hcitool commands for test purposes

Hi Hans,

i will forward that question to Bluetooth expert and you will get answer soon

Regards,

Eyal

I too have tried to find this information. It would seem that a document would be available with this information, however, I have not been able to find the document. Can you copy me on the answer please.

Hi Miguel,

According to the latest info on that page, I can get continuous transmit with

<setup BT>

hcitool cmd 0x3F 0x0184 0x00 0x00 0x00 0x0F 0x00 0x00 0x00 0x00 0x00 0x00 0x00 0x00 // CW, PN9, freq index=0, power level 15

hcitool cmd 0x3F 0x0180 0xFF 0xFF 0xFF 0xFF 0xFF 0x01  // internal setting #2, skipping internal setting #1 because the page says not required for CW

I am still seeing brief transmit, not continuous transmit.  Are there additional setup required?

Paul

Paul,

        Please verify that no other RF activity is running. You can execute  "hciconfig noscan" to disable page and inquiry scans. 

Regards,

~Miguel

Hi Miguel,

WIFI is off, and a "hciconfig noscan" is performed, I am still getting only a brief Tx.

One thing that I am doing different from the page is that I use the Android menu to turn on bluetooth (followed by noscan), since I do not have the script file BT_init.sh.  Could you provide me with the script, or otherwise outline the equivalent hciconfig/hcitool commands to run?

Thanks,

Paul

Paul,

     The BT_Init.sh only turns on the BT controller, performs the hciattach and enables the bluetooth deamon. Nothing else. Could you run  'hciconfig'  to see the hci0 summary? It may be that some Inquiry activity may be happening. Please post the screenshot on this thread.

Regards,

~Miguel

Thanks,

Paul

Paul,

     If using Android, you only need to execute the following command: "hciconfig hci0 up". That will take care of the BT controller initialization. You may need to do "start bluetoothd" if you need to interact with other BlueZ utilities.

     From your screenshot, the FCC commands should work after you send the "hciconfig hci0 noscan" command. Let us know.

Regards,

~Miguel

Hi Miguel,

Yes we verified the continuous Tx works after noscan is properly applied.  Thank you for the prompt support.

Another question:  The power setting through hcitool (settings from 0x10 to 0x1f) seems to work on ~5dBm increments, are there other ways to fine tune the output?  We have a need to attenuate the output by about 1dBm.

Thanks,

Paul

Paul,

        These settings are predefined in the Bluetooth Service Pack. The reason to keep a 5dBm step size is to comply with Bluetooth specifications. Do you mean that you want to attenuate each of the levels by 1dBm? or you mean to have a step size of 1dBm.

~Miguel

Hi Miguel,

I meant having a step size of 1dBm.  Not necessarily through hcitool either - I was thinking of something similar to tiwlan.ini where the boost can be adjusted.

Our device has one case where the power is over by ~1dBm, which if not addressed would require us to go through additional testing.  I am looking for a way to decrease overall power output, but not @ 5dBm increments.

Thanks,

Paul

Paul,

        As mentioned earlier, the recommended step size is 5dBm. This is in order to comply with the BT specs (2dBm to 8dBm).

        The levels are defined inside the BT Service Pack file (BTS file). At this moment, we do not provide the option to modify them unfortunately. 

Regards,

~Miguel

Hi Miguel,

I'm working with Paul here on getting this Bluetooth FCC testing done.  Another thing we're still working on is demonstrating frequency hopping for the FCC test.  According to the FCC, the test's purpose is:

 "to ensure that the RF energy of frequency hopping systems is hopping at a minimum defined rate.  This helps ensure sufficient time off to enable other frequency hopping devices to co-operate within this allocated band."

Therefore, we need a command in which to have Bluetooth continuously transmit and hop through all the available channels.  From what I understand of the information on the TI wiki: http://processors.wiki.ti.com/index.php/Bluetooth_FCC_Mode#WL1271_.2B_ARM_Platforms ,  the commands provided there only transmit on one defined channel, but we need to have transmission hopping through all channels.

This is what we expect to see in the frequency domain on the spectrum analyzer (Here we're zoomed into 3 channels).

Could you provide us with a command to satisfy this test?

Thanks,

Andre

I just went through this and use a scrpt "bt-spectrum":

#!/bin/sh

while [ 1 ]; do
l2ping 00:11:22:33:44:55
l2ping 00:11:22:33:44:A5
l2ping 00:11:22:33:44:5F
l2ping FF:FF:FF:FF:FF:FF
l2ping FF:FF:FF:FF:FF:A5
l2ping FF:FF:FF:FF:FF:F5
done

This caused it BT to x-mit on all channels,

Hi Bill,

Thanks for you reply,  it got us very close!  We definitely have hopping on all channels now,  but since it's pinging (instead of transmitting) it's only doing short blips on each channel:

It's only occupying each channel for a few microseconds at a time, but to demonstrate the FCC frequency occupancy requirement, we'll need to have each channel occupied for 675 microseconds as per Bluetooth spec.

Would anyone happen to know how to make the Bluetooth chip "send" in the blind, without connecting to a real other device first?  I tried using l2test with the "connect and send" mode but since I had nothing to connect to, I couldn't get it to send.

Thanks,

Andre

Andre,

     Please look at this thread: http://e2e.ti.com/support/low_power_rf/f/307/p/124304/521798.aspx#521798

Regards,

~Miguel

Hi Miguel,

Thank you for your reply, it was very helpful!

The commands that we are using now are:

HCI_VS_DRPb_Tester_Packet_TX_RX for hopping (0x00), disabled RX, DH5, PRBS9 Random, max packet length, power level 15, whitening disabled, and PRBS9 init to 0x0000

hcitool cmd 0x3F 0x185 0x00 0x00 0xFF 0x05 0x00 0x00 0x00 0x00 0x0F 0x01 0x00 0x00

followed by

Send_HCI_VS_Write_Hardware_Register 0xFF01, 0x0019180c, 0x0101

hcitool cmd 0x3F 0x0301 0x0C 0x18 0x19 0x00 0x01 0x01

I have several observations here:

1. The instruction to  set Frequency Mode is (0x00 - Single Freq, 0x03 - Hopping) but I found that 0x03 produced no transmission, but 0x00, 0x01, and 0x02 produced hopping.  Could you please clarify this?
2. I tried setting Packet Data Length to 0xFFFF, but that crashed the device.  I suppose 0 for Max Data Length would be the option we should use?

However, the two questions that we have are:

1. From what we see right now, the psuedo random sequence of hopping through all channels doesn't restart every cycle.  What we would like is for every hopping cycle through all the channels, we would like the pseudo-random sequence to start again from the same seed (init).  This would let us calculate the hopping frequency by measuring the duration between two visits of the same channel ( = time for one complete cycle of hoping all channels).  For this to work, the channels need to be visited in the same order every cycle. 
2. The pulse width at each frequency (duration for which each channel is held), was observed to be 50 - 100 microseconds.  We would like to have the pulse width set to 300 to 650 microseconds as per a typical Bluetooth device.

Could you let us know whether the two questions I just mentioned could be solved through the setting of some parameter?

Thanks,

Andre

Thanks Miguel!  We finally got it to work with DH1 and Packet Data Length set to 27:

hcitool cmd 0x3F 0x185 0x00 0x00 0xFF 0x01 0x01 0x00 0x1B 0x00 0x0F 0x01 0x00 0x00