimes the oscillations can behave as an RF transmitter and cause problems in unconnected other devices.

lidnariq wrote:Oddly, I can't coax google into giving me any good visualization of it. So, here's what it won't do:
It won't catch on fire
It's exceptionally unlikely to cause permanent damage

The problem is "not connected to anything" isn't actually true; there's actually a little bit of capacitance between everything. And that little tiny bit of capacitance (roughly 1pF), when combined with the intrinsic inductance of the leads of the IC (about 3nH) can cause oscillations which drain huge amounts of power and cause massive interference. Sometimes either can be big enough to cause the game to stop working. Sometimes the oscillations can behave as an RF transmitter and cause problems in unconnected other devices.

I've experienced this before I wired up a circuit that divides a clock signal by 1/6, it worked but it was picking up stuff like an rf antenna and going in and out of working. I wired some floating pins to vcc and now it's completely fixed. Anyway as long as there won't be any damage that's good to know. These sunsoft 5b chips are really hard to find so the last thing I'd want to do is damage it somehow.