MAX3232E: max3232E

Hi Li,

1. T2IN cannot be left floating. It must be pulled to VCC or GND if unused. This is to help prevent shoot through current (increased power consumption) and oscillations due to shoot through current (which can cause emissions issues depending on system setup). This needs to be fixed 1st. 

2. Why are you using a 2-channel device? We have 1T1R devices - so you wouldn't have unused ports. In general you shouldn't be adding ports unnecessarily. 

3. Are you connecting the PC to RS-232 bus pins - i.e.  it sounds like it goes PC -> cable B -> cable A -> CON1 on schematic. If that is the case this isn't how RS-232 should be used - RS-232 bus pins should be connected to other RS-232 devices. PC control should be on TxIN and RxOUT pins - not the other way around. If this was intentional there is no reason to use RS-232 - RS-232 bus is RS-232 only - no PC signals. Designers use RS-232 to get the additional drive strength on the RS-232 bus - if you are using a PC on the bus then you probably need to redesign without RS-232.

Ultimately based on how you are using this device I am very skeptical that you chose the correct interface here in the first place and I really don't think what you are trying to do is going to be met by RS-232. The issue that you are seeing is because you have multiple cable connections - they  are very capacitive and will create impedance discontinuities on the bus, by having current setup you don't have the standard 2 mechanical interface discontinuities but 6, plus the cable impedance may also be a discontinuity (there are 4 boundaries between cables and 5 different transmission mediums). Basically this setup is destined to create noise and fail - adding filtering most likely will not overcome the system setup. This system needs redesigned. 

Please let me know if I have misunderstood your issue in my reading of the problem or if you have any additional questions. But basically - RS-232 bus pins should only communicate with RS-232 bus pins on a different RS-232 transceiver - anything else and RS-232 is probably not what you want.

Best,

Parker Dodson 

Hi Dodson,

1. The image below shows the pin function of the MAX3232, PC control should be on 7/14 and 8/13 pins.

2. In yesterday's test, the computer only had USB ports, so I found a USB to serial cable B. Today I found a PC with a serial port to test, PC <-> cable A <-> CON1.  I found that there were noise glitches in the phase noise of some chips on the board. Is there any way to remove noise glitches.

Please show an oscilloscope trace that shows these glitches on the RS-232 bus lines.

The picture on the right is trace of rs232_tx signal, PC -> cable B -> cable A -> CON1. The picture on the left is trace of rs232_tx signal, only cable A connecting to CON1.

Noise glitches are not seen on the RS232, but on the phase noise curve of the output signal of a chip on the board.

Li,

Can I please have full diagram of your test  system.  USB-PHY  and RS-232 are not directly compatible (due to voltage levels)  and I am not sure what you are doing or why - so having a diagram would help. As when I asked something similar in my last response you just circled all the RS-232 comms pins- which I am well aware of - and not what I was asking. 

The reason I am asking is because you're system description doesn't make sense as to why you are using this setup - it seems like you have 1 RS-232 transceiver - not 2 which is what an RS-232 system is supposed to look like. If you don't do that - there is no point in using RS-232.

For your phase noise problem once again you have multiple mechanical connections - if you have more than 1 cable for RS-232 you are doing something wrong and any design with that principle is a bad design and bad designs will perform poorly. . However I think the main issue is the system design at the very beginning. 

This is why I need to see a full block diagram of the test - one transceiver node doesn't show the full system you are testing - and by the way your circuit is also problematic please see points in my first reply. 

Best,

Parker Dodson

Hi Parker

As far as I know, the RS-232/USB cable also contains an RS-232 transceiver.

RS-232 transceivers typically use charge pumps to generate the bus voltages. This switching power supply can generate some noise; it would also be possible that the PC's USB power lines are noisy. I doubt that this noise goes through the MAX3232E; the power rails are more likely. Please check your boards GND and VCC voltages with an oscilloscope.

Hi Li,

Thank you for the update. 

Do you have a part number for your serial to USB connector cable - I need to verify there is a transceiver on there as Clemens alluded too so I can check to ensure there is a transceiver. Honestly if you have the part number for cable A and B that would be great because if everything is okay architecture wise then I do think the problem is cabling. 

What is the length of cable A and length of cable B as well. 

The noise could be generated by the power supply - but from the sounds of the issue -  you have jitter (phase noise) issues - which is usually the result of the actual layout / system setup moreso than power supply noise (it would still be a good idea to get a scope of the GND and VCC lines to see if there is noise present there). The issue is you have multiple connection interfaces - every mechanical interface is going to add capacitance and create impedance discontinuities for higher harmonic content of the signal - which could result in worse jitter. The other factor is what your data rate and the length of the system (i.e. why I am asking about the cabling).

Basically, if the USB to Serial cable has transceiver integrated into it (please provide part number so I can verify) then most likely the issue is due to the system setup. Ideally you want to minimize capacitance on the bus so the fewer mechanical interfaces the better - i.e. there should be one cable link and depending on cable parameters, total length of cabling, and data-rate of system will ultimately determine how much jitter is possible. 

Please confirm:

Cable A part number

Cable B part Number

Total Bus length

Data rate of system. 

Most likely the issue is the setup and depending on system needs the system design most likely needs to be simplified. 

Best,

Parker Dodson

I bought cable on an internet e-commerce company and there were not part number. The cable is 1m length. The baud is 115200.