Dash Backlighting -- The Full Story
09-06-2013, 04:11 AM
#1
Race Director
Thread Starter
Dash Backlighting -- The Full Story
I started getting random changes in the brightness of the dash backlighting during the daytime. I decided to look into how the backlighting works so I could fix this problem.
I got several surprises as I investigated this circuit. The first was that the backlighting is controlled by the microcontroller in the dashboard. I assumed when I started looking at this that the dash light control was a simple DC voltage, but that turned out to be incorrect. The actual control is by PWM (Pulse Width Modulation). What that means is that the voltage switches on and off and the average voltage is determined by how wide the ON pulses are.
I drew a vastly simplified schematic of the dashboard that only shows the dimmer circuitry. I drew this schematic by tracing out the circuit visually and by making ohmmeter readings to verify the connections. Very tedious and time consuming. I also used a schematic and layout diagram that came from Bryan A. Thompson's http://www.batee.com web site. Unfortunately it is missing a pretty large area in the lower right corner, which just happens to be where some of the circuitry of interest is located. The reference designators are what I found on my 1986 Corvette PC board. The reference designators on the schematic are different, so apparently they were renumbered from year to year (I HATE when they do that -- very inconvenient for people who get used to a certain numbering scheme and then suddenly it's different). The LCD PC board did not have reference designators marked on it so I just used random numbers there.
Starting at the headlight switch. The headlight switch sends a signal to the dash that tells it when the parking lights or headlights are on. The dash uses this signal to dim the backlights because it assumes that the ambient is dark so not much light is needed. A rheostat on the headlight switch controls how bright the backlighting is at night.
The lights-on signal and the rheostat setting are connected to the dashboard in the upper left part of the schematic below. The parts inside the dotted boxes are portions of resistor networks. The purpose of this circuitry is to protect the rest of the dashboard circuitry from static electricity and other electrical problems.
During the daytime, the brightness of the backlighting is determined by a photocell that responds to the intensity of the ambient light. The photocell also overrides the lights-on signal from the headlight switch if the ambient light is bright so the parking lights or headlights can be on during the daytime and the dash will still be bright enough to read.
The signals from the headlight switch and photocell are applied to an ADC (Analog to Digital Converter). That is an integrated circuit that compares the input to a reference voltage and puts out a digital number that corresponds to the ratio of the input to the reference. An input of zero volts will result in an output of an 8-bit binary number equal to zero: 0000 0000. An input equal to the reference voltage will cause a binary output equal to 255 (the largest number that can be represented in 8 bits): 1111 1111.
The output of the ADC is connected to the microcontroller (usually abbreviated uC -- that "u" is actually a Greek letter mu [for micro] but we use u because it's much easier to type than mu). The uC tells the ADC which channel it wants to read and the ADC responds with the current reading from that channel. The three channels we're dealing with here are labelled IN4, IN5 and IN6 on the ADC. The programming in the uC uses the inputs from the ADC to determine how bright the dash backlight should be. It creates a signal that is proportional to the ambient light and sends that signal out on port 21 (P21 in the uC in the schematic). This signal is "filtered" to delay sudden changes caused by driving in and out of shady areas (which would cause the backlighting to flash), so it takes several seconds for the brightness to change when there is a change in the amount of light reaching the photocell.
The P21 signal goes to some circuitry and to a test connector. I used the test connector to tap into the DIMMER signal to create the oscilloscope screen shots below. There is an inverter and a mystery chip that is labelled "LEVEL SHIFTER" on the component layout diagram I have. The purpose of the level shifter seems to be to interface the uC that has a 5 volt power supply to the light bulb dimmer circuit, which uses 12 volts.
The backlight bulb dimming circuitry is shown in the lower left corner of the schematic. Basically it's just a driver transistor Q1 that controls power transistor Q2, which then controls the brightness of the back lights by varying how much current flows through them.
Now the bad news. This circuit is pretty much not repairable except for replacing the bulbs and some people have replaced the photocell. The main problem is that the parts have AC Delco or GM part numbers and there are no cross references to standard parts. Parts like the mystery IC are just that and we have no way to determine what it actually is. I thought at first it might be a ULN series part, but those are all inverters and the signal is not inverted there. The AC Delco web site is long gone and there are no data sheets for these parts. In many cases the GM part numbers are abbreviated versions that only show the last 4 numbers of an 8 digit number, so there is no way to determine what the full number is. I put the part numbers that I could find in the schematic. There is a fuse shown as CLSTR, which is the label on the fuse panel. This fuse is shown in the FSM (Factory Service Manual) dash backlighting diagram as INST.
The simplified schematic:
Now for some waveforms taken from my oscilloscope. The microcontroller puts out a square wave that determines how bright the bulbs get. The upper part of the waveform is the ON part (12 volts) and the lower part is the OFF part (zero volts). The ratio of the ON time to the OFF time determines how bright the bulbs get. This screen shot shows the waveform when the headlight switch rheostat is in the minimum position (fully clockwise). This was recorded in a dark garage. This waveform shows that the signal is on about a quarter of the time. The bulb filament responds relatively slowly to this waveform and you don't see it flickering.
This screen shot shows the waveform when the headlight switch rheostat is turned to maximum brightness (fully CCW). In this case the waveform is ON about 2/3 of the time and the bulbs are therefore brighter. Again, captured in a dark garage.
I shined an LED flashlight into the photocell at about a 45 degree angle from about a foot away to get the most ON time I could get that still had some OFF time:
This last screen shot shows what happened when I shined the LED flashlight directly on the photocell. There is no OFF time at all and the bulbs are at maximum brightness.
The photocell (more accurately known as an LDR for Light Dependent Resistor because it changes its resistance with changes in the amount of light that is falling on it) has a part number marked on it: VT6861. I couldn't find any info on this part number and it appears to be a custom part made for GM. Back in the years when these were used, there was a company known as Vactec that made photocells. They were bought by EG&G (Edgerton, Germeshausen, and Grier, Inc.) and became known as EG&G Vactec. That company no longer exists, but a company named Excelitas Technologies is currently manufacturing these parts with the old VTxxx numbers. The main difference I see is that the new parts do not have the option of the black plastic light shield that the VT6861 parts have. The photocells appear to be glued into the light shield and I was not able to get it out. I determined that part number VT90N1 is a suitable replacement. They are available from Newark for less than a dollar each: http://www.newark.com/excelitas-tech...ar%20Resistors I don't believe that the light shield is necessary. Some people have handled a bad photocell by putting a blob of solder across the photocell pins. This will cause the dash backlighting to be at maximum brightness all the time.
The 882 halogen backlight bulbs were replaced with 7073 xenon bulbs in 1989. They have about the same power output but are physically larger. They require a different socket because they are wedge base. The 882 part number includes the socket. I don't know the part number of the bulb (or even if it's available) without the socket.
There are two ways to replace the bulbs. The first is to remove the instrument cluster from the dashboard. The cluster bezel must be removed. To do that the headlight **** must be removed. It is held in place by a spring clip that can be released by putting a small screwdriver in a slot and pulling outward. The bezel is held in place by several screws, some a little difficult to access. After the bezel has been removed the cluster itself is held in place by four 7mm screws, two in the upper corners and two at the bottom near the steering column. There are two large cable harnesses that connect to the cluster and they must be unplugged to get it out. The connectors have latches on them that must be released to unplug the connectors. Early connectors have a metal latch that hooks on the mating connector. The latch must be pried upward to release it. A small screwdriver works well for this. The later latches are plastic and have a "handle" that must be pressed down to release the connector. After the cluster is out the rear metal shield must be removed. It is held with several 5.5mm screws. After the shield is removed the microcontroller board will be seen. Remove more 5.5mm screws to get it out. Now the LCD display board with the bulb sockets can be seen and the bulbs can be replaced from the rear. This involves twisting out the bad bulb sockets and twisting in new bulbs. Many times you will find that the sockets are discolored from the heat of the bulb. The PC board will also have dark areas around the bulbs that are caused by the heat. It is important to avoid touching the glass part of the bulbs with your fingers because skin oil can cause uneven heating of the bulb which can cause it to break. They can be cleaned with alcohol.
The second method for replacing the bulbs is to remove the cluster bezel as described above and then replace the bulbs from the front. This involves removing the replacement bulbs from their sockets and inserting them in the dashboard sockets. To access the bulbs, a silver metal light shield over each bulb must be removed. Use a small tool and pull it out. The bad bulbs must be removed. Multiple methods have been used for this. One is to use longnose pliers and grab the bulb and pull it out. I don't recommend this method because of the possibility of breaking the bulb. Another method is to wrap tape around the bulb and pull it out. Many people use a small piece of rubber hose. There are various tools that are designed to remove bulbs like these. Here are a few that I know of.
A Jonard S-339, used to remove and replace bulbs in telephone switchboards (this is what I use):
The "official" tool, as listed in the FSM, Kent-Moore J-33885:
Here's a fun one. The "Mill Bulb Tool", designed for changing bulbs in pinball machines:
I got several surprises as I investigated this circuit. The first was that the backlighting is controlled by the microcontroller in the dashboard. I assumed when I started looking at this that the dash light control was a simple DC voltage, but that turned out to be incorrect. The actual control is by PWM (Pulse Width Modulation). What that means is that the voltage switches on and off and the average voltage is determined by how wide the ON pulses are.
I drew a vastly simplified schematic of the dashboard that only shows the dimmer circuitry. I drew this schematic by tracing out the circuit visually and by making ohmmeter readings to verify the connections. Very tedious and time consuming. I also used a schematic and layout diagram that came from Bryan A. Thompson's http://www.batee.com web site. Unfortunately it is missing a pretty large area in the lower right corner, which just happens to be where some of the circuitry of interest is located. The reference designators are what I found on my 1986 Corvette PC board. The reference designators on the schematic are different, so apparently they were renumbered from year to year (I HATE when they do that -- very inconvenient for people who get used to a certain numbering scheme and then suddenly it's different). The LCD PC board did not have reference designators marked on it so I just used random numbers there.
Starting at the headlight switch. The headlight switch sends a signal to the dash that tells it when the parking lights or headlights are on. The dash uses this signal to dim the backlights because it assumes that the ambient is dark so not much light is needed. A rheostat on the headlight switch controls how bright the backlighting is at night.
The lights-on signal and the rheostat setting are connected to the dashboard in the upper left part of the schematic below. The parts inside the dotted boxes are portions of resistor networks. The purpose of this circuitry is to protect the rest of the dashboard circuitry from static electricity and other electrical problems.
During the daytime, the brightness of the backlighting is determined by a photocell that responds to the intensity of the ambient light. The photocell also overrides the lights-on signal from the headlight switch if the ambient light is bright so the parking lights or headlights can be on during the daytime and the dash will still be bright enough to read.
The signals from the headlight switch and photocell are applied to an ADC (Analog to Digital Converter). That is an integrated circuit that compares the input to a reference voltage and puts out a digital number that corresponds to the ratio of the input to the reference. An input of zero volts will result in an output of an 8-bit binary number equal to zero: 0000 0000. An input equal to the reference voltage will cause a binary output equal to 255 (the largest number that can be represented in 8 bits): 1111 1111.
The output of the ADC is connected to the microcontroller (usually abbreviated uC -- that "u" is actually a Greek letter mu [for micro] but we use u because it's much easier to type than mu). The uC tells the ADC which channel it wants to read and the ADC responds with the current reading from that channel. The three channels we're dealing with here are labelled IN4, IN5 and IN6 on the ADC. The programming in the uC uses the inputs from the ADC to determine how bright the dash backlight should be. It creates a signal that is proportional to the ambient light and sends that signal out on port 21 (P21 in the uC in the schematic). This signal is "filtered" to delay sudden changes caused by driving in and out of shady areas (which would cause the backlighting to flash), so it takes several seconds for the brightness to change when there is a change in the amount of light reaching the photocell.
The P21 signal goes to some circuitry and to a test connector. I used the test connector to tap into the DIMMER signal to create the oscilloscope screen shots below. There is an inverter and a mystery chip that is labelled "LEVEL SHIFTER" on the component layout diagram I have. The purpose of the level shifter seems to be to interface the uC that has a 5 volt power supply to the light bulb dimmer circuit, which uses 12 volts.
The backlight bulb dimming circuitry is shown in the lower left corner of the schematic. Basically it's just a driver transistor Q1 that controls power transistor Q2, which then controls the brightness of the back lights by varying how much current flows through them.
Now the bad news. This circuit is pretty much not repairable except for replacing the bulbs and some people have replaced the photocell. The main problem is that the parts have AC Delco or GM part numbers and there are no cross references to standard parts. Parts like the mystery IC are just that and we have no way to determine what it actually is. I thought at first it might be a ULN series part, but those are all inverters and the signal is not inverted there. The AC Delco web site is long gone and there are no data sheets for these parts. In many cases the GM part numbers are abbreviated versions that only show the last 4 numbers of an 8 digit number, so there is no way to determine what the full number is. I put the part numbers that I could find in the schematic. There is a fuse shown as CLSTR, which is the label on the fuse panel. This fuse is shown in the FSM (Factory Service Manual) dash backlighting diagram as INST.
The simplified schematic:
Now for some waveforms taken from my oscilloscope. The microcontroller puts out a square wave that determines how bright the bulbs get. The upper part of the waveform is the ON part (12 volts) and the lower part is the OFF part (zero volts). The ratio of the ON time to the OFF time determines how bright the bulbs get. This screen shot shows the waveform when the headlight switch rheostat is in the minimum position (fully clockwise). This was recorded in a dark garage. This waveform shows that the signal is on about a quarter of the time. The bulb filament responds relatively slowly to this waveform and you don't see it flickering.
This screen shot shows the waveform when the headlight switch rheostat is turned to maximum brightness (fully CCW). In this case the waveform is ON about 2/3 of the time and the bulbs are therefore brighter. Again, captured in a dark garage.
I shined an LED flashlight into the photocell at about a 45 degree angle from about a foot away to get the most ON time I could get that still had some OFF time:
This last screen shot shows what happened when I shined the LED flashlight directly on the photocell. There is no OFF time at all and the bulbs are at maximum brightness.
The photocell (more accurately known as an LDR for Light Dependent Resistor because it changes its resistance with changes in the amount of light that is falling on it) has a part number marked on it: VT6861. I couldn't find any info on this part number and it appears to be a custom part made for GM. Back in the years when these were used, there was a company known as Vactec that made photocells. They were bought by EG&G (Edgerton, Germeshausen, and Grier, Inc.) and became known as EG&G Vactec. That company no longer exists, but a company named Excelitas Technologies is currently manufacturing these parts with the old VTxxx numbers. The main difference I see is that the new parts do not have the option of the black plastic light shield that the VT6861 parts have. The photocells appear to be glued into the light shield and I was not able to get it out. I determined that part number VT90N1 is a suitable replacement. They are available from Newark for less than a dollar each: http://www.newark.com/excelitas-tech...ar%20Resistors I don't believe that the light shield is necessary. Some people have handled a bad photocell by putting a blob of solder across the photocell pins. This will cause the dash backlighting to be at maximum brightness all the time.
The 882 halogen backlight bulbs were replaced with 7073 xenon bulbs in 1989. They have about the same power output but are physically larger. They require a different socket because they are wedge base. The 882 part number includes the socket. I don't know the part number of the bulb (or even if it's available) without the socket.
There are two ways to replace the bulbs. The first is to remove the instrument cluster from the dashboard. The cluster bezel must be removed. To do that the headlight **** must be removed. It is held in place by a spring clip that can be released by putting a small screwdriver in a slot and pulling outward. The bezel is held in place by several screws, some a little difficult to access. After the bezel has been removed the cluster itself is held in place by four 7mm screws, two in the upper corners and two at the bottom near the steering column. There are two large cable harnesses that connect to the cluster and they must be unplugged to get it out. The connectors have latches on them that must be released to unplug the connectors. Early connectors have a metal latch that hooks on the mating connector. The latch must be pried upward to release it. A small screwdriver works well for this. The later latches are plastic and have a "handle" that must be pressed down to release the connector. After the cluster is out the rear metal shield must be removed. It is held with several 5.5mm screws. After the shield is removed the microcontroller board will be seen. Remove more 5.5mm screws to get it out. Now the LCD display board with the bulb sockets can be seen and the bulbs can be replaced from the rear. This involves twisting out the bad bulb sockets and twisting in new bulbs. Many times you will find that the sockets are discolored from the heat of the bulb. The PC board will also have dark areas around the bulbs that are caused by the heat. It is important to avoid touching the glass part of the bulbs with your fingers because skin oil can cause uneven heating of the bulb which can cause it to break. They can be cleaned with alcohol.
The second method for replacing the bulbs is to remove the cluster bezel as described above and then replace the bulbs from the front. This involves removing the replacement bulbs from their sockets and inserting them in the dashboard sockets. To access the bulbs, a silver metal light shield over each bulb must be removed. Use a small tool and pull it out. The bad bulbs must be removed. Multiple methods have been used for this. One is to use longnose pliers and grab the bulb and pull it out. I don't recommend this method because of the possibility of breaking the bulb. Another method is to wrap tape around the bulb and pull it out. Many people use a small piece of rubber hose. There are various tools that are designed to remove bulbs like these. Here are a few that I know of.
A Jonard S-339, used to remove and replace bulbs in telephone switchboards (this is what I use):
The "official" tool, as listed in the FSM, Kent-Moore J-33885:
Here's a fun one. The "Mill Bulb Tool", designed for changing bulbs in pinball machines:
Last edited by Cliff Harris; 09-19-2013 at 04:01 AM. Reason: Corrected factual error.
09-07-2013, 03:31 AM
#2
Safety Car
Interesting analysis of the circuit, but did you figure out your problem? I would guess your photo sensor is failing. Anyone getting into this depth to repair their display should know about a few other tips:
1. If you need to replace the photo sensor, don't try to unsolder the defective sensor from the p.c. board. The circuit pads are so small, and they are "plated thru" holes, so there is a very high risk that you'll lift a pad. Snip the bad sensor right down the middle of the device, then cut each leg as long as possible. Lap solder the new sensor to these protruding legs. I used Philmore #10807 to fix my '84. (I bought a bunch of them.)
2. Replace the 882 light bulbs from the front of the display rather than removing the display module from the dash. Slide short pieces of heat shrink sleeving on your needle nose pliers, then pull the bulb out of its socket. The 882 light bulbs are available on eBay. The dealer prices are absolutely insane!
1. If you need to replace the photo sensor, don't try to unsolder the defective sensor from the p.c. board. The circuit pads are so small, and they are "plated thru" holes, so there is a very high risk that you'll lift a pad. Snip the bad sensor right down the middle of the device, then cut each leg as long as possible. Lap solder the new sensor to these protruding legs. I used Philmore #10807 to fix my '84. (I bought a bunch of them.)
2. Replace the 882 light bulbs from the front of the display rather than removing the display module from the dash. Slide short pieces of heat shrink sleeving on your needle nose pliers, then pull the bulb out of its socket. The 882 light bulbs are available on eBay. The dealer prices are absolutely insane!
09-07-2013, 04:09 AM
#3
Race Director
Thread Starter
1. If you need to replace the photo sensor, don't try to unsolder the defective sensor from the p.c. board. The circuit pads are so small, and they are "plated thru" holes, so there is a very high risk that you'll lift a pad. Snip the bad sensor right down the middle of the device, then cut each leg as long as possible. Lap solder the new sensor to these protruding legs.
The first time I took my dash apart all 4 bulbs were working. When I put it back in 3 were out. I couldn't believe it. I had two on hand and ordered a set of 4 from somebody on eBay (which is what I did several years ago). I figured buying 4 at a time is much cheaper in the long run.
Last edited by Cliff Harris; 09-07-2013 at 04:12 AM.
09-23-2013, 12:27 AM
#4
1st Gear
Member Since: Sep 2013
Posts: 1
Likes: 0
Received 0 Likes
on
0 Posts
power inverter
The power inverter can produce square wave, modified sine wave, pulsed sine wave, or sine wave depending on circuit design. The two dominant commercialized waveform types of inverters as of 2007 are modified sine wave and sine wave.There are two basic designs for producing household plug-in voltage from a lower-voltage DC source, the first of which uses a switching boost converter to produce a higher-voltage DC and then converts to AC. The second method converts DC to AC at battery level and uses a line-frequency transformer to create the output voltage.