Sampling 24V one wire serial with a voltage divider?

Question

I would like to sample 24V one wire serial with a voltage divider, wouldn't grounding the resistors in series potentially affect the data being transmitted?

I was considering using an optocoupler but I am uncertain about how to select the correct type for a 115200 baud, half-duplex, 8-N-1, UART signal that is responsive enough.

According to documented reverse engineering efforts, I was able to find Inactive level is 24V, Active level is 0V.

Therefore:

"Mark" would correlate with the active level, which is 0V. "Space" would correlate with the inactive level, which is 24V

If an active state is being represented by a lower voltage level (0V in this case) and an inactive state by a higher voltage level (24V), it represents inverted logic.

With "eight data bits, no parity, one stop bit", a "BREAK" condition would involve the transmission line remaining in the "Space" state (24V) for longer than the period required to transmit a single byte plus the stop bit.

Given a baud rate of 115200, line can change state 115200 times a second, that is 115.2 kilohertz.

T = 1 / f gives us about 8.7 ms per state (high or low). If I recall correctly we have 1 start bit, 8 data bits, no parity bit and a stop bit so thats 10 bits.

So, that's a total of 10 bits, and 87 ms of time to transmit them.

I see modules online with this circuit design that seem like a solution to interface with controller at 24V but I am not certain about it being the correct solution. It appears on the out side you would set Vcc at 3.3V/5V depending on your MCU Rx voltage and IN+ to 24V. C1 provides noise filtering and/or decoupling functions.

115200 baud = 115200 cycles per second is a frequency and it is equivalent to 115200 Hz, and 6N137, is a 10MBit/s rated Optocoupler in theory.

In my opinion, this is the correct solution, but there might be better lower cost solutions that I'm not aware of.

Answer 1

The 6N137 is a good fit for 5V systems, if you have a 5V supply. If you only have 3.3V available, you'll need a different opto-isolator. At speeds of 115kBaud, jelly-bean devices like the 4N25 aren't really fast enough.

I did a search for devices that would work in this scenario, but stopped searching as soon as I found one, so my suggestion here is lazy. I found the H11L1M by Onsemi. It will work at 3.3V or 5V (or more) and handle up to 1Mbaud (they claim). This device also has a comparatively low LED forward current requirement, which might be important, as I'll discuss later.

If the document you linked to is to be believed, that the output you are sensing is open-collector/drain, then you are not as interested in voltage levels as you are the state of the output transistor (off or on). The LED of your opto-isolator will be a load connected high-side, between some positive potential and the transistor's collector/drain. This is easy if you have access to the remote +24V supply.

In this role, it is LED current that will flow or not, and you will size the LED's series resistor to pass the current necessary to switch on the opto-isolator's output, plus a little margin. For the H11L1M, that current is about 1.6mA; I will aim for 2mA to be safe:

^{simulate this circuit – Schematic created using CircuitLab}

$$ R_2 = \frac{24V - V_{D1}}{2mA} = 11.3k\Omega $$

Presumably your system already has R1 installed, and so I show it as part of that remote system. It's important not to overload output transistor Q1, by drawing excessive current through LED D1, which is why I mentioned the using an opto-isolator with low input current requirements. The 6N137, for example, requires 5mA LED current to "switch on". How much is too much depends entirely on the remote system, and that's all I can say without more information.

LED current flows when the COMM output is low (0V), which corresponds to a low output from the opto-isolator, so from a voltage level perspective, there is no signal inversion.