On Tue, 23 Aug 2016 14:18:38 +1000, "David Eather" <eather@tpg.com.au>
wrote:

>On Tue, 23 Aug 2016 05:24:45 +1000, Jim Thompson  
><To-Email-Use-The-Envelope-Icon@on-my-web-site.com> wrote:
>
>> On Mon, 22 Aug 2016 11:52:39 -0700 (PDT), Lasse Langwadt Christensen
>> <langwadt@fonz.dk> wrote:
>>
>>> Den mandag den 22. august 2016 kl. 20.44.09 UTC+2 skrev  
>>> edward....@gmail.com:
>>>> Where do they connect auto immobilizer?  Ignition wire is simply, but  
>>>> too simple and obviously easy to by-pass.  For my old Ford, i disable  
>>>> it with the crankshaft sensor, since it would not start without it.
>>>>
>>>> I am using Intel Edison or Raspberry Pi to control an relay.  Both can  
>>>> do the job.  Edison is better because of the small size and embedded  
>>>> flash; However, it is also worse because of the embedded flash.  I had  
>>>> to reload the OS many times when it just disappear, due to the way  
>>>> they reload the OS on-board.    Fortunately, this would not be a  
>>>> problem with it power up all the time anyway.  RPI uses a separate  
>>>> uSD, loaded from PC.
>>>
>>> why on earth would you use something like an edison or RPI to control a  
>>> relay?
>>>
>>>
>>> -Lasse
>>
>> Some people can only "design" using a uC/uP ;-)
>> 		
>>                                         ...Jim Thompson
>
>Yeah, but they need ridiculously big and growing bigger uC/uP 's

Transistors are free.  Might just as well use 'em.

Yes, I believe they are using something similar.  In addition, we want to limit the places and routes of where they can go.  That's why we need the immibilizer, if it's driven out of range.

Jasen Betts <jasen@xnet.co.nz> wrote:
> On 2016-08-22, Don Y <blockedofcourse@foo.invalid> wrote:
>> On 8/22/2016 12:38 PM, Rob wrote:
>>> edward.ming.lee@gmail.com <edward.ming.lee@gmail.com> wrote:
>>>> On Monday, August 22, 2016 at 12:02:59 PM UTC-7, Don Y wrote:
>>>>> On 8/22/2016 11:44 AM, edward.ming.lee@gmail.com wrote:
>>>>>> Where do they connect auto immobilizer?  Ignition wire is simply, but too
>>>>>> simple and obviously easy to by-pass.  For my old Ford, i disable it with
>>>>>> the crankshaft sensor, since it would not start without it.
>>>>>
>>>>> Modern vehicles disable the "engine" by preventing the engine control
>>>>> app from *running* the engine (i.e., no "wires" have been cut; the control
>>>>> signals are just inhibited -- no injector controls, etc.).  You'd have to mimic
>>>>> the functionality of the entire ECU to work-around it.
>>>>
>>>> So, is the RFID controller inside or close to the ECU?
>>>
>>> Probably yes.  I have a car which has this feature as standard and
>>> I think there is just an RFID readout that is connected to the ECU.
>>> The ECU has to be programmed with the keys it has to recognize, and when
>>> the readout does not read one of those numbers it will simply not run
>>> the engine.  It is still possible to run the starter motor.
>>> A message appears on the LCD telling about it.
>>>
>>> Each key has a passive RFID chip embedded.  This is separate from the
>>> active (battery-powered) transmitter used to open and close the doors.
>>
>> Are you sure the doors are dependant on battery power in the fob?
>> E.g., SWMBO's vehicle can unlock the doors/windows/etc. "from afar"
>> using the (battery powered) transmitter.  But, the doors unlock
>> when the fob is proximate to the vehicle *and* the door handle
>> grasped -- i.e., no need to fiddle with the fob in your pocket).
>>
>> Likewise, the engine starts with the fob tucked away in your pocket.
>>
>> [OTOH, I recall some fine print that said when the battery dies you hold
>> the fob right up against the START button to start the vehicle.  So,
>> perhaps the battery is used in ALL cases -- just used MORE when trying to
>> unlock remotely?]
>
> The range limit for inexpensive passive RFID is about 100mm, 
> for about $1000 (receiver/antenna cost)  you go to 1m 
>
> so yeah, that fob is almost certainly using the battery if it's in your
> pocket.

The explanation given earlier (battery powered device woken by local
emission of a low-frequency field) is the correct one for normal
operation.  The passive RFID pill is a backup.  In my car, that is
the only mechanism during engine start, but in the Renault cars that
is different.

edward.ming.lee@gmail.com <edward.ming.lee@gmail.com> wrote:
>> These fobs ("Driver 1" and "Driver 2" so the car can recognize *who*
>> is driving it) have a key inside that can be extracted to manually
>> unlock the doors.  It can also be withheld from a parking attendant
>> so they can't access the glove box, etc. (manual lock).
>
> We have more than two drivers.  We probably won't bother with removing the key.  It will just stay in the key slot.  However, we will enable/disable the engine and/or doors remotely.  That's why we need a mobile web server hooking up to the engine immobilizer.

We have a "car sharing system" operating here that does things like that.
(intended for people that need a car only infrequently, probably unthinkable
in the USA)

You pay a small monthly fee to get a logon ID and card and then you can
logon to a website, book a car near to where you are, it will tell you
where it has one available (they are on many block corners on a reserved
parking place) and when you present your card (contactless card of course)
it will recognize the fact that you have booked it and open the door
and enable the engine.   When you have returned and locked the car, some
rental amount for time and covered distance is charged to your account.

Probably a custom-made device for the company running that system, it will
communicate over 2G/3G/4G I think.

On Monday, August 22, 2016 at 9:18:42 PM UTC-7, David Eather wrote:
> On Tue, 23 Aug 2016 05:24:45 +1000, Jim Thompson  
> <To-Email-Use-The-Envelope-Icon@on-my-web-site.com> wrote:
> 
> > On Mon, 22 Aug 2016 11:52:39 -0700 (PDT), Lasse Langwadt Christensen
> > <langwadt@fonz.dk> wrote:
> >
> >> Den mandag den 22. august 2016 kl. 20.44.09 UTC+2 skrev  
> >> edward....@gmail.com:
> >>> Where do they connect auto immobilizer?  Ignition wire is simply, but  
> >>> too simple and obviously easy to by-pass.  For my old Ford, i disable  
> >>> it with the crankshaft sensor, since it would not start without it.
> >>>
> >>> I am using Intel Edison or Raspberry Pi to control an relay.  Both can  
> >>> do the job.  Edison is better because of the small size and embedded  
> >>> flash; However, it is also worse because of the embedded flash.  I had  
> >>> to reload the OS many times when it just disappear, due to the way  
> >>> they reload the OS on-board.    Fortunately, this would not be a  
> >>> problem with it power up all the time anyway.  RPI uses a separate  
> >>> uSD, loaded from PC.
> >>
> >> why on earth would you use something like an edison or RPI to control a  
> >> relay?
> >>
> >>
> >> -Lasse
> >
> > Some people can only "design" using a uC/uP ;-)
> > 		
> >                                         ...Jim Thompson
> 
> Yeah, but they need ridiculously big and growing bigger uC/uP 's

But all we care about is price & power consumption.

For $50, we get quad cores (RPI), 1G RAM, Wifi and bluetooth.  That's really pretty cheap.

Power is around 5W, mostly for the radios (including external USB cellular modem).

On Tue, 23 Aug 2016 05:24:45 +1000, Jim Thompson  
<To-Email-Use-The-Envelope-Icon@on-my-web-site.com> wrote:

> On Mon, 22 Aug 2016 11:52:39 -0700 (PDT), Lasse Langwadt Christensen
> <langwadt@fonz.dk> wrote:
>
>> Den mandag den 22. august 2016 kl. 20.44.09 UTC+2 skrev  
>> edward....@gmail.com:
>>> Where do they connect auto immobilizer?  Ignition wire is simply, but  
>>> too simple and obviously easy to by-pass.  For my old Ford, i disable  
>>> it with the crankshaft sensor, since it would not start without it.
>>>
>>> I am using Intel Edison or Raspberry Pi to control an relay.  Both can  
>>> do the job.  Edison is better because of the small size and embedded  
>>> flash; However, it is also worse because of the embedded flash.  I had  
>>> to reload the OS many times when it just disappear, due to the way  
>>> they reload the OS on-board.    Fortunately, this would not be a  
>>> problem with it power up all the time anyway.  RPI uses a separate  
>>> uSD, loaded from PC.
>>
>> why on earth would you use something like an edison or RPI to control a  
>> relay?
>>
>>
>> -Lasse
>
> Some people can only "design" using a uC/uP ;-)
> 		
>                                         ...Jim Thompson

Yeah, but they need ridiculously big and growing bigger uC/uP 's

On 8/22/2016 7:36 PM, edward.ming.lee@gmail.com wrote:
>> These fobs ("Driver 1" and "Driver 2" so the car can recognize *who* is
>> driving it) have a key inside that can be extracted to manually unlock the
>> doors.  It can also be withheld from a parking attendant so they can't
>> access the glove box, etc. (manual lock).
>
> We have more than two drivers.  We probably won't bother with removing the
> key.  It will just stay in the key slot.  However, we will enable/disable
> the engine and/or doors remotely.  That's why we need a mobile web server
> hooking up to the engine immobilizer.

Knowing who is driving only really is significant if the car
can *do* something, different, for each driver.  E.g., alter
the position of seat, mirrors, music preferences, etc.

IMO, current vehicles don't do *enough* of this as there are
many "preferences" that could easily be tracked that aren't
(and that doesn't begin to address driving *styles*!)

> These fobs ("Driver 1" and "Driver 2" so the car can recognize *who*
> is driving it) have a key inside that can be extracted to manually
> unlock the doors.  It can also be withheld from a parking attendant
> so they can't access the glove box, etc. (manual lock).

We have more than two drivers.  We probably won't bother with removing the key.  It will just stay in the key slot.  However, we will enable/disable the engine and/or doors remotely.  That's why we need a mobile web server hooking up to the engine immobilizer.

On 8/22/2016 1:40 PM, kevin93 wrote:
> On Monday, August 22, 2016 at 12:43:50 PM UTC-7, Don Y wrote: ..
>> Are you sure the doors are dependant on battery power in the fob? E.g.,
>> SWMBO's vehicle can unlock the doors/windows/etc. "from afar" using the
>> (battery powered) transmitter.  But, the doors unlock when the fob is
>> proximate to the vehicle *and* the door handle grasped -- i.e., no need to
>> fiddle with the fob in your pocket).
>>
>> Likewise, the engine starts with the fob tucked away in your pocket.
>>
>> [OTOH, I recall some fine print that said when the battery dies you hold
>> the fob right up against the START button to start the vehicle.  So,
>> perhaps the battery is used in ALL cases -- just used MORE when trying to
>> unlock remotely?]
>
> This is commonly done by the car transmitting on a low frequency such as
> 125kHz, this can be very well localized so that the car can tell whether the
> fob is outside the car or inside and which door is close to the fob.
>
> The fob has a 125kHz receiver which has a very low power consumption when
> idle (it has to run for years on a con-cell battery).  When it sees the
> 125kHz signal it replies on the 390/418/433MHz frequency to the car with its
> code.  When the car sees that together with the door handle button press it
> knows to unlock the doors etc.

That would make sense -- assuming the fob expects a certain *coded*
125KHz signal (e.g., Car_ID) and doesn't just spill its guts to *any*
such signal.

[I suspect the details of these protocols are available online with
a bit of looking.  OTOH, I found "homelink" to be pretty *opaque*]

> It is not easy to make a high frequency receiver (e.g. 433MHz) with the
> required low power consumption.
>
> This is often completely separate from the RFID like immobilizer function
> that does not require any battery.

That doesn't agree with the experience and documentation we've seen.
I.e., the car gladly starts with the fob "far from" the START button.
Yet, the owner's manual suggests you have to position the fob right
up against the START button when the battery (in the fob) has failed.
I think there's a coil in the fob that inductively couples power
into the fob from a companion coil in (near) the START button.

There's a "fob low battery" warning which leads me to believe the
fob transmits battery health to the vehicle (as trying to deduce this
from field strength would be prone to error when the fob is physically
distant from the normal driving position?)

On 8/22/2016 1:09 PM, Rob wrote:
>> E.g., SWMBO's vehicle can unlock the doors/windows/etc. "from afar"
>> using the (battery powered) transmitter.  But, the doors unlock
>> when the fob is proximate to the vehicle *and* the door handle
>> grasped -- i.e., no need to fiddle with the fob in your pocket).
>
> That is a Renault, I think?  My car does not have that feature.
> It is just a "standard" key which can open the door mechanically,
> with buttons to send an RF signal (433.920 MHz) received by the car
> to operate the electrical door locks when you are with 20 meters or so
> from the car when pressing the button.  I always open/close it using that.

Previous car was like that -- key with buttons on it.  "Chip" in
the key that chatted with something in the car WHILE inserted in
the lock.

This one has a fob (with buttons) -- and "emergency key" INSIDE it.

>> Likewise, the engine starts with the fob tucked away in your pocket.
>
> I need to start it with the mechanical key that also unlocks the steer
> lock and powers the car.  The RFID is only there to prevent that thieves
> mechanically force the lock and start the engine.

That sounds more like the older style she had.  I.e., the key
"talked" to the car to allow it to be started.

>> [OTOH, I recall some fine print that said when the battery dies you hold
>> the fob right up against the START button to start the vehicle.  So,
>> perhaps the battery is used in ALL cases -- just used MORE when trying to
>> unlock remotely?]
>
> The key can be opened, and recently I needed to open it as the buttons
> had worn out.  I bought a replacement keycase and could swap the
> electronics (transmitter with lithium battery), RFID pill, and key beard
> separately from eachother.

These fobs ("Driver 1" and "Driver 2" so the car can recognize *who*
is driving it) have a key inside that can be extracted to manually
unlock the doors.  It can also be withheld from a parking attendant
so they can't access the glove box, etc. (manual lock).

When the battery fails, you are instructed to hold the fob right up
against the START button so the car can talk to the RFID chip inside
Normally, it can do so from a few feet away (indeed, if I climb in
the passenger side before she gets in on the driver side, the car will
think *I* will be driving -- a frequent annoyance).  No idea if this was
a technical limitation or a "security feature".

> During the time I have the car (12 years) it has happened once that it
> displayed the message which I only noticed when it did not start the
> engine when cranked.  Turning the key back to zero and re-trying fixed
> it (pffewww).