Skip to main content ↑ ↓
Topic: [solved: hub problem] Logic Sniffer doesn't work (Read 2731 times) previous topic - next topic

[solved: hub problem] Logic Sniffer doesn't work

Hello,
the ComPort driver is installed correctly. When I try to capture data, I get the message "Error while trying to communicate with device: "Device not found"". I made an firmware update with pump-program.bat which worked fine.
When I try to update the ROM , the comport is opened succesfully but I get the message "Error writing to PUMP".

What's going wrong?

Regards,
sparky

Re: Logic Sniffer doesn't work

Reply #1 –
There was a problem with the USB-Hub. When I connect the Logic Sniffer directly to the PC it works fine.

Re: Logic Sniffer doesn't work

Reply #2 –
Sparky, thank you for reporting how you solved the problem.

After some testing with different USB hubs we must assume the following:

1. OLS on unpowered USB hubs (w/o an external power supply) does not work in many/most cases!
2. OLS on powered hubs seems to work in most (all?) cases (I have performed successful tests with 4-port and 8-port powered D-Link and some other brand hubs)
3. OLS works on PCs and most laptops when connected directly to a USB port. However there may be problems when running the OLS on some laptops even when connecting directly to a USB port)

In most cases (USB) connection problems seem to be caused by insufficient power supplied by the USB port/hub the OLS is connected to or a failed configuration of the high power bus powered function during enumeration. See USB in a NutShell (or the USB specification) for details.

The power consumption of the OLS is well within the USB specification (< 5 loads or 500mA; 1 load = 100mA) but it can be higher than 100 mA in some cases ... I can't find Jack's power calculation atm (the power consumption of the FPGA depends on various factors and must be calculated with a Xilinx tool). I suggest to Ian to look at the PIC code and check if the OLS is configured for high power bus powered configuration during enumeration as by standard it will be configured for 1 load (100mA) only! I won't be able to perform any further tests until later tomorrow or Monday but it would be interesting to know the actual voltages on the 3.3V (VR1), 2.5V (VR2) and 1.2V (VR3) supply rails under (USB) communication problem conditions.

Re: Logic Sniffer doesn't work

Reply #3 –
If I recall correctly, the minimum voltage of the 5V rail on the end of the USB cable is something like 4.25V according to the spec (worst case with a minimum voltage from the computer, passive hub and a long cable), maybe something is dropping out of regulation?

Re: Logic Sniffer doesn't work

Reply #4 –
Hi all,

I tried my OLS on my macbook pro 15", directly to the USB port, cable length 6 ft.
both using the software for windows in a VMWare session as well as directly for mac
I have no success whatsoever. I see tha driver installed (win) and giving a COM4,
but then device not found. Also on my mac natively I see the device created, but
same saga... Device not foud, See the msgs in the mac OSX thread.

Next I will test on a hub with power.

Checked he supply voltages (device connected to USB) ... (No activity...)
According to my recently calibrated Fluke 112:

1.2 V rail : 1.204 V
2.5 V rail : 2.495 V
3.3 V rail : 3.310 V
5.0 V rail : 4.919 V

BTW, on startup of the board, I see the act led flashing to program the fpga,
so the board is doing tricks :-)

Rudi

Re: Logic Sniffer doesn't work

Reply #5 –
[quote author="alm"]If I recall correctly, the minimum voltage of the 5V rail on the end of the USB cable is something like 4.25V according to the spec (worst case with a minimum voltage from the computer, passive hub and a long cable), maybe something is dropping out of regulation?[/quote]You are correct. As little as 4.01V is considered a valid Vbus level. The minimum function supply voltage is 4.35V, but I recall seeing 4.1-something in a spec somewhere.

During transient conditions, the supply at a hub can drop from 4.4V to 4.070V, but that is the output from the hub - I haven't bothered to calculate how low the supply might drop at the end device, although local capacitance might smooth that out.

These numbers are all taken from the USB 2.0 specification, search for "4.0"

Re: [solved: hub problem] Logic Sniffer doesn't work

Reply #6 –
I thought of a possible issue with USB->serial latency for boards that have intermittent response problems. I noticed a few people in this thread had this issue (with or without a hub). If possible, could someone please test this solution and see if it helps at all:
http://dangerousprototypes.com/forum/in ... opic=549.0
Got a question? Please ask in the forum for the fastest answers.

Re: [solved: hub problem] Logic Sniffer doesn't work

Reply #7 –
Hi Ian,

Tried your idea, but no go here (note that my communication has never succeeded
yet, I do not have intermittent issues). ....

Will see if I can get something out of the OLS using different PC/mac hardware ...

Rudi