[groom]

Originally Posted by DerAlte

In your case, it was just a coincidence of a certain mass deflecting the sensor,

Not a coincidence, purely calibration grey zone!! ideally should not deploy.
The calibration should take care of "certain mass deflecting the sensor" kind of misuses.

Originally Posted by DerAlte

Airbag activation is unrelated to Force or Deformation.

Originally Posted by DerAlte

The primary decision
* depends only on deceleration profile
* is taken well before any deformation can happen

Airbag deployment is totally based on newtons second law of motion. Force - Mass X acceleration, because the mass of the vehicle is always constant, the deceleration or negative acceleration measured by the vibration sensor is nothing but forces acting on the car (deceleration caused by braking, dog hitting the sensor / bumper).

The decision must be taken while the deformation is about the happen or while the deformation has just started, not before the deformation can happen. We never know, the deformation may never happen (like dog case).

like i said before
d^2s/ dt^2 = deceleration or negative acceleration or force.

And btw this logic is very company dependent. Autoliv / Delphi may use only deceleration profile, TRW may use only rate of acceleration, Bosch may use force, displacement velocity. All of these companies have different patents in these methods. All of them work.

[PapaBravo]

Originally Posted by vikram_d

Doesn't matter if the car is expensive or cheap. All airbag systems calculate rate of deceleration for airbag deployment.

http://en.m.wikipedia.org/wiki/Airbags

When we are talking about deceleration, I could resist but crunch some numbers.

Glossary :

v = final velocity
u = initial velocity
a = acceleration
s = distance
t = time

Example 1 (Hitting a Dog in this case) :

Assumption made that the car deceleration was not affected by Dogs hit. (since mass of dog is less than 50 times the mass of car)

Now, stopping distance for a car travelling at 80 kmph(22.2m/s) is approx 53 meters ( Some Googling would get you this). Of this 53 meters 15 meters is thinking distance & 38 meters is braking distance, therefore, deceleration happens in 38 meters only.

On application of acceleration equation

v^2 = u^2 + 2*a*s

0^2 = 22.2^2 + 2*a*38

a = -6.5 meters per second square

Example 2 (Crash) :

In a crash a car comes to standstill situation in approx 0.03 seconds.

On application of acceleration equation

v = u + a*t

0 = 22.2 + a * 0.03

a = -740 meters per second square (or 75G)

Inference : The deceleration in a crash is more than ten times as in a dog hit case. Now how do airbags designed for deceleration in the range of 500 or more meter per second square open up in a deceleration condition of 6.5 meter per second square.

There should be a limit on the minimum deceleration on which the airbags get deployed otherwise they can open in any heavy braking instance & do more harm.

Just my two cents.

Regards

[DerAlte]

Originally Posted by groom

Not a coincidence, purely calibration grey zone!! ideally should not deploy.
The calibration should take care of "certain mass deflecting the sensor" kind of misuses. ...

Thankfully the companies that make Airbags don't conduct public polls to figure out what they should do. This is not a matter of 'should' kind of speculation, nor is it explainable in 'reverse engineering' terms!

Originally Posted by groom

... And btw this logic is very company dependent. Autoliv / Delphi may use ... TRW may use only ... Bosch may use ... All of them work.

Sure they do.

But, it would be illuminating for you to study what they do (there is a lot of public information beyond Wikipedia, should you chose to seek) - saying 'may' clouds the issue.

[vikram_d]

Originally Posted by PapaBravo

There should be a limit on the minimum deceleration on which the airbags get deployed otherwise they can open in any heavy braking instance & do more harm.

I completely agree with your numbers, but you have to look at it from the airbag controller's point of view. The airbag ECU does not know that you are about to hit a wall or a dog. Also there is no preset deceleration range for airbag deployment.

The airbag ECU is programmed with some amount of logic and this logic is based on the number of things it is getting feedback from. In my opinion what has happened in this case is as follows. Some people have mentioned that this particular Mazda is an inexpensive car. So taking this into consideration I am going to assume that it was only getting feedback from the accelerometer based sensors that are the primary feedback source in all airbag systems.

That being said I would just like to repeat what I said earlier in this thread. The airbag ECU just detected a high rate of deceleration and based on it's programmed logic it decided that the car was about to hit something. It didn't know if the car was about to hit a dog or a wall. When this decision was made by the ECU, it started to put the car into crash mode (don't know what else to call it). It pulled in the seat belts to make the sure the driver was held tight in his seat. After this happened, the ECU sensed an impact and decided to deploy the airbags.

I agree with everybody who says that the airbags should not have deployed, but the car/ECU did what it was supposed and what it had been programmed with. Is this particular logic in the ECU right or wrong, I don't know.

Now everybody imagine a scenario where the very same thing had happened and the car had hit the wall. We would all be praising the car instead of figuring out why airbags deployed at those speeds.

@PapaBravo - While I agree with your numbers, they don't take into consideration that the dog was not anchored to anything like a wall will be. The dog after taking the hit would have probably flown a few feet into the air. This act of the dog flying into the air would have prevented any major damage to the front bumper.

I have been in a similar situation where I have also hit a dog. I was travelling in my Swift doing about 80 - 90kmph on a highway and was on the left lane. A dog darted across the road from behind some bushes and I didn't even have the time to react. My seeing the dog and hitting him was almost simultaneous. The airbags in my car did not deploy and the only damage to my car was a small 1.5 inch cut/crack.

[PapaBravo]

Originally Posted by vikram_d

That being said I would just like to repeat what I said earlier in this thread. The airbag ECU just detected a high rate of deceleration and based on it's programmed logic it decided that the car was about to hit something. It didn't know if the car was about to hit a dog or a wall. When this decision was made by the ECU, it started to put the car into crash mode (don't know what else to call it). It pulled in the seat belts to make the sure the driver was held tight in his seat. After this happened, the ECU sensed an impact and decided to deploy the airbags.

That means there would have been a difference in time of tightening of seat belt & deploying of airbag. But since these things happen so fast it would be difficult to notice.

I totally agree with you on the deceleration part. I just needed to be sure that the airbags deploy in certain range if deceleration. There would be some limit of minimum deceleration crossing which would deploy the Airbags.

Originally Posted by vikram_d

I have been in a similar situation where I have also hit a dog. I was travelling in my Swift doing about 80 - 90kmph on a highway and was on the left lane. A dog darted across the road from behind some bushes and I didn't even have the time to react. My seeing the dog and hitting him was almost simultaneous. The airbags in my car did not deploy and the only damage to my car was a small 1.5 inch cut/crack.

Hope everyone in the car came out safe.

Regards

[vikram_d]

Originally Posted by PapaBravo

That means there would have been a difference in time of tightening of seat belt & deploying of airbag. But since these things happen so fast it would be difficult to notice.

I totally agree with you on the deceleration part. I just needed to be sure that the airbags deploy in certain range if deceleration. There would be some limit of minimum deceleration crossing which would deploy the Airbags.

Hope everyone in the car came out safe.

Regards

Yes there would have been a really tiny amount of time difference and as you said it would be almost impossible to notice.

I don't know if there are any minimum numbers as I have never heard about them, but that doesn't mean that there aren't any. There must be some sort of number on which the controller logic is based.

Oh yes, I was alone in the car. I stopped the car as soon as I hit the dog to check on the dog and spotted him still running madly across to the other side of the road. Thankfully s/he was ok.

[groom]

Originally Posted by DerAlte

But, it would be illuminating for you to study what they do (there is a lot of public information beyond Wikipedia, should you chose to seek) - saying 'may' clouds the issue.

Being in the crash safety industry from nine years and worked on more than 30 five star rated vehicles around the world, i choose to say "may" is because there are multiple algorithms available. Hence finally it is left to engineers to choose what suits the vehicle - profile / force transfer / displacement / velocity.

But the abuse set more or less should remain same. This abuse set should have included cases like deer hit / dog hit. Such a car may also deploy a bag if the hood is slammed hard.

[vikram_d]

Originally Posted by groom

Such a car may also deploy a bag if the hood is slammed hard.

I highly doubt that. I haven't seen a single car in which airbags have deployed when the car was stationary and if you are referring to the video on the previous page then I would like to inform you that this particular video is an ad made by one of Mercedes competitors. Will post the original video here once I find it.

Edit: Found it and the ad was made by Ikea and not by any competitor.

http://www.youtube.com/watch?v=u2zgr_Gg11U

[groom]

Originally Posted by vikram_d

I highly doubt that. I haven't seen a single car in which airbags have deployed when the car was stationary .

Absolutely, because engineers make sure "Hood Slamming" is part of the abuse tests and ensure that such cases don't deploy during the calibration of the ECU.

[DerAlte]

Originally Posted by groom

Being in the crash safety industry from nine years and worked on more than 30 five star rated vehicles around the world ...

Having the actual subject knowledge, and thinking that one knows what it is, is quite different. Perhaps your presumptions / ego, as you have stated above, is preventing you from seeing it as it is? Because ...

Originally Posted by groom

... Such a car may also deploy a bag if the hood is slammed hard.

belies that! There is no 'may' here - totally wrong axis in the context!

And, the video @vikram_d posted demonstrates what I had written earlier - a judiciously placed strike with a mallet / hammer will fire the airbag. Been there, done that! (That lady's bag was weighted - one can make out)

@vikram_d, if I may comment on your close-to-perfect explanation:

Originally Posted by vikram_d

... there is no preset deceleration range for airbag deployment. Correct, and it is not a 'range' at all, nor is it a simple number comparison (greater than). It is a 'profile', i.e. a sequence of acceleration numbers w.r.t. time (milliseconds) which leads to the 'crash' inference

... The airbag ECU just detected a high rate of deceleration and based on it's programmed logic it decided that the car was about to hit something. It didn't know if the car was about to hit a dog or a wall. When this decision was made by the ECU, it started to put the car into crash mode (don't know what else to call it) Correct. But, if the 'decision' is made it is already 'crash mode' - simple linear logic! . It pulled in the seat belts to make the sure the driver was held tight in his seat. After this happened, the ECU sensed an impact and decided to deploy the airbags. Actually, these are independent. SBPT gets the same 'crash mode' decision at the same time as the airbag ignitor. The explosion is faster than the SB pull-back (by a couple of hundred milliseconds)

... Is this particular logic in the ECU right or wrong, I don't know. It is not, that's how it should be. @PapaBravo rightly wondered "how do airbags designed for deceleration in the range of 500 or more meter per second square open up in a deceleration condition of 6.5 meter per second square". It is quite simple: the trigger point is related to the lower value, as waiting to calculate how big it is going to be is not a safe 'safety' practice. If one sees an avalanche start, one runs to safety - one doesn't wait to check the likely size of the avalanche!
...

[groom]

Originally Posted by DerAlte

Perhaps your presumptions / ego, ...

this is too personal and i rest my case in this forum.

Originally Posted by DerAlte

belies that! There is no 'may' here - totally wrong axis in the context!

It is may because the hood can be slammed at different speeds 1m/s. 2m/s or 5m/s,

[DerAlte]

Originally Posted by groom

this is too personal ...

Personal? Weren't you the person who said "Being in the crash safety industry from nine years and worked on more than 30 five star rated vehicles around the world"??? Not very difficult to understand, isn't it? And I had asked a question, I didn't make a statement so I wonder what the consternation is!

Originally Posted by groom

... It is may because the hood can be slammed at different speeds 1m/s. 2m/s or 5m/s,

I give up. I talk of you referring to the wrong axis in the airbag sensor context, and you talk of hood slamming speed?

[vikram_d]

Originally Posted by DerAlte

@vikram_d, if I may comment on your close-to-perfect explanation:

Thank you sir. Now my understanding of airbags and their deployment is pretty much complete.

[Tejas Ingle]

I went through the complete thread. Even I am not satisfied that hitting a dog should deploy the air-bags. I agree that deceleration caused due to braking puts a car in crash mode. Seat-belt pre-tensioners kick in, windows are rolled up, seats are straightened and all that. But deceleration due to braking and a following slight impact should not trigger the airbags. If that was the case either air-bags were out of India or the speed-breakers were out. I say this because hitting a speed-breaker at high speed makes much more impact than hitting a dog. And there are hundreds of them out there.
It is only the intensity of the impact that fires the air-bags. Companies perform thousands of tests and trials to make sure that the air-bags dont fire due to bumps or due to jumping over something. Hitting a dog while braking wouldn't even be felt inside the car since the dog is a light animal and its not anchored.
The possibilities of this deployment would be that the bumper detoriated sufficiently during the impact to set off the sensors behind it or the control unit was over sensitive which is a manufacturing defect. I have heard of a case where air-bags of a certain SUV were deployed after going over a bump at good speeds. The repairs were done free of cost by the company.
Please see this video of the Honda City and notice the extent of damage caused before the deployment of the air-bags. That is the intensity required to set off air-bags.

http://www.youtube.com/watch?v=d_uU7...yer_detailpage

[Jeroen]

i could be wrong, but I don't think heavy braking will put a car in "crash mode".

Slamming on the brakes should not tension the seat belts etc. There is some minimum deceleration rate per time unit to be met as pointed out earlier.

As far as I know, but I could be wrong, the seat belt tensioner and airbag deployment always come together. They do not occur separately. It's part of the same sequencing logic based on deceleration rate per time unit. Once that threshold is reached the belt tensioners fire and next the airbags. I'm not aware of situation or design that would only fire the belt tensiorners. (at least not on a straightforward forward speed deceleration)

Still, from all I read my two cents would be that the airbags should not have deployed in this particular case. So it was probably some sort of a technical glitch. The sensors or the computer or something.

The only other scenario I can think of is that whilst under heavy braking the "bump with the dog" pushed the system through the threshold. Possible, but not very likely I think unless it was a very big and heavy dog.

Jeroen