C5 Mk1 - over-volting - a challenging problem for electrical experts !!

[aspire_helen]

Here is a challenge for those knowledgeable on Citroen/Peugeot electrics! Good luck. Under what fault condition could the following occur, both intermittent and neither reproducible at will; both on battery power, NOT on alternator?
1. The front accessories socket (supplied by the battery), continued to show 3.6V even after removing its fuse. Replacing the fuse recovered the voltage back to battery voltage. Later, it measured battery (fully charged -12.5V) and alternator voltage (14.5V) correctly.
2. At a different time, when turning the key to position 1, the socket began at battery voltage for a few seconds, then began to climb to 14v, 15v, 16v, 18v over 3 to 5 seconds until the voltmeter (set at 20V range) gave up. I turned off the key rapidly. I reset the voltmeter to 200V range, and tried again. This time, it displayed battery voltage and stayed there.

This happened a second time, but this time when I was measuring the input voltage to the 22SY451 (the colour MFD screen / display ecu). On all other of many attempts the input was correct, both on battery and alternator.

It would appear the entire power supply is intermittently over-volting, WHEN ON BATTERY ! So, how can the car deliver 20volts plus (and climbing), when on a battery of 12.5V ?

I suspect this has caused a couple of 22SY451 ecus to fail, each one lasting several months / few thousand miles. After months of perfect operation, the symptoms started with an intermittent distorted screen picture. This often occurred on turning on ignition, before engine start. Indeed, on several occasions it could be cleared by several attempts of tuning the ignition off and on again. However, eventually the screen goes blank, terminally. It may be that the ecu can withstand several episodes of random over-volting before finally giving up. But then again, I am unaware of any other system having problems, even those using the same supply fuses and earth points. Perhaps the 22SY451’s over-voltage and transient voltage suppression (TVS) is not as good as other systems.
As a “sticky plaster” solution, I will fit a suppressor across the battery (eg Kemo M168). I will also fit a voltage limiter across the power supply pins to the 22SY451 itself eg varistor or TVS diode to clamp the its input. That may protect the 22SY451, but there is clearly a wider fault somewhere!! Any guidance would be most welcome. Pretty please

[PaulC5]

Possibly the alternator voltage regulator is faulty. For the accessory socket at first I thought there might be wiring coiled near it that was acting like a transformer or maybe some damaged wiring but then thought about the alternator.

[xantia_v6]

I suspect that your volt meter is dying. There is literally nothing that can cause the voltage to be more than 14.5V when the engine is not running.
Overvoltage with the engine running is usually caused by a faulty regulator in the alternator.

[CitroJim]

I suspect that your volt meter is dying. There is literally nothing that can cause the voltage to be more than 14.5V when the engine is not running.
Overvoltage with the engine running is usually caused by a faulty regulator in the alternator.

Absolutely so. Modern digital multimeters often have such high input impedances that can cause them to give erroneous and misleading readings.

As can RF interference from nearby items such as mains power supplies and mobile phones. Even being near a mains lead can cause some meters to read alarmingly high voltages...

If in doubt, use an old-school voltmeter with a real moving hand or shunt the input of your digital meter with a 1K resistor.

[aspire_helen]

Thanks for the thoughts. Yup, I thought measuring error initially. I used 2 different multi-meters throughout the whole process. I have experienced the effect of a low multi-meter battery.......some while ago, I spent hours trying to diagnose a measurement of 10v instead of 12v. I replaced the 9v battery and it suddenly read 12V (no low battery warning - grr).

Having measured the 3.6v at the accessories socket with one, I checked it with the second...3.6v again, remembering that the voltage should have been zero with the fuse removed...and then both read battery voltage with the fuse back in. That said, I did think the over-voltage "runaway" was a measuring equipment issue. But, I cannot reproduce it at will, with either multi-meter. Nevertheless, I will replace both batteries and re-check as you advise.

[aspire_helen]

Nope. The 3.6v is not a measurement issue. There is enough standing voltage to light an LED. Ideas still welcome!!!

To confirm, there is a standing voltage of 3.68v on the accessories socket (connected to car battery) with its fuse (F10) removed, in all states eg key on ACC, key on ignition, engine running etc. With fuse fitted, it measures battery and alternator voltages as expected.
Setting aside any multi-meter issue (see below), I connected an LED to the socket. It lit, but very dimly, suggesting it could not draw enough current to illuminate normally (even 10mA should be enough). So, I measured the current it was drawing as a mere 0.056mA. This was explained when I measured the voltage across the LED, which had dropped to 2.64v. In other words, there is a standing voltage of 3.68v but its source has a very, very low capacity. Such a low capacity might suggest it originates from an electronic circuit???? I have a Garmin dashcam connected to the same fuse (F10), but disconnecting that makes no difference…….so I am thinking the dear old BSI is throwing “un tantrum” ….grrr.
So, I also monitored the socket voltage as I switched on various consumers, eg side lights, radio, full beam etc (on car battery only, no key, F10 removed). On each occasion a consumer is powered, the socket voltage rises slightly by 10 to 20mV (eg 3.68v to 3.69v), and falls back when the consumer is turned off. The same happens with no consumers except the BSI (I assume it’s the BSI). When the BSI shuts down (nominally 60secs after no activity, as evidenced by a brief whir from the instruments, followed 5 secs later by an audible relay click), the socket voltage dips back. Usually, when power consumers are turned off, the source battery voltage should rise slightly.

All the above was repeatable. Ref my two multi-meters. Their batteries both registered 9v/”full” on a battery tester, but I replaced them anyway. After more testing, one (a cheap modern one) was definitely inconsistent on its 20V range, so that is scrapped. The other, 30 years old and higher quality, I deem OK.

As to the other issue of runaway over-voltage, on car battery, I am cautiously putting that down to the bad multi-meter, pending further testing of the voltage supply to the 22SY451 ecu, which started all this in the first place.

[moizeau]

I've never delved into car electrikery to the extent that some have. Having tested mains house wiring I know that an LED screwdriver tester can pick up a 'residual power source'. The circuit has been knocked off on the breaker (on both live and neutral) yet it still glows dimly. This happens when multiple circuits are sent down the same conduit. I've never tested it with a meter but you can touch it with no kick at all.
Someone with more electrickery knowledge will will explain I'm sure.

[CitroJim]

Having tested mains house wiring I know that an LED screwdriver tester can pick up a 'residual power source'. The circuit has been knocked off on the breaker (on both live and neutral) yet it still glows dimly. This happens when multiple circuits are sent down the same conduit. I've never tested it with a meter but you can touch it with no kick at all.
Someone with more electrickery knowledge will will explain I'm sure.

Yes, very much the case on house wiring Pete. What you see is 'transformer action' where live cables running close to those you have isolated induce small voltages in them and as you've seen, enough to illuminate an LED, the parallel wires are behaving like a transformer. This can cause modern LED lights to remain on dimly even even when switched off.

This is due to the supply to a house being Alternating Current (AC). A car runs exclusively on DC and with DC there can be no transformer action. One place where you can see a bit of transformer action is on HT leads where one can induce enough voltage into one running parallel to cause a cross-fire and fire a spark plug out of time. V8s can suffer this and to avoid this, HT leads running parallel are often twisted and crossed over to cancel out induced voltages and prevent cross-fire.

Helen, I can see why you have 3.8v on your accessory socket under those conditions as numerous 'leakage paths' exist all via the multitude of ECUs and wiring. It is nothing to worry about and it's a bit of a red herring. I've seen the same effect on many cars.

I think your biggest issue is your alternator. I had the alternator fail in my Pixo in such a way that it would often go massively over-voltage just for a few moments at a time and ended up cooking the battery. On a meter its output voltage looked a bit high but more or less OK. I needed an oscilloscope across the battery to see exactly what was happening. If an alternator gives out much more than 14.4v then its suspect.

As said, to ensure your meter is no affected by surrounding electric fields of whatever sort and to prevent them inducing confusion into your readings, shunt your meter input with a 1K resistor in parallel with a 100nF capacitor. Or find someone like me who still has an old-school AVO Meter with a proper moving hand. I would also expect measuring the accessory socket with an AVO would read zero...

[aspire_helen]

Thanks, again. Ref the alternator. Yup, I suspected the alternator and replaced the original alternator a year (December 2022) and 3,965miles ago. I was going to just change the regulator, but having dismantled it I found the slip rings were almost worn through, and a replacement alternator was only marginally cheaper than a new regulator. Lexia records its output as 14.4V.
But, I have had two second hand 22SY451 ecus fail since replacing the alternator! Both worked perfectly to start. Both developed intermittent screen distortion and eventually went blank. From other posts (under 22SY451) you will see I have focused on providing additional cooling for the ecu, such that I can monitor the ecu’s CPU temperature and keep it below 40degsC. However, such cooling did not protect the ecus!
I intend to insert a stable and protected voltage regulator into the ecu’s supply to eliminate any voltage transients or over-volting. I just hope the BSI wont object…..moody cow.

[CitroJim]

I intend to insert a stable and protected voltage regulator into the ecu’s supply to eliminate any voltage transients or over-volting. I just hope the BSI wont object…..moody cow.

That's a plan It should prima-facie be OK but shouldn't be needed. It's hiding the symptom rather getting to the root cause.

The Lexia will only show the charging voltage averaged out over time. It won't show short duration fluctuations and any high spiky peaks that may account for the killing of your ECUs.

Have you checked the integrity of every earthing point? A bit of resistance in an earthing point, or a disconnected one, can cause all sorts of strange issues...

Keep us posted, this is a most interesting problem...