hello, I have "novice level" electrical knowledge and am doing some self learning stuff here - wanted to use transistors not FETS.
have tested on everycircuit and does seem to work.
I made a transistor based circuit that cuts current when voltage drops below 10.8 or so volts (the intent is to cut voltage when 12v supply starts to drop).
you can see I have a differential circuit here such that a voltage range doesnt chatter the circuit on and off rapidly - once off, voltage has to be 11.8 to come back on or so.
I'm curious what the methodology would be for sizing transistor resistors to the base - both for pulldowns and for the main current limiting resistor to each base.
I did some reading, but for scenarios like this I didn't quite understand.
probably a solid difference between "works" functionally and WORKS correctly.

Several of these transistors are setup to be darlington units if that matters.
would like to size for a theoretical max amp draw of 5 amps from the main load

 

Why no FETs?
Seems like a circuit that would be perfect for a P-channel FET, driven by a TL431, and three resistors.R4 is optional and adds a bit of hysteresis so that it doesn't jitter around the threshold point.

 

To provide a switch-off as a battery supply voltage drops can also be done with a comparator circuit that will be far simpler to adjust to switch quickly at exactly the voltage you choose for it to switch at.
Controlling higher power transistor switches will add a bit of complexity, but the design effort will be rather simple.

 

Why no FETs?
Seems like a circuit that would be perfect for a P-channel FET, driven by a TL431, and three resistors.View attachment 322676R4 is optional and adds a bit of hysteresis so that it doesn't jitter around the threshold point.

would love to, but this circuit is also on a system with an oldschool automotive coil - FETS are great until they are close to those scenarios, then if anything goes sideways they just blow up where a transistor can hang with it.

 

would love to, but this circuit is also on a system with an oldschool automotive coil - FETS are great until they are close to those scenarios, then if anything goes sideways they just blow up where a transistor can hang with it.

These days automotive coils are driven by IGBTs which have the same gate structure as a FET, and modern automotive coils are much the same as the old ones. Perhaps on a different shaped core, but same principle.

 

JFTR FETs are transistors.

 

These days automotive coils are driven by IGBTs which have the same gate structure as a FET, and modern automotive coils are much the same as the old ones. Perhaps on a different shaped core, but same principle.

I just want some info on transistor resistor sizing vs me doing it trial by fire style which is my typical testing MO.
This is a product called "carb cheater" for classic cars.
there are several thousand beta versions out there doing well - the only thing I have encountered is problems when someone loses an alternator or a voltage regulator, if they keep driving and battery voltage drops, issues arise - attempting to mitigate that.
I use an IGBT for rev limiter things - thats fine. your typical mosfets do have issues, the board was redesigned around transistors which mitigated any issues from ignition voltage spikes that the mosfets couldn't handle.

hence the transistorized nature of the above circuit -
I'm a superior programmer to my PCB design skills - was just curious about the methodology behind what I'm looking for when sizing pulldown and base resistors for the sake of full saturation vs more than I need for full saturation (if thats even the right term - again, self taught trial by fire guy)

 

I just want some info on transistor resistor sizing vs me doing it trial by fire style which is my typical testing MO.
This is a product called "carb cheater" for classic cars.
there are several thousand beta versions out there doing well - the only thing I have encountered is problems when someone loses an alternator or a voltage regulator, if they keep driving and battery voltage drops, issues arise - attempting to mitigate that.
I use an IGBT for rev limiter things - thats fine. your typical mosfets do have issues, the board was redesigned around transistors which mitigated any issues from ignition voltage spikes that the mosfets couldn't handle.

hence the transistorized nature of the above circuit -
I'm a superior programmer to my PCB design skills - was just curious about the methodology behind what I'm looking for when sizing pulldown and base resistors for the sake of full saturation vs more than I need for full saturation (if thats even the right term - again, self taught trial by fire guy)

I see. Could it be an over voltage situation that upsets it? If batteries or regulators fail the alternator output could go too high. Also, the “load dump” phenomenon (where a large load is suddenly removed and it takes a little time for the alternator to get back in regulation) could cause problems, especially in cars where there is sometimes very few loads on the electrical circuit.
What devices fail when itgoes wrong?

 

I see. Could it be an over voltage situation that upsets it? If batteries or regulators fail the alternator output could go too high. Also, the “load dump” phenomenon (where a large load is suddenly removed and it takes a little time for the alternator to get back in regulation) could cause problems, especially in cars where there is sometimes very few loads on the electrical circuit.
What devices fail when itgoes wrong?

I seem to function okay even up past 24volts - several buck converters for 5v 3.3v, and 12v.
oddly enough, most of the time no issue arises even with these anomalous vehicle charging system failures.
Although, I have had 3 times now where someone's battery voltage is down in the 10v range sustained where the processor associated with the wideband o2 fails (which is run off of 12v) - enough for me to look into.

my thought was to slap a PTC on my main voltage line, and employ a simple under voltage latching shutdown circuit - not sure if I would ever actually integrate it, even if testing proved clean - one of those "increased susceptibility to failure via complexity" vs "reliability through simplicity" conundrums.

which again, is kind of the idea of my test circuit in my original post - was hoping to glean some transistor resistor sizing knowledge over my current what I would consider to be "effective but novice" understanding.