Hi -- I've read a lot, seen many circuits from LarryD, and learned much. My project is similar to others, but I still have specific requirements and questions. I'll try to keep it simple and straightforward:
I want to control devices typical in autos (lights, motors, etc). A heater blower motor is about the highest current draw, steady state close to 10A. Since all the things are grounded to the chassis, that means high-side switching is required. I need to level shift the 3.3v logic, and I'd like the circuit to be simple, reliable, and usable across a wide variety of loads. Here's my initial idea:
R2 and R3 are orders of magnitude too high, and and you need a flyback diode across the motor. Also, R2 should be connected from port pin to GND, not gate to GND. As shown, it forms a voltage divider to the gate, reducing the gate drive.
Automotive electrical systems are extremely harsh environments for consumer and DIY electronics, and you will need to include protection circuitry (at the very least, TVS diodes) to guard against possible voltage reversals and voltage spikes of hundreds of volts.
Thank you for the help. Yes, as I approach a final design, it will need to include protection circuitry. Also, I didn't show a flyback diode just for simplicity sake. However, the diode internal to the MOSFET may not be up to the task?
I was looking at R2 and R3 just as pull-down and pull-up resistors, so I didn't think too much about their values. Maybe you or someone else can help explain how to properly choose the resistance of these.
cedarlakeinstruments, not ignoring you -- the ICs look good, but 10-11 months is a bit much.
Assuming your circuit is not PWM it is good but needs some refinement. You are correct it is a Avalanche rated MOSFET so a flyback diode is not needed. Remove R2, change it to about 10K and place it on the port pin rather then the gate, this eliminated a voltage division. R3 should be changed to something in the 10K range. If it turns off to slow decrease that value. You should also place something in the 1K range in the drain of M2, this is needed to dissipate the load dump. The MOSFET is fine however I suggest you place a Zener to limit the VGS to a -15V max. Next comes the tricky part, what happens if the polarity is reversed? What will the loads do and is it allowed? What is the battery range, I am use to 6-24V survival for 60 seconds and 6-18 operating. If you want the easy out look at the Infineon BTS428L2, this is a SMD part and one of many available today.
A rule of thumb I use is two times the RDSon resistance during flyback. This is what you can use to calculate the heat generated during flyback. Note this peaks then decreases rapidly. The biggest problem with PWM is the switching is to slow causing it to stay in the linear reagan getting hot. The faster the switching the cooler it will be.
It is a serious mistake, and simplifies nothing, to not show all the details on a schematic, especially something as critical as a flyback diode. The body diode in the MOSFET is in the wrong place, and is not a substitute for the required flyback diode.
I was looking at R2 and R3 just as pull-down and pull-up resistors
There is much more to it than that, which is why successful EEs start by taking design courses at a university. Low values for pullup and pulldown resistors are required for noise immunity, which is especially important in such a noisy environment as automotive electronics. 1K is not unreasonable for R2 and R3. Don't forget to reposition R2 as mentioned in post #2.
Don't try to fool us, we know you are just guessing
Thank you for the constructive criticism--it's so much more enjoyable than the destructive kind.
I've made suggested improvements to the circuit as shown below. I am still a bit confused as the the placement of a load dump dissipating resistor. That would go where? Parallel to R3, or between the drain of M2 and the gate of M1? I put a "1k?" label to mark the spot in the schematic.
Easy out? Yes, that sound preferable, and what cedarlakeinstruments was trying to get me to look at. The BTS428L2 is great, but looks like it can't quite handle the nominal current I need. NCP45770 seems like it might work, and is currently available even if not quite as hobbyist-friendly as though-hole packages.
It is true that I am not an EE, but if I were I likely wouldn't be posting such a simple question to an Arduino forum.
And while many autos, especially more modern ones, use ground/earth switching, it is far from unusual to switch hot. I'd venture to say that almost all pre-80's vehicles simply bolted lights and motors to the chassis connected to a single switched and fused hot wire.
Yes the resistor would go from M2 Drain to R3,D1, and Gate of M1. The reason is to protect D1. Assuming you get a 60V load dump and M2 is on that is a lot of power. The zener may survive but I would rather be safe. I have more years in automotive electronics design then I want to mention and wanted to save you some future problems. Wrongly many think the car is simply 12VDC, actually it is one of the worst environments available for semiconductors. You are a fast learner, great job. This should last as long as your car but if you go into production there is more you have to do.
I would not put critical items like lights through these circuits - a switch is reliable and immune to noise . You need to consider safety .
I recall in days gone by when home made windscreen wiper delay circuits were in fashion . When you indicated you often got a free windscreen wiper sweep too .
The purpose of doing this is to bring modern controls to fans and lights, smart and adjustable wiper control, and a few nice modern upgrades to a classic auto.
Safety is a valid concern. For my purposes, most of the devices to be switched are accessories. For headlights, the plan is to have the circuit provide a level of smart control over DRL and full lights at dusk, but also allow for a parallel mechanical switch, just in case the system fails.
That's interesting. Are you documenting this anywhere?
I have been working on something similar: Bluetooth audio for the stereo system on a classic muscle car but with a "retro" display appearance.
I take it that shot was sent in my direction.
In reply I can only say that those that do not consider all possibilities as helpful are doomed to fail at some point.
Good luck getting any help then.
Tip...look up "Norm...I know boats".
No, there was no specific shot in your direction. Your comment that most automobiles are controlled via earth return, while not completely accurate, got me to google it and learn that many applications are.
Some of the help offered to me was kind and technically helpful, while other help brought sarcastic ridicule along with it. Both can offer lessons, but one is much more enjoyable.
