Quad Mosfet switch.

Attached is the circuit I am using:

Look at a CD4504 IC, that's a hex voltage translator.



Note that the RX input, whether to a FET or IC, needs protection diodes to prevent excess voltages.


What are the "12V" TX and RX connecting to?

If you are trying to produce an RS232 serial port, the signals are inverted relative to the MCU logic and the transmit should go to a negative voltage not 0V.

A MAX232 is the simplest solution for that instance, it has protected drivers and receivers plus charge pumps to create the +/- supplies from 5V input.

Thanks for the info.
I was hoping for just 4 or more FETs in a small package. I might just have to get a circuit board made with discreet SM fets. It would be easier for me to just solder in 1 DIP and some resistors, especially if I knock out 10 or so boards at a time.

My TX and RX are not for RS232, they are slave/master and can be linked, so a number of TX and RX can operate on the same line. Because of this I need the open collector/open drain.

The 12V will have motors etc. on it, so I will have to put some more protection on it once I move past working on the comms.

All the multi-FET devices I can find are power devices in such as half bridge config.

And note the receiver threshold should be roughly around half the total signal voltage for noise immunity

If it's a shared one wire and ground open-collector bus, a K line interface could work, as that's exactly how that works.

They also have all the protection built in.

This one looks to be readily available:





Alternatively you could use CANBUS transceiver ICs, even though you are not using CAN. It's still a single shared bus with high noise immunity etc. for multipoint applications.

There are a few ICs form different makers that have compatible pinouts, also.

It isn't what I was expecting, but the K Line looks nearly perfect. I have not heard of K line before. A quick google says it is for very slow data rates.

My 4 Fet works (in ideal conditions) tested at 115200. For what I am doing I could probably get away with 9600. I should really look at the waveform of the transmitted data at this speed to see how fast It can reliably go.

I had nothing to do at work so I was playing with it. I had a series of mishaps today, starting with losing a screw while I modified a circuit board. I found it later when I turned on the system and smoke came out of one of my arduinos. After modifying all the circuit boards, my Master started talking to itself and getting stuck in a loop. I swapped circuit boards, now I have 2 boards acting up. Not sure where the problem is yet...

Tarsier79 said:

K Line looks nearly perfect. I have not heard of K line before. A quick google says it is for very slow data rates.

Click to expand...

The automotive system is rather low speed.

However, from that device datasheet, the delay is somewhere around 2 - 4uS with 500pF capacitive load, so I'd expect them to work at quite high speeds? Quite possibly 115200 depending on cable lengths and loading, or 57600 should be fine.

Here is a short list of dual and quad packaged n-channel mosfets.

I would suggest adding some protection to your gate pins. Both a series resistor into the gates, and a gate-source resistor. As they are, they're very susceptible to ESD damage.

https://www.digikey.com/en/products...lZRVUNIgBaFOh9KHkAVxNSSApGIXVNHjMAW2IcNDxJJaA