Hello everyone,

I have taken a circuit from breadboard to PCB and I am having some issues with a few components. The device is a power controller designed to monitor 2 batteries. 1 battery is a backup battery and the other is the main battery. The main battery is also a starting battery for an engine. The power controllers job is to make sure that the main battery never drains past the point where it would no longer have the ability to start the engine. At the threshold where the main battery get's cutoff, say 12v, the backup battery is now the source of power down to it's own cutoff threshold, also 12v, at which point the power controller will shutdown the PC that is being used to log data. Hopefully that in a nutshell provides enough information to help me with an issue I am encountering.

I had 4 of these PCB's created with the limited knowledge I have in this area as well as leaning on various information available online. 1 of the completed PCB's does exactly as it is supposed to do and the other 3 do not. To make matters worse or maybe better in terms of problem solving, the other 3 exhibit the same exact behavior. I have 2 p-channel mosfets (FQP27P06) and 2 darlington transistors (BC517) used to turn on/off the mosfets.

On the working unit I am observing the following voltages:
Mosfet 1:
G: 0.658v
D: 14.65v
S: 14.65v

Transistor controlling Mosfet 1:
Collector: 0v
Base: 1.407v
Emitter: 0.656v

Mosfet 2:
G: 0.657v
D: 14.65v
S: 14.65v

Transistor controlling Mosfet 2:
Collector: 0v
Base: 1.405v
Emitter: 0.655v

On the other 3 units I have a variation of the following with the exact same behavior:

Mosfet 1:
G: 8.1v
D: 13.82v
S: 13.82v

Transistor controlling Mosfet 1:
Collector: 0v
Base: 0.77v
Emitter: 8.09v

Mosfet 2:
G: 8.11v
D: 0.001v
S: 13.82v

Transistor controlling Mosfet 2:
Collector: 0v
Base: 0.77v
Emitter: 8.11v

What I am hoping is that someone could look at my novice built PCB's and tell me what may be causing this from a layout or component choice perspective as well as if I am breaking any major dos or don'ts for PCB design. I hope I have provided enough info to do this but if not, please feel free to let me know what other info would be helpful, and lastly, please feel free to tear this apart, so I can get better at this stuff.

Thanks!

 

Welcome to AAC, @diablodude64.
The BC517 is NPN and so its collector should be positive of the emitter, but both collectors are connected to GND.

 

Welcome to AAC, @diablodude64.
The BC517 is NPN and so its collector should be positive of the emitter, but both collectors are connected to GND.

So after 180'ing the transistors, everything seems to be working. Thank you so much!

On another note, do you have any pointers about the board design that you would do differently?

 

i would consider different routing to avoid vias where possible, specially on power traces.
also some of the clearances are rather tight.
rest is less important, like 90deg turns. if you are ordering this from a PCB shop, check the prices, 2,4, and 6 layer PCBs are normally same cost. that can make the routing neater. you can move signals to inner layers and use outer layers for power.
i am not big fan of vertically mounted components on automotive circuits. things should be lying flat or be bolted to handle vibrations, otherwise all top heavy components will fail sooner or later. if you continue watching video or skip to 21min you will see what happens.

 

and i would highly recommend to work on some schematic layout, show annotation, avoid zillion crisscross connections, label wires or perhaps avoid using wires. labeled nets are very useful in PCB layout allowing you easily select significant ones (like power) and create different class (different trace width, clearance, via settings etc.). also automotive environment is very harsh, i would spend some time to protect it. in this circuit it is not obvious where the power is coming from, there is no protection of any kind (reverse polarity, transient suppression, filter...)

 

and i would highly recommend to work on some schematic layout, show annotation, avoid zillion crisscross connections, label wires or perhaps avoid using wires. labeled nets are very useful in PCB layout allowing you easily select significant ones (like power) and create different class (different trace width, clearance, via settings etc.). also automotive environment is very harsh, i would spend some time to protect it. in this circuit it is not obvious where the power is coming from, there is no protection of any kind (reverse polarity, transient suppression, filter...)

I really appreciate the feedback. Since everything seems to be working now, i'll update the PCB. Is there a rule of thumb for trace spacing? Does it change based on power vs signal trace?

 

primary criteria for trace spacing is voltage differential. other factors could be inductive and capacitive coupling. so the higher the voltage the larger gap need to be. when designing for even higher voltage, it is common to strategically place slots in the PCB to reduce creepage.

for increasing ampacity of a trace, sometimes solder mask may be slotted. this allows adding more metal (solder). this is not as good as copper but it helps somewhat...

 

primary criteria for trace spacing is voltage differential. other factors could be inductive and capacitive coupling. so the higher the voltage the larger gap need to be. when designing for even higher voltage, it is common to strategically place slots in the PCB to reduce creepage.

for increasing ampacity of a trace, sometimes solder mask may be slotted. this allows adding more metal (solder). this is not as good as copper but it helps somewhat...

View attachment 257158

Thanks for taking the time to put this information out there. I am working on the reverse polarity protection currently and I was hoping I could get some input. I am using FQP27P06 pmos simply because I have them already and maybe they are not the best for this so that might be my first question, but second, in my research I am reading that if the supply voltage is less than Vgs of the mosfet than I don't need a zener diode or pull down resistor. My project will never see more than 15V supply max. Would this work?

Thanks

 

My project will never see more than 15V supply max.

why do you think that? is the car going to be switched off at the moment you use your device?
How do you plan on ensuring that? manually plugging it in, taking readings then disconnect?

if this is supposed to be permanently installed in a vehicle, you do need to anticipate and take care of number of concerns.
12V battery is not really 12V, cars are packed with devices and many are inductive (starter, ignition, fans, seat window and mirrors motors...). all of those may be causing significant deviations from nominal 12V (even if briefly). add to that vibrations in cars operation, and possibly corroded or wet contact, nothing is in a way of intermittent connections. in other words you have a perfect storm.




So what is reasonable? I don't know, you need to be aware of your target consumers and what they drive. what study and data collection did you conduct?

maybe you are lucky and one unit works great for couple of months or longer. maybe you get bad luck and 20 others fail in one week. what do you think life span and warranty should be appropriate with your creations?

there are plenty of articles on bulletproofing designs for harsh automotive products.
some simple options include using series diode to endure correct polarity. then adding more components can help deal with other problems. for example in supply you could have an LC filter and transient suppression. but that is only power. to make your product really robust, you need to take care of every single signal.

 

why do you think that? is the car going to be switched off at the moment you use your device?
How do you plan on ensuring that? manually plugging it in, taking readings then disconnect?

if this is supposed to be permanently installed in a vehicle, you do need to anticipate and take care of number of concerns.
12V battery is not really 12V, cars are packed with devices and many are inductive (starter, ignition, fans, seat window and mirrors motors...). all of those may be causing significant deviations from nominal 12V (even if briefly). add to that vibrations in cars operation, and possibly corroded or wet contact, nothing is in a way of intermittent connections. in other words you have a perfect storm.

View attachment 257323


So what is reasonable? I don't know, you need to be aware of your target consumers and what they drive. what study and data collection did you conduct?

maybe you are lucky and one unit works great for couple of months or longer. maybe you get bad luck and 20 others fail in one week. what do you think life span and warranty should be appropriate with your creations?

there are plenty of articles on bulletproofing designs for harsh automotive products.
some simple options include using series diode to endure correct polarity. then adding more components can help deal with other problems. for example in supply you could have an LC filter and transient suppression. but that is only power. to make your product really robust, you need to take care of every single signal.

Again thank you for all of the explanations. This has been great info and suggestions. Assuming these things will not be an issue, and I am sacrificing robustness for cost to build or some ratio of it, would you have any suggestions on which PMOS to use?

With the config I mentioned above I noticed that everything powers on as it should with correct polarity and does not with reverse polarity however, I do have a 0.3A current draw in reverse polarity that is not so obvious to me why. I have my gates wired to ground through 100k resistors, drains wired to each voltage source (backup battery & starting battery respectively) and sources wired to there respective inputs to the power controller. Am I missing something here?

 

Okay, found one of my p-mosfets was bad. Replaced and now everything is working as it should.