So far, after wiring a HUGE capacitor combination across my battery, the appearance now and then of "restricted performance" error has not shown.

I'm wondering if anyone has done a comprehensive glitch-supression routine across starter motor, alternator, and battery using 15volt VDRs?

I'm suspecting that due to a long lead from the battery to starter motor with its impedance effect, power-glitches are appearing which affect the microprocessors in ECM and TCM during cranking, to show seconds later when engine is running. With the massive current flowing through the brushes in the starter motor, it's just possible that big glitches are being created here.

Anyone done any work on this one? Be grateful to know, might save me some work...

Leedsman.

 

I've not seen any posts along those lines.

I'm puzzled: if you've put the capacitor across the battery, why is it better than a battery usually is at suppressing hypothetical glitches? Sounds like a faulty battery, doesn't it?

Also, are these hypothetical glitches voltage drops or do you mean upward spikes? Got any measurements? (Probably need a DSO, which most people don't have!)

If there's a single feed going forward from the battery, as I think there is, wouldn't any glitches still occur at the front end and thus upset modules there?

Could it not be that your cap is helping hold the voltage up of a battery that otherwise droops i.e. is on its way out?

I guess another possible issue would be poor connection (power or ground) which then causes a voltage offset under heavy load. Have you cleaned and re-made the heavy-duty ones?

(I'd go for failing battery, 95%, but I'm guessing you don't think it's that.)

 

So far as I know, no-one has tested the IMPEDANCE of battery and its wiring/implements etc. at a typical glitch frequency of, say, around 100Khz. to 1Mhz. -- and I don't have such facilities! Not any more.
Glitches unf. aren't hypothetical, anytime something sparks in an electrical circuit, glitches travel along the wiring, and radiate out in the air as EMI (ElectroMagnetic Interference). There is a huge raft of european legislation called the EMC laws (ElectroMagnetic Compatibility) governing this very situation. In my last position, I was responsible for ensuring our equipment in the electronic security industry did not infringe these laws.

You cannot predict where and when glitches will affect your microprocessor based equipment, or why. It's a bit of a black art. It all happens so quickly, usually in microseconds. All you can do is suppress the likely origins, using experience and see what happens. The empirical approach. It would be nice if it were different, but that's how it is. Glitches can have an upswing and/or downswing, it depends on the high-frequency electrical characteristics of the circuits concerned. If you ever get the chance, check out the glitch recorder used by power supply authorities/companies. It records time, duration and amplitude of glitches on the electricity mains supply. The one I saw used a paper roll as readout. It's well known in England that glitches of 750volt. 'spikes' very often appear on the otherwise 230volt. single phase supply. Britain is famous for its 'dirty mains'. Check the glitch situation near a 25Kv. railway line! Especially if there is ice on the overhead power line.

My own vehicle battery is not faulty, when it cranks the (diesel) engine, it sounds like a sewing machine on heat. Cranking is hugely faster than average. The stored energy of any capacitor across a car battery would be about a millionth of the energy requirement of a starter motor, it would have no effect in that area. However, if the battery's IMPEDANCE was high at typical glitch frequencies, the capacitor, if big enough, could have a suppressing effect on glitches, as would a VDR if large enough. The VDR and capacitor would be better fitted to the offending item with short leads, chiefly I suspect, the starter motor.
So has anyone done any real work on this? I'd be grateful indeed for anything discovered.
Leedsman.

 

I already know that stuff, but hopefully it will be interesting to others.

Have you done the work I mentioned on power & ground? The thing is that you have a problem which others do not appear to have, so you're really looking for something wrong with your specific car (whether it's power/ground-related or other issue).

Others haven't done anything and posted about it over the last 2 or so years (i.e. since I've monitored the forum).

People will definitely have monitored the impedance of batteries across the frequency spectrum.

 

interesting info, but I do have to agree with Jagv8, a properly functioning Jaguar electrical system doesnt have these issues. Now Im not talking new cars because Im begging to feel,think, and hypothesize that many ecu gremlins may be caused by spikes, voltage drops etc... The cars comeup with too many weird non duplicatable "gliches" screen lockups, blackouts, no starts etcc have already seen issue with 3G,4G phones interferring with smart keys and starting ability. Of course we cannot replicate cause I dont have that phone, but 1 of our advisors has already shown it in a customer car when they dropped off without issue for oil change. Went to start it and no start. took his phone off hhis waist and put inside the building and then the car started. Brought the phone back out and again no start...I think there are other manufactures experiancing this too from other posts Ive read. As technology maches on it brings a whole new era of never before seen issues doesnt it

 

And the impedance of a typical 90a/h 12volt lead-acid battery at 500KHz is...?
Leedsman.

 

Unlikely to be posted on a Jgauar S-Type forum.

 

I agree: therefore I am pursuing this matter elsewhere. Although an automotive problem, it's an area where automotive experience doesn't apply.
Leedsman.

 

If your worried about suppressing RF noise on the power supply, what is the ESL and ESR of a large cap? Its hard to get a large cap that has a low impedance at high freq and typically smaller caps are used to shunt the high freq noise while a larger cap in parallel can help even out the lower freq voltage variation. Large caps that are used in power supplies typically don't look very good at high frequencies.

If it's a voltage drop issue due to high current load, a large cap by the electronic module giving the fault codes would be better then by the battery. Makes you wonder if some of the connectors have corroded so your seeing large voltage variations with current load changes effecting either the module (less likely since the ICs are tyically operating at 1.2, 3.3, or 5V - regulated from 12V) or its effecting one of the sensors connected to the module.

It terms of EMI, most electronics modules are designed to be pretty robust on the power supply, many times the glitches that cause the uP glitch are induced on the signal wires that go into and out of the shielded module.

A good test of how advanced the EMI design of a car is to use a GSM phone and the car stereo. The GSM phone will periodically send packets to the base station (1.9 GHz). Many audio systems will down convert this and you can hear a loud interferece noise (chirp) in the audio system in any mode - cd, am, fm, etc. My cars from '97 and '02 are very sensitive to this especially when the phone is in close juxtaposition to the receiver. My 04 jag, and 08 cars don't have an issue. Although this just tests the EMI suppression of the Audio system, it gives a good indication that the car may have been designed to supress RF emission from cell phones.

 

Interesting comment CCC. The ESR of an electrolytic is def. finite, and gets in the way of effectively "squashing" a glitch*. Hence I've used an additional plastic foil capacitor across mine -- not brilliant I agree. Further thought on the Jaguar situation makes me think about that very long battery lead being a possible impedance and even "aerial" (US antenna) that can pick up EMI from mobile 'phones etc. Although the broadcast from a mobile 'phone is 'supposed' to be GHz., it obviously isn't just GHz., otherwise audio and medium wave radios would not see it as a signal. This long battery lead then effectively sends the EMI to all devices connected to it! But I do appreciate the metal body will at least partially shield the lead from EMI OUTSIDE the car -- but then, what about EMI from INSIDE?

Elsewhere on the 'net, discussion re. these problems seems to fall into two categories, interference from OUTSIDE the car and interference from INSIDE it. The Toyota problem in that respect doesn't seem to be resoved at all. That one appears to be the engine control processor getting a wide-throttle demand, but not from the driver. If your analog sensor input from the pedal is A/D converted in, say, an 8-bit system (likely because it's cheap and has enough resolution) then IF the most significant bit in the latch/register is nudged by a glitch from 0 to 1, that would produce a near full-throttle binary number -- and off you jolly-well go!

My current thinking about that long Jaguar battery lead would be to put half a dozen giant ferrite beads on it at the starter motor end, lifting it's impedance to HF glitches from said motor. Also strategically placed 15volt VDRs at positions various in the 12volt supply lines could be indicated. I know these do effectively squash glitches. But I cannot prove where glitches might be coming from in this case, it's all experience-guesswork really, can't be anything else.

BTW., with the big capacitor across my battery, I still haven't had the "restricted performance" errorcode yet. But time will tell.

Leedsman.

*Terminology problem here, so I'm defining a "glitch" as an impulsive spike on a line that shouldn't be there. Often people say glitch when they simply mean a fault.
I use the word "bug" to mean a software error or conflict, unintended by the program writer.

 

The Toyota acceleration problem is not present on the jag because it's designed out (look at circuit to confirm, if interested). Toyota failed to design it out (details on the net, if interested google).

Maybe you just have a failing starter, as this doesn't appear to affect S-Types in general.

 

"Although the broadcast from a mobile 'phone is 'supposed' to be GHz., it obviously isn't just GHz., otherwise audio and medium wave radios would not see it as a signal"

The broadcast from cell phones are in the GHz range and there are a lot of expensive filters to make sure it stays in its bandwidth. The GHz signal enters the audio system and due to the non-linear behavior of some of the circuits in there, it gets down converted or mixed down to baseband audio noise that you hear.

The power lines are expected to be long so there is quite a bit of EMI filtering in most systems so I would not expect too much of the EMI effects going throught the power line, its typically the other signal wires.

In your system, since most cars don't show the fault, the key question is whats different about your system. I may be a voltage drop or spike issue when you start the car. This may cause some sensor to report a fault condition and hence the restriced error. There may be some high resistance point in your wire harness that the cap may be helping to filter out. Just my guess...