2
\$\begingroup\$

I've ordered a custom planned PCB which contains an FT2232D chip. Circuit looks like this: enter image description here

VCC and USB_VCC are connected on another page. GND and USB_GND are connected on another page.

Datasheet of FT2232D: http://www.ftdichip.com/Support/Documents/DataSheets/ICs/DS_FT2232D.pdf Datasheet of EEPROM: http://lomex.hu/pdf/(stm)_m93c46-m93c86.pdf

Although it seems starting upon connecting, no further steps could be done:

[23980.186603] usb 5-1: new full-speed USB device number 91 using xhci_hcd
[23980.277845] usb 5-1: Device not responding to set address.
[23980.569101] usb 5-1: Device not responding to set address.
[23980.769169] usb 5-1: device not accepting address 91, error -71
[23980.880898] hub 5-0:1.0: unable to enumerate USB device on port 1

I'm using self powered configuration, so the chip gets its power from USB port directly. A 6 MHz XTAL was applied, according to datasheet. And I also applied an EEPROM (although it's not a must I think).

On Windows I got unknown device and then Windows disables it because thinks it works wrong. FT_Prog doesn't recognize it either.

On Linux the dmesg shows lines above. Device is not responding.

How could I move on? Do you guys seen anything like this?

\$\endgroup\$
8
  • \$\begingroup\$ When you say it's almost the same it'd be worth including a schematic of what you built. I believe that part occurs very early during enumeration so it could be something wrong at a fairly low level. It doesn't sound like missing device drivers or anything similar to me. \$\endgroup\$
    – PeterJ
    Dec 20, 2013 at 8:51
  • \$\begingroup\$ Yes, you are right, here it is: 93.190.5.65/usb.png \$\endgroup\$
    – Daniel
    Dec 20, 2013 at 9:44
  • \$\begingroup\$ Quarz parameters are missing. \$\endgroup\$
    – Turbo J
    Dec 20, 2013 at 17:44
  • \$\begingroup\$ Have you not heard of power net symbols? \$\endgroup\$ Dec 21, 2013 at 8:43
  • \$\begingroup\$ Knee-jerk response would be it looks like you may not have enough decoupling. Default decoupling practice should be one 0.1 uF cap per Vcc pin, and only reduced once that's verified as working. \$\endgroup\$ Dec 21, 2013 at 8:48

2 Answers 2

1
\$\begingroup\$

You have used the wrong pin for GND on your mini USB socket. Pin 5 is GND, pin 4 is ID and most likely not connected at all (in the cable).

You can try to solder bridge the pins 4 and 5 in order to get a stable ground on the USB. This should work with most cables.

EDIT: The VCC-USB_VCC connections makes this configuration bus powered. Thus you need to connect RESET directly to VCC without the resistor divider. See above FT2232 datasheet, schematic in chapter 6.0.

\$\endgroup\$
4
  • \$\begingroup\$ That was exactly what I've done. Without this there is no power at all on the pins of the chip. Now I'm wondering what else could I check. These FTDIs won't be my favourites for sure.. \$\endgroup\$
    – Daniel
    Dec 20, 2013 at 16:01
  • \$\begingroup\$ What do you think, is the layout correct? Can I measure some pins for getting more information on what's going on? \$\endgroup\$
    – Daniel
    Dec 20, 2013 at 16:08
  • \$\begingroup\$ @Daniel - We can't tell you anything about your layout unless you actually post your layout. \$\endgroup\$ Dec 21, 2013 at 8:49
  • \$\begingroup\$ sorry, I meant the schematic. \$\endgroup\$
    – Daniel
    Dec 22, 2013 at 8:55
0
\$\begingroup\$

I don't remember if it's the very same error, but I remember similar errors when D+ and D- were switched. You might want to double check that.

\$\endgroup\$
12
  • \$\begingroup\$ It would detect a low speed device in this case. \$\endgroup\$
    – Turbo J
    Dec 20, 2013 at 12:09
  • \$\begingroup\$ I created a schematic based on this layout, and it does not showed any errors. In addition, I checked each connection if right or not on the PCB, so I'm sure D- and D+ are not switched. (Unless the layout plan above is not correct either) \$\endgroup\$
    – Daniel
    Dec 20, 2013 at 12:16
  • \$\begingroup\$ But I guess D+ should connect to D+ on the other side, and D- to D- on the other side, right? \$\endgroup\$
    – Daniel
    Dec 20, 2013 at 12:35
  • \$\begingroup\$ @TurboJ: Not necessarily. On Windows I got "This device cannot be started" and on Linux some similar errors (but no successful connect). \$\endgroup\$
    – Tom L.
    Dec 20, 2013 at 12:52
  • \$\begingroup\$ @Daniel: D+ must be connected to D+, that is correct. It is NOT like a UART RX/TX pair. Did you check all the way to your connector and from that through the cable? \$\endgroup\$
    – Tom L.
    Dec 20, 2013 at 12:53

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.