3h. When things don't work: Infrared Optic Switches
Make sure to check fuses F601 and F602 (all WPC games). F601 is used for +62 volts, and F602 is used for -113, -125 volts. On WPC-S and before, these are 3/8 amp fast-blo 1.25' fuses. On WPC-95, these are T0.315 amp 5x20mm fuses. It's easier to test voltages at the dot matrix display itself than at the controller board. A Bally Serial number identifies the year the Bally Walk-In was manufactured. This information is useful in determining correct model parts, as many parts have evolved in design over the years. The Serial number is required to determine warranty coverage. If you ever need assistance, your Serial number allows the Bally Representative to quickly.
Several different optos used in Williams games. The 'U' shapedoptos are used for Fliptronics flippers, Twilight Zone clocks, etc.These consist of a transmitter and receiver in one package. The stand-upoptos are two parts: the green board opto stand-up is the transmitter,and the blue board opto stand-up is the receiver. The transmitterLED is larger and protrudes further from its case. The single diodeshown is a replacement LED transmitter for the stand-up optos, andfor opto boards used in ball troughs, etc.
Optos are used on fliptronics
flipper switches. Note the plastic
activator arm that moves between
the 'U' shaped optos. Originally
Williams made these from metal,
but switched to plastic to save
money. The plastic version can
often warp so they don't clear
the opto, causing a flipper not
to work.
Two parts to a opto switch.
Each opto switch has two parts; a transmitter and a receiver.The transmitter is a infra-red LED (light emitting diode).The receiver is a light sensitive transistor. The transmitteris always on when a game is powered on. If the light beamfrom the transmitter is interrupted, then this registers theswitch as 'open'.
Because the transmitter is always on and producing light (andhence heat), the transmitter is the part that fails 98% of the time in a opto switch.
Testing Opto Switches.
Testing infrared optos switches is no different than testingmechanical switches (to a point). Just use the WPC internal test software. Press the 'Begin Test' button inside the coin door, and go to the Test menu. Select the 'switch edge' test. Activate an opto switch by passing somethingin front of it. The display will indicate if the switch works.
If an opto switch doesn't work, first check that your +12 volts isworking. If you have blown the +12 volt fuse, there will be no powerto the optos, and they won't work.
If there is +12 volts going to the opto, there is a good chancethe transmitter has failed. To test the receiver, firstput the game into the 'switch edge' test. Then block the optotransmitter with a piece of tape. Now shine a penlight flash lightinto the receiver. The switch should 'close'. When you removethe light, the switch should 'open'. If the receiver is workingproperely, there's a good chance the transmitter has burned out.
If you have +12 volts in your game (Hint: do other optos work?),and the opto switch doesn't register in the diagnostic test,your opto transmitter is probably burnt. The receiverside of an opto switch almost never dies. That's because itonly senses light, and doesn't produce light. The transmitterwill be the offending unit 98% of the time. But thereis a good way to test the transmitter side of an opto switch.
Optos on Newer WPC games.
Older WPC games use optos with straight resistive photocells. Newer WPC games use a transistor gate photocell. This means the internaltransistor can die, even if the photocell part of the opto is OK. Keepthis in mind; if an opto transmitter tests good (with your Radio Shacktest card), the opto could still not function properely. Replacing theopto is the only thing that will fix it.
The Opto Receiver and Transmitter Tests Good, now what?
If the receiver tests good with the penlight flashlight, andthe transmitter tests good with the infrared sensor card, thereis one more thing that could be wrong. This would be the LM339voltage comparitor on the opto board. If everything else checksout, replace ALL the LM339 chips (and use sockets!) on theopto board (there are usually two of these chips on the opto board).
WPC Fliptronics Flipper Optos.
If a WPC Fliptronics flipper doesn't work, and it's not a coil, transistor orwiring related problem, you should suspect the flipper opto board. This boardhas two 'U' shaped optos that detects the flipper button. These boards are all madewith two optos, even if the game only has two flippers instead of four.
Use your infrared sensor card to determine if the opto is working on the flipper board. If you suspect a problem with this opto (and don't havea infrared sensor card), you can swap the left and right flipper opto boards, and see if the problem moves to the opposite flipper. Note: you musthave both flipper opto boards plugged in for this test to work! Opto board power is jumped through the left opto board. Both opto boardsmust be plugged in for the right one to work!
If indeed one of the flipper optos is bad, and your game only has twoflippers, you can reverse the two optos on the bad flipper opto board.One of the optos will be unused since the game only has two flippers,instead of four. Mark the bad opto, and its position on the opto board.Then unsolder both optos, and throw the bad one away. Then re-solderthe good opto into the marked position on the flipper opto board.
Weak Flippers and Bad LM339's on the Fliptronics Board.
On WPC fliptronics to WPC-S board, chips U4 and U6 (LM339) on fliptronics board can fail. On WPC-95, these LM339 chips are on the CPU boardat locations U25 and/or U26. This will makethe flipper opto boards seem like they are not work. Swap the twoflipper boards to test this. If the problem doesn't change, suspectthe LM339 chip(s). These LM339 chips canalso become 'leaky'. This will make flippers seem very weak.A bad LM339 can also give the indication that the EOS switch is bad.
If there is a marginal flipper switch reading, this causes the high powered sideof the flipper to rapidly oscillate between on and off. The holding side of the flippercoil never engages. This problem will cause the flipper coil to getvery hot in a short time.
Replacement Optos.
Unfortunately, optos are fairly expensive (compared to micro-switches).For example, if you are repairing your Twilight Zone clock (which meansreplacing all eight of the 'U' shaped optos), this can get expensive.The cheapest I've found these optos is from Competitive Products Corp (800-562-7283)for $5 each (long leads too!). For non 'U' shaped opto applications, you can get infrared opto LED transmitters cheaply (get a transmitter voltage of 1.0 to 1.4 volts). Note the color of the opto doesn't matter, just as long as it's aninfrared emitting LED. Radio Shack sells the infrared LED, part number 276-143C(replaces Williams A-14231). Radio Shack also sells an infrared transistor (receiver), part number 276-145A (replaces Williams A-14232).
The industry part number for the 'U' shaped optos is QVE11233,with a standard sensitivity of .0110. Unfortunately, Williams requiresa higher sensitivity opto for their applications. This meansthe cheap $1 optos from most electronic supply houses won't work, as theirsensitivity rating isn't high enough. If youare shopping for these 'U' optos, keep this in mind. You should be lookingfor part number QVE11233.0086, where .0086 is the increased sensitivity rating.
That is all that is required to adjust an eddy sensor. To test thesensor, put the game into WPC diagnostic's first switch test. Thenmove a pinball over the playfield area where the eddy sensor islocated. The switch should activate on display.
Twilight Zone Eddy's.
Note eddy sensors were used as early as Twlight Zone. The eddysensor in TZ are different than the later sensors, and does NOT have an adjustment pot (they also are called a different name, likethe 'Trough Proximity' board). The eddy sensor that causes the most trouble in Twilight Zone is the sensor by the ball trough (switch# 26). Usually the problem is the molex connector. Just taking the two pin molex connectoroff and putting it back on its header pins will usually the problem. If not,this small board often needs to have its molex header pins resoldered. Also, it is possible for the TDA0161 (Williams part number 5370-13452-00) chip to die on this board. If you don't want to replace just this chip, the wholeproximity board is available for under $15.
Starting in 1993 with Indiana Jones, a new ball trough design was used thatinstead relied on gravity to feed the ball into the trough. This saved one coil(the outhole coil was no longer needed). The new designalso used opto switches instead of mechanical switches. This allowedone ball trough design to be used in all Williams games, regardless ofthe number of balls used in the game. The ball trough could nowcomfortably hold from one to seven balls (depending on the game).
To fix this problem, Williams redesigned the attachmentpoints for the two opto boards. Instead of being bolteddirectly to the trough, the mounting holes on the optoboards were enlarged (and one hole moved). Then special rubber gromets where inserted into the holes. Lastly,1/4' metal tubes where inserted through the rubber gromets.When the opto board bolts where tightened down, they tightened onthe metal tubes. This allowed the opto boards to 'float'on the rubber gromet, reducing vibration considerably.
Ball Trough Upgrade Kit.
If you want to upgrade your Indiana Jones ball troughto the current design (and solve your random multi-ballproblems), you can order an upgrade kit, part# A-18244.This includes two new opto boards, and all the mountinghardware needed (the mounting hardware is absolutelynecessary). At $50, this is an expensive kit! Withsome thought and a good hardware store, you can modify your original opto boards to mount like the newer style. It willrequire enlarging the mounting holes on your existing optoboards. You will also need to find some rubber gromets and small metal tubing(hobby shops often carry this).And if you use the newer metal trough #A-16809-2 (as currently sold by Williams), you will alsoneed to move one of the large blueresistors to the back of the board, and drill a new mounting hole in theopto board. If you want to just be done with it (and spending $50 is not an issue), just order the upgrade kit from Williams, part# A-18244, and get the twonew trough opto boards and the mounting hardware.
To fix this, you can often use a Dremel tool and grindthe divots out of the metal. If this doesn't work, youcan order a new ball trough, part number A-16809-2. Thisnewer design of the ball trough should last longer anddivot less.
More Random Multiball: the Ball Trough Optic Resistors.
The ball trough optic boards have several large blue resistors mounted to them.Since these board get a fair amount of shock and vibration from balls,often these resistors can break. If this happens, you'll can get random(and continual) multiball. Check these large blue power resistors for breaks or cracks. Usually the resistor leads break right where theyconnect to the circuit board.
WPC Alpha Numeric Score Display Problems.
The first three WPC games that used AlphaNumeric displays have a commonproblem. The resistors R48 and R49 (39k ohm) on the AlphaNumeric Display boardoften fail and go open, or go out of spec. This can cause all the score displaysin the game to work very weak, or not work at all. Before replacing a scoredisplay, replace BOTH of these 39k resistors with 'flame proof' 1 or 2 wattversions. See the Williams System 11 repair guide athttp://marvin3m.com/sys11/index3.htmfor more informationon repairing AlphaNumeric score displays. All the information there appliesto these three WPC games (though the component label numbers will bedifferent).
But the really bad news about DMD's that are failing is how they die.When a DMD starts to get blurry or displays gaps, the power requirementsfor the display dramatically increases. This stresses the dot matrixcontroller board. If the displayis not replaced, the controller board can fail too. Sometimes thecontroller board burns beyond repair.
The moral to this story is to replace a marginally bad DMD witha new display as soon as possible. Don't postpone the inevitable. You can get a new DMD from Competitive Products Corp (800-562-7283)or Pinball Resource (914-473-7114). The whole assembly is about $115.
Buy an entire DMD display glass and board, or just a new Glass?
You can just buy a new dot matrix glass only, which will also solvethe 'outgas' problem. These are available for about $65, which is almost half the price of buying both the display and its attached circuit board.Don't be cheap; just spend the extra $50 and get both the display andits attached circuit board. Installing a new glass intothe surrounding board is A LOT of work. And games produced in 1993 and laterdon't have 'pin' style glasses, so these display glasses alone are NOT replacable.Even if you have a 'pin' style DMD, it's just not worth the troubleto unsolder 132+32 pins, install the new glass, and resolder all thosepins again. It's a solid two hours worth of eye straining work, andit's very easy to make a mistake. It's just not worth the trouble.
Diagnosing Other Problems.
If you are sure the display itself is working, there are someother things to check when a DMD doesn't work.
Make sure to check fuses F601 and F602 (all WPC games). F601 isused for +62 volts, and F602 is used for -113, -125 volts. On WPC-Sand before, these are 3/8 amp fast-blo 1.25' fuses. On WPC-95,these are T0.315 amp 5x20mm fuses.
Display and fuses good, but still missing or low voltages.
If you have checked the fuses, and you know the display itself isgood (tested in another game), you will need to rebuild the Dot matrixcontroller board. Usually it's either or both of the MJE power transistors (the ones with big heat sinks) at locations Q1 (WPC-95) or Q3 (WPC-S andprior), and Q6. If you want to rebuild the entire power sections, hereare the components to replace.
Also you will need to check resistor values. They should bewithin 10% of spec:
Generally either Q1/Q3 or Q6 will go bad. Ifeither of these fail, both should be replaced(make sureyou use white heat sink compound when replacing these, andmake sure you have them screwed tightly to their heat sinks). The MJEat Q7 comes after Q6, so it usually survives. Also the smallerswitching transistors that are connected to Q1/Q3 and Q6 should be replaced (MPSD02 and MPSD52). Finally the zener diodes 1N4759 (62 volts) should also bereplaced. The diode 1N4742 (12 volts, D6, D8 on WPC-S and prior and D3, D5 on WPC-95) generally do not go bad. On WPC-S and prior, check the 120 ohm 1/2 watt resistors at R4, R5and 47k ohms 1/2 watt at R3, R6, R12, R13. Also check the4.7 ohms 5 watt and 1.8k ohms, 5 watt resistors. Replace anyresistors that are out of tolerence or that appear burnt. Always mountthe resistors slightly above the board to allow air flow below them.
Cloudy Dot Matrix Display.
This is usually caused by heat related problems. Fixing this couldbe as simple as adding new white heat sink compound to Q1/Q3, Q6, Q7.Also make sure they are tight to their heat sink.
Wavy Hum-bar in the Dot Matrix Display.
If you WPC-S or earlier dot matrix display has a 'wavy hum-bar', try replacingcapacitors C6, C9 and C10 on the dot matrix controller board. Theseare .1 mfd 500 volt caps that filter the DMD high voltages. If theycaps fail, you will get the hum bars.
Crystallized Solder Joints.
If a DMD display is not displaying correctly, and the voltagesseem Ok, also check this. It's common for the solder jointson the zener diodes in the power section to crystallize, causingheat damage, excessive resistance, and finally a lost ofvoltage regulation. This can then lead to a failed DMD and damagedpower circuits. These diodes are D3, D4, D5, D6 on the dot matrixcontroller board.
Replacing Q2/Q10 and Q4/Q5 on the Dot Matrix Controller board.
On WPC-S and prior games, the dot matrix controller board usestwo MPSD02 and MPSD52 transistors. More common transistorsare available for replacement. A 2N5551 can be substituted for the MPSD02,and a 2N5401 can be used for the MPSD42.
CPU Flash Codes WPC-S and Prior.
D19 is labeled 'blanking', D20 is labeled 'diagnostic' and D21 is labeled '+5vdc'.
Problem Power-On CPU D20 (diagnostic) Flash Codes:
Sound Board Error Beeps pre WPC-DCS (WPC alpha-numeric, WPC dot-matrix and WPC fliptronics.
Sound Board Error Beeps WPC-DCS and WPC-S.
Driver Board Flash Codes WPC-S and Prior.
WPC-95 CPU Flash Codes.
LED201 is labeled 'blanking', LED202 is labeled 'power' and LED203 is labeled 'diagnostics'.
Problem Power-On CPU LED203 (diagnostic) Flash Codes:
WPC-95 Audio/Video LED.
Problem Power-On Audio/Visual Board Beep Error Codes:
WPC-95 Driver Board LED's.
Most often, this error occurs because the three 'AA' batteries on the CPU board havedied. These batteries should be replaced every year with good quality alkaline batteries(batteries are cheap, battery damage is expensive). The three batteries must keep at least +4 volts of power to the U8RAM chip for it to remember. When power goes below +4 volts, memory reset can occur(and you get the 'Factory Settings Restored' error message).
Is power getting to the U8 RAM chip?
With your game off and batteries installed, put your DDM on DC volts andput the black lead on the backbox ground strap. Then put the red leadon diode D2 on the CPU board. The banded side of the diode should showabout .5 volts less than the non-banded side (which should be about 4.3 volts).If only one side of the diode shows voltage, this 1N4148 diode is bad.
Next test for voltage at the CPU U8 RAM chip. With the game off, you should get about 4.3 volts DC at pins 26, 27 or 28 ofchip U8. If you don't, the battery voltage is not getting to the U8 RAM chip,and the game will boot up with the 'Factory Settings Restored' error. Notepin 28 of the 28 pin U8 chip is in the same position as pin 1 ofthe chip, but on the opposite row of pins. Pin 1 is designated with an impressed 'dot' right on the top of the chip.
You can still have problems even if you installed new batteries and all the voltages check out. If your game is still giving 'Factory Setting Restored' or 'Set Time and Date' errors, you may have a bad CPU U8 RAM chip. But make sure you double check that battery holder. Even minor corrosion can cause this problem. The voltages may all check out, but the corrosion may be enough to limit CURRENT, and cause this problem. The U8 RAM chip is a 6264-L or 2064 RAMchip.
If you install new batteries with the game turned on, the machine will notforget the old option settings or bookkeeping totals.
The MOV is the green disc soldered across the
lugs of the radio frequency interference filter.
Volume up FULL and Can't turn it Down.
The volume control on all WPC games is electronic. On pre-WPC95 games, thisis controlled by an electronic prom pot. This E-pot is a X9503, at location U5 on the sound board. If turning the volume up or down has no effect, and the volumeis stuck on full blast, this is the first component that should be checked.Also the capacitor C18 (47 mfd, 25 volts) that connects to the E-pot can fail too, and shouldbe checked.
Lots of Static.
Problem sound boards can produce a large amount of static. The TL084 quad Op-Amp (U21 on WPC-S and prior)can be the cause of this. Also the TDA2030 amp can also cause this. Finally the large filtering1000 mfd 35 volt capacitors can also be the problem. Finally check for cracked solder joints on these 1000 mfdcaps (solder jumper wires, as done to the bridge rectifiers explained earlier).
Most test reports refer to switches that are tagged as defective. Often thisis not the case. If a switch hasn't be used in 30 games, it will be listedas bad. But it could be the switch is working, yet positioned in a place that itjust doesn't get activated much during game play.
If you do get a test report about a possibly defective switch, go to the 'switch edge' test and manually activate the switch. This willindicate if the switch is working. If it does work, this will reset the 30 gamecounter for this switch and the switch will not be reported in the testreport.
Prototype ROM Software and Bad Switches.
If your game has early prototype U6 CPU EPROM software, sometimesnon-existant switches can show up in the test report. This happenedin early versions of Twilight Zone and Judge Dredd games. There is noway to correct this but to upgrade to the lastest U6 CPU EPROM software.A new EPROM will need to be 'burned' (using an EPROM programmer). The software for this is available at Williams' home page at thehttp://www.pinball.wms.com/tech/roms.htmlwebsite.
The Diagnostic Credit Dot.
If you are checking out a game that is being operated, look for aperiod after the number of credits shown on the display duringattact mode. If there is a period (dot) after the number of credits, this means there is a test report for the game. If there is no period after the credits, there's no diagnostic test report and the game is probably functioning correctly.
It doesn't happen often on WPC games. You have power(+5 and +12 volts) getting to the CPU board. The +5 LED (lower of the three)is on, as it should be. But the middle diagnostic LED is not flashing constantly (indicatingthe CPU is dead). And the blanking LED (the top one) is doing nothing (no flashes when the game is turned on). You have a dead CPU.
WPC-S and Prior CPU Flash Codes .
D19 is labeled 'blanking', D20 is labeled 'diagnostic' and D21 is labeled '+5vdc'.
Problem Power-On CPU D20 (diagnostic) Flash Codes:
If D20 blinks at power-on, and then stops blinking, here is whatthe blinks translate to:
WPC-95 CPU Flash Codes.
LED201 is labeled 'blanking', LED202 is labeled 'power' and LED203 is labeled 'diagnostics'.
Problem Power-On CPU LED203 (diagnostic) Flash Codes:
If LED203 blinks at power-on, and then stops blinking, here is whatthe blinks translate to:
Dead CPU Step One: Remove the Ribbon Cables.
Before you do anything, turn the game off and remove all the ribbon cablesfrom the CPU. This will issolate the CPU from the driver board, the dotmatrix display board, the sound board, the fliptronic board (if yourgame has one), and any other connecting boards. The ribbon cables areat connectors J201, J202, J211, and J204 (on some games). While you'reat it, you might as well remove the switch connectors at J205 to J209,and J212. The only connector still attached is J210 (the power connector).
After everything is removed but connector J210, turn the game on. Ifthe CPU board boots correctly, the lower LED (+5 volts) should beon, the middle LED (diagnostics) should be blinking continually, and the top LED (blanking)should be off. If this is the case, turn the game off and replace theribbon cables, one at a time, and turn the game back on. When the CPUdoesn't boot, you have found the connector (and hence the board) thatis dragging the CPU down.
Move to the Work Bench.
If the above 'step one' didn't get you anywhere, don't worry.Now it's time to remove the CPU from the game.Don't try and fix a dead CPU while it's still in the game. You aremuch better off fixing it on your workbench. Fixing it on the workbenchmeans you have issolated the bad CPU from the rest of the game (includingit's power supply!).
With the CPU on the workbench and issolated from the game, youcan test the board much easier.
Re-seat the U9 WPC chip.
You would be amazed at how often this works. A dead CPUcan suddenly come to life after removing and reinsertingthis large U9 chip. You will need a special tool to remove thisbig, square chip. You can buy this tool at Radio Shack,part number 276-2101, $9.99. Do NOT try to remove this chipwithout this tool! Note one corner of this chip is 'notched',so you can only re-insert the chip one way.
Bad socket at U9.
The large U9 WPC square chip can have a bad socket.It's not much fun to replace this 84 pin socket!Radio Shack sells replacement sockets, part numberRSU 11354453, $1.99, but they may not stock it.Use your DMM and check for continuity with the chipinstalled before you replace this socket.
Shotgun Approach.
The chips are U1, U2 (74LS244) and U3 (74LS245) arethe ones that affect on a dead CPU the most. If replacing those yeilds nothing, then try replacing U5 (74LS14).Also check resistors R95 and R99 (1 meg ohms) to make surethese are the correct value. Finally U10 (a MC34064 transistor that is part of the startup circuit) can be replaced.Using a logic probe, also check for a good clock signal on pins 34 and 35 of U4 (6809).
The Clock won't keep Time.
Problem: The internal time clock appears to be running very slow, only about 25% of realtime speed. Numerous spot checks show that it advances about 6 hours per day. The batteries, which when weak cancause the clock to lose time, are brand new.The clock function is handled by U9 (the ASIC chip) and U21 (a CMOS 4584), and the 32.768KHz crystal.
Answer: Both legs of crystal X1 were soldered to the same spot! It looks like it came from the factory that way. After removing the crystal and putting both legs in the correct locations, the time is tracking correctly.
The 32.768 KHz crystal is very common and used in everything from wrist watches to computers to anything that keeps time. The reason for that particular frequency is 2 to the 15th powerequals 32,768. The frequency is very easy to divide by two, fifteen times, using flip-flops or some other form of divider network. This nets a one second time increment. Since your crystal was shorted, the oscillator was free running at a RC-determined frequency that undoubtedly drifted with temperature and miniscule voltage changes, hence the accumulated errors.
I can't enter my high score initials on Funhouse.
Problem: game works fine, but won't let player advance through the initialsby pressing the flipper buttons when a high score is achieved. The start button works correctly as 'enter',and the flippers work fine in game play.
Answer: there are two optocouplers on the power driver board at U7 and U8 that are numbered 4N25. If these go bad, they will prevent the flippersfrom moving through the high score initials. Since this game does nothave fliptronic flippers, these optocouplers don't effect the flippersthemselves. If this was a fliptronics game, the flippers wouldn't haveworked either.
The flippers and dot matrix display died while playing a game.
Problem: The flippers on my Indy Jones died. The dot matrix display only has one vertical linewhich is always lit. The GI lamps are fine, as are the controlled lamps.I turned the game off and back on, the game continually launched ballsfrom the ball trough.
Answer: the +12 volts has died, probably from a bad fuse at F116, or maybea bad BR5 bridge. Some dot matrix power is derived from the +12 volts,and the +12 volts also powers the optos (hence the auto ball launchingproblem and no flippers). If the +12 volts is good, unplug the fliptronicsand sound board ribbon cable, leaving just the dot matrix display pluggedin to the ribbon cable. Now see if the display clears up and you can see the error report.
Strange Error Message.
Problem: When I turn my Creature from the Black Lagoon on, I get the errormessage 'check switch #F6 U.R. Flipper'. But this game doesn't have anupper right flipper.
Answer: Every flipper opto board has two optos. One is wired to the lower andthe other to the upper flipper switch inputs. This is true even on games withjust lower flippers. If the flipper opto board has a dirty opto, you can getthis error, even if your game doesn't have the flipper reported in the errormessage. Clean your flipper opto board optos with a Qtip. Replace the optoif the problem doesn't resolve.
Dot Matrix Display Got Blurry.
Problem: When I was playing my Twilight Zone, the dot matrix displaystarted to become very blurry. Within 5 minutes the display becamealmost unreadable. The dots to the left and right of the activeones started to flicker.
Answer: the ASIC chip on the CPU board was not making good contact toits socket. The ASIC chip is the large square chip on the CPU board. After removing the chipand cleaning all of its pins, and reseating the chip in the socket,the problem went away. Another thing to try is reseating the ribboncables in their sockets.