MAX3221: RON is not Hi-Z when deactivating the IC

Hi Alexandre,

Could you conduct some lab tests to gather a scope capture of COM1-RX-B, USART1-RX, USART1-PRCTL, and /INVALID? Thanks!

Regards,

Tyler

Hi Tyler,

Here are the requested captures, sorry I have only 2 channels on my Fluke 225C, so I repeated capture by always including COM1-RX-B as channel A. Also note that there is USART1-PRCTL and /USART1-PRCTL (I've not captured the later since they are static in the test, it is a simple BJT/resistors inverter).

In my tests, I've noted that signal coming from MAX13487 pin RO is arround 3.9V so I suspect a long term issue there, but tried powering it at 3.3V, same behaviour but the chip is not rated for this power rail. It will be replaced by THVD1406DR which will be at 3.3V, they are ordered but will receive them only in about a week.

A: COM1-RX-B B: USART1-RX

A: COM1-RX-B B: USART1-PRCTL

A: COM1-RX-B B: /INVALID

Thank you,

Alexandre.

Hi Alexandre,

One more scope request. Can you gather CH1 = USART1-RX, CH2 = USART1-PRCTL?

Remind me again, /USART1-PRCTL is the same signal as USART1-PRCTL just inverted? 

Regards,

Tyler

Hi Tyler,

Here is the new capture :

CH1 = USART1-RX, CH2 = USART1-PRCTL

Yes /USART1-PRCTL is just the inverted signal of USART1-PRCTL, currently USART1-PRCTL is tied to 3.3V and there is a 2N2222 with two resistors of 10K doing the inversion to generate /USART1-PRCTL as this schematic include :

Also, I've found this application note including a similar circuit : https://www.ti.com/tool/TIDA-00540 I saw the FORCEON is always at VCC, I tried it, no difference.

Hi Alexandre,

Given that USART1-PRCTL = /EN is HIGH, it doesn't make sense that there is still an output signal on USART1-RX = ROUT. However, looking at your schematic, USART1-RX shares the output of RO on the MAX13487EESA. Are we sure that the 13487 is not generating an output signal that could be mistaken for the output of the MAX3221? 

Could we try another test. Cut the 5V supply to the MAX13487, assure that it is in the off state. Make sure the /EN pin is LOW. Then input a valid RS232 signal on Pin 8 = COM1-RX-B = RIN, see if you read the correct digital output on pin 9 = USART1-RX = ROUT. Then repeat the test ensuring that /EN is HIGH. 

The /EN pin logic is directly associated with the receiver portion of the device. 

What we expect is that when /EN = HIGH, ROUT will be in High-Z. We should not see data on ROUT when valid RS-232 signal is present on RIN. 

We expect that when /EN = LOW, we should see digital signal on ROUT when valid RS-232 is present on RIN.

Basically, I want to see if we can get the MAX3221 working first alone, then see if powering the MAX13487 changes our outcome.

Please let me know what you think of these tests.

Regards,

Tyler

Hi Tyler,

I find the behavior as odd as you. I've designed this circuit part according to both datasheets. I've ran some tests :

1. Connect only MAX13487 RO to USART1-RX, works as indented, no echo in RS-485 mode in both directions

2. Connect only MAX3221 ROUT to USART1-RX, works as indented, no echo in RS-485 mode in transmit only of course

3. Keep connections, but leave MAX13487 VCC floating, works as indented in RS-485 mode (in both directions!!!!), but I see spikes on USART1-RX, still they are too short and I receive nothing on serial port, see capture

CH1 = USART1-RX, CH2 = USART1-PRCTL

There is a conflict between MAX3221 ROUT and MAX13487 RO, for now I would put this on hold, wait reception of THVD1406DR that I would use instead of MAX13487 and also power it at 3.3V, that's my hope for the moment. I should receive them next week so I will give you update about results!

Thank you!

Alexandre

Hi Alexandre,

Since this design is going to be replaced with the THVD1406DR, I will hold off on support for this question. However, it would still be good to understand why this design is failing for knowledge sharing purposes. 

I will move this question to "waiting for customer." Please respond back when you get the THVD1406DR implemented in your circuit. I would like to know the results of the new design to see if it was something in the MAX13487 causing issues.

Regards,

Tyler