Forums

I am losing my mind here! I bought my first Master System off eBay. I ended up getting a deal on a console with a slight issue. Here is the eBay description;

"Here is a listing for a sega master system with 2 pads, gun, and rf cable.
The console has been tested and it powers on and plays games with no problem. The problem with this console is that control port #1 does not work properly which causes the character or object to move to the left by itself without pressing anything and with any controller. The controllers and gun work properly. No ac adapter is included. Item sold as is for repair or parts only."

Sounds pretty simple to fix right, or at least that's what i thought. I got the system, powered it up, booted into the snail maze game and sure enough, both controllers were showing the same issue. The snail could not go right at all (i assume, because the left button was constantly being pressed as per the eBay listing.)

I tried out a half broken 3rd party genesis controller, and lo and behold, it worked. I could move my little snail in all directions freely. After testing the wiring on the genesis controller, i decided a cable transplant would be the best bet, just use the cable from the genesis controller and no big deal. I wired everything up, using my multimeter to make sure the pinouts matched the diagrams and making sure my soldering was okay and...nothing.....same problem with the d-pad.

Okay, maybe its the soldering of the ports. I used solder wick and cleaned up all the pads on both controller ports, then resoldered them, making sure to check for bridges. STILL NOTHING. I remove both ports completley, and soak them in 99% isopropyl (and cleaned every contact and membrane on the controllers), and swap them, meaning i put the one that was the player two port into the spot for the player one port. Still nothing. I finally cut the trace from the pin for dpad left, hoping to stop the erroneous input once and for all. Still nothing. As i said im losing my mind!

After a bit of research, it could be something to do with IC4, but i don't have a donor board to test. I'm thinking of re flowing the solder on the pins of that IC just to try something else.

TLDR; Controllers go left constantly, no matter what i try.

Anyone have any advice? Maybe something i've overlooked. I'm going crazy here!

lazyjackal wrote

I tried out a half broken 3rd party genesis controller, and lo and behold, it worked.

I wonder how could that have worked!

sverx wrote

I wonder how could that have worked!

I have no clue. The d pad worked flawlessly and i was able to hold down A and C in order to boot the snail maze game.

did you try with a multimeter to test if the port pin #3 is grounded? If it is, that's the reason why 'left' is pressed.

the only pins with continuity to ground are pins 5 and 8. Is that normal for the Master System?

lazyjackal wrote

the only pins with continuity to ground are pins 5 and 8. Is that normal for the Master System?

Pin 5 is supposed to be 5V, not ground.

Could be IC4 or an EMI filter.

Which IC4 do you have, the 315-5216 or the 315-5237?

asynchronous wrote

Could be IC4 or an EMI filter.

Which IC4 do you have, the 315-5216 or the 315-5237?

I have the 315-5216. Do those little EMI filters fail often?

Not often that I know of, and neither does IC4.

The EMI filters are the only thing between the controller ports and IC4. There aren't even external pull-up resistors from memory. Each EMI filter is 2 ferrite beads in series with a capacitor in the middle to ground. You could test it for shorts to ground to try de-soldering it completely for the "L" button lines and see if that solves your problem.

Well i desoldered the EMI filter, lifted a pad doing so (not the first one ive lifted doing this). The issue still persists.If it is indeed IC4 then i guess i will need to keep my eyes peeled for another parts system.

As a last test, try connecting KILLGA (315-5216 pin 23, or pin 43 on the cart slot) high. If the problem goes away when KILLGA is connected high, then IC4 is faulty.

asynchronous wrote

As a last test, try connecting KILLGA (315-5216 pin 23, or pin 43 on the cart slot) high. If the problem goes away when KILLGA is connected high, then IC4 is faulty.

Forgive my ignorance, do you mean connect pin 23 of IC4 to +5v?

EDIT: I went ahead and tried connecting the KILLGA to +5v via a 1k resistor, then tried booting the console. no inputs worked at all.

EDIT 2: Okay so putting KILLGA high while playing the game makes the snail freak out as if all directional inputs are being pressed. but it doesnt solve my problem.

I had this problem and resolved it by making jumper wires from the db-9 connector the EMI filters on the offending controller port. Worth a try if you're desperate. Probably just needed solder reflowed for the EMI filter.

Ah, sorry my mistake. I forgot the data bus is not terminated on the master system. There is a resistor array on the board but, double checking the schematics, it's not on the data bus. So disabling IC4 would have meant a floating data bus when the CPU read the IO port.

If your soldering is good and the PCB tracks and EMI filters are in good nick then it does sound like IC4 is stuffed.

But.....it's strange that the problem went away with the broken Genesis controller. The Genesis controller multiplexes the L,R,A,B,C,Start buttons so the Left line is driven high when the D-pad is not pressed left. Maybe the Left inputs on IC4 are stuffed but when they're driven high they work OK?? In that case maybe, just maybe, soldering on external pull up resistors on the Left inputs may fix it?

asynchronous wrote

Maybe the Left inputs on IC4 are stuffed but when they're driven high they work OK?? In that case maybe, just maybe, soldering on external pull up resistors on the Left inputs may fix it?

Holy crap that would be a weird but simple fix. I am going to try that tonight, there were some resistors on that 3rd party genesis controller. I will check them for poops and giggles and report back their values.

Okay, so this is the controller in question:

Front

Back

All the resistors are 47k Ohms, the electrolytic capacitor is 10uf, the little ceramic decoupling capacitor is .1uf.

I tried a single 47k, and two 47k resistors in series. The results were erratic. I could move right but the left input would intermittently and quickly move the character on screen back left.

Dropping down to 20k though, yields different results. I can freely move up, down, and right but cannot move left at all. Same results all the way down to 5.1k where it becomes erratic again.

So the saga continues.

OK, what are you connecting the resistor to exactly?

I know it's an old thread again, only confirms information doesn't spread well...

Back in 2017 I fixed a MS2 with a similar issue (IIRC 2 directions weren't working). It was due to an internal fault in the 315-5237 custom chip. I reproduced the faulty part of the chip with an external PLD chip (GAL16V8), it was well documented on my blog (now gone).

The good news is both the 315-5126 and 315-5237 are easy to fully reproduce in a CPLD.

P.S.: Here's a picture of my bodge repair

20190201_084459.jpg (2.99 MB)

Just for "fun" I sucessfully reproduced both the 315-5216 and 315-5237 I/O chips in a CPLD but...

I wanted to use something 5V compatible to avoid voltage shifting, so made a first implementation in an ATF1504 (64 macrocells, both chips fit easily). Problem is the TQFP44 package (that could have fitted between the two rows of the dip footprint, tightly yes...) only allows 36 I/Os and we need 40 for the 315-5216 and 46 for the 315-5237.

So for testing I used a big a** PLCC84 chip, overhanging over the nearby chips...

Now I'm looking at the Cool Runner CPLDs from Xilinx, 3.3V but with 5V compatible I/Os (so would only need an extra 3.3 regulator) and allowing 41 I/Os for the TQFP44 package. Good enough for the 315-5216 but still not good for the 315-5237. I'm exploring more options.

BTW I was shocked to discover the WebPack toolkit from Xilinx was almost 16GB! Seriously?!

Nice.

I found the same issue regarding the lack of I/O pins when looking the ATF CPLD series for this. I was thinking of using a couple of 74LS157's or 74LS241's next to the CPLD to handle the port DC and DD reads. The TQFP44 and two SOIC20 would fit.

I also thought of using TTL chips or two TQFP44 CPLDs but couldn't resolve to it, once you get into CPLD territory you expect total integration.

After a bit of research I chose to target the XC9572XL CPLD from Xilinx (72 macrocells, overkill, I need only half of those) in VQ64 package (12x12mm, 56 I/Os, 3.3V with 5V tolerant I/Os, will need an extra regulator, no biggie)

Waiting for the chips to come now.

P.S.: I found other chips that could do the job but in BGA package only if you want enough I/Os, not hobbyist friendly IMO.

Author	Message
lazyjackal Joined: 06 Jul 2017 Posts: 8	SMS controller (port?) issue Posted: Thu Jul 06, 2017 6:15 am
lazyjackal Joined: 06 Jul 2017 Posts: 8	I am losing my mind here! I bought my first Master System off eBay. I ended up getting a deal on a console with a slight issue. Here is the eBay description; "Here is a listing for a sega master system with 2 pads, gun, and rf cable. The console has been tested and it powers on and plays games with no problem. The problem with this console is that control port #1 does not work properly which causes the character or object to move to the left by itself without pressing anything and with any controller. The controllers and gun work properly. No ac adapter is included. Item sold as is for repair or parts only." Sounds pretty simple to fix right, or at least that's what i thought. I got the system, powered it up, booted into the snail maze game and sure enough, both controllers were showing the same issue. The snail could not go right at all (i assume, because the left button was constantly being pressed as per the eBay listing.) I tried out a half broken 3rd party genesis controller, and lo and behold, it worked. I could move my little snail in all directions freely. After testing the wiring on the genesis controller, i decided a cable transplant would be the best bet, just use the cable from the genesis controller and no big deal. I wired everything up, using my multimeter to make sure the pinouts matched the diagrams and making sure my soldering was okay and...nothing.....same problem with the d-pad. Okay, maybe its the soldering of the ports. I used solder wick and cleaned up all the pads on both controller ports, then resoldered them, making sure to check for bridges. STILL NOTHING. I remove both ports completley, and soak them in 99% isopropyl (and cleaned every contact and membrane on the controllers), and swap them, meaning i put the one that was the player two port into the spot for the player one port. Still nothing. I finally cut the trace from the pin for dpad left, hoping to stop the erroneous input once and for all. Still nothing. As i said im losing my mind! After a bit of research, it could be something to do with IC4, but i don't have a donor board to test. I'm thinking of re flowing the solder on the pins of that IC just to try something else. TLDR; Controllers go left constantly, no matter what i try. Anyone have any advice? Maybe something i've overlooked. I'm going crazy here!

sverx Joined: 05 Sep 2013 Posts: 3827 Location: Stockholm, Sweden	Posted: Thu Jul 06, 2017 8:24 am
	lazyjackal wrote I tried out a half broken 3rd party genesis controller, and lo and behold, it worked. I wonder how could that have worked!

lazyjackal Joined: 06 Jul 2017 Posts: 8	Posted: Thu Jul 06, 2017 9:21 am
lazyjackal Joined: 06 Jul 2017 Posts: 8	sverx wrote I wonder how could that have worked! I have no clue. The d pad worked flawlessly and i was able to hold down A and C in order to boot the snail maze game.

sverx Joined: 05 Sep 2013 Posts: 3827 Location: Stockholm, Sweden	Posted: Thu Jul 06, 2017 9:29 am
	did you try with a multimeter to test if the port pin #3 is grounded? If it is, that's the reason why 'left' is pressed.

lazyjackal Joined: 06 Jul 2017 Posts: 8	Posted: Thu Jul 06, 2017 9:36 am
lazyjackal Joined: 06 Jul 2017 Posts: 8	the only pins with continuity to ground are pins 5 and 8. Is that normal for the Master System?

ApolloBoy Joined: 30 Jan 2007 Posts: 313 Location: San Jose, CA	Posted: Fri Jul 07, 2017 5:45 am
	lazyjackal wrote the only pins with continuity to ground are pins 5 and 8. Is that normal for the Master System? Pin 5 is supposed to be 5V, not ground.

asynchronous Joined: 14 Aug 2000 Posts: 742 Location: Adelaide, Australia	SMS controller (port?) issue Posted: Fri Jul 07, 2017 6:07 am
	Could be IC4 or an EMI filter. Which IC4 do you have, the 315-5216 or the 315-5237?

lazyjackal Joined: 06 Jul 2017 Posts: 8	Posted: Fri Jul 07, 2017 6:15 am
lazyjackal Joined: 06 Jul 2017 Posts: 8	asynchronous wrote Could be IC4 or an EMI filter. Which IC4 do you have, the 315-5216 or the 315-5237? I have the 315-5216. Do those little EMI filters fail often?

asynchronous Joined: 14 Aug 2000 Posts: 742 Location: Adelaide, Australia	SMS controller (port?) issue Posted: Fri Jul 07, 2017 6:52 am
	Not often that I know of, and neither does IC4. The EMI filters are the only thing between the controller ports and IC4. There aren't even external pull-up resistors from memory. Each EMI filter is 2 ferrite beads in series with a capacitor in the middle to ground. You could test it for shorts to ground to try de-soldering it completely for the "L" button lines and see if that solves your problem.

lazyjackal Joined: 06 Jul 2017 Posts: 8	Posted: Fri Jul 07, 2017 9:38 am
lazyjackal Joined: 06 Jul 2017 Posts: 8	Well i desoldered the EMI filter, lifted a pad doing so (not the first one ive lifted doing this). The issue still persists.If it is indeed IC4 then i guess i will need to keep my eyes peeled for another parts system.

asynchronous Joined: 14 Aug 2000 Posts: 742 Location: Adelaide, Australia	SMS controller (port?) issue Posted: Fri Jul 07, 2017 10:01 am
	As a last test, try connecting KILLGA (315-5216 pin 23, or pin 43 on the cart slot) high. If the problem goes away when KILLGA is connected high, then IC4 is faulty.

lazyjackal Joined: 06 Jul 2017 Posts: 8	Posted: Fri Jul 07, 2017 7:11 pm
lazyjackal Joined: 06 Jul 2017 Posts: 8	asynchronous wrote As a last test, try connecting KILLGA (315-5216 pin 23, or pin 43 on the cart slot) high. If the problem goes away when KILLGA is connected high, then IC4 is faulty. Forgive my ignorance, do you mean connect pin 23 of IC4 to +5v? EDIT: I went ahead and tried connecting the KILLGA to +5v via a 1k resistor, then tried booting the console. no inputs worked at all. EDIT 2: Okay so putting KILLGA high while playing the game makes the snail freak out as if all directional inputs are being pressed. but it doesnt solve my problem.

Maraakate Joined: 05 Jun 2010 Posts: 757 Location: Pennsylvania, USA	Posted: Sat Jul 08, 2017 8:57 am
	I had this problem and resolved it by making jumper wires from the db-9 connector the EMI filters on the offending controller port. Worth a try if you're desperate. Probably just needed solder reflowed for the EMI filter.

asynchronous Joined: 14 Aug 2000 Posts: 742 Location: Adelaide, Australia	SMS controller (port?) issue Posted: Sat Jul 08, 2017 10:05 am
	Ah, sorry my mistake. I forgot the data bus is not terminated on the master system. There is a resistor array on the board but, double checking the schematics, it's not on the data bus. So disabling IC4 would have meant a floating data bus when the CPU read the IO port. If your soldering is good and the PCB tracks and EMI filters are in good nick then it does sound like IC4 is stuffed. But.....it's strange that the problem went away with the broken Genesis controller. The Genesis controller multiplexes the L,R,A,B,C,Start buttons so the Left line is driven high when the D-pad is not pressed left. Maybe the Left inputs on IC4 are stuffed but when they're driven high they work OK?? In that case maybe, just maybe, soldering on external pull up resistors on the Left inputs may fix it?

lazyjackal Joined: 06 Jul 2017 Posts: 8	Posted: Sun Jul 09, 2017 12:20 am
lazyjackal Joined: 06 Jul 2017 Posts: 8	asynchronous wrote Maybe the Left inputs on IC4 are stuffed but when they're driven high they work OK?? In that case maybe, just maybe, soldering on external pull up resistors on the Left inputs may fix it? Holy crap that would be a weird but simple fix. I am going to try that tonight, there were some resistors on that 3rd party genesis controller. I will check them for poops and giggles and report back their values.

lazyjackal Joined: 06 Jul 2017 Posts: 8	Posted: Mon Jul 10, 2017 2:37 am
lazyjackal Joined: 06 Jul 2017 Posts: 8	Okay, so this is the controller in question: Front Back All the resistors are 47k Ohms, the electrolytic capacitor is 10uf, the little ceramic decoupling capacitor is .1uf. I tried a single 47k, and two 47k resistors in series. The results were erratic. I could move right but the left input would intermittently and quickly move the character on screen back left. Dropping down to 20k though, yields different results. I can freely move up, down, and right but cannot move left at all. Same results all the way down to 5.1k where it becomes erratic again. So the saga continues.

asynchronous Joined: 14 Aug 2000 Posts: 742 Location: Adelaide, Australia	SMS controller (port?) issue Posted: Tue Jul 11, 2017 11:38 am
	OK, what are you connecting the resistor to exactly?

Apocalypse Joined: 26 Feb 2021 Posts: 159	Posted: Sat Feb 12, 2022 3:51 pm
Apocalypse Joined: 26 Feb 2021 Posts: 159	I know it's an old thread again, only confirms information doesn't spread well... Back in 2017 I fixed a MS2 with a similar issue (IIRC 2 directions weren't working). It was due to an internal fault in the 315-5237 custom chip. I reproduced the faulty part of the chip with an external PLD chip (GAL16V8), it was well documented on my blog (now gone). The good news is both the 315-5126 and 315-5237 are easy to fully reproduce in a CPLD. P.S.: Here's a picture of my bodge repair 20190201_084459.jpg (2.99 MB)

Apocalypse Joined: 26 Feb 2021 Posts: 159	Posted: Thu Feb 17, 2022 7:42 pm
Apocalypse Joined: 26 Feb 2021 Posts: 159	Just for "fun" I sucessfully reproduced both the 315-5216 and 315-5237 I/O chips in a CPLD but... I wanted to use something 5V compatible to avoid voltage shifting, so made a first implementation in an ATF1504 (64 macrocells, both chips fit easily). Problem is the TQFP44 package (that could have fitted between the two rows of the dip footprint, tightly yes...) only allows 36 I/Os and we need 40 for the 315-5216 and 46 for the 315-5237. So for testing I used a big a** PLCC84 chip, overhanging over the nearby chips... Now I'm looking at the Cool Runner CPLDs from Xilinx, 3.3V but with 5V compatible I/Os (so would only need an extra 3.3 regulator) and allowing 41 I/Os for the TQFP44 package. Good enough for the 315-5216 but still not good for the 315-5237. I'm exploring more options. BTW I was shocked to discover the WebPack toolkit from Xilinx was almost 16GB! Seriously?!

asynchronous Joined: 14 Aug 2000 Posts: 742 Location: Adelaide, Australia	SMS controller (port?) issue Posted: Fri Feb 18, 2022 2:36 am
	Nice. I found the same issue regarding the lack of I/O pins when looking the ATF CPLD series for this. I was thinking of using a couple of 74LS157's or 74LS241's next to the CPLD to handle the port DC and DD reads. The TQFP44 and two SOIC20 would fit.

Apocalypse Joined: 26 Feb 2021 Posts: 159	Posted: Sat Feb 19, 2022 8:53 am
Apocalypse Joined: 26 Feb 2021 Posts: 159	I also thought of using TTL chips or two TQFP44 CPLDs but couldn't resolve to it, once you get into CPLD territory you expect total integration. After a bit of research I chose to target the XC9572XL CPLD from Xilinx (72 macrocells, overkill, I need only half of those) in VQ64 package (12x12mm, 56 I/Os, 3.3V with 5V tolerant I/Os, will need an extra regulator, no biggie) Waiting for the chips to come now. P.S.: I found other chips that could do the job but in BGA package only if you want enough I/Os, not hobbyist friendly IMO.

Forums

View topic - SMS controller (port?) issue