Tomorrow, if it's not raining, I'm going to strap a HUGE capacitor of 30,000microfarads across my battery to see if it has any effect on what I think could be a spurious response from one of my car's microprocessors.
I'll post results, but it could take some time to check fully. For those of an electronics 'bent', I'm thinking there could be "hash" on the 12volt line from somewhere. I've met this one on a vtr. in the past. The 5volt line from the switchmode power supply had hash on it causing spurious effects.
Leedsman.

 

leedsman, I take it you are getting random noise into the electrical system and is causing you all sorts of grief with the interference. I can't say that I have heard of noise called "hash".

With that being said, you may find that you will get a better result if you wire the capacitor between the power wire at the fuse box and chassis ground. The idea behind this is that you are filtering only the power going to the computer and associated stuff, not the whole car. This will have a more dramatic effect than simply wiring it across the battery (especially since the battery is effectively something on the 100 farad scale).

I would also tell you to do some looking at your local stereo shops and see what you can find there. 30,000 microfarad isn't much really. You can get some caps at the local stereo shop that are in the neighborhood of 1-5 farad. Even a 0.5 farad cap would be tons more than you are looking at right now. I think I picked up my 1 farad cap for like $30 through e-bay.

On a side note, if you are getting a lot of electrical noise, especially enough to cause problems with your computers, it sounds like to me that one of the phases of your alternator has died. This is causing excessive ripple in your DC system. Essentially there are moments where the power in the car is dropping from 14 VDC down to 12 and then back up. This is too much of a change for the computer to compensate for.

 

Thanx for your contribution Thermo; The expression "hash" refers to what is seen on an oscilloscope screen when you monitor a power line and turn up the gain of the Y-channel. The resultant random noise appears as hash, called "grass" in the old days since the display is always in green. If you do this with your 'scope, make sure the X-input is a/c connected, thusly blocking the dc. of the power line. Use 50mv/cm. or better.
High capacitance capacitors were used in vtrs as a substitute for battery backup, generally being around 1.5farad BUT of only 5volt. It's hard to find these capacitors in higher than 5volt. While on with capacitors, there is a problem with internal resistance, called ESR, (Effective Series Resistance), and ESI (Effective Series Impedance), mainly to do with electrolytics. This is not a problem with other types, tantalum being the type most used when ESR/ESI is a problem. Tantalum are quite expensive. My electrolytic capacitors in the experiment are shunted by a 20microfarad of the plastic and foil type. There is now an organic electrolytic which is much better than plain ones, they were started by Sharp, again expensive. I just happened to have in stock the capacitors I'm using!
As you mention, the internal resistance of a lead-acid 12volt car battery is astonishingly low, prob. about .01Ohm. or less, but do we know the internal IMPEDANCE of such a battery? And what do we know about its stability with temperature cycling, being bounced around in a moving car etc? Your point re. one of the phases missing from the alternator is a good one, I might look into that. I'm strapping my capacitors across the battery 'cause it's easy to do, you're quite right about shunting capacitors being better when close to the sensitive load.
Leedsman.

 

Leedsman, from my playing with batteries, the internal resistance is like you stated, about 0.01 ohm. I base this on the fact that I tend to notice around a 0.1 VDC drop for every 10 amps of load.

As for the farad caps, that is why I mentioned car stereo shops. Most of them are rated at 15-20 VDC. So, they would do what you are after pretty nicely.

I have only been involved in electrical stuff for around 25 years. I started out with wiring up model trains, but have since moved on to new things. Most of my time has been spent on the repair side, not so much with the design end. But, I am still a young dog and can be taught many new tricks. I will have to keep the "hash" and "grass" terms handy. I like them.

Wish you luck in getting your problem resolved. I guess I am a big fan of overkill when it comes to problems like this. Why filter it some when you can filter it much more?

 

The one big snag with running down microprocessor-based faults/problems is the time element. By this I mean that the fault has happened long before you get your hands on it. By "long before" I mean electronically speaking, say a second. This is where logical fault-finding techniques as taught in technical colleges don't work. You have to use what seem almost like 'black arts' to run down a problem. Plus a big dollop of experience. The one I'm currently trying to fix can only be tried and tested as to its efficacy over time, say in months. Otherwise known as empiricle methods, "suck it and see" if you like. Sheer experience has shown me that plain old logic circuits are quite reliable, it's far more likely to be the power supply arrangements, or where there is a mechanical connexion (pots, sliders, constant flexing, etc.). If something gets hot and moves a lot, look there! Eg. throttle position sensors.
I remember with the Rowe-Ami compact-disc jukeboxes I used to service we had an enormous intractable problem in UK with these american machines, constantly going faulty in service, and most expensively so. They were ruining the Philips CDM-3 player, to a tune of £300 each. There were also massive software errors, such as coin-check registering in millions when the item had only been in service a couple of months, bridge rectifiers in the power amp constantly going dead-short (when you NEVER had to replace these kind of components usually). In the US, there were no such problems! It took me months to find the problem, which sounds so simple when the solution is written down like now.
It turned out to be caused by the fluorescent lamp in the jukebox's window! At that time, the starter for the lamp was the simple heat contactor, still in use even now. The lamp has a choke ballast to limit current in the lamp in operation (negative resistance characteristics remember) and when the contacts open, the lamp is "fired" due to the back EMF of the choke, and stays on until switched off. So why did the american version cause no problems? Very, very simple. The US versions were wired DIRECTLY across the american 110volt 60Hz mains/powerline supply, thusly any high-voltage glitches from the lamp's starter switch were 'clamped' be the very low internal resistance/impedance of that same supply. The british version was not quite the same. The UK version had a HUGE power transformer in the bottom of the cabinet to transform the 230volt UK mains supply to the 110volt US machine. SO.... because of the internal impedance introduced by that transformer, the lamp starter glitches were allowed free reign to cause havoc all over the place, there being NO very low impedance to clamp those naughty glitches.
When you get microprocessors in cars, you have a machine with a petrol engine which uses high-voltage sparks to fire the mixture. Not good bedfellows! Much care has to be taken to ensure that sparking anywhere in the car is prevented from affecting the control microprocessors (Jags have two, TCM and ECM). Apart from long leads between them, the ZF autobox manufacturer had actually put the TCM INSIDE the box! So it is subject to huge temperature cycling, quite unnecessarily. Capacitors, if electrolytic, have a limited life (1000h to 5000h depending on price and quality), which is severly shortened in high-temperature conditions. They dry out. This may or may not cause problems, you just don't know. Hence the randomness of faults around the microprocessors. Microprocessors in cars with an errorcode system are often used by mechanics as a substitute for thinking. The fault code produced should only be regarded as a suggestion from a lookup table. It is the microprocessors "best guess". It's still a guess though. Oh, it was so much simpler and easier to fix cars in the 1970s! And a LOT cheaper.
Leedsman.

 

Amen brother. I lvoe the cars from the 60's and 70's. Gotta love that simplicity and being able to see what is going on. Haven't seen one tech that could say he could see the power flowing through the wire (ok, there have been a few instances, but it didn't last long and had lots of people very excited, gotta love wires that are glowing white hot).

I find these digital problems as an interesting challenge. But then, some call me sadistic because I take on challenges like this. I'm sure we could talk shop for long, long periods of time with some of the fixes we have had to come up with. I used to fix the portable CD players when they first came out. When tapped in the middle of an ocean on a submarine, when your CD play breaks, you don't have a lot of options. I had to get very creative at times to make them work. Oh what you can do with superglue and strands of wire.

HOpefully whatever happens, you can come to a solution with your problem and make something that will let you enjoy your kitty.