Overhead Console Computer

I've been reading this thread with great interest, although I know absolutely nothing about electronics beyond putting a piece of black tape over the flashing clock on my stove (it works great on "Check Engine" lights, too!)

I have a '96 XJ with an overhead console. Two years ago, it started sometimes working, sometimes displaying gibberish, and sometimes not working at all. After several months of this, the display went completely dark.

Three months ago, the display suddenly came to life, working perfectly! That continued for about a month, then it went back to cycling between perfect, gibberish and dark again. It's been dead almost all the time for the past several months. Every once in awhile, it will come to life for a moment, either displaying gibberish, or working perfectly, displaying the correct temperature and direction before going dark again.

What puzzles me is, I thought when resistors or capacitors burned out, the electronics would be dead. Does it make sense that mine is intermittent?

I've applied the universal repair of giving it a good whack several times to no avail.

Some devices (resistors, capacitors and crystals for example) will erode before they break. This means that a 100 Ohm resistor will erode down to only providing 80 ohms of resistance and a 50uF capacitor might drop to only providing 40uF of capacitance. Circuits are often designed to accommodate 10-20% variance in a component but if a part erodes beyond that point you can see flaky behavior like you describe. It is important to note that the parts aren't going to heal themselves, they are absolutely going to fail if you keep using them so take this as a sign that it is time to fix them.

I have seen the gibberish before and it almost always relates to degradation in the the clock circuit (crystal). Since it is temperature related, a circuit might work on a cool day but get flaky once it gets hot. It is an easy fix (described in this thread) but once you are in there, it would be a good idea to change out the capacitors as well.
HTH

kastein said:

Crystals can be fairly sensitive to heat but I'm not sure how sensitive. The crystal (depending on whether it's parallel or series resonant) requires "loading capacitors" commonly in the range of 1-25 picofarads from each pin to ground, which I think are the ones you're talking about. A 3 pin crystal typically is actually just a 2 pin with a ground pin for the can included. There are also ceramic resonators, which typically have a ground pin as well as the 2 active pins, and have somewhat different temperature coefficients and available frequency ranges.

I'd try replacing the caps that connect to the crystal wiring if they are looking crispy or chemically destroyed.

And yeah I agree on the heat. Usually what kills electronics is heat, overvoltage, reverse voltage, or dried up electrolytic capacitors.

I believe that's just a linear voltage regulator, which is why it gets so damn hot. Converting 13.8 volts to 5 volts by just dumping the extra power as heat (your heat dissipation will be (13.8V - 5V) * current output, plus a little bit) is very very inefficient. A modern design would use a switch mode power supply rather than a linear, but back in '87 implementing those for every embedded system in a vehicle was a fairly pricey proposal and switchmode converters were nowhere near as efficient or cheap as they are today. It shouldn't be too difficult to retrofit with a switch mode converter if you wanted to - might have to put a daughterboard in to make it fit well, though. I'd start with an LM2575 Simple Switcher(r) and the 3-4 support components it requires if you want to take that route. There are probably others available that are more efficient (the LM2575 is usually around 77% efficient, a significant improvement over a 5 volt linear in this case, which is around 36% efficient!) but that's the first one that comes to mind.

Click to expand...

I finally had some time to mess around with these PCBs again. I figured it out. Got them connected to my scope and noticed that the clock signal was extremely unstable. Turns out when I replaced the quartz crystals, I needed to change the associated ceramic capacitor to one that matched the load capacitance spec on the datasheet. Did that and they both powered up instantly.

This thread is fantastic. I have been working in my overhead console install in my 96 today. I am having somewhat of an issue. When I power on the unit, I can see red lines in the display. But the unit will not come on. As I was running some quick voltage checks I found that if pin 4 on the regulator is jumped to ground, the unit works perferctly. And i dont have to hold it to ground,just a quick touch and its like it resets. I can cycle the key and it will come on. I can press the left button and turn it off and on. Its beautiful. However, if it sits for an hour or more, it will not power on. If I jump pin 4 to ground, it again comes on and works amazingly. I performed tge checks mentioned earlier and my circuits seem to be intact to the best of my knowledge. What am i missing?

Update: I was sitting in the Jeep preparing to write down some info. I turned the key on and let it sit for about 5 minutes. All of a sudden the module powered on without me having to manually reset.

Update II:So what i found was that pin 2 on the regulator is 4.76v. When power is put to the unit the voltage at this pin is about 2v then it slowly rises to ~4.76v. After it hits 3.09v the unit will power up but voltage will slowly rise to top out about 4.76v. The unit powers on fine its just slow to get that voltage.

if i had to guess it sounds like a capacitor slowly charging.

96crow said:

This thread is fantastic. I have been working in my overhead console install in my 96 today. I am having somewhat of an issue. When I power on the unit, I can see red lines in the display. But the unit will not come on. As I was running some quick voltage checks I found that if pin 4 on the regulator is jumped to ground, the unit works perferctly. And i dont have to hold it to ground,just a quick touch and its like it resets. I can cycle the key and it will come on. I can press the left button and turn it off and on. Its beautiful. However, if it sits for an hour or more, it will not power on. If I jump pin 4 to ground, it again comes on and works amazingly. I performed tge checks mentioned earlier and my circuits seem to be intact to the best of my knowledge. What am i missing?

Click to expand...

I have seen that before. IIRC, it pointed to a ground issue around the capacitor that sits between the 12V regulator and the 5V regulator. Again, IIRC, I fixed it by placing a wire between the ground on that capacitor and one side of the switch that sits on that side.

90xj06 said:

if i had to guess it sounds like a capacitor slowly charging.

Click to expand...

If that is the case, I would change all the electrolytic capacitors as soon as possible because when it eventually fails, it will create a lot of damage.

I have new caps ordered. I tried to post a picture but I couldnt. On the board, on the middle 100uf cap to the right of the small surface mount caps, I noticed that one time when it powered on the trace going to the left side of this cap glowed and smoked momentarily before the unit powered on.

96crow said:

I have new caps ordered. I tried to post a picture but I couldnt. On the board, on the middle 100uf cap to the right of the small surface mount caps, I noticed that one time when it powered on the trace going to the left side of this cap glowed and smoked momentarily before the unit powered on.

Click to expand...

I have seen that before. That trace will now be "open" and you will have to repair it using the steps I documented earlier in this thread.

I have all the parts and have done this a dozen times so if you prefer, we can exchange boards. I have a couple of good ones with new caps and crystals and can send you one in exchange for your board. Total cost would be simply shipping. Send me a PM if you choose this route.

so im back to working on my compass. i replaced all the caps and removed the little ceramic ones. all the traces are good but it wont turn on on its own. i have to ground the reset pin and then it will turn on. im not sure where to look now.

Having just read the ENTIRE thread...:wow:

sounds like an issue with the clock. but that's just my layman's observation.

I did swap the crystal out. It was a 3 pin blue blob. And I swapped it with a 2 pin metal casing. I only have another blue blob I could try. Don't know if it's good

I replaced the crystal and it turned on. But now it does not again. Going to try another crystal and see what happens.

HI! Know it's a old thread.
The pictures in the thread is gone.
This reset thing, where do I find that on the pcb?
I have tried change all caps, and the two voltage regulators and the christall.
Still dead, the L487 gets very hot.

(sorry bad English, I'm from Sweden)

I'm winging this from memory. But I think the country reset is pin 5 and goes to (pin 3 or pin 4) on the 5 pin regulator. Did yours have the blue 3 pin crystal or the silver metal one? I will try and find a spare one to look at tomorrow.

My have the blue 3 pin crystal.

Swecherokee said:

HI! Know it's a old thread.
The pictures in the thread is gone.
This reset thing, where do I find that on the pcb?
I have tried change all caps, and the two voltage regulators and the christall.
Still dead, the L487 gets very hot.

(sorry bad English, I'm from Sweden)

Click to expand...

I stopped using the L487 because it gets too hot. Instead, I now use an L4947 (Jeep moved to this part on the late model ZJ's) which gets warm to the touch but doesn't generate any where near the heat of the L487. They are available on eBay for about $3.00.

Sorry about the picture removal. Photobucket decided that I needed to pay them if I wanted to keep using their service and I decided that I didn't want to pay them so I had to remove the pictures.

The 3 pin crystal has an integrated capacitor that fails. I find it easier to solder in a standard 2 pin (4MHz) crystal and then add a 47pF capacitor between one of the leads on the backside (I use the one closest to the middle of the board) and the center hole (now unused - which ties to ground).

Not sure what the "reset" discussion is about. I have fixed a dozen of these things and haven't seen that. These regulators do have a "reset" function build into them but it is all hardwired and is triggered by a capacitor and a signal from the CPU. This function is only used by the CPU and there isn't any way to mess with it (other than soldering a wire between pin 5 and pin 2 which essentially bi-passes it). If after power on you have 4.5V-5V on pin 2, the reset function is working properly (if it is 0, you have a problem and may need to solder a 100nF capacitor between pin 4 and ground - this capacitor is what triggers the reset). Also, make sure that you have a good connection between pin 5 of the regulator and the middle 100uF capacitor (just above the crystal). This capacitor stabilizes the 5V signal and without a good connection, the reset function won't work (and you will see 0 volts on pin 2).

Note that the left most button on the console will turn the unit off so after applying power to the board, always press that button first to see if it comes on.

If you are still having problems, send me a PM with your email and I will send you the pictures.

HTH
Todd

90xj06 said:

so im back to working on my compass. i replaced all the caps and removed the little ceramic ones. all the traces are good but it wont turn on on its own. i have to ground the reset pin and then it will turn on. im not sure where to look now.

Click to expand...

You will find the datasheet for the L4947 at this link: http://www.mouser.com/ds/2/389/stmicroelectronics_cd00000074-329631.pdf

It is the same function as the L487 you have now but runs MUCH cooler!

What the datasheet describes is that the "reset" function supervises the output (on pin 5) and then charges up a 100nF capacitor (one of the tiny surface mount ones on the board) on pin 4 (to ground) to create a small delay. Once the cap is charged (microseconds), you should have 4.5V-5V of power on pin 2 of the regulator. If this isn't happening, you need to check two things:

1) the 100uF capacitor (sits in the middle of the three 100uF capacitors) is directly wired to pin 5. If not, the voltage on the regulator won't stabilize and pin 2 will never get energized. BTW, it goes without saying that you need to make sure that the capacitor is still good and that the negative side still connects to ground (the ground connection is one of the traces that frys so you need to double check it)

2) there is a small surface mount 100nF (0.1uF) capacitor that sits between pin 4 and ground. This can sometimes fry due to heat and if so, the pin will never charge and the board will never light. If you have a good meter that can check capacitance, you can try to check it but they are VERY cheap ($3.00 will buy you 100 of them) so I usually just change them out or (if the board is burned in this area), I switch to a ceramic "through-hole" capacitor and solder it on the backside of the board between pin 4 and ground.

Same offer as my earlier post - if you are still having trouble, send me a PM with your email address and I will send you some pics and further guidance.

Also note, as per my earlier post, you can "kinda" work around this by soldering a wire between pin 5 (5V out) and Pin 2 (reset out) on the regulator but I don't recommend this because it makes it impossible for the CPU to reset itself if it gets confused. Having said this, it has the effect of removing the reset function from the equation and the board should start up. I often do this when I start working on a board just to figure out what I am dealing with but I always complete the repair properly to avoid other issues.

HTH
Todd