Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Dave DND]

TB,

Somehwere on the circuit board will be some form of diode protection for the charging circuit. I`m afraid that I do not know what shape / type / form it will take.

If there is any way to dig around and locate it, and if so, do you have a multimeter to take a forward and reverse reading from it ?

In theory, fire should not have altered the electrical characteristics of this diode component, so if it reads as short circuit or leaky, then that can only have happened pre-fire.

 :-?

[Kevin Wood]

I suspect the fire reports are probably where, over time, the batteries have leaks or vented, causing PCB damage?

But I fear that is a very easily reached conclusion as to what has happened, without looking a bit more in depth as to the cause, and every time another one goes up, we (myself included) are all too quick to blame the powersounder and add this to the growing list of torched cars instead of looking into it further.

Old rechargeables simply do not leak or vent on such a scale as we see here on the forum, and even if they did, as the gooey acidic mess comes out, there would be little or no current left inside to cause a spark. Find a leaking nicad that still holds a good charge and prove me wrong !!

But you are right that there are various types of Powersounder that all do the same thing, and as they are all having the same fault, the problem MUST lie with something else on the car that is causing this to happen.

I know this doesn't go anywhere towards explaining your observation of them combusting on the shelf, but when connected to an Omega they have access to an 80 A fused permanent 12v supply to aid combustion.

I have a PCB that I can connect up and experiment with variable supply voltages while watching the battery charging behaviour. I'll see if that turns anything up.

The link to the main alarm ECU is able to report back status of the power sounder as well as to activate / deactivate the alarm sounder. When a dying power sounder activates it gets logged in the main alarm ECU and a Tech 2 is also able to access diagnostic functions on the power sounder. (a test sounding can be commanded, and main and backup battery charge level can be shown IIRC).

EDIT:

Disagree - surely the transformer there could still theoretically allow it to happen ?

The transformer just steps up the squarewave output of the microcontroller on the PCB to a level suitable to drive the piezo sounder. I can't see a mechanism by which it would be rectified to generate a higher DC voltage.

Kevin

[Kevin Wood]

Another question:

Of the ones that have combusted - do we know what type of internals they had? I have seen at least 2 variants of sounder which aren't potted, each with different batteries, and I believe there is more than one type of potted internals based on the pictures I've seen posted.

Just thinking that, while they all fail given time (what product relying on NiCd batteries doesn't?) it's quite likely that only one design has the really fatal flaw that can start a fire.

Kevin

[Dave DND]

Kevin,

thanks for the transformer identification, I knew there was one in there, just not sure what for.

 

If you can, monitor the voltage and current on the rechargeable pack, and see what hapens as the main battery voltage is turned down -

Also, what is there in the way of diode protection?

How do the above readings react if the diode is shorted out ?

 :-?

[Dave DND]

Another question:

Of the ones that have combusted - do we know what type of internals they had? I have seen at least 2 variants of sounder which aren't potted, each with different batteries, and I believe there is more than one type of potted internals based on the pictures I've seen posted.

Just thinking that, while they all fail given time (what product relying on NiCd batteries doesn't?)
Kevin

I`ve also only seen 2 non potted versions, but have also seen 2 potted versions both with slightly different shaped outer casings - very subtle differences around the mouth of the siren, but unsure how different the internals are

it's quite likely that only one design has the really fatal flaw that can start a fire.

No, there are at least two diferent styles on this forum from the pictures I have seen
And thats what prompted me to start digging further . . . .

[TheBoy]

TB,

Somehwere on the circuit board will be some form of diode protection for the charging circuit. I`m afraid that I do not know what shape / type / form it will take.

If there is any way to dig around and locate it, and if so, do you have a multimeter to take a forward and reverse reading from it ?

In theory, fire should not have altered the electrical characteristics of this diode component, so if it reads as short circuit or leaky, then that can only have happened pre-fire.

 :-?

Not sure its worth it on mine, as mine wasn't the usual failure.  I suspect mine was just getting tired after 10yrs use.  It actually still works, but is prone to sounding the siren for now reason.

The pic was just to show that type, and the batts...

[Kevin Wood]

Had a prod at a power sounder board that still shows signs of life after the battery crud has been cleaned off.

2 12V supplies. Permanent and switched.

7.2V Ni Cd battery  (2 x 3.6V)

Battery is charged from the permanent 12V input at a rate of 15mA. Regulated between 14 and 16mA over the range of 10V - 16V input. Above that the charging is cut off. Almost as if there's an overcurrent/voltage protection circuit.

Current pretty much stable down to a short circuit where the NiCd battery should be.

Charging diminishes below 10V input and by 8V there's none, as expected. Alarm sounds below about 6V input, and whenever the power is disconnected. Alarm is silenced when both 12V supplies are restored.

The feed to the sounder (and most of the other circuitry) is fed via a pair of diodes so that it is fed from either the battery or the permanent 12V input, whichever has greatest voltage. Failure of one of these diodes could therefore result in 12V into the rechargeable battery or the rechargeable battery trying to drive the car electrics.

Alarm sounder transformer is indeed driven by what looks like a MOSFET straight from the junction of the 2 diodes. Transformer primary DC resistance is about 1.2 ohms. Feed to the MOSFET comes from a pin on the microcontroller which is toggled up and down to generate the tones.

I hope there isn't a DC path from the microcontroller to the MOSFET as this pin ending up in the wrong state due to a firmware glitch would result in combustion without a doubt.

I must double check the above.

The area around the MOSFET is one of the areas that was affected by battery leakage on this unit.

Kevin

[Dave DND]

Full set of Brownie points to Kevin !!!!

Battery is charged from the permanent 12V input at a rate of 15mA. Regulated between 14 and 16mA over the range of 10V - 16V input. Above that the charging is cut off. Almost as if there's an overcurrent/voltage protection circuit.

Current pretty much stable down to a short circuit where the NiCd battery should be.

About what I would have expected

Charging diminishes below 10V input and by 8V there's none, as expected. Alarm sounds below about 6V input, and whenever the power is disconnected. Alarm is silenced when both 12V supplies are restored.

Again as expected, but surprised that the low voltage "Beep" didn`t kick in - although a feature of the powersounder, its consistant with others who say that they have never heard it.

The feed to the sounder (and most of the other circuitry) is fed via a pair of diodes so that it is fed from either the battery or the permanent 12V input, whichever has greatest voltage. Failure of one of these diodes could therefore result in 12V into the rechargeable battery or the rechargeable battery trying to drive the car electrics.

Thats worrying !!
Do they read ok though?
Cheesy / cheap components ?

I hope there isn't a DC path from the microcontroller to the MOSFET as this pin ending up in the wrong state due to a firmware glitch would result in combustion without a doubt.

If you mean a software fault on the main Alarm control box, then you can pretty much rule that one out - the units are virtually indestructable and apart from a few dirty or worn out indicator relays, seldom give any cause for concern. The programming that is done to them for various alarm features etc, does not affect the firmware in any way whatsoever. The comminication between the two is done at around 5v IIRC. However, taking on board your comments, can you elaborate your theory any further with regards to this

The area around the MOSFET is one of the areas that was affected by battery leakage on this unit.

So another problem could occur from here once the battery had drained onto it?



Very interesting results . .   

[Kevin Wood]

Didn't check for beep. May well have missed it because I had the sounder disconnected to avoid the din but you can hear some feint screeching from the transformer when it kicks off.

Thats worrying !!
Do they read ok though?
Cheesy / cheap components ?

They were fine on this example but an example of what might go wrong. No sign of anything to mitigate against this (I would have liked to see a fuse on the PCB or  perhaps a zener or VDR across the input to protect against transients, fast chargers, etc.).

If you mean a software fault on the main Alarm control box, then you can pretty much rule that one out - the units are virtually indestructable and apart from a few dirty or worn out indicator relays, seldom give any cause for concern. The programming that is done to them for various alarm features etc, does not affect the firmware in any way whatsoever. The comminication between the two is done at around 5v IIRC. However, taking on board your comments, can you elaborate your theory any further with regards to this

I'm not referring to the main alarm unit. On the power sounder board is a small microcontroller and an E² memory chip. When the alarm sounds you get a squarewave which sweeps up and down in frequency on one of the output pins of the microcontroller. In other words, the firmware is toggling the pin up and down. This drives the MOSFET and via the transformer the alarm sounder so there is no oscillator outside the chip. The microcontroller directly generates the alarm sound. Now, let's say that microcontroller can enter a state where it leaves this pin active for longer then it should (due to low battery, voltage dipping during startup, etc.).This might burn out the transformer and cause a fire. Just a theory. I would have put a cap in series to break the DC path from the microcontroller myself but I will check that's the case.

So another problem could occur from here once the battery had drained onto it?

Yes, again, another theory but the gunge on this one had caused the mosfet to partially turn on, leading to a high drain on the battery. Nothing worse than that in this case but maybe...

Kevin

[Dave DND]

Interesting.

What blows most most of our theories out of the water however, is that there appears to be no other internal component faults except the battery itself, and so far replacement of that seems to bring the unit back to life.

Now, let's say that microcontroller can enter a state where it leaves this pin active for longer then it should (due to low battery, voltage dipping during startup, etc.).This might burn out the transformer and cause a fire. Just a theory.

Not sure the transformer would be so delicate, but wouldn`t this cause a continual drain on the battery and cause that to overheat instead?  Or, could continual drain / high useage of a rechargeable then cause it to leak? But either way, if that output was left in a high state wouldn`t that give some sort of audible tone?

Hang on . .  The sirens are know to "screetch" when they fail, aren`t they ?

the gunge on this one had caused the mosfet to partially turn on, leading to a high drain on the battery. Nothing worse than that in this case but maybe...

That is indeed one of the tell tale indications that the siren is knackered, but doesn`t help as to why the battery leaked on it in the first place.

Or maybe, this IS simply down to a bad design of battery pack that does start to leak over time.

Might be worth seeing if we can get hod of some other dead ones for you to perform an autopsy on, as you certainly have mastered the theory of operation I think, and would be good to make some comparisons.

Has anyone out there got any dead / blown / burnt out ones that we could disect ?

[Jimbob]

mine went in the bin sorry (it stank)

[Dave DND]

mine went in the bin sorry (it stank)

Is it still there? Or are you one of those who give blokes a bad name by actually emptying your garage bin once in a while?

 

[Jimbob]

lol it had to go, the acid stink got right up my nose whenever  I went in.

Sorry 

[Andy H]

When my main car battery required replacing last year (due to one dead cell) the power sounder was being triggered in full siren mode when the volts went low during a cold start. I can't remember whether it beeped as well  :-/

[Andy H]

7.2V Ni Cd battery  (2 x 3.6V)

Battery is charged from the permanent 12V input at a rate of 15mA. Regulated between 14 and 16mA over the range of 10V - 16V input. Above that the charging is cut off. Almost as if there's an overcurrent/voltage protection circuit.

Are there any markings indicating the capacity of the batteries? IIRC a safe charge rate would charge a NiCd in 10 hours so I wouldn't be surprised if the batteries were rated at 150 mAh. Continued charging at the 10 hour rate would destroy the cells in a few days so I presume the charging circuit must maintain a float voltage once they are fully charged.

My suspicion is that when the batteries get old then one cell will die first and the charging circuit will stay on at 15mA. What happens next depends on what the batteries do when they fail (ie where the acid goes )