Safe Speed Forums

Gatso speed cameras inaccurate in cold weather?

Motorist challenges speeding fine after discovering expert evidence on camera reading reliability
Posted: 27 May 2010
by Visordown News

Gatso: Not suitable for sub-zero temperatures?

A MOTORIST is challenging a speeding fine after discovering Gatso speed camera units may provide inaccurate readings in cold weather.

Andrew Fowler, 48, is to take his case to court after the dad-of-three was snapped allegedly travelling at 41mph in a 30mph limit, on a road he uses every day on his way to work.

The electrical engineer claims to have expert evidence showing Gatsos can give false readings in sub-zero temperatures.

Mr Fowler believes that vital timing circuitry inside the unit is affected by cold weather, making speed readings inaccurate.

If successful, Mr Fowler's case could throw 1000s of speeding fines already handed out into question.

"It's me versus the system, there's no two ways about it, and it is pretty daunting to be honest," said Mr Fowler.

"I'm nowhere near being a rich man but I don't mind how much this costs me because it is for a good cause, this is for all motorists.

"If I'm successful everyone will be at their crown court with speeding cases. Everyone will be able to appeal."

and at lower temperatures vital timing circuitry is slowed inside the unit.

He added: "If it's so cold the capacitor on the circuit can't hold its charge or is damaged the flash will not go off in time.

I'm sure somebody could work out his speed.

I don't buy it

However, i could accept that the camera might have been falsely triggered by the oncoming traffic on the other side - I've seen that happen many times.

How do you know it was falsely triggered?

Speaking as a seasoned electronic design engineer: I find it inconceivable that anyone could design a frequency reference that significantly slows at -5C (significant means more than 1%) into such a critical piece of equipment (evidence capture). For example, crystal oscillators usually have a drift of up to 100ppm over the operating temperature range, which is 0.01% (or 1% of 1%). That's certainly nowhere near the claimed 37% error.

I didn't say that it was falsely triggered, just that I do accept that it could happen - not that it did happen in this case.

...I've seen that happen many times.

Speaking as a seasoned electronic design engineer: I find it inconceivable that anyone could design a frequency reference that significantly slows at -5C (significant means more than 1%) into such a critical piece of equipment (evidence capture). For example, crystal oscillators usually have a drift of up to 100ppm over the operating temperature range, which is 0.01% (or 1% of 1%). That's certainly nowhere near the claimed 37% error.

He added: "If it's so cold the capacitor on the circuit can't hold its charge or is damaged the flash will not go off in time.

I was talking about the false triggering that you saw (not of this reported offence):

Capacitors fail at high temperature,s not low.

A timing anomaly might have been more plausible. I don't know what's inside the Gatso, but I do know that several microcontrollers have PLLs which are set by pin configuration, and that a bad solder joint on one of the pins (which could be affected by temperature) could conceivably change the clock frequency by 3/4, leading to both the doppler decoder reading 4/3 high and the camera timer being 4/3 slow. But I'm just speculating.

I prefer my speculation that the white lines on the road move closer together.

As an equally seasoned electronic designer I can only concur with that, Steve. I have just specified a crystal oscillator module which is accurate to 20ppm from -20 to 100 degC. And that isn't anything special - costs less than $10.

Capacitors fail at high temperature,s not low.

I prefer my speculation that the white lines on the road move closer together.

...and that a bad solder joint on one of the pins (which could be affected by temperature) could conceivably change the clock frequency by 3/4, leading to both the doppler decoder reading 4/3 high and the camera timer being 4/3 slow.