COMMODORE VIC20 COMPUTER - DIAGNOSTICS AND REPAIR latest additions and corrections 4-9-2023 There are two basic versions of the VIC20. I've only seen one of the very early 1980 model designated as the VIC20-E, so it must be rather rare. Therefore, I'll not discuss that one any further here. That board and the later standard VIC20 are easily recognized from the outside by the two pin connector used for 9VAC power. The power supply for that model is a 9 volt transformer rated at 3 Amps AC. It's about half the size of the C64 PS. Earlier boards are larger and have two large black heat sinks for the bridge rectifier and 5 volt power regulator. Since most of the power supply is inside the computer, those early VIC20's run warmer inside. The later 1981/82 VIC20 is different than the earlier ones. The board is smaller and uses fewer, more integrated chips (CR or cost reduced) but still uses the same basic chips as the earlier board. It has a 7 pin DIN connector and uses the same power "brick" as a Commodore 64 but draws more power from that supply than a C64. The computer itself runs cooler than earlier versions because the power regulators and their heat sinks are in the PS and not in the computer. Because of the differences in boards and some chip ID numbers, I will list the two versions separately in the chips vs symptoms portion of this article. Although TTL (so-called "glue logic") chips are rather robust, any part can fail at any time. Two IC's that do seem to fail most often are the VIC and Kernal ROM. Note that all semiconductor devices (transistors, diodes and ICs) that normally run hot fail more often than those that run cooler. I noticed the VIC chip inside the metal cover runs hot in all models and should therefore have a metal heat sink that touches the metal cover of the enclosure above it. I've seen many failures of that IC due to overheating. That's why I always install a heat sink on it when I get a VIC20 on the bench for repair. The sink consists of a thin strip of metal epoxied to the top of the IC and with ends bent upwards to touch the shield cover when it's reinstalled. (See the photos) That draws the heat away from the IC and transfers it to the metal cover. Another common failure is the Kernal ROM UE12, a Commodore 901486-06. Either chip failure will cause a blank screen. Note that a completely dead VIC will not display any change on the monitor screen when the computer is switched on. If the VIC is working and the Kernal is bad, the screen will darken (blank screen) when the computer is turned on. It's advisable to install a heat sink on any IC that normally runs hot to make it last longer. The BASIC ROM next to the Kernal also runs pretty hot but I've yet to see one of those fail. That said, a heat sink is cheap insurance. Other than that, the IC's directly connected to external ports such as the VIA and interface logic can be damaged due to external voltage "spikes" and surges. That's why you should avoid plugging or unplugging anything with the computer powered up. The last part of this article will tell how to troubleshoot problems and offers some diagnostic procedures you can do with little or no test equipment. Careful observation is important. As with other Commodore computers, the easiest way to test suspected chips is by installing them, one at a time, in a working board. If a chip is soldered in (no socket), it makes it that much more difficult to swap them for tests. Minimum equipment for board rework includes a vacuum desoldering iron, a volt-ohmmeter and basic hand tools. Since the VIC20 has no internal RF modulator, you need to either use the factory external RF converter box that plugs into the 5 pin A/V port on the computer (to feed the RF input of your TV), or use a direct A/V connection to a monitor or TV A/V inputs. The USA version outputs a signal on TV channel 3 or 4 (switchable) and the European version on TV channel 36. Note that the RF converter itself (the ones made in Japan) suffers a common problem. See the article regarding that repair. There is a not-so-obvious problem that may look like a failure but it's simply a factory setting to allow use of the RF unit. In the days when the VIC20's came out, few people had video monitors, just TV sets. The external RF adapter outputs a signal on channel 3 or 4 so the TV sees it as a normal broadcast channel. The apparent fault will be low video or perhaps no color when running the VIC20 with a two-wire A/V cable into a video monitor such as a 1702 or 1084. The video level was factory set that way to accommodate the RF adapter. If the video level is set too high into the adapter, it will overload the adapter and produce a distorted image. So, running an A/V cable will produce a normal picture if the video level pot inside the VIC cage is adjusted up (counter-clockwise) a bit. Too high and the color will start to distort, so a correct setting is just below that level. There are photos of early and late boards on this website to indicate which pot to adjust. Don't touch any of the others! The VIC20 opening screen should appear within about four seconds after power up. It has a cyan border and blue letters on a white background. The full screen is 22 characters wide by 23 high. At power up, it should look like this: * * * * CBM BASIC VERSION 2 * * * * 3 5 8 3 B Y T E S F R E E R E A D Y [] <---(flashing cursor) *********************************************************************** EARLY VERSION BOARD: ASSY #324003 FAB #324002-01 REV D 1980 (note large black heat sinks and two pin power connector) UAB1 6522 VIA INTERFACE, KEYBOARD-SERIAL Startup screen normal, but no cursor. Keyboard doesn't work at all. Partial failure: some keys "stuck" or don't work... may be a whole row or column. Also can cause serial port (drive access) problems. Can produce blank screen if shorted (remove to check). UAB3 6522 VIA INTERFACE, SERIAL-JOY-USER-CASS Startup screen normal, but drive access problems ("SEARCHING FOR.. " forever). One or more joystick positions don't work. No cassette or user port access. Can produce blank screen if shorted (remove to check). UB4 7406 LOGIC SERIAL PORT BUFFERS Blank screen. Partial failure: serial port (disk drive) access problems. UB6 LM555 TIMER POWER-ON SYSTEM RESET Will not reset on power up. May produce garbled or random screen characters-graphics, or screen may start up "frozen". If drive resets when computer is powered up, this chip is OK. UB7 6560 (NTSC) or 6561 (PAL) VIC VIDEO/AUDIO White or blank screen, garbled or no video, screen full of (or few random) "garbage" characters. "Blind" disk commands may still work. Partial failure: Dark or smeary image, loss of color, garbled or no sound, game paddles or light pen doesn't work. UB9 7402 (M53202P) OSC MASTER CLOCK OSC White screen... no RF interference in AM radio (see text). Partial failure can produce missing color or incorrect colors, slight or severe tearing of picture or diagonal lines, sounds play at incorrect pitch. Symptoms may only appear after warmup. UC2 74LS04 LOGIC Screen character colors incorrect (scrambled). If very bad, entire screen becomes garbled with flashing characters, vertical bars and random colors. UC3 74LS02 LOGIC UC4 74LS138 LOGIC MEMORY CONTROL Blank screen. UC5 74LS138 LOGIC BLOCK CONTROL Blank screen. UC6 74LS138 LOGIC I/O CONTROL UD1 CD4066 QUAD SWITCH (LOGIC GATES) (UC2 in very early version) Screen characters have incorrect colors. UD2-UD6 UE1-UE6 2114 or TMM314A SRAM MEMORY Blank screen. Partial failure: less than normal "3583 BYTES FREE" at startup, sometimes will produce "garbage" screen or screen freeze after warmup. Shorted chips may get very hot. NOTE: UE1 is used to generate character color. UD7 901460-03 ROM CHARACTER Startup screen characters missing (just blocks or flashing lines where characters should be) and alphanumerics missing from games. Game carts that use mostly graphics characters may look normal while running. UD8 and UE8 MPS65245 or MOS65245 or 74LS245 TRI-STATE LOGIC Blank screen. Partial failure: garbled video or audio. see UB7 UD9 74LS133 13 INPUT LOGIC GATE Blank screen. UE3 7406 LOGIC Serial port problems or blank screen. UE10 6502 MICROPROCESSOR Blank screen. Partial failure: computer may run for awhile, then freeze. UE11 901486-01 ROM BASIC Startup screen with borders but no characters. Disk commands do not work, but game carts still work. UE12 901486-06 ROM KERNAL Blank screen on startup, no drive access, game carts don't work. Partial failure: games may still work. This is a very common failure IC! UF8 MPS65245 or MOS65245 or 74LS245 TRI-STATE LOGIC Blank screen. Q3 2SD880 TRANSISTOR, CASS. MOTOR DRIVER Most common problem when cassette motor will not run. VR1 LM323K +5V REGULATOR, 3 AMP Computer appears "dead" with no red power indicator, or wavy picture with hum in sound. Shorted regulator will damage chips in the computer and blow the fuse. KBPC802 BRIDGE RECTIFIER, 5A 50V Computer appears "dead" with no red power indicator. Partial failure may produce hum in audio and wavy picture. Partially shorted bridge may make the power pack (transformer) get hot and may blow the main fuse (3.15A) inside the computer. Bad solder connections may make this part look intermittent or open. The block will run too hot if the heat sink it's attached to doesn't contact it properly. See text below. F1 3.15 AMP FB FUSE, MAIN POWER Computer "dead"... no red power light. If new fuse blows, suspect short in bridge rectifier, regulator, or computer chips, in that order. Make sure the fuse clips are clean and tight or fuse and holder will get hot and the new fuse will eventually blow for no apparent reason. Y1 CRYSTAL, 14.31818MHz in USA version See UB9 (7402) *********************************************************************** LATEST VERSION CR BOARD: ASSY #250403 FAB #251040-01 REV D 1981/2 (note this model uses the same power pack as a C64) UAB1 6522 VIA INTERFACE, KEYBOARD-SERIAL Startup screen normal, but no cursor. Keyboard doesn't work at all. Partial failure: some keys "stuck" or don't work... may be a whole row or column. Also may cause serial port (drive access) problems. Can produce blank screen if shorted (remove to check). UAB3 6522 VIA INTERFACE, JOY-USER-SERIAL-CASS Startup screen normal, but drive access problems ("SEARCHING FOR... " forever). One or more joystick positions don't work. No cassette or user port access. Can produce blank screen if shorted (remove to check). UB4 7406 LOGIC Blank screen or serial port (disk drive) access problems. UB6 LM555 TIMER POWER ON RESET Will not reset on power up. May produce garbled or random screen characters-graphics, or screen may start up "frozen". If drive resets when computer is powered up, this chip is OK. UB7 6560-101 VIC VIDEO/AUDIO White or blank screen, garbled or no video, screen full of (or few random) "garbage" characters. "Blind" disk commands may still work. Partial failure: Dark or smeary image, loss of color, garbled or no sound, game paddles or light pen doesn't work. UB9 7402 (M53202P) OSC MASTER CLOCK OSC White screen... no RF interference in AM radio (see text). Partial failure: can produce missing color or incorrect colors, slight or severe tearing of picture or diagonal lines, sounds play at incorrect pitch. Symptoms may only appear after warmup. UC2 74LS04 LOGIC Screen character colors incorrect (scrambled). If very bad, entire screen becomes garbled with flashing characters, vertical bars and random colors. UC3 74LS02 LOGIC UC4 74LS138 LOGIC MEMORY CONTROL Blank screen. UC5 74LS138 LOGIC BLOCK CONTROL Blank screen. UC6 74LS138 LOGIC I/O U14-U15 MB8416A RAM MEMORY Blank screen. Less than normal "3583 BYTES FREE" at startup. Sometimes will produce "garbage" screen or screen freeze after warmup. Shorted chips can get very hot. UD1 CD4066 QUAD SWITCH (LOGIC GATES) Incorrect color on screen characters. UD2 2114 SRAM UD7 901460-03 ROM CHARACTER Startup screen characters missing (just blocks or flashing lines where characters should be) and alphanumerics missing from games. Game carts that use mostly graphics characters may look normal while running. UD8 & UE8 MPS65245 or MOS65245 or 74LS245 TRI-STATE LOGIC Blank screen. Partial failure: garbled video or audio. see UB7 UD9 74LS133 LOGIC UE1 2114 SRAM UE2 2114 SRAM UE10 6502 MICROPROCESSOR Blank screen. Partial failure: programs may run for awhile, then freeze. UE11 901486-01 ROM BASIC Startup screen with borders but no characters. Disk commands do not work, but game carts still work. UE12 901486-06 ROM KERNAL Blank screen on startup, no drive access, game carts don't work. Partial failure: some game carts may still work. This IC is a very common failure! UF8 MPS65245 or MOS65245 or 74LS245 TRI-STATE LOGIC U13 74LS08 White garbled screen or blank screen. Q3 2SD880 TRANSISTOR, CASS. MOTOR DRIVER Cassette motor doesn't run. Common failure. CR2 S10B BRIDGE RECT, 1A, 50V F1 1 AMP FB FUSE, 9 VOLT AC SOURCE Opening screen normal and all functions work except cassette motor and 9VAC source to user port. Y1 CRYSTAL, 14.31818MHz See UB9 (master oscillator chip). ******************************************************************* Diagnosis of chip and other failures is often difficult, but there are some things you can try to narrow the diagnosis to a specific fault. Blank screen is the most common symptom, and it can be the most difficult because there are many potential causes, including power supply (internal in the early version VIC20) faults. Note the difference in the two terms: blank screen and white screen. Blank refers to a dark screen (called a "raster") with no border and no characters. White screen is likewise blank with no border and no characters, but brighter than a "blank" screen. An AM radio tuned to the lowest point on the dial (off station) can be used to see if the computer is generating RF interference (which is normal). The digital signals inside a computer are like little transmitters, and an AM radio held nearby can pick up those signals. Keep the TV or monitor off for this test as it generates strong interference of its own. The VIC20 startup sequence to opening screen takes about four seconds and produces characteristic sounds at power up. If you are used to the sounds it normally makes, that can be used during diagnostics. For example, if the computer generates normal interference sounds in the radio, it indicates the microprocessor and support chips are working, and a blank screen may be a VIC chip problem. Of course a total failure of the VIC chip will produce a white screen and little noise in the radio because nothing is running in the computer except the master oscillator. Clock signals go through the VIC chip to feed the rest of the computer. Early VIC20 computers have a rather high input current and you need to make sure the fuse is tight in the fuse holder. If it's loose in the clips, the fuse and socket will get hot from the slight bit of resistance in the contacts. The resulting B+ voltage may be too low to run the computer properly and the fuse may eventually open due to the heat. Make sure the clips are clean of dirt and corrosion, and squeeze them together to make a tight fit before the fuse is installed. Likewise, make sure the power plug is tight in the receptacle. If necessary, squeeze the plug end gently with a pair of pliers to crimp its internal contacts (be careful not to crush them), to make a tight fit before plugging in the supply to the computer. Loose connections at the socket will cause intermittent operation or no power to the computer. Another common trouble spot in the early version is the solder connections on the large bridge rectifier under one of the black heat sinks. I often resolder it just to make sure it's making good contact. Also, be sure to tighten the heat sink on that bridge block. Every early version VIC20 I've worked on had a loose screw mounting the diode to the sink, which was making the diode run very hot. If left that way, the diode will actually get hot enough to un-solder itself and/or fail (usually by shorting and blowing the fuse) eventually. Rather than removing the entire bottom metal shield, I cut it from hole to hole at the diode area with snips, then cut away the plastic to get at the mounting screw underneath. As preventive maintenance, tighten the screw and apply a dab of epoxy to the threads to keep it from getting loose again. A nut (metric) will keep it tight forever. Speaking of contacts, one preventive (and diagnostic) step you can take is to reseat all the socketed chips. The contact point between chip and socket is very small and a bit of corrosion or dust can cause intermittent operation or a "dead computer" symptom. Lift up each end of the chip with a small screwdriver and press it back down to wipe the contacts. Don't pry against the PC board or you may damage the traces (printed "wires"). No need to remove chips for this procedure, but if you do, make sure the pins are straight when you reinstall them. A bent pin is sometimes impossible to see (unless the chip is removed again) and the resulting failure or intermittent adds yet another problem. One diagnostic check you can make on a "dead" computer is to feel each of the chips and see if any are getting hot. Some chips (the large ones like the microprocessor and the VIC) will run very warm to the touch after the computer has been running for awhile. There is quite a bit of variation in normal chip temperatures, so a bit of experience helps in diagnosis. Some chips run almost cold and it's perfectly normal, but if you find a cold VIC after running the computer for ten minutes, it may be bad. If a replacement also runs cold, the chip is either not getting power (no +5 volts on pin 40) or the ground has opened up (pin 20). I had such a failure in one board. The white screen and "radio check" told me the computer was nearly dead. The master oscillator was running (checked with an oscilloscope) and the chip had +5 volts on pin 40. The chip was replaced and it still didn't work. It turns out the socket was at fault. It was likely damaged because of a bent IC pin on a previous VIC chip installation. The ground return (pin 20) for the chip was open circuit. I was able to repair the socket by bending the receptacle for pin 20 back so it made proper contact with the new chip when it was inserted. I only mention this to show how a "simple" fault can be VERY hard to find. Unexpected or illogical faults like that one really burn up bench time and frazzle a technicians nerves. It's not a "normal" failure, but it happens often enough to keep us humble. As a diagnostic, the computer will still start up with a select few of the chips removed, such as the two VIA chips, the character ROM, & UC2. If you suspect one of those chips is shorted and producing a blank screen, you can remove them to see if the computer will then start up. Obviously, without a character ROM you will not have a normal screen display but a normal border should appear. Swapping the two VIA chips can help diagnose which one is bad. Faults will follow the bad chip. A stuck (shorted) keyboard key can be isolated by unplugging the keyboard and powering up the computer without it, or substituting another keyboard. C64 keyboards work on VIC20s and (for some versions) internal parts can be interchanged as well. Keys that require a hard press to make them work are probably worn out (or dirty). If you don't have spare parts handy, the conductive rubber plunger can be swapped for a little-used key to get that one working again. If several keys are bad, you might try wiping the contacts gently with a clean sheet of paper. Of course that requires the keyboard be taken completely apart (many tiny screws and some unsoldering). Power supply problems account for quite a few failures. As I mentioned before, there are two different versions of the VIC20 and each requires the appropriate power pack. Early version boards use a single 9VAC source (transformer-only inside the power pack) at a rather hefty 2 Amps. That kind of supply rarely fails. The rectifier, filter capacitors and regulator are all inside the computer. You'll notice the two large black metal heat sinks when you open it up. That version is easy to identify by its two pin power connector instead of the round DIN socket of the later VIC20CR. Later version VIC20s use the same supply as a C64, so there are two voltages: +5VDC at 1.5 Amps and 9VAC at 1 Amp. It's helpful to know that a VIC20 draws more power from the supply, so if the power pack is marginal, it may still work with a C64 but not with your VIC20. It's best to use one of the "beefier" after-market repairable supplies, especially if you plan to use accessories like RAM expansion or game cartridges. They draw power from the computer supply. If the 5 volt source is low or missing, the computer will not work at all and the red power light will not come on or will be dim. If the 9VAC is missing because of a bad supply or open 1 Amp fuse in the computer, the red power light will come on and the computer will work except for no cassette power and no 9VAC to the user port. It might therefore go unnoticed. If the fuse is blown, look for a short on one of those ports. Accidental mis-contact while connecting to a port (with the computer turned on) is usually the cause of an open fuse. Avoid connecting or disconnecting -anything- with equipment turned on! The above information may help get your computer going again, but keep in mind that actual computer faults can be a challenge even for experienced techs. There is a limit to what the novice can repair him/herself. I've been in the repair business for over 40 years and I still sometimes damage a chip or board trace when removing components from a Commodore computer or drive. It's just not that easy to do. So, if you have a major problem with your VIC20, it's probably best to start looking for another one or contact a repair tech. If a computer is uneconomical to repair, it's advisable to keep the bad one for parts. Chips are scarce now, and we need to salvage whatever we can to keep our systems running as long as possible. Ray Carlsen CET Carlsen Electronics... a leader in trailing-edge technology. Questions and comments are always welcome, especially if you spot a mistake here. Thanks!