How an EMC problem can kill people and a company

[eugenenine]

There are a lot of people who don't understand ABS.  I've heard people telling how they swerved around another car and "ABS saved their life" then later having to make a panic stop on a wet road and tell about how the pedal went to the floor and the car made noises and don't understand that was the first time they actually engaged the ABS.  Or others claiming how ABS can pump the brakes millions of times a second and fail to understand that the fluid will dampen the pulses so the pads aren't moving a million times a second.  Or that ABS is still reactive in that it takes a short amount of time to detect wheel stoppage (depending on how long of a delay is programmed to prevent false positives) so that it can't beat good driving where the driver can be pro-active.  Its been shows that people with ABS simply drive faster since the abs will; save them (simplifying here for brevity).  As vehicles become more and more appliance like we are going to have more and more accidents where the vahicle manufacturer is to blame since they didn't make it idiot proof enough.

[cdev]

I just wanted to mention a thought that I had. People on sites like this don't realize it but user communities like this one are really important in the feedback loop - the process of getting manufacturers to build safe equipment everywhere. If it wasn't for the voices of people saying "look at this horrible design" when one is found, and peer pressure, even across international and language borders, the situation would get totally out of control.

This thing with train crossings, there is a good chance that this is an occult cause of a fair number of fatal accidents, and it needs to be told, because the people who experienced it may not have gotten or get to tell the story because they are dead.

Very strong magnetic fields may exist wherever roads cross energized tracks- The smaller electric trains or trolleys usually have the power in overhead wires, they are farther away so likely not a major problem, but sometimes they go below and then they may be, I dont know, it seems quite plausible to me.

In the pre-MP3 era, I used to ride electric trains (The Bay Area's 'BART', SF's Muni Metro) wearing various Walkman style tape players - a lot - and often heard a whine in my headphones from the magnetic fields in the environment.

A strong magnetic field combined with bad design (especially grounding and intercommects between different parts of an e-vehicle) might cause an error in a microcontroller severe enough to make them reboot on a train crossing.

[helius]

With overhead catenary power, the return path is through the rails (and you will see short jumpers welded to the track to ensure electrical continuity across track switching points). Even in railways that are not electrified, there is current passing through the rails for signaling.

[cdev]

I've seen those small wires but there is absolutely no way that they could carry enough current for a trolley car by themselves. The track sections are always connected by metal plates with (bolts? rivets?) too, but they always look dirty and/or rusty and likely not perfect conductors either.

[Kjelt]

Today the results of an independent investigation by TNO has been published.
Although I haven't seen or read the final report, the minister has continued the roadban for this vehicle.
The media say (translated from quote below)

TNO has conducted an investigation of the stint vehicle on six important vehicle safety points.
On all six points the vehicle failed the safety norms.
Problems can occur with the brake of the vehicle, the brakedistance en an uncontrollable acceleration of the vehicle can occurr where by the driver no longer has control over the vehicle.

Looks like the company can be blamed for not taking care of these issues.

TNO heeft de stint gecontroleerd op zes punten die belangrijk zijn voor de veiligheid van een voertuig. De stint blijkt op al die punten niet veilig. Zo komt het onderzoeksinstituut tot de conclusie dat er problemen kunnen optreden met de rem van de stint en de lengte van de remweg. Daarnaast kan er een "onbeheersbare versnelling" optreden, waardoor de stint als het ware op hol slaat.

[Jeroen3]

Publications. https://www.tno.nl/nl/over-tno/nieuws/2018/12/presentatie-veiligheidsanalyse-stint/
TNO-2018-R11484
TNO-2018-R11485

Conclusion points, translated:
- None of the investigated vehicles brakes meet the 4 m/s² deceleration.
- None of the investigated vehicles offer a seat for the driver, increasing the risk of the driver falling of the vehicle. Losing control.
- Losing the "nuldraad", (negative wire) of the throttle results in undesirable acceleration. This can only be prevented by moving the throttle to 0%. However, it is deemed unlikely a driver in this situation would perform this action.
- The automatic parking brake can result in dangerous situation when activated spontaneously. Many failure modes activate the parking brake.
When on a dangerous location the vehicle can only be moved when manually lifting the parking brake.

De EMC testen tonen aan dat de aandrijving van voertuig H304 niet ontoelaatbaar beïnvloed wordt door elektromagnetische straling, ook niet bij hogere veldsterktes of bij andere oriëntatie van het voertuig ten opzichte van de antenne dan de norm voorschrijft. Het LCD display is wel gevoelig voor elektromagnetische straling. Dit laatste voldoet niet aan de norm.

EMC tests conclude the drivetrain of vehicle H304 is not unacceptably influenced by electromagnetic radiation, including high field strengths or other orientations than described in the standard.
The LCD is affected by electromagnetic radiation. This does not comply to the standard.

There is also a recommendation document. But I'm not a translation service.

[Ice-Tea]

In other words: it's junk, but it did not accelarate due to radiation related to the train stuff, right?

[Jeroen3]

Above report is an investigation into the safety of the type of vehicles by TNO. Not a forensic investigation in the crashed vehicle.

They have found that, using standardized electromagnetic interference testing, the drivetrain functions as specified. (Appendix 6 VN/ECE Reglement 10).
The display stops functioning between 47 tot 57 MHz.

There still is no report on the technical cause of the crashed vehicle.

[Berni]

Tho EMI immunity testing never exposes it to a strong 50Hz magnetic field. They stay in the MHz range for immunity testing because that's where its possible to produce lots of RF power with those antennas.

[metrologist]

BTW, has anyone tried turning off their non electric car in any gear while it's moving.

Yep. Everything was fine until the turn...

[RobK_NL]

Looks like the company can be blamed for not taking care of these issues.

Agreed, but our erstwhile government (f-ing Camiel Eurlings ... again!) is equally to blame for circumventing European legislation in order to "get these things on the road!"

[jmelson]

The ABS is typically pneumatic, while the main brakes are hydraulic. So the calipers have both fluid and air (or vacuum) actuators. Even removing the ABS module will leave the hydraulic brakes functional (any car with a brake fluid tank has that manual system, assisted with a pressure booster).

No.  The brake booster is pneumatic (or vacuum-actuated, to be pedantic).  But, the ABS system is all hydraulic.  There are CERTAINLY no air/vacuum cylinders or lines in the brake calipers.  Just look at one, sometime, there's just ONE pipe/hose to the caliper.  The pneumatic booster is on the firewall, attached between the brake pedal and the master cylinder.

There are failures in the ABS control valve that will cause one wheel to lose all braking, but since the ABS system has to have individual control of each wheel, then it should only affect one wheel.  These failures should be rare, as this is a safety-critical system, but due to corrosion and contaminants, it can happen.

Today, most cars have VSA as well as ABS, so the system can not only interrupt braking pressure to any wheel, it can also apply the brakes on any wheel that is spinning due to low traction.  So, there is a lot more stuff between the brake pedal and the wheel cylinders.

Jon

[Kjelt]

Looks like the company can be blamed for not taking care of these issues.

Agreed, but our erstwhile government (f-ing Camiel Eurlings ... again!) is equally to blame for circumventing European legislation in order to "get these things on the road!"

Yes saw the news they were really not paying attention back then 

[Halcyon]

"This device complies with Part 15 of the FCC rules. Operation is subject to the following two conditions: (1) this device may not cause harmful interference, and (2) this device must accept any interference received, including interference that may cause undesired operation."

I've never quite understood this rule.

How does a device not "accept" interference? If it were a choice, why wouldn't you design a device to reject interference, particularly of the type which causes undesired operation?

[Mr. Scram]

I've never quite understood this rule.

How does a device not "accept" interference? If it were a choice, why wouldn't you design a device to reject interference, particularly of the type which causes undesired operation?

In their definition of "accept" seems to mean "take" or "endure".

[nctnico]

De EMC testen tonen aan dat de aandrijving van voertuig H304 niet ontoelaatbaar beïnvloed wordt door elektromagnetische straling, ook niet bij hogere veldsterktes of bij andere oriëntatie van het voertuig ten opzichte van de antenne dan de norm voorschrijft. Het LCD display is wel gevoelig voor elektromagnetische straling. Dit laatste voldoet niet aan de norm.

EMC tests conclude the drivetrain of vehicle H304 is not unacceptably influenced by electromagnetic radiation, including high field strengths or other orientations than described in the standard.

But that doesn't say the standard is adequate and nothing can go wrong due to a pulse which has more energy. We'll have to wait until the accident report is ready to see if they have been able to find the root cause of the accident and if they did find the cause, what is the cause.

[Halcyon]

I've never quite understood this rule.

How does a device not "accept" interference? If it were a choice, why wouldn't you design a device to reject interference, particularly of the type which causes undesired operation?

In their definition of "accept" seems to mean "take" or "endure".

Sure and that's how I interpreted it, but if a device "accepts" interference but doesn't operate correctly because of it, how is that considered "accepting" or "enduring". To me if a device accepts interference, it performs more or less the same as it would without interference.

For example, if a computer suffers from interference but continues to operate, that to me is endurance. If a computer locks up or reboots due to interference, that product doesn't "accept" interference.

The wording is just very strange in my opinion.

[nctnico]

I agree. My guess is that an external electrical field shouldn't damage the device. For a machine this probably means going into a safe state and/or recover into a safe state.

[T3sl4co1l]

The FCC doesn't care about function, only emission.  They recommend you do the usual tests (immunity to conducted, radiated and transients), but do not require it.

CE does however require both.

Tim

[coppercone2]

Tho EMI immunity testing never exposes it to a strong 50Hz magnetic field. They stay in the MHz range for immunity testing because that's where its possible to produce lots of RF power with those antennas.

the military does this test with Helmholtz coils. I have some threads on this.

[Berni]

Tho EMI immunity testing never exposes it to a strong 50Hz magnetic field. They stay in the MHz range for immunity testing because that's where its possible to produce lots of RF power with those antennas.

the military does this test with Helmholtz coils. I have some threads on this.

I would like to see one these glorified segways put into such a coil.

Tho in the case of this incident the train was likely running on 750V DC. So there would be a constant field, but the contacts bouncing as the train moves probably causes some very sharp transients in the current. Such sharp transients would be particularly nasty for interference since the sharper the change the more voltage the field will induce in conductors.

Out of curiosity do these trains make compass needles go crazy around them when they go full throttle?

EDIT: Yes it seams to be 1.5kV instead of 750V for that train.

[nctnico]

The trains in the NL run from 1500V DC.

[Berni]

Ah yeah my mistake. 750V is only used in some towns there. But it is not the new 25kV 50Hz AC system (Likely not as bad magneticaly due to lower currents).

[eugenenine]

Also the lab tests are probably;y done on normally working samples.  The one that hit the train could have had a failed filter cap which let more EM into the system causing the failure.

[coppercone2]

scary shit that vehicles can become unsafe to fields like that now with robots and increased electronics.

I wonder if car inspections in the future will feature EM test chambers.