door triggers, diodes

My company installs GPS tracking devices in vehicles. I have completed 25 installs into 10 2010 Lincoln TownCars, 10 2010 Cadillac DTS's, 5 Cadillac Escalades for a Limo company. I am trying to monitor the Rear Doors and when they open it sends the signal to my tracking device which then throws a message to my server so the user on my program knows the vehicles door is open. On the Towncars it is one wire for the two rear doors so it is one splice and adding of a relay and you are good to go. With the Cadillac vehicles there is one wire for rear each door, and I wanted to monitor BOTH doors together. This requires use of dioides to prevent backflow into the BCM which could make the car think two doors are open if only one door is actually open....So I am using a 1amp diode to isolate each wire...I spliced into each door wire, put the dioide in line and then connected BOTH diodes to the one wire that feeds the relay then sends the signal to my tracking device. Out of the 15 Cadillac vehicles that were installed...2 DTS's and 1 Escalade got a fried BCM after about 7-10 days of using the vehicle and everything working properly. After 7-10 of the vehicles being used ALOT and everything working 100% including my tracking device and door sensing, the cars started having issues, headlights wouldn't turn off, blinkers didn't work, trunks wouldn't pop ETC... THe Cadillac tech looked at my companies install and said everything looks to be installed right yada yada yada....750 in labor and parts on ONE DTS...So I cannot have this happen to anymore vehicles....Any ideas, please feel free to PM me with any questions...The door triggers on the vehicles are negative door triggers so my install looks something like this...The cathode of the diodes are FACING the BCM which should not allow backflow into the BCM, only allowing flow into my relay then tracking device... AND here is more on dioides... https://www.the12volt.com/diodes/diodes.asp Please someone give me some insight...I have two installers working on these vehicles who have done alarms and other vehicle monitoring for over 5 years and I really cannot figure out why it would take 7-10 days for symptoms to appear...What do you guys think?!?

If I had to put money on anything, my money would be that they did not parallel a diode across the coil of the relay.  When the relay is deenergized there is a phenomenom that generates a severe voltage spike.  It is very severe when connected to voltage sensitive devices such as vehicle computers. 

http://voconversionbasics.websitetoolbox.com/post/show_single_post?pid=34889913&postcount=3

i am an idiot wrote: If I had to put money on anything, my money would be that they did not parallel a diode across the coil of the relay.  When the relay is deenergized there is a phenomenom that generates a severe voltage spike.  It is very severe when connected to voltage sensitive devices such as vehicle computers.  http://voconversionbasics.websitetoolbox.com/post/show_single_post?pid=34889913&postcount=3

But if that phenomenom did happen, wouldn't it shoot back out of the relay and hit my two diodes before getting to the BCM?

i am an idiot wrote: Without the diode across the coil of the relay, when the spike occurs, it may and probably is going to the BCM via the positive wire.  The power wire also feeds power to the BCM.

I am having trouble finding relays w/ diodes for the application I need...I am switching me relay on with GND, the only power wire is the constant power feeding the relay...That is grabbed from the IGN harness...if I put a dioide in that wire...Would that be suffice? I am connected a GND to the relay, the input is already diode isolated from the doors, the output goes into my tracking device, and the constant is grabbed from the IGN harness in the vehicle...

join to each door wire, between your join and the door module place a diode in line (1N4004) with the band towards the join then bring the wires together using 2 more diodes in line with the bands away from your relay. Last thing diode across 85 and 86 with the band towards the pos terminal. With this the signal will reach your relay, sleep circuit pulses will not and your circuitry will not affect the vehicle.

Sorry for my late answer, I missed Mr. Idiot's learned post but it's interesting that we independantly came up with the same answer. You can buy relays with built in diodes but that's expensive or Mr. I can link you to a photo showing the way he and I diode the relays. I'll try and show you how it's done on smaller PCB relays which are far handier to use when you are switching less than 10 amps. No problem constantly feeding the relay, it's only going to be triggered when a rear door is opened. Just a thought, if your tracker has a neg signal (or event) input, why do you even need a relay? P.S.The latest CAN/data wire fed units would pick all of this up without having to wire into the doors, we are using them on bus and taxi fleets.