This approach of going back and forth isn't going to work. The layout needs to be re-done to keep the spikes and RF coming in on every pin contained near the connectors.
For instance J1 B2 & B3 go right into the middle of the circuits with no input filtering.
Maybe if you think of the input capacitors as (water) seals. And the rest of the vehicle is being sprayed with water. If any of the seals lets water leak in then the unit will fail.
And I don't know what is coming in on the J2 pins. Your tranzorbs are "huge". What is coming in on those pins.
I've reworked a couple of my own circuit diagrams, so the component labels dont match yours.
J2-2 seems to be an analog voltage input, and I'm assuming you've calculated your voltage divider suitably. This probably just needs some simplification such as:
The BZT03C18 is a fast surge-rated zener, probably a bit of overkill, depends where that analog voltage comes from.
Or an alternative:
This should still protect the Arduino, and the zener wont be affecting the analog voltage unless the schottky conducts.
J2-3 seems to be the main power supply to the circuit.
I would move the TVS D2 as close to the connector as possible.
Add a thermistor as a auto-resettable fuse, alternatively use an old fashioned fuse.
Change D3 to be in series with the supply.
Sounds like I need three more tranzorbs near pins B1, B2 and B3. Those would be different because they are on 3.3V circuits used for three RTD sensors. I will have to find something from a different series of diodes. I chose the TP6KE18CA for the three power circuits because the Vr was just above automotive operating voltage and Vc was below the absolute maximum 28V operating limit of the Arduino voltage regulator. The three power circuits are from the Automotive battery C30, Ignition switch C15 and the radiator fan, which is a 14V signal pulled down to zero using the PWM transistor circuit.
Where each block represents an area of the board. These areas should not be mixed and should be as "tight" as reasonably possible.
Now your signal list:
J1:
B1, B2 & B3 are input/output signals that do not connect to 12V. No need for heavy transient protection for these signals. A 0.1 and a small low voltage TVS should be fine.
A4 goes to the High current controller input. I assume it goes to a 12V or 5V signal from the high current controller. Protection is the same as for the B inputs as this signal does not go directly to the Vehicle 12V battery.
J2-2 is a low current sense pin. So you take 1 or 2 K from R14. Put it in series with J2-2 with a low power TVS.
J2-3 is the only circuit coming directly from the battery. However you only need 6 volts going into the nano. So put a resistor in series with J2-3 to limit the current and put a smaller TVS after the resistor.
Also know at startup on a cold day the battery voltage is a sine wave that can go as low as 6 V.
See post #23 for my description on how the J1 and J2 connector is used and why I choose the tranzorbs. For the voltage divider I needed to get down below 3.3V for the analog input pin, therefore ~ 10:1 voltage reduction was used. It gives me ~ 1.2V at the pin. I used a large resistance to keep the current in check. I believe I can reduce the resistance to 10KΩ and 1KΩ but it currently works as is.
Currently I have everything the same as your first diagram with the exception of your R50 resistor and I am using a 0.1µf cap in stead of 1µf. I dont understand what your R50 does in the first diagram?
Your second diagram is getting a little too complex for my mechanical engineering background to comprehend so I would like to stick with something more like diagram 1.
You third diagram is also a little complex for me. You have to understand I'm working with a system that has been working perfectly for three years without any of this. So I'm just looking for the first layer of protection for the moment.
It looks like my board has no high frequency or ESD protection. I will say that the wires that connect to the RTD signals on J1 are about 3' to 5' long and are twisted going to the sensors, if that counts.
Sounds like 6 tranzorbs sized correctly will cover the high voltage transients, but probably not reverse voltage?
For bulk filtering, how does the placement of my caps look?
The recommended Nano 33 BLE runs to 21V, that is the main reason I selected it. The Nanno 33 BLE uses a voltage regulator with an absolute Vin max of 28V. I've considered the automotive voltage to be between 14 and 15 Volts, based on I have never seen a car running properly over 15V. If I need to consider 15 to 16V as the operational range, I will need to use a TP6KE20CA tranzorb in stead of the TPKKE18CA. Is 16V max the standard used in the automotive industry?
OK, only looked at the nano, didn't realize the BLE was different.
Is 16V max the standard used in the automotive industry?
Yes, but there is also a requirement for 24 volts (which you don't care about) for when folks put two batteries in series to get a stubborn engine started.
@JohnRob
This is my latest board. I swapped pins 2 and 3 on connector J2 to get a cleaner board layout. I tried adding lower voltage tranzorbs to pins B1, B2 and B3 but I ran into 2 problems. First my free version of Eagle Cad would not allow me to make more room on the board do to freeware limitation. Second I could not find a tranzorb with through hole mounting and the SMD I found were of 3.3V or 5V Vr. which seem too low and too high.
But in thinking about it, these are DC voltage outputs to 3 thermisters that go under the hood. I was thinking if anything maybe the tranzorb is really required at the 3.3V pin of the Arduino. I have included a blown up picture of the area in question so you can see the trace labels.
