Author Topic: Super Nintendo - Unique problem  (Read 17206 times)

0 Members and 1 Guest are viewing this topic.

Offline RaellzTopic starter

  • Contributor
  • Posts: 23
  • Country: us
Super Nintendo - Unique problem
« on: January 03, 2016, 03:46:12 pm »
I've been trying to troubleshoot a Super Nintendo Motherboard all night and after replacing all the caps, testing for shorts and reading a lot of forum posts I decided to lift Pin 4 on the lockout chip and then tie it to ground. This has worked, to a degree. One of my games work, while another does not. When using the one that does work, the system updates the video about once every 2 seconds or so making title screens take half an hour. The sound usually works and is not slowed down at all.

I put some of the original caps back in after testing them on the multimeter. I'm no expert, but I'm wondering if it is a bad cap still since... and forgive this comparison... but if you change the capacitors in a 555 circuit then you change the cycle of the output, such as blinking an LED faster or slower. I only got this far after lifting pin #4 on the lockout chip so I don't know what my next step should be.

Has anyone heard of this problem or know how I can resolve it? Thanks.  :-+
 

Offline Simon

  • Global Moderator
  • *****
  • Posts: 17729
  • Country: gb
  • Did that just blow up? No? might work after all !!
    • Simon's Electronics
Re: Super Nintendo - Unique problem
« Reply #1 on: January 03, 2016, 03:55:13 pm »
what was the failt in the first place ? caps these days won't be controlling clock spedds or anything like that. It will be a crystal clock.
 

Offline RaellzTopic starter

  • Contributor
  • Posts: 23
  • Country: us
Re: Super Nintendo - Unique problem
« Reply #2 on: January 03, 2016, 03:59:20 pm »
what was the failt in the first place ? caps these days won't be controlling clock spedds or anything like that. It will be a crystal clock.

Ah, I forgot to mention that =/ It was a black screen with no static and no sound. The signal was picked up by my TV, but it went straight to black.
 

Offline amyk

  • Super Contributor
  • ***
  • Posts: 8240
Re: Super Nintendo - Unique problem
« Reply #3 on: January 03, 2016, 04:03:18 pm »
I'm not familiar with the hardware but a timer running at the wrong frequency could cause the symptoms you're describing. There's a schematic here:

http://wiki.superfamicom.org/snes/files/snes_schematic_color.png

You can find the clock crystals and check them. It seems the sound is based on a separate crystal (24.576MHz) than the rest of the system (21.47727MHz, 6x NTSC colourburst), which is why it could remain operational.

There are also test ROMs available. If you can find a cartridge to put it on, that might be a good option.
 

Offline RaellzTopic starter

  • Contributor
  • Posts: 23
  • Country: us
Re: Super Nintendo - Unique problem
« Reply #4 on: January 03, 2016, 04:17:23 pm »
I'm not familiar with the hardware but a timer running at the wrong frequency could cause the symptoms you're describing. There's a schematic here:

http://wiki.superfamicom.org/snes/files/snes_schematic_color.png

You can find the clock crystals and check them. It seems the sound is based on a separate crystal (24.576MHz) than the rest of the system (21.47727MHz, 6x NTSC colourburst), which is why it could remain operational.

There are also test ROMs available. If you can find a cartridge to put it on, that might be a good option.

That link doesn't seem to work. Is this the same thing? http://wiki.superfamicom.org/snes/show/Schematics,+Ports,+and+Pinouts

I'm currently looking into how to make a test cartridge.

So should I check the crystals?
 

Offline stmdude

  • Frequent Contributor
  • **
  • Posts: 479
  • Country: se
Re: Super Nintendo - Unique problem
« Reply #5 on: January 03, 2016, 04:50:17 pm »
Check the crystals. If you have access to an oscilloscope, you can probe one of the pins of the crystal and you should see the oscillations in the frequency specified on the schematic. ( Yes, it's the same schematic you linked you. Apparently, they don't allow direct linking)

However, that's a really wierd failure scenario for a SNES. I've repaired _tons_ of them, and haven't seen a "slow" SNES ever. Plenty with the black screen, but I don't think any of them had sound either.

Anyhow, I'm hoping you did the obvious things before jumping to caps?  As in, cleaning the cartridge connector, as well as lifting it up, and cleaning the connector-to-the-connector (or whatever it's called). Some might require re-tensioning of the "blades" in the connector as well, depending on how its been used.
This easily fixes 9 out of 10 SNESes that passes by my desk..

Also, verify the output from the 7805. I've seen a few of those go bad as well, and cause wierd and wonderful failure modes.
 

Offline RaellzTopic starter

  • Contributor
  • Posts: 23
  • Country: us
Re: Super Nintendo - Unique problem
« Reply #6 on: January 03, 2016, 05:21:07 pm »
Check the crystals. If you have access to an oscilloscope, you can probe one of the pins of the crystal and you should see the oscillations in the frequency specified on the schematic. ( Yes, it's the same schematic you linked you. Apparently, they don't allow direct linking)

However, that's a really wierd failure scenario for a SNES. I've repaired _tons_ of them, and haven't seen a "slow" SNES ever. Plenty with the black screen, but I don't think any of them had sound either.

Anyhow, I'm hoping you did the obvious things before jumping to caps?  As in, cleaning the cartridge connector, as well as lifting it up, and cleaning the connector-to-the-connector (or whatever it's called). Some might require re-tensioning of the "blades" in the connector as well, depending on how its been used.
This easily fixes 9 out of 10 SNESes that passes by my desk..

Also, verify the output from the 7805. I've seen a few of those go bad as well, and cause wierd and wonderful failure modes.

I do not have access to an oscilloscope unfortunately. I have cleaned all the connections and checked the solder points. I reattached the lockout chip's pin #4 to it's original position, now that I see it was already at ground  |O Same issues still.

I tested the 7805 and the results were 10.7v on the left pin and 5.02 on the right pin. According to what I'm reading, the left pin should be a couple volts less than the output of the power supply. I'm using the original SNES power supply which is 10v 850ma that works fine on all other SNES systems. Does this sound like a bad 7805? EDIT: Nah, it makes sense to me now. I been up all night so I'm not thinking right, haha. The 7805 seems good. Now I'm back to suspecting the caps.
« Last Edit: January 03, 2016, 05:36:43 pm by Raellz »
 

Offline marshallh

  • Supporter
  • ****
  • Posts: 1462
  • Country: us
    • retroactive
Re: Super Nintendo - Unique problem
« Reply #7 on: January 03, 2016, 07:34:09 pm »
99% chance it's the cartridge connector. Always is. One time though I did see a bad WRAM (128k PSRAM).
Timing is via crystal, though it is varactor-tuned. If you get a color video signal at any point that is fine (It will mostly affect the NTSC colorburst which is a bit sensitive)
Verilog tips
BGA soldering intro

11:37 <@ktemkin> c4757p: marshall has transcended communications media
11:37 <@ktemkin> He speaks protocols directly.
 

Offline stmdude

  • Frequent Contributor
  • **
  • Posts: 479
  • Country: se
Re: Super Nintendo - Unique problem
« Reply #8 on: January 03, 2016, 07:47:00 pm »
99% chance it's the cartridge connector. Always is. One time though I did see a bad WRAM (128k PSRAM).

Yup, seen WRAM as well as VRAM failures, but they are rare. I've also seen mod-chip installations that have deteriorated over time.
Busted internal fuse is not unheard of either.

I've actually never seen a broken trace, even though some of the systems I've repaired looks like they've been stored in a submarine.

If you know your pseudo-history, you'll appreciate that I've never seen a soda-spill on/in any of the systems I get to see either (European SNES and Super Famicom). :)

By the way. You guys are american. I actually haven't got an North-American Super Nintendo. Anyone want to trade cases with me? I've got plain-old european ones, as well as Super Famicom ones. No need to ship the electronics. They're the same.
 

Offline marshallh

  • Supporter
  • ****
  • Posts: 1462
  • Country: us
    • retroactive
Re: Super Nintendo - Unique problem
« Reply #9 on: January 03, 2016, 08:04:45 pm »
By the way. You guys are american. I actually haven't got an North-American Super Nintendo. Anyone want to trade cases with me? I've got plain-old european ones, as well as Super Famicom ones. No need to ship the electronics. They're the same.

Trust me, you don't want 'em... It's garbage. Though, if you want a sickly yellowed one, those are easy to get.
Verilog tips
BGA soldering intro

11:37 <@ktemkin> c4757p: marshall has transcended communications media
11:37 <@ktemkin> He speaks protocols directly.
 

Offline stmdude

  • Frequent Contributor
  • **
  • Posts: 479
  • Country: se
Re: Super Nintendo - Unique problem
« Reply #10 on: January 03, 2016, 08:15:34 pm »
Trust me, you don't want 'em... It's garbage. Though, if you want a sickly yellowed one, those are easy to get.

Well, they're unique garbage at least. :)

It would look nice next to its brethren over here..
 

Offline RaellzTopic starter

  • Contributor
  • Posts: 23
  • Country: us
Re: Super Nintendo - Unique problem
« Reply #11 on: January 04, 2016, 05:53:59 am »
99% chance it's the cartridge connector. Always is. One time though I did see a bad WRAM (128k PSRAM).
Timing is via crystal, though it is varactor-tuned. If you get a color video signal at any point that is fine (It will mostly affect the NTSC colorburst which is a bit sensitive)

I get color video. It just slowly updates the video.
 

Offline poot36

  • Frequent Contributor
  • **
  • Posts: 678
  • Country: ca
Re: Super Nintendo - Unique problem
« Reply #12 on: January 04, 2016, 06:44:06 am »
The 21.4Mhz clock is fed into the CPU, PPU1 and the cartridge slot.  It exits the CPU at I think around 4Mhz or 2Mhz depending on what the game tells the CPU to output.  This then goes to PPU2 and another pin on the cartridge slot.  If you have a frequency counter you could check this.  What version of SNES do you have?  If it is the first version then the clock generation circuit will be different from the aforementioned schematic and there is a separate 8Mhz clock for the CIC (lockout chip).  If your multimeter has a frequency function see if it will read the clocks correctly.  I am beginning to suspect a bad CPU or solder joint.  When I get to my computer with the schematic on it I will double check what I posted is correct.
 

Offline RaellzTopic starter

  • Contributor
  • Posts: 23
  • Country: us
Re: Super Nintendo - Unique problem
« Reply #13 on: January 04, 2016, 07:55:54 am »
The 21.4Mhz clock is fed into the CPU, PPU1 and the cartridge slot.  It exits the CPU at I think around 4Mhz or 2Mhz depending on what the game tells the CPU to output.  This then goes to PPU2 and another pin on the cartridge slot.  If you have a frequency counter you could check this.  What version of SNES do you have?  If it is the first version then the clock generation circuit will be different from the aforementioned schematic and there is a separate 8Mhz clock for the CIC (lockout chip).  If your multimeter has a frequency function see if it will read the clocks correctly.  I am beginning to suspect a bad CPU or solder joint.  When I get to my computer with the schematic on it I will double check what I posted is correct.

SHVC-CPU-01 (1990) is the SNES model.

My multimeter does have Hz, but I'm not sure how to test what you're talking about. What do I probe exactly?
 

Offline RaellzTopic starter

  • Contributor
  • Posts: 23
  • Country: us
Re: Super Nintendo - Unique problem
« Reply #14 on: January 04, 2016, 01:21:18 pm »
I removed the crystal then added another one, 16k, didn't start up. I put it back then swapped some capacitor values to see if anything would change, they did not. I then decided to reflow everything on the board and I still have the same slow video issue. I looked for solder bridges both before an after reflowing. I guess it's just a bad IC at this point  :-//
 

Offline poot36

  • Frequent Contributor
  • **
  • Posts: 678
  • Country: ca
Re: Super Nintendo - Unique problem
« Reply #15 on: January 05, 2016, 05:01:49 am »
Ok, I screwed up.  The 21.4Mhz clock goes to the CPU, and both PPU1 and PPU2.  The slower SYSCLOCK goes to the SRAM for the CPU and the cartridge slot.  I have attached the correct partial schematic for the version 1 SNES that includes the main clock generator circuit and the CIC (lockout chip) circuit.  One of the resistors in the clock generator circuit is wrong but I do not remember which on (I think it is R4).  To use your multimeter to measure the frequency set it to the Hz setting and connect the negative probe to any grounded part on the SNES (I prefer the metal band surrounding the front of the SNES) and the positive probe on the pin that has the clock signal you are trying to measure in this case pin 57 of the cartridge slot for the SYSCLOCK (the slow one) and pin 1 of the cartridge slot for the 21.4Mhz clock.  I would also check also that all reset pins are at 5V (no lower then 4.3V) or thereabouts to prevent random reseting of various components.  You can also check out the website www.NESDev.com for additional help.  That is how I figured out that my SNES had two bad registers in the CPU that was causing it to crash trying to play any normal game.  I also wound up buying a logic analyzer for that as well (I already had a oscilloscope).  You can view my adventures in that discovery here: http://forums.nesdev.com/viewtopic.php?f=12&t=13088.  If anybody in Calgary has a broken SNES that has another fault (bad RAM, PPU1, PPU2) that the CPU is fine I would be interested in the CPU to fix my SNES.
« Last Edit: January 05, 2016, 05:04:46 am by poot36 »
 

Offline true

  • Frequent Contributor
  • **
  • Posts: 329
  • Country: us
  • INTERNET
Re: Super Nintendo - Unique problem
« Reply #16 on: January 08, 2016, 05:22:36 am »
The PPU was known for an odd assortment of failures.

Does the game actually play at this reduced speed, or does video output just sporadically work? If the latter, you probably have a bad PPU.
 

Offline Mephitus

  • Regular Contributor
  • *
  • Posts: 174
  • Country: us
  • Dashing, no?
Re: Super Nintendo - Unique problem
« Reply #17 on: January 08, 2016, 11:02:52 pm »
Actually, it could be even simpler. What is the game? Several have alternate processors on them that could be causing a problem. Especially if other games work properly. Also, Raellz, what state do you live in? If you are in utah, you are welcome to use my scope.
A true gentleman must be prepared for anything. - Pepe le' Pew
 

Offline RaellzTopic starter

  • Contributor
  • Posts: 23
  • Country: us
Re: Super Nintendo - Unique problem
« Reply #18 on: January 23, 2016, 02:32:44 am »
Actually, it could be even simpler. What is the game? Several have alternate processors on them that could be causing a problem. Especially if other games work properly. Also, Raellz, what state do you live in? If you are in utah, you are welcome to use my scope.

I have Super Mario World 2: Yoshi's Island which does not play at all and I have Super Ghosts n Ghouls. SGnG works at the first logo screen, but updates very slowly. Sound plays without issue. Once it's beyond that first title screen the video goes crazy. Random textures, but the movement of the textures is correct, just not the textures themselves.

I'm in Kansas, but I appreciate the offer!

Sorry for the lack of updates. I've replaced the crystal and all the caps then double checked that the caps are good, still nothing.
 

Offline RaellzTopic starter

  • Contributor
  • Posts: 23
  • Country: us
Re: Super Nintendo - Unique problem
« Reply #19 on: January 23, 2016, 02:37:35 am »
The PPU was known for an odd assortment of failures.

Does the game actually play at this reduced speed, or does video output just sporadically work? If the latter, you probably have a bad PPU.

It plays at the title screen, but I haven't tried using a controller yet. The textures are all broken though.
 

Offline RaellzTopic starter

  • Contributor
  • Posts: 23
  • Country: us
Re: Super Nintendo - Unique problem
« Reply #20 on: January 23, 2016, 02:39:25 am »
Ok, I screwed up.  The 21.4Mhz clock goes to the CPU, and both PPU1 and PPU2.  The slower SYSCLOCK goes to the SRAM for the CPU and the cartridge slot.  I have attached the correct partial schematic for the version 1 SNES that includes the main clock generator circuit and the CIC (lockout chip) circuit.  One of the resistors in the clock generator circuit is wrong but I do not remember which on (I think it is R4).  To use your multimeter to measure the frequency set it to the Hz setting and connect the negative probe to any grounded part on the SNES (I prefer the metal band surrounding the front of the SNES) and the positive probe on the pin that has the clock signal you are trying to measure in this case pin 57 of the cartridge slot for the SYSCLOCK (the slow one) and pin 1 of the cartridge slot for the 21.4Mhz clock.  I would also check also that all reset pins are at 5V (no lower then 4.3V) or thereabouts to prevent random reseting of various components.  You can also check out the website www.NESDev.com for additional help.  That is how I figured out that my SNES had two bad registers in the CPU that was causing it to crash trying to play any normal game.  I also wound up buying a logic analyzer for that as well (I already had a oscilloscope).  You can view my adventures in that discovery here: http://forums.nesdev.com/viewtopic.php?f=12&t=13088.  If anybody in Calgary has a broken SNES that has another fault (bad RAM, PPU1, PPU2) that the CPU is fine I would be interested in the CPU to fix my SNES.

Thanks very much! I will give that a shot. I'm still really new to this, but I've fixed a few things before  :box:. I'm not sure what a logic analyzer is. I'll look that up.
 

Offline RaellzTopic starter

  • Contributor
  • Posts: 23
  • Country: us
Re: Super Nintendo - Unique problem
« Reply #21 on: January 23, 2016, 03:07:25 am »
Ok, I screwed up.  The 21.4Mhz clock goes to the CPU, and both PPU1 and PPU2.  The slower SYSCLOCK goes to the SRAM for the CPU and the cartridge slot.  I have attached the correct partial schematic for the version 1 SNES that includes the main clock generator circuit and the CIC (lockout chip) circuit.  One of the resistors in the clock generator circuit is wrong but I do not remember which on (I think it is R4).  To use your multimeter to measure the frequency set it to the Hz setting and connect the negative probe to any grounded part on the SNES (I prefer the metal band surrounding the front of the SNES) and the positive probe on the pin that has the clock signal you are trying to measure in this case pin 57 of the cartridge slot for the SYSCLOCK (the slow one) and pin 1 of the cartridge slot for the 21.4Mhz clock.  I would also check also that all reset pins are at 5V (no lower then 4.3V) or thereabouts to prevent random reseting of various components.  You can also check out the website www.NESDev.com for additional help.  That is how I figured out that my SNES had two bad registers in the CPU that was causing it to crash trying to play any normal game.  I also wound up buying a logic analyzer for that as well (I already had a oscilloscope).  You can view my adventures in that discovery here: http://forums.nesdev.com/viewtopic.php?f=12&t=13088.  If anybody in Calgary has a broken SNES that has another fault (bad RAM, PPU1, PPU2) that the CPU is fine I would be interested in the CPU to fix my SNES.

Alright I tested it a few times (while turned on of course) and I get 0.00 back on the Hz setting. Flipping over to Voltage I get 3.1 mV from pin 57, but that's all.

I tested pin 1 and I got 21.47 khz (Edit: Mhz not Khz) so that's good.

The reset pin gives me a whopping 314 mV. Well under the 5v target.

We've narrowed it down!  :box: Now what can I do about that?

« Last Edit: January 24, 2016, 09:57:31 pm by Raellz »
 

Offline poot36

  • Frequent Contributor
  • **
  • Posts: 678
  • Country: ca
Re: Super Nintendo - Unique problem
« Reply #22 on: January 24, 2016, 05:27:13 pm »
Wait Khz or Mhz?  If it is khz your meter probley cant read this high and is reading a harmonic of the frequency (mine read 300 Hz!).  On pin 57 on my SNES with my meter set to DC volts I got 2.3V or so (my meter on on Hz mode read 29 Khz because it cant read this high).  The reset pin on the cartridge slot should not be below 4.3V so yours is currently in reset.  You must preform the tests with a cartridge in the slot or else the CIC (lockout) chip will keep the SNES in reset unless you disable the CIC chip.  Because your SNES is currently in reset you will not get a signal on pin 57 as this is controlled by the CPU.
 

Offline RaellzTopic starter

  • Contributor
  • Posts: 23
  • Country: us
Re: Super Nintendo - Unique problem
« Reply #23 on: January 24, 2016, 09:56:49 pm »
Wait Khz or Mhz?  If it is khz your meter probley cant read this high and is reading a harmonic of the frequency (mine read 300 Hz!).  On pin 57 on my SNES with my meter set to DC volts I got 2.3V or so (my meter on on Hz mode read 29 Khz because it cant read this high).  The reset pin on the cartridge slot should not be below 4.3V so yours is currently in reset.  You must preform the tests with a cartridge in the slot or else the CIC (lockout) chip will keep the SNES in reset unless you disable the CIC chip.  Because your SNES is currently in reset you will not get a signal on pin 57 as this is controlled by the CPU.

My mistake, Mhz. 21.47 Mhz oscillating crystal is what I installed and that's what it reads.

I did the test with a cartridge in and I got just over 5 volts on the Reset pin and I got 2.729 Mhz on pin 57.

I found this information on the SNES CPU specs. What I'm getting off that pin seems to be about the same speed as the "slow access" speed for the ROM and RAM. How would I go about testing the "quick access" speeds? I'm a newbie, but I assume I would check the Hz between the CPU and the RAM?

"The SNES's CPU is a 65c816 based processor. While its clock speed is at about 21 MHz, it's effective speed is considerably lower, at 3.58 MHz for quick access (i.e. hardware registers at $2100-$21FF in banks $00-$3F), 2.68 MHz for slow access (i.e. ROM and RAM) and 1.79 MHz for very slow access."
« Last Edit: January 24, 2016, 10:16:14 pm by Raellz »
 

Offline poot36

  • Frequent Contributor
  • **
  • Posts: 678
  • Country: ca
Re: Super Nintendo - Unique problem
« Reply #24 on: January 25, 2016, 12:06:11 am »
Those are way better readings.  To test fast access you need a cartridge that uses fast access rom chips.  The speed of that clock signal is programmed by the code in the game so only some games used the faster speed.  At this point I think you will need to make a test cartridge with a simple program on it to verify the basics of the SNES are working correctly.  Find a cheap game that you can take the non re- programmable rom chip out of and place a re-programmable chip in a socket to run various test programs.  Try and find a Lo-rom based game not a Hi-rom game as this will make loading code a lot easy'er for you.  I made that mistake and was restricted to 32Kb programs not fun.  You will have to rewire the cart to use a standard eeprom chip as the orignal rom chip pinout is nonstandard.  You will also need an eeprom programmer as well.  Good luck on fixing the SNES.  The most likely reason Super Mario World 2: Yoshi's Island does not work at all is because it uses what is called the super fx chip which uses all of the cartridges pins and directly ties into the SNES's graphics chip.  If any of the connections are bad ore the graphics chip is not working correctly the game will not play.  Super Ghosts n Ghouls does not use this super fx chip so it has a higher chance of running correctly even on a broken SNES.  If you could make a video of the corruption you are seeing that would be nice.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf