FT230XS: not recognised ... sometimes!

[DefekC]

So, the general problem is that FTDI's FT230XS chip is not talking to my laptop.

So, what I got and what I've tested:
I got Asus S46CB laptop. Being desperate at one point I even went through the pain to re-install fresh Windows 7 Pro N x64 on it (no VMs).

FT230XS chip(s ... I got 5 of them, all of them doing the same) is not dead, that is for sure, 'cause it is talking to several other laptops and PCs (Win7 Home, Win7 Pro, Win 10 Pro & Arch Linux OS included) I had access to. And by "talking" I mean FT_Prog is able to read and write data to this chip (or in Linux case able to at least recognize PID and VID).

No other USB device has any issues what-so-ever (including USB mice, keyboards, flash drives, external HDDs, FT232R, CH340, PL2303HX - that is all I tried). The most weird part is that I am able to get this chip to talk to my laptop if USB hub is used in between them. I used dedicated USB hub (China brand) and Tracer Gender TRK-272 Polish keyboard with built-in USB hub. Only when I plug it directly to my PC I get "USB Device Is Not Recognized" message with error code 43 (Windows has stopped this device because it has reported problems) and devices IDs "UNKNOWN".

Tested FT230XS test platform changes: USB cable shortened from 50cm (20") to ~15cm (6"); traces from USB wires to FTDI chip are 20mm (0.8") in length at most; D+ & D- lines were tested with 22?, 30?, 100? & 1k? series resistors, all combinations were tested with and without bypass 22pF capacitors (layout was as recommended in datasheet examples). All (except 1k? versions) worked with other PCs... again, except mine.

Test platform (without bypass caps):

Break-out allows me to configure it in both self-powered and externally powered.

FT230XS chips are genuine... well, I hope because they came from Farnell.

I am lost here...
I really don't want to migrate to other FT232__ versions due to price and other USB-UART bridges due to accessibility.

Am I only one out there having this problem?
Any suggestions?

[bktemp]

Did you use VBUS_Sense voltage divider in self-powered configuration?

[DefekC]

... bktemp, You (and some other people) gave me a hint on what might be the problem...
I will check it tomorrow (if I will be able to keep myself awake long enough to make another break-out board) and report my results.

[DefekC]

Sooo, yea... problem solved!
Reason: datasheet, rookie mistake? Anyway, something to watch out for.

So, the datasheet has this info in it (page 7, table 3.1. "Power and Ground"):

(Pin 10, VCC**, POWER Input) 5V (or 3V3) supply to IC
(Pin 8, 3V3OUT**, POWER Output) When VCC is 3V3; pin 8 is an input pin and should be connected to pin 10.

... and right under this table "**" are explained more...

** If VCC is 3V3 then 3V3OUT must also be driven with 3V3 input

My first impression was: "Ough, You have to drive output? That's ...weird! But if datasheet says so, who am I to argue with it..." 
My system had exactly these conditions: power from system (not from USB VBus), 3.3V and without option to power it otherwise.

This is exactly the mistake I made - I jumped Vcc, 3V3out and VccIO (not that it matters) on single 3.3V bus. Also in datasheet, page 13, "+3.3V LDO Regulator" section it says:

It also provides +3.3V power to the 1.5k? internal pull up resistor on USBDP. The main function of the LDO is to power the USB Transceiver and the Reset Generator Cells rather than to power external logic. However, it can be used to supply external circuitry requiring a +3.3V nominal supply with a maximum current of 50mA.

Solutions:
1) don't use 3V3out (but decouple it with small capacitor for stability);
2) use it only for VccIO; connect Vcc to external power source and don't connect Vcc with 3V3out;
3) ...or just follow any pre-drawn examples in datasheet "as-is"!

But why?
My best guess is that by providing 3.3V to 3V3out pin I messed up USB initialization "ping" of FT230X: two power sources meet, of course, they are not perfectly matched, and the most powerful wins. In this case, external regulator takes over and pushes FTDI chip's LDO to some safety cutout mode (50mA max current?) or pulls to GND or .. some random voltage thus prohibiting normal internal USB initialization when plugged.

Why only my PC did not recognize it and others did? And what about weird USB hub-thing (read above)?
Again, my best guess is Southbridge on PC and the way how it handles USB "requests" upon plug-in: chip in my PC might have had either lower response time (faster reaction), different error handling strategy or something along these lines. FTDI is not dead... just pulling internal circuitry to initialize startup took longer to handle excessive voltage/current inrush into 3V3out. This is why it generally worked with more distant USBs (front panel hubs on stationary PCs or external hubs) and did not work properly when kept as short as possible (case of laptop, where even internal USB bus lines are very short).

P.S. - this is how looks my home-brewed FTDI FT230XS v3 module just for good ending.