RS232 folks,
I have a customer wanting to parallel 3 SN75185 outputs for their customer load needs.
Here is the actual email verbaige:
We use SN75185 RS-232 transceiver on a few of our legacy serial interface products. They are supplied with +/-12V and 5V power supply. These products have 16- serial ports and each serial port has its own RS-232 transceiver.
A large customer has an application where he’s connecting eight RS-232 devices into one serial port on our product. That’s a lot of load for one driver output on the transceiver. At each of the eight receiver input, they have a 6.5 Kohms in series (between the connector pin and the RX input on their RS-232 transceiver). Basically, the equivalent diagram for their application is attached. The internal pull-down for each of the RX input is ~ 5 Kohms.
Their requirement is that the driver output voltage must be > 8.5VDC when the total load on it is 1.15 Kohms. Which is close to 6mA.
I made some measurements today to see what each driver output’s output voltage on the SN75185 is with different output load. Here’s what I had measured. With a single driver output, the output voltage falls short of customer’s requirements. ‘75185 has 75.8Ω and 320Ω in series between the VDD and DYx output per the schematic of the driver. With the voltage drop calculation at those resistors, the numbers below are close to what the calculation shows what the output voltage should be. Will the transceiver chip and circuit operate reliably if I paralleled the 3- driver outputs together? I need to send a prototype unit to the customer for a proof of concept. Will it be reliable if I did this rework, tested it, and sent it to the customer for their evaluation?
Are there other alternative solutions to accomplish what I am trying to do? Are there other TI RS-232 transceivers that have lower series resistance or higher output load capabilities?
I have attached a PDF with the customer's test data and drawing of the original 1 driver into 8x receivers.
Thoughts to his questions above? My main concern would be heat generation with 1 TX to 8x RX loads. If he parallels TX outputs - I think this would be OK, as I don't think they will 'fight' each other - plus yo get the benefit of dispersed power dissipation.
Thanks!
