This weekend, I got a couple of projects done on my car. The first one (and the biggest pain!) was to properly align the driver side window and set the up stops correctly. The driver side door was replaced and the shop did not get it adjusted quite right (it was too far in and the stops were too high), so it clunked hard every time it went up.

That fix involved rolling the window up and down A LOT getting every thing set up correctly. I know I was hard on the battery because I end up leaving the ignition in the on position while I was testing which had everything powered up. And I had to test a LOT!

Got that project done (window is MUCH better aligned and the window seats like it used to!)

The next project was to install my interface circuit so I could control my driving lights (high and low beam) on/off with the factory fog light switch. To make a long story short, you have to have some kind of high impedence interface circuit so that you don't load down the SJB IC with the light on/off relay. Any way, I got everything working late Sunday night and called it quites.

The next morning, I still had to put parts of the dash and the door back together. Before I wrapped up the interface circuit, I wanted to give it one more test. Just to be sure...

Well, when I turned on the parking lights, everything worked normally. As soon as I turned on the head lights, all heck broke loose! My dash lights started flashing in time with the blinker flasher (but the turn lights were not coming on and the blinker relay wasn't clicking) and I could hear relays all over the car turning on and off opposite the dash lights!

I (fairly) quickly deduced that what was causing all the commotion was the CPU resetting itself (and the car with it). That surprised me because the CPU shouldn't reset until less than 10V or so... The trusty Fluke 87 came out and verified that the battery was at 9.6V.

I hooked up the batter charger and let it charge for 1/2 an hour then started the car. It fired right up so I let it idle for about 5 minutes with a couple of minutes at a high idle (yes, I disconnected the charger once the car started...) to get a bit of a charge back in the battery. After that, I put everything back together and all appeared well. The doors worked correctly, the lights worked perfectly, car was running right, radio was completely reset, etc.

It was only after I fired it back up and pulled it out of the garage that I noticed the "problem": If I kill the car and take the keys out, when I open the door it does NOT shut off the stereo. At all. I had to kill it with the power **** on the radio!

Everything else appears normal. It turns the interior lights on when you unlock the doors and then fades them out when you lock them back up. The alarm works perfectly. The window droop is working as well, so I KNOW the car understands when the door is opened and closed (I first thought I forgot to connect some wire in the door). The passener side door works exactly like the drivers side door and I didn't touch that door at all during this project.

Right now, I have the battery disconnected. In 1/2 hour or so, I'll go reconnect it and then retrain the ECU and see if all is well. Hopefully, the full reboot will do the trick on the car and get accessory delay working properly.

Does anyone have any ideas on what the problem could be? I have a feeling that if I can't get it working with a reboot, it is going to have to go to a dealer for a connect to a scanner so they can reprogram the SJB.

 

The root cause of the problem was running the battery down and causing mutiple computer resets through out the car. The mod itself works perfectly once the battery was charged. As does everything else on the car.

And just in case anyone else has any questions of what this mod does, here is the run down:

1) I tapped into a switched power and (fixed) ground from the automatic tranny connector under the center console. I have no use for that circuit as I have a manual tranny...

2) The power feeds a very simple circuit consisting of a CD4012 dual quad input NAND (not AND) gate. I am using one one of the NAND gates and actually have all four inputs tied together (which means I am using it as an inverting buffer). This chip draws VERY little current. Micro amps. The output of this chip goes through a current limiting resistor (to 20 mA) that drives the base of a transistor that conducts to ground. The transistor is fully saturated by 20 mA, so it is not dissipating much power when it is on. The On/Off coil is powered by a fused tap from the battery. The ground for this coil goes back to the collector of the transsistor. The emmiter on the transistor is grounded.

3) Operation is simple: When there is NO signal (aka: the switch is "on" and the signal on the OR/LG wire is grounded), the 4012 inverts and buffers that signal and causes the transistor to switch fully on. This causes current to flow through the coil of the relay and turns the driving lights on. When the switch is open, a pull up resistor in the car (not my circuit) pulls the OR/LG wire to 12V. This tells the car I want the fog lights off and it also causes my circuit to turn the driving lights off by sutting off the flow of current through the relays coil via the transistor.

This circuit can't pull enough current from the SJB because the 4012 is a buffer. It only pulls micro amps from the OR/LG wire to get the signal. It also can't pull much current from the automatic tranny shifter harness because it is only pulling about 20 mA (that's 0.020 Amps for those not familiar with electronics). That would be barely enough to turn on an LED, let alone an small light bulb. The real power to turn the relay on is provided by a seperate power feed that is external to the car's "normal" systems.

 

Oh, and resetting everything by pulling the ground for 1/2 an hour did not make any difference. Other than forcing me to reprogram my radio presets. Again.

 

RR,

I don't know if this would be of help to you but Scott Hoag said in another thread about cats and CEL light, to reset the car's computer the headlight switch must be switched on/off during battery disconnect

 

Thanks for the info Blackhat. I'll give it a try tonight when I get home. I have a feeling that I need to figure out a way to reset the Smart Junction Box. Hopefully, the light switch thing WILL tell the SJB to reset. The light switch is directly connected to the SJB and not the EC, so maybe...

Of course, that same feeling tells me I am going to have to take it to a dealer so they can hook up to the SJB and reset it directly. But I will definitely give this a try first!

 

Anytime RR

Good luck, I hope it works.

 

More testing done. And the interface circuit was the caust of the problem. It appears even with it's high input impedence, it is drawing too much current from the SJB and loading it down. That is causing it to be unable to do it's job and turn off the accessories. As soon as I cut the signal line from the fog light signal wire (Orange with light green stripe), it started working correctly, no reset required.

I think I need to double check the parameters on the CD4012. It looks like it is pulling more current than I thought. Or, the SJB has VERY small pull up resistors on it's inputs (probably internal) and can't source that little bit extra current that I need.

If the CD4012 is pulling more than I thought it should, then I will need to build a new circuit and add in a super high input impedence buffer of some kind. If it's an issue with the pull ups in the SJB, then I guess I should just need to add an external pull up resistor on my circuit.

More testing to come! I'm just SO GLAD that I didn't mess anything up in the SJB! And thanks for the suggestions.

 

Update:

I finished building and installing Rev 1 of the driving light interface circuit today. There were only a couple of differences between Rev 0 and Rev 1:

1) I used 47k ohm current limiting resistors instead of 4.7k ohm resistors. I reread the datasheet on the CD4012 and I probably could have used a 500k ohm resistor if I wanted too! The inputs REALLY don't require much current.

2) I configured each input (one for driving lights on/off and one for flash to pass) to use a 47k ohm pull up resistor attached to Vdd (12V accessory supply).

Basically, these chanes reduced the current draw of the part in total (larger resistors so 1/10 the current) but it also basically completely seperated my circuit from the SJB as the SJB does not supply ANY current to this circuit even though the inputs are common to both my circuit and the SJB.

Best of all, it all works perfectly! I have completely control over the lights and the accessory delay is working exactly like it should. Well, before I say that, I SHOULD test to see if the 10 minute delay works, but I KNOW it turns everything off exactly like it should when the door is opened.

So my battery SHOULD stay charged now!

 

Way over my head!!

 

Interesting read.