1. Have been fiddling with the board. Found that, 22 ohms separating two grounds + 100 ohms in RX/TX line + 10 ohms between MAX ground and F232 ground helps to reduce the frequency of hanging to almost zero. (I am still able to make it hang, by switching on/off the UPS/serial input source at rapid pace.). Instead of resistor, I tried ferrite beads. Did not seem to be making any difference. Also tries Ls of 100 muH, 200 myH. A bit surprisingly, this also did not help. Please comment............

2. I am modifying the lay-out to a 2 layer one as per your suggestions. I have noted down every important recommendation made by you guys (thanks a million!!). I have asked FTDI support whether they have any recommended boards / lay-out samples for a simple converter with MAX232 + F232R which has a good immunity. Waiting for their reply before finalizing my lay-out.

3. Chip hangs regardless of application software. In VCP mode, hyper terminal will hang if I attempt to close the port and re-open the port.

4. Chip cannot be reset at all from application, in hanged state. Hanging F232R, is an un-resettable F232R.

5. In application, we can easily handle other issues - taking out the port and plugging the port as many times as you want, taking it out and plugging it into another port etc. This is not a critical area. You only need to sense an error condition, force-close the device handle, and try to keep opening the device as soon as a device is available.

6. Had sent the circuit and layout to FTDI support engineer as per his request. His reply that came today did not indicate any lay-out issue. I guess he suspects the application software, and he has asked me to send the application source code to him. Guess he had not noticed that I had already mentioned that even without the application software running or with windows hyper terminal running on VCP driver, chip hangs.

7. Even though the decoupling cap was placed a bit far in the lay-out, we had soldered a cap right on he chip when the problem cropped up. We also tried more decoupling caps at other points. Nothing worked.


As you all pointed out, the issue seems to be directly connected to noise, ground impedance, decoupling etc. But, I still have the same complaint - Single side boards and double-sided boards without any fancy isolators/chokes using other chips seem to be working in the same condition. I have been collecting many boards, some directly from existing UPSs with USB port and have been feeling that F232R is a bit more of a spoiled kid, when it comes to being fussy in noisy environments. It needs all the protections and encouragements to work without pulling the rug once in a while.


- Raj Nambiar