CC3000MOD problems

Terje Nikolaisen

Just assembled a new PCB with the CC3000MOD from TI. I've used LSR TIWI earlier.

When running same software on both boards, the CC3000MOD behaves strangely.

Logging with wireshark shows strange behaviour. I get retransmitted packets and duplicate ACK's. Also everything is kinda slow.

Both Wifi modules have the latest patch installed.

Is there anything I should be aware of regarding the CC3000MOD?

Short Wireshark log:

27 0.881825 10.0.0.1 10.0.0.53 TCP 66 53458 > http [SYN] Seq=0 Win=8192 Len=0 MSS=1460 WS=4 SACK_PERM=1

28 0.890886 10.0.0.53 10.0.0.1 TCP 60 http > 53458 [SYN, ACK] Seq=0 Ack=1 Win=1460 Len=0 MSS=1460

29 0.890927 10.0.0.1 10.0.0.53 TCP 54 53458 > http [ACK] Seq=1 Ack=1 Win=64240 Len=0

30 0.891034 10.0.0.1 10.0.0.53 HTTP 395 GET /net HTTP/1.1

31 0.897767 10.0.0.53 10.0.0.1 TCP 60 http > 53458 [ACK] Seq=1 Ack=342 Win=1460 Len=0

39 1.253484 10.0.0.53 10.0.0.1 HTTP 89 HTTP/1.0 200 OK

43 1.454289 10.0.0.1 10.0.0.53 TCP 54 53458 > http [ACK] Seq=342 Ack=36 Win=64205 Len=0

45 1.476085 10.0.0.1 10.0.0.53 TCP 54 [TCP Dup ACK 43#1] 53458 > http [ACK] Seq=342 Ack=36 Win=64205 Len=0

60 2.036672 10.0.0.53 10.0.0.1 TCP 1454 [TCP Retransmission] http > 53458 [PSH, ACK] Seq=1436 Ack=342 Win=1460 Len=1400

61 2.036695 10.0.0.1 10.0.0.53 TCP 54 [TCP Dup ACK 43#2] 53458 > http [ACK] Seq=342 Ack=36 Win=64205 Len=0

62 2.090010 10.0.0.53 10.0.0.1 TCP 1454 [TCP Retransmission] [TCP segment of a reassembled PDU]

63 2.090040 10.0.0.1 10.0.0.53 TCP 54 53458 > http [ACK] Seq=342 Ack=2836 Win=64240 Len=0

66 2.114369 10.0.0.53 10.0.0.1 TCP 743 [TCP segment of a reassembled PDU]

67 2.120798 10.0.0.53 10.0.0.1 TCP 958 [TCP segment of a reassembled PDU]

68 2.120820 10.0.0.1 10.0.0.53 TCP 54 53458 > http [ACK] Seq=342 Ack=4429 Win=64240 Len=0

69 2.133445 10.0.0.53 10.0.0.1 TCP 1232 [TCP segment of a reassembled PDU]

71 2.241096 10.0.0.53 10.0.0.1 TCP 1242 [TCP segment of a reassembled PDU]

72 2.241133 10.0.0.1 10.0.0.53 TCP 54 53458 > http [ACK] Seq=342 Ack=6795 Win=64240 Len=0

78 2.356401 10.0.0.53 10.0.0.1 TCP 400 [TCP segment of a reassembled PDU]

82 2.556796 10.0.0.1 10.0.0.53 TCP 54 53458 > http [ACK] Seq=342 Ack=7141 Win=63894 Len=0

142 3.673357 10.0.0.53 10.0.0.1 TCP 60 http > 53458 [FIN, ACK] Seq=7141 Ack=342 Win=1460 Len=0

143 3.673393 10.0.0.1 10.0.0.53 TCP 54 53458 > http [ACK] Seq=342 Ack=7142 Win=63894 Len=0

144 3.673549 10.0.0.1 10.0.0.53 TCP 54 53458 > http [FIN, ACK] Seq=342 Ack=7142 Win=63894 Len=0

145 3.681245 10.0.0.53 10.0.0.1 TCP 60 http > 53458 [ACK] Seq=7142 Ack=343 Win=1460 Len=0

Terje

over 12 years ago

0 made4engineering over 12 years ago

TI__Mastermind 27565 points

Hi Terje,

Did you design your own PCB? If not, where did you get it from? Retransmissions could indicate a signal problem.

Thanks,

Aaron

0 Terje Nikolaisen over 12 years ago in reply to made4engineering

Prodigy 185 points

You're probably right. Im going to connect a network analyzer and do the impedance matching properly.

I was just kinda surpised since the pcb was only 1 meter away from the AP. I've also got new antennas from Johnson now.

The CC3000MOD has a 50ohms antenna input i guess, and what is the recomended max VSWR?

Terje

0 Matt Dotson over 12 years ago in reply to Terje Nikolaisen

Prodigy 20 points

I'm seeing something similar with the FRAM Experimenter Board + CC3000EM. The CC3000EM seems to duplicate packets. I see the same slowness, and it seems to be caused when the driver gets stuck waiting for responses to dup packets in hci_event_handler (evnt_handler.h). Did you ever solve your issues?

Wireshark Trace:

(192.168.1.3 is the CC3000)

No. Time Source Destination Protocol Length Info

2060 73.114512000 192.168.1.3 192.168.1.10 TCP 58 cft-1 > http [SYN] Seq=0 Win=1460 Len=0 MSS=1460
2061 73.114490000 192.168.1.3 192.168.1.10 TCP 58 cft-1 > http [SYN] Seq=0 Win=1460 Len=0 MSS=1460
2064 73.191894000 192.168.1.10 192.168.1.3 TCP 58 http > cft-1 [SYN, ACK] Seq=0 Ack=1 Win=8192 Len=0 MSS=1460
2065 73.196470000 192.168.1.3 192.168.1.10 TCP 54 cft-1 > http [ACK] Seq=1 Ack=1 Win=1460 Len=0
2067 73.196487000 192.168.1.3 192.168.1.10 TCP 54 [TCP Dup ACK 2065#1] cft-1 > http [ACK] Seq=1 Ack=1 Win=1460 Len=0
3643 133.045016000 192.168.1.3 192.168.1.10 TCP 54 cft-1 > http [FIN, ACK] Seq=1 Ack=1 Win=1460 Len=0
3644 133.045120000 192.168.1.10 192.168.1.3 TCP 54 http > cft-1 [RST, ACK] Seq=1 Ack=2 Win=0 Len=0
3646 133.045044000 192.168.1.3 192.168.1.10 TCP 54 cft-1 > http [FIN, ACK] Seq=1 Ack=1 Win=1460 Len=0
3772 136.717108000 192.168.1.3 192.168.1.10 TCP 58 cft-2 > http [SYN] Seq=0 Win=1460 Len=0 MSS=1460
3773 136.717230000 192.168.1.10 192.168.1.3 TCP 58 http > cft-2 [SYN, ACK] Seq=0 Ack=1 Win=8192 Len=0 MSS=1460
3774 136.721212000 192.168.1.3 192.168.1.10 TCP 54 cft-2 > http [ACK] Seq=1 Ack=1 Win=1460 Len=0
3775 136.717133000 192.168.1.3 192.168.1.10 TCP 58 cft-2 > http [SYN] Seq=0 Win=1460 Len=0 MSS=1460
3776 136.721234000 192.168.1.3 192.168.1.10 TCP 54 cft-2 > http [ACK] Seq=1 Ack=1 Win=1460 Len=0
3777 136.847772000 192.168.1.3 192.168.1.10 HTTP 174 POST /api/values HTTP/1.0 (application/json)
3778 136.853977000 192.168.1.3 192.168.1.10 TCP 54 [TCP ACKed unseen segment] cft-2 > http [ACK] Seq=121 Ack=220 Win=1460 Len=0
3779 136.847734000 192.168.1.3 192.168.1.10 HTTP 174 [TCP Retransmission] POST /api/values HTTP/1.0 (application/json)
3780 136.849831000 192.168.1.10 192.168.1.3 HTTP 272 HTTP/1.1 204 No Content
3781 136.853958000 192.168.1.3 192.168.1.10 TCP 54 cft-2 > http [ACK] Seq=121 Ack=220 Win=1460 Len=0
5559 193.055172000 192.168.1.3 192.168.1.10 TCP 54 cft-2 > http [FIN, ACK] Seq=121 Ack=220 Win=1460 Len=0
5560 193.055149000 192.168.1.3 192.168.1.10 TCP 54 cft-2 > http [FIN, ACK] Seq=121 Ack=220 Win=1460 Len=0
5561 193.055231000 192.168.1.10 192.168.1.3 TCP 54 http > cft-2 [ACK] Seq=220 Ack=122 Win=64240 Len=0

0 Terje Nikolaisen over 12 years ago in reply to Matt Dotson

Prodigy 185 points

I have been busy with other things, but now I'm back to programming the CC3000 again.

I have not solved the problems with retransmissions, not with the latest pack V1.11 either. I've using two prototypes, one with LSR, and one with TI module.

The LSR module seems to be working better than the TI module. But, both has a tendency to just give up transmitting TCP packets after a while, and the WEB page cannot be displayed. I'm currently trying to decode the RSSI, to see if the strength is ok, (see other post).

I've also tried different solutions in the code, but nothing will improve this performance... I have 3 different Wifi routers, and none of them shows any differences.

The TI module also has trouble connecting to the AP sometimes.

I've ordered a new experimenter board just to try TI's own code, and see if the problems are there also.

Anyone with solutions to this???

Terje

0 Terje Nikolaisen over 12 years ago in reply to Terje Nikolaisen

Prodigy 185 points

After doing some research, I cannot see why the retransmissions, duplicate ACK and out of order packets are happening...

I've tried sending 100 TCP packets of different sizes from 500 bytes to 1400 bytes, and this has nothing to do with this case. I've also tried different software solutions, with and without RTOS.

I've finally got the Signal strength by using DD-WRT software on the TP link, and this shows better than -50 dBm signal from the LSR module. The TI module probably has something wrong with the soldering since it has less than -80dBm no matter what I do with the antenna. (Network analyzer shows good match on the antenna)

The wlan_ioctl_get_scan_results does not give me the RSSI correctly. I can make out all of the other data there, but RSSI just sticks on the same value not consistent with what the TP-Link reports. I want the RSSI printed out from the LSR module by debug once every 10 second.

So the signal strenght should have nothing to do with the problems I experience...

Terje

Wi-Fi

Wi-Fi forum

CC3000MOD problems