Robotron infinite rug pattern & ram error

opt2not

Well-known member
Joined
May 24, 2009
Messages
1,580
Reaction score
241
Location
Orange County, California
Alright, I'm stumped.
Been working with Ken on this one, but I think now it's time to get some more eyes on this thing.

My Robotron Cabaret was playing awesome till it froze while I was having a nice 5 million run ....grrrrr!
It froze then went straight to the rug-pattern and gave me the initial "Factory reset" error regarding the cmos.

So, I figured it was a battery thing...turned the cab off, went into the back, noticed two things:

1. original battery holder and the AA batteries was all crusted up and fuzzy looking. I ended up pulling the main board out, and installing a new lithium battery holder I got from bob roberts. While I was on that board, I re-flowed the solder on all the power headers...and did the Ram upgrade to 4164's with power adapter plug, again purchased from bob.

2. the power supply had a few caps swollen. Now I just recapped this thing a year ago, caps should last longer than that! So I was concerned that perhaps the caps I got from bob had a couple defective ones that were on their last legs...so I recapped the Power Supply with new ones. Also swapped out the bridge-rectifier at BR1, heat-sink transistors with new ones, AND replaced all the power headers and re-crimped new IDC connectors to and from the Power supply.

Practically a full shotgun.

Now, when I power up the machine, I get RAM error 1-2-L and the rug pattern persists and doesn't progress past that. It just keeps repeating the rug pattern.

I swapped around RAMS at that location, but am still getting the same error at that spot. I thought perhaps the socket is gone, so I pulled the board again, and replaced the socket.
Still the same error.

Another thing I noticed is that the F2 fuse on the power supply keeps blowing. I believe for compact games, this line is used for coin lights/switches and general illumination. Don't know if this line is used anywhere else for operation...
I checked voltages at the 1J1 header: the 5v line is coming in at 4.8v, and the 12v line is sitting at 11.9v.

Ken suggested that I re-crimp the ribbon cable from the Rom to the CPU board. I did so, and made sure to checked the continuity of all pins, first on the ribbon cable to the ROM board, then from the ROM connection header pins to the CPU board's headers. Re-crimping was successful, but I still get the same 1-2-L RAM error.

So I went back and metered the Test Points (from the schematics) on the Power Supply Board just to make sure I'm not going crazy.

With 4J2 plugged-in I get:
TP1 = 4.8
TP3 = 11.5
TP4 = -4.8
TP5 = 11.2
TP6 = -15.7

With 4J2 unplugged I get:
TP1 = 4.8
TP3 = 13.6
TP4 = -4.8
TP5 = 14.6
TP6 = -15.0

What's interesting is the voltage drop under-load on TP3 and TP5...is this normal? I just want to eliminate the possibility that my PS board is at fault here. I've pulled out that board so many times now and have practically checked every component and pin on it.

Ken also suggested I re-flow the fuse holders and squeeze the clamps together so they make good contact with the fuse. Did that, no change. Fuses are new and clean, they're fitted snugly into the clips and are making good contact.

I really don't know what else to try now. I have a logic probe and multi-meter, and am handy with the soldering iron.

Anyone have any suggestions?
 
What's interesting is the voltage drop under-load on TP3 and TP5...is this normal?

Voltages will drop under load, that is normal.

Your voltages look fine too, 4.8 the board should boot.

Hopefully you cleaned up all the battery acid and checked for damage. If not, that's the first thing I would do.

Start at Pin 37 of the CPU, if that's going from High to Low when the rug pattern repeats then the Watch Dog is barking.

You could disable the Watch Dog (pad between 5I and 5H) and see if the game gets past the Rug Pattern. If not, then you need to follow the circuit back from the reset line and check which component is causing the CPU to reset.

IC's to check in the circuit would be 6F, 5H, 5I, etc...

You should also check your ROMS to make sure they are good.

If swapping the RAM around didn't change the problem, then it could be an address decoder issue or data access IC.
 
Hopefully you cleaned up all the battery acid and checked for damage. If not, that's the first thing I would do.

Yup, cleaned it up good. The acid was only on the metal battery holder, and didn't leak anywhere on the board. I checked all components around and below the batter holder and they were clean.

Start at Pin 37 of the CPU, if that's going from High to Low when the rug pattern repeats then the Watch Dog is barking.

You could disable the Watch Dog (pad between 5I and 5H) and see if the game gets past the Rug Pattern. If not, then you need to follow the circuit back from the reset line and check which component is causing the CPU to reset.

IC's to check in the circuit would be 6F, 5H, 5I, etc...

Thanks, this is exactly what I thought I should be moving onto now...probing IC's for damage. I'll grab some schematics for those IC's and start poking around. Anything in particular in for those IC should I be looking for? High pins, Low pins etc?

You should also check your ROMS to make sure they are good.

Logic probe 'em?

If swapping the RAM around didn't change the problem, then it could be an address decoder issue or data access IC.

I believe that's what Yellowdog initially guessed:
YellowDog said:
What you are seeing is probably an issue with memory addressing. During normal operation in the screen memory addresses reads come from the ROM and writes go to the RAM. During the RAM test, it is switched into a mode where READS and WRITES go to the RAM. What sometimes happens is that the second bank doesn't go into this mode. It stays in Write->RAM Read<-ROM and so the RAM test fails.

I know I have fixed this at least once before, but I haven't found my notes on what causes this and the fix. I actually need to find this, because I have ROM card that is doing this now.
 
Did you replace the UA728 chip on the PS? It's a voltage regulator IC. I bet if you run the game on a switcher it'll make it past the rug pattern.
 
Did you replace the UA728 chip on the PS? It's a voltage regulator IC. I bet if you run the game on a switcher it'll make it past the rug pattern.

Yep, I replaced the voltage IC's on the PS along with the full shotgun of caps, BR, and headers. (came with the bob roberts deluxe PS kit)

Thanks for the suggestion, but I'm no-way going to throw a switcher in this cab, for 1. I want it original, and 2. the Linear PS is more reliable in the long-run and doesn't mess with my hi-score saving. Switchers are prone to do that.
 
Thanks, this is exactly what I thought I should be moving onto now...probing IC's for damage. I'll grab some schematics for those IC's and start poking around. Anything in particular in for those IC should I be looking for? High pins, Low pins etc?

Depends on the chip you're looking at. You'll have to look at the schematics to see what they should be doing.

Logic probe 'em?

No, you need to read them with an EPROM Programmer and compare them to their images from MAME.

I believe that's what Yellowdog initially guessed:

I would start with Yellow Dog's suggestion, he's knows a lot more about the Williams stuff then I do.
 
Not saying it is the problem, but I had a similar problem and it ended up being a faint cracked solder joint on one of the 5V pins of 4J2 on the power supply board. As the game got warmer while playing the resistance of the solder joint increased causing a larger voltage drop. Eventually the drop was high enough to cause a reset or random RAM error. Replaced 4J2 and it hasn't done that since. Imagine you could just reflow it but typically the pins have oxidized over the years and need cleaning beforehand. I just replaced it to be sure of a good connection.
 
Start at Pin 37 of the CPU, if that's going from High to Low when the rug pattern repeats then the Watch Dog is barking.

Probed it and found that Pin 37 stays on High till the pattern repeats, flips to Low for half-a-second, then back to High. Repeat.


You could disable the Watch Dog (pad between 5I and 5H) and see if the game gets past the Rug Pattern. If not, then you need to follow the circuit back from the reset line and check which component is causing the CPU to reset.

I disabled the Watch Dog, and after the rug pattern it goes to a black screen and stays that way. No game...

IC's to check in the circuit would be 6F, 5H, 5I, etc...

Here are my probing results in that circuit (am I missing any IC's here?), I linked the datasheets I'm referencing. I found dead pins at the 5I and 6H IC's, but I'm not sure if they are the cause of this issue. In the Robotron schematics I don't see where these lines are connected too. I'll have to go over the schematics more, but if anyone sees anything weird here, please let me know...

6F (7408 Quad And Gate):
1 - Low
2 - Low
3 - Low
4 - High
5 - High
6 - High
7 - GND
8 - Pulse Low
9 - Pulse Low
10 - Pulse Low
11 - Pulse Low
12 - Pulse Low
13 - Pulse Low
14 - VCC

5H (7402 Quad 2-input NOR Gate)
1 - Pulse Low
2 - Pulse Low
3 - Pulse Low
4 - Pulse Low
5 - Pulse Low
6 - Pulse Low
7 - GND
8 - Pulse High
9 - Pulse High
10 - Pulse Low
11 - Low
12 - Low
13 - High
14 - VCC


5I - (74LS393 Dual 4-bit bin counter)
1 - High
2 - Pulse Low
3 - Rapid Pulse Low
4 - Low
5 - Low
6 - Low
7 - GND
8 - Low
9 - Low
10 - Low
11 - Low
12 - DEAD (should be active High pulse?)
13 - Low
14 - VCC


6I (7410 Triple 3-input NAND)
1 - Low Pulse
2 - Low Pulse
3 - Low Pulse
4 - Low Pulse
5 - High, but flutters during rug reset
6 - High Pulse
7 - GND
8 - High
9 - Low
10 - Low
11 - Low Pulse
12 - Low Pulse but no beeping sound
13 - Low Pulse
14 - VCC


5G (74LS133 Dual 2 to 4 Line Decoder)
1 - Low + High Pulse
2 - Low + High Pulse
3 - Low Pulse
4 - Low + High Pulse
5 - Low + High Pulse
6 - Low + High Pulse
7 - Low Pulse
8 - GND
9 - Low + High Pulse
10 - Low Pulse
11 - Low Pulse
12 - Low Pulse
13 - Low Pulse
14 - Low Pulse
15 - Low Pulse
16 - VCC

6E (7432 Quad 2-input Or Gate)
1 - High Pulse
2 - High
3 - High
4 - Low
5 - High w/ rapid activity during rug pattern
6 - High w/ rapid activity during rug pattern
7 - GND
8 - High
9 - High w/ rapid activity during rug pattern
10 - Low Pulse
11 - High, but pulsing during rug pattern
12 - High
13 - Low Pulse
14 - VCC

6H (74LS139 Dual 2 to 4 Line Decoder)
1 - High Pulse
2 - Low Pulse
3 - High, but flips at reset, then back
4 - High
5 - High
6 - High
7 - Pulse High
8 - GND
9 - High
10 - High
11 - High
12 - High
13 - Low
14 - Low
15 - DEAD (should be pulsing Low?)
16 - VCC


tas said:
Not saying it is the problem, but I had a similar problem and it ended up being a faint cracked solder joint on one of the 5V pins of 4J2 on the power supply board. As the game got warmer while playing the resistance of the solder joint increased causing a larger voltage drop. Eventually the drop was high enough to cause a reset or random RAM error. Replaced 4J2 and it hasn't done that since. Imagine you could just reflow it but typically the pins have oxidized over the years and need cleaning beforehand. I just replaced it to be sure of a good connection.

Aye, I replaced those headers with new ones...it was one of the first things I did. I think the issue may be deeper...
 
Did you replace the UA728 chip on the PS? It's a voltage regulator IC. I bet if you run the game on a switcher it'll make it past the rug pattern.

No such animal The voltage regulators are LM723's

Yep, I replaced the voltage IC's on the PS along with the full shotgun of caps, BR, and headers. (came with the bob roberts deluxe PS kit)

Thanks for the suggestion, but I'm no-way going to throw a switcher in this cab, for 1. I want it original, and 2. the Linear PS is more reliable in the long-run and doesn't mess with my hi-score saving. Switchers are prone to do that.

If you did not replace the 18KuF cap, it may be part of your problem. It does not come in the cap kit from Bob. You also need to verify that the connector at 4J1 is not burnt up at pins 1, 3, 5, as well as the connector at 4J4 all pins.

Do NOT use the test points on the power supply for assuming the voltages are good. Test for voltages on the boards themselves at the chips and or headers to see what you are actually getting on the board. 4.8vdc at the power supply will usually leave you in rug, due to it being less than 4.8vdc at the chips on the boards.

In addition remove R10 from your power supply altogether, and test it again.
Dok
 
Fuse 2 is blowing because your coin door lights are grounded to either the coin door or the coin door lockout coils. Remove the coin door lamps and the fuse should quit blowing.
 
Thanks Dok, I appreciate your input on this too.

If you did not replace the 18KuF cap, it may be part of your problem.
It does not come in the cap kit from Bob.
Actually the Deluxe PS Repair Kit from bob comes with the big 18,000uF cap as well, it's just radial not axial, so I attached two small 20 gauge wires to the leads and soldered them to their holes.

You also need to verify that the connector at 4J1 is not burnt up at pins 1, 3, 5, as well as the connector at 4J4 all pins.
I changed the header pins with new ones, as well as re-crimped the pins on the wire-connectors plugged into them. But I was hearing a slight crackle from 4J1, so I checked it again and pin 5 was burnt. Changed that out too, but still get the same RAM error and reset rug.

Do NOT use the test points on the power supply for assuming the voltages are good. Test for voltages on the boards themselves at the chips and or headers to see what you are actually getting on the board. 4.8vdc at the power supply will usually leave you in rug, due to it being less than 4.8vdc at the chips on the boards.
Good to know. I just checked the voltages at the CPU board, checking each IC (and RAM) and they do actually have 4.8v.

Checked the header connections at the CPU power input, and have a solid 12v on that line, and again 4.8v coming in.
4.8v at the chips should be good right?

In addition remove R10 from your power supply altogether, and test it again.
Dok

Same RAM 1-2-L error and rug-reset came up after removing this. My 5v line was at 6.1v, so I turned the machine off for fear that I'd burn out anything and returned the resistor back to R10.

Fuse 2 is blowing because your coin door lights are grounded to either the coin door or the coin door lockout coils. Remove the coin door lamps and the fuse should quit blowing.
Yup, that's it!
One problem solved, and the big whoppin' one to go.

Any advice on where to look from here? In the meantime I'll pour over these schematics more...
 
You may have crapped the CMOS RAM, it is pretty common. Usually it will get very hot to the touch when it is bad. Run the game for about 3 minutes and touch the CMOS RAM with the game powered up.
 
You may have crapped the CMOS RAM, it is pretty common. Usually it will get very hot to the touch when it is bad. Run the game for about 3 minutes and touch the CMOS RAM with the game powered up.

Be quick with the first touch! I burned the phuque out of my fingertip probing for bad ram on my Defender once. That damn chip was 15 billion degrees, at least! =)
 
I appreciate your guy's help. :)

The CMOS chip at 1C is cold after the power stays on for over 5 mins. It has no heat what-so-ever, shouldn't it be slightly warm at least?

I checked the voltage and probed it's pins:

1C (5114 CMOS Static RAM)

1 - High + Low lights up, pulsing
2 - High + Low lights up, pulsing
3 - High + Low lights up, pulsing
4 - High + Low lights up, pulsing
5 - Low pulse
6 - Low pulse
7 - High + Low lights up, pulsing
8 (CS) - Stuck High
9 - GND
10 (WE)- High Pulse
11 - Low pulse
12 - Low pulse
13 - Low pulse
14 - Low pulse
15 - High Pulse
16 - Low pulse
17 - Low pulse
18 - 4.7v

I'm suspecting this CMOS is malfunctioning, but I'm not a electronics wiz by any means.
The first thing I noticed about this chip is that it's running at 4.7v, whereas all other IC's on the board are at a steady 4.8v.

I did a bit of searching and found this thread, talking about testing the 5114. I quoted channelmanic's info here:


Chip Select and Output Enable lines are active high.

If they have a bar over them then they are active low. (In text it's written like CS* and OE* for active low)

The chip is only selected if the lines are set to an active state.

WE (WE* for active low), also called R/W* (read / write*) is in read mode if not active and write mode if in the active state.

Those are all input lines for the chip - used for control.

To understand more, look at this post on my repair log site: http://newlifegames.net/nlg/index.php?topic=18.0

RJ


They should read high until the RAM needs to be read.

When reading, CS* goes low and WE* stays high.

When writing, both CS* and WE* go low.

When not being accessed then both CS* and WE* stay high.

If you are getting a RAM error on the board and the logic probe is never seeing CS* or WE* going low then the RAM is not being written to. When that happens you have to look at the address decoding circuitry and the R/W signals to see where the failure is.

Going off this, I checked pin 8 (CS*) and it was stuck High, and never goes low. Pin 10 (WE*) seems to be pulsing high, then flutters high activity during the rug-reset. So it sounds to me that the IC isn't switching CS* to low, which might be my issue.

Any of this sound sane?
 
Last edited:
Anyone know of a good place to easily source these 5114's?

I can't seem to find them in my local electronic shops.
Arcadechips has a minimum of $15 to complete the order, and bobroberts only accepts bank drafts or money orders from "international" customers...MO's costs me $7 at the bank. :(
 
Thanks Ken, I just placed an order with them.

Hoping I didn't jump-the-gun here, but from what I've been reading the CMOS sounds like it's the culprit.

We'll see when I swap it out. ;)
 
What does it say about the cmos in test mode? I'd spend the extra $$$ and get the NV ram replacement since you're going to have to de-solder that chip anyway.
 
Back
Top Bottom