RS422/RS232 (5V/3.3V) --> DS8921 + TXB0104

Hi Riccardo,

I just want to make sure I understand the application.

Sensor (5V) Outputs RS422 signals ->  MCU Input requires RS232 signal

I would verify the electrical characteristics the MCU is looking for on the RS232 interface.  The RS232 interface can have signal levels as high as +/- 25V!

Conversion from a RS422 differential output to RS232 input can be handled in two steps.

A DS485 Transceiver (used as a receiver only in this case) can receive the RS422 signals from the Sensor.  The DS485 output is TTL. 

A MAX3221 Transceiver (used as a driver only in this case can receive TTL signals from the DS485 and output RS232 signals to the MCU.

Hope this helps,

Regards,

Lee

Thank you Lee for the quick reply.

I didn't find the answer on the datasheet of the MCU and I'm still waiting for a reply from the customer service for that.

I've been looking at the datasheet of both DS485 and MAX3221 and I wonder how do you estimate the power consumption of such a system.

I can't find any data about Cpd.

How would you proceed?

Best,

Riccardo

Riccardo,

Power in the DS485 will be very small since only the receive function is operating.  Most power in these devices comes from internal heating due to output load current in the RS485 driver side.  For the DS485 device less than 1mA is needed.  Dynamically there will be some additional current due to Cpd, this will also be quite small due to the minimal device loading and low operating frequency.  I would estimate << 5mA.

Power in the MAX3221 will be determined by the load current from the MCU.  Using the load  resistance at the MCU, you can estimate the internal power which will be dissipated in the DOUT output.  This will be relatively low as well.   

Regards,

Lee

Riccardo,

For MAX3221.


The power dispassion is simply VCC * ICC when the device in shutdown or the output is unloaded (RS232 cable disconnected). When the RS232 cable is connected and transmission is active, it is more complicated.
The receivers won't add much to the power consumption so I’ll focus on the drivers.
For each driver the DC output current is: I = DOUT / RL = 1.08mA (RL is load of remote RS232 receiver, normally 5k ohms. DOUT typical is +/-5.4V)
For each driver the switching (AC) output current is: I=F*C*10.8V^2/5.4V this is simplifies to I=F*C*21.6V. F is the driver frequency which is not the same as bits per second. With a continuous alternating bit pattern (worst case), F=bps/2. C is capacitive load on the driver line. 10.8V is the peak to peak voltage and 5.4V is the charge pump voltage. Also only one driver toggles at a time.
Add both AC & DC currents (for all drivers) calculated above to create a total current required of the charge pump. Now ICC can be calculated. ICC = 0.3mA + 2.3 * I[total]. The 0.3mA value is the typical static ICC, the 2.3 is the measured charge pump efficiency factor. If supply power is needed, multiply ICC by VCC.
This will provide a good estimation of typical active current or power consumption.
Temperature rise is P * TJA.

For example. 38400 baud, continuous alternating bit pattern, 2500pF 5k load, VCC=3.3V.
DC current is 1*5.4V/5k = 1.08mA
AC current is 1*38400/2*2500pF*21.6 = 1.04mA
Total ICC is 0.3mA + 2.3*(1.08mA+1.04mA) = 5.18mA
Power is 3.3V*5.18mA= 17.1mW

Thank you a lot Lee and Ron,

I still have to redo the calculation on my own.

I am looking at the DS485 and I don't understand why the lines of rx and tx are connected together.

Is the only reason to reduce the number of pins?

In my system the sensor is streaming data continuously toward the MCU. The MCU sends sometimes a command to the sensor:

reset, zeroing, selftest..

Is it ok to use such an interface? Do the signals not short-circuit?

Best,

Riccardo

Riccardo,

Yes it reduces pins, but most importantly it reduces cable to a single twisted pair. because of the single line anyone on the line can talk. The problem arises when two talk at the same time.

Yes, they will short out. Damage should not occur, but the data will be bad.

So I better use something like the DS8921, right?

Best,
Riccardo

Yes, that is better.

If you want RS232 and RS422 to share same lines then it won't work. The DS8921 doesn't allow negative voltages on driver outputs.
RS485 drivers do allow that.

You are right Ron.

I kept studying the datasheet and I see now what you and Lee meant.

I actually didn't describe the problem in the proper way, assuming this would be the easier one.

The MCU that I'm using has an integrated UART (which is compatible with RS-232).

And I need to interface it with this sensor running on RS-422 standard.

1) Is it ok to connect only the data signals (between the UART/RS-232 and the RS-422), leaving the control signals floating?

2) Is there anything like DS8921 but with negative voltages on driver outputs?

Best,

Riccardo

Riccardo,

UART's are not locked into voltage standards like RS232. Since the goal is RS422, there is no reason to use any RS232 devices.

Tell me about the RS422 data / interface.
How many TX and RX channels are used?
What is the bit rate?