Open dtheriault opened 9 years ago
Dump of lsusb:
Bus 001 Device 014: ID 0bda:2838 Realtek Semiconductor Corp. RTL2838 DVB-T Bus 001 Device 013: ID 1c04:2074 QNAP System Inc. Bus 001 Device 012: ID 1c04:2074 QNAP System Inc. Bus 001 Device 007: ID 0bda:2838 Realtek Semiconductor Corp. RTL2838 DVB-T Bus 001 Device 011: ID 0bda:2838 Realtek Semiconductor Corp. RTL2838 DVB-T Bus 001 Device 010: ID 0bda:2838 Realtek Semiconductor Corp. RTL2838 DVB-T Bus 001 Device 009: ID 0bda:2838 Realtek Semiconductor Corp. RTL2838 DVB-T Bus 001 Device 008: ID 0bda:2838 Realtek Semiconductor Corp. RTL2838 DVB-T Bus 001 Device 006: ID 1c04:2074 QNAP System Inc. Bus 001 Device 005: ID 0bda:2838 Realtek Semiconductor Corp. RTL2838 DVB-T Bus 001 Device 004: ID 0bda:2838 Realtek Semiconductor Corp. RTL2838 DVB-T Bus 001 Device 003: ID 1c04:2074 QNAP System Inc. Bus 001 Device 002: ID 05e3:0610 Genesys Logic, Inc. 4-port hub Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub Bus 002 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
And looks like rtl_hpsdr can enumerate 8 ports:Parsing config file /home/doug/configs/rtl_hpsdr.conf
Found 8 RTL device(s), using 8. RTL base sample rate: 1536000 hz
Global settings: config file: /home/doug/configs/rtl_hpsdr.conf ip address: 192.168.2.24 number of rcvrs: 8 hpsdr output rate: 48000 hz signal multiplier 100 calibration freq: 15000000 hz up_xtal freq: 125000000 hz sound device: none
Unfortunately, SkimSvr stops at 7 ports.
-- Would rtl-sdr software care if 'n' ports are spread across 'n' USB hubs? NO -- Have you tried a configuration with a couple hubs? YES, it works -- Do you think there's any performance improvement if we can spread dongles across multiple hubs? My concern here is, don't we want to only have one USB port that the user will have to hook to their PC? If so, it HAS to be USB 3.0. If they only have USB 2.0 ports then we'll have to give them two ports to connect to their PC.
I have tried at least 5 or 6 different USB 2.0 hubs and none of them will get past the 7th dongle. At least in operation. It will enumerate them, i.e. lsusb and see 8, but try to run all 8 in tandem and it will not work. -- Assume should find chip that handles both 2.0 and 3.0 speeds (future proof). -- What's the chipset of your hub? Can you read it/send to me? I love this hub. It uses VL812 chips, three of them, see: http://www.via-labs.com/product_show.php?id=41
Tried two hubs last nite and your code enumerates fine but skimmer would not let me set eighth port. Didn't try cudasdr but will tonight.
I'll crack open my hub...:-( there's no screws and didn't want to destroy it in process of peeling off plastic laminate. But it's for good cause.
I did see some 2.0/3.0 combo chips and can see to get 8 ports uniform speed into single upstream port ya need 3 chips. 15$ or so for 3 so not bad. Hoping connectors are cheap.
-dt-
On Feb 24, 2015, at 15:41, n1gp notifications@github.com wrote:
-- Would rtl-sdr software care if 'n' ports are spread across 'n' USB hubs? NO -- Have you tried a configuration with a couple hubs? YES, it works -- Do you think there's any performance improvement if we can spread dongles across multiple hubs? My concern here is, don't we want to only have one USB port that the user will have to hook to their PC? If so, it HAS to be USB 3.0. If they only have USB 2.0 ports then we'll have to give them two ports to connect to their PC.
I have tried at least 5 or 6 different USB 2.0 hubs and none of them will get past the 7th dongle. At least in operation. It will enumerate them, i.e. lsusb and see 8, but try to run all 8 in tandem and it will not work. -- Assume should find chip that handles both 2.0 and 3.0 speeds (future proof). -- What's the chipset of your hub? Can you read it/send to me? I love this hub. It uses VL812 chips, three of them, see: http://www.via-labs.com/product_show.php?id=41
— Reply to this email directly or view it on GitHub.
I like the block 1 diagram better. I hadn't been thinking that you were including the processing core on the board. What do you think of using a SOM sys on a module ARM? I like the idea of less cabling. Your intent is to house both in the same enclosure?
Was not thinking of integrating the CPU board such as C1 onto the processing board per se' the diagram just shows foot print there but in reality, maybe just a blank space and 4 holes for standoff's. Didn't want to get into recreating a new CPU or anything.
Had not thought about some SOM sys on a module ARM; do you have any recommendations there? I was just thinking C1 price point.
As for housing, can go two ways, one is single housing for both boards, other would be two separate housings, one per board. The later might be more flexible but not really clear on which way to go.
Hub I am using contains 2x ASMEDIA ASM1074L's, 2x SENSMEDIA 25L4006E 1x SX17 3.3v regulator, 2x 25Mhz xtals. Couldn't find anything on the SENSMEDIA device. ASMEDIA chip mfg at: http://www.asmedia.com.tw/index.php
Oh the actual product Mfg for the hub is: http://www.broway.com.cn/infor.asp?id=3
Been looking for USB hub chips that provide 8 ports. They seem to come in two varieties, 4 port and 7 port. That's a bit surprising. I am wondering following:
-- Would rtl-sdr software care if 'n' ports are spread across 'n' USB hubs? -- Have you tried a configuration with a couple hubs? -- Do you think there's any performance improvement if we can spread dongles across multiple hubs? -- Assume should find chip that handles both 2.0 and 3.0 speeds (future proof). -- What's the chipset of your hub? Can you read it/send to me?
Cypress HUX3 hub chip seems like a player, but its 4 ports. Cheap, $2.50ea. TI and Microchip seem to have bunch of options too.
Thinking we develop board that takes 8 USB ports on one side for the dongles, and two ports up stream to the embedded CPU. Board also takes 5v supply and generates the 1.2v and 3.3v for the chipset. This way we can change the embedded CPU w/o any impact on the hub. We would have better margins by not having to integrate a COTS hub into a box.