Unable to run more than two RTL-SDR dongles at the same time.

[hipp...@gmail.com]

Hi all,
I am running a websdr server using two R820T. One at 2Mhz wide and the other 2.8Mhz.
The server works well and can be seen here http://ccr.linkpc.net:8901
The trouble is I would like to add more bands, using more RTL dongles, but once I add a third dongle and start a third instance of rtl_tcp I start to get signal caught messages from rtl_tcp and the websdr server is forced to wait for me to restart the rtl_tcp instance.
This seems to happen even if I reduce the bandwidth and sample rates.

I know what you're thinking, too much data for the USB port and a slow PC, wrong! It's high-end AMD 8 Core system with a dozen USB ports, 16GB Ram and an SSD. The OS is a recent install of Ubuntu 14.04 64bit.

If anyone can give me a clue what's up I'd be grateful. I've checked pretty much everything I can think of.

Thanks for taking time out to read this.

Glenn

[jdow]

FWIW I am running one SDRSharp (2.4 ksps), two SDRConsoles (2.304 ksps via TCP),
and an ADSBSharp at the moment. It's not likely the dongles. You suggest the
hardware is fine. Check to see if the port for your third dongle is USB 3.0. On
some (many? most? all?) machines the dongles don't work with USB 3.0.

Aside from that you're basically left with pilot error, USB implementation
issues, OS issues, or dongle installation issues it would seem. (I am mildly
surprised you're able to run at 2.8 ksps. Usually the dongles start throwing
rampant dropped packets above 2.4 ksps.)

{^_^} Joanne

On 2014-08-26 12:11, hipp...@gmail.com wrote:
> Hi all,
> I am running a websdr server using two R820T. One at 2Mhz wide and the other 2.8Mhz.
> The server works well and can be seen here http://ccr.linkpc.net:8901

> <http://ccr.linkpc.net:8901/>

[jdow]

Please pardon me - one other thought came to mind. Perhaps you have them all
plugged into an underpowered USB hub. These dongles subscribe for 500 mA each.
They seem to take on the order of 375 mA of that. You might check how big your
hub's power supply is and how many things are plugged int it.

{^_^} Joanne

On 2014-08-26 12:11, hipp...@gmail.com wrote:

> Hi all,
> I am running a websdr server using two R820T. One at 2Mhz wide and the other 2.8Mhz.
> The server works well and can be seen here http://ccr.linkpc.net:8901

> <http://ccr.linkpc.net:8901/>

[Glenn Pearson]

Hi Joanne,
The two existing dongles are connected directly to individual motherboard ports and the extra dongle to a powered hub. So it's unlikely but I'll double check.

Thanks,
Glenn

Cheshire Computer Repairs Ltd.
Hoole
Chester
CH2 3EW

Tel: 01244 458779
Tel: 01606 339308
Tel: 01928 798161
Tel: 0151 353 9982
Mob: 078 456 45870

www.cheshirecomputerrepairs.co.uk

> --
> You received this message because you are subscribed to a topic in the Google Groups "Ultra Cheap SDR" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/ultra-cheap-sdr/RH4neus2dX0/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to ultra-cheap-s...@googlegroups.com.
> To post to this group, send email to ultra-c...@googlegroups.com.
> Visit this group at http://groups.google.com/group/ultra-cheap-sdr.
> For more options, visit https://groups.google.com/d/optout.

[jdow]

That leaves the other options including installation error. For example, does
the dongle get the correct driver? Is there an OS limitation of some crazy sort?
Is the hub a 3.0 hub?

{^_^}

[Siegfried Jackstien]

And a question is ... did he use rtl eeprom to change the serial numbers?!?

Dg9bfc

Sigi



> -----Ursprüngliche Nachricht-----
> Von: ultra-c...@googlegroups.com [mailto:ultra-cheap-
> s...@googlegroups.com] Im Auftrag von jdow
> Gesendet: Dienstag, 26. August 2014 20:09
> An: ultra-c...@googlegroups.com
> Betreff: Re: [ultra-cheap-sdr] Unable to run more than two RTL-SDR dongles
> at the same time.

> --
> You received this message because you are subscribed to the Google Groups
> "Ultra Cheap SDR" group.
> To unsubscribe from this group and stop receiving emails from it, send an

[Glenn Pearson]

That's an interesting question.
I'll take a look when I'm finished work.

Thanks,
Glenn

Cheshire Computer Repairs Ltd.
Hoole
Chester
CH2 3EW

Tel: 01244 458779
Tel: 01606 339308
Tel: 01928 798161
Tel: 0151 353 9982
Mob: 078 456 45870

www.cheshirecomputerrepairs.co.uk

> You received this message because you are subscribed to a topic in the Google Groups "Ultra Cheap SDR" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/ultra-cheap-sdr/RH4neus2dX0/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to ultra-cheap-s...@googlegroups.com.

[Adam Nielsen]

> I know what you're thinking, too much data for the USB port and a
> slow PC, wrong!

Maybe not. Since you're on Linux, try running "lsusb -t" and it will
show you which USB bus each device is plugged in to. Try to
unplug/replug until you get each one on a different bus as that will
give you the most available bandwidth.

The reason I say it is because lots of motherboards have USB ports in
groups of two, and often these are connected to the same USB bus so I
believe that means they share bandwidth, even though they are ports
directly attached to the motherboard.

If all else fails you can buy a cheap PCI USB2.0 add-in card and this
will give you some more ports on independent busses.

Cheers,
Adam.

[Glenn Pearson]

Thanks Adam,
I have looked at lsusb -t already but will take another look.
I must be missing something.

Cheers,

Glenn

Cheshire Computer Repairs Ltd.
Hoole
Chester
CH2 3EW

Tel: 01244 458779
Tel: 01606 339308
Tel: 01928 798161
Tel: 0151 353 9982
Mob: 078 456 45870

www.cheshirecomputerrepairs.co.uk

[Leif Asbrink]

Hello Glenn,

You write: "I know what you're thinking...." but that is not
right for all of us;-)

Unfortunately I can not tell you what the problem is, but I can tell
you what the problem is not.

Have a look at this video:
https://www.youtube.com/watch?v=7tRI9RmJu5M

You can see 4 dongles on a laptop with 4 USB ports and 8 dongles
on a desktop machine which has 6 USB2 and 2 USB3 ports.

The limitation is the number of USB ports, not librtlsdr or the
USB hardware. This leaves rtl_tcp and the server. I would
guess rtl_tcp is the problem. The program is less than 600 lines of
code and you can see the cause of the error by checking
the line above the signal caught message. Locate the corresponding
text in rtl_tcp.c.

You could use Linrad to send data to the server. That means that
the server would have to understand the header in the packets Linrad sends.
It is a fairly uncomplicated thing however, but I guess you do not want
the CPU load from Linrad computing FFTs for all the dongles...

Regards

Leif / SM5BSZ

> --

> You received this message because you are subscribed to the Google Groups "Ultra Cheap SDR" group.

> To unsubscribe from this group and stop receiving emails from it, send an email to ultra-cheap-s...@googlegroups.com.

[Glenn Pearson]

Interesting video, thanks.
Ii think you may be right about rtl_tcp. I run three or more instances and they all start fine, waiting for a connection. It's only when the websdr server attempts to connect to the third dongle that the signal caught message appears on one of the already connected rtl_tcp processes.
I have made sure they're on different USB controllers but the problem persists.

I will investigate further.

Thanks,

Glenn

Cheshire Computer Repairs Ltd.
Hoole
Chester
CH2 3EW

Tel: 01244 458779
Tel: 01606 339308
Tel: 01928 798161
Tel: 0151 353 9982
Mob: 078 456 45870

www.cheshirecomputerrepairs.co.uk

> You received this message because you are subscribed to a topic in the Google Groups "Ultra Cheap SDR" group.
> To unsubscribe from this topic, visit https://groups.google.com/d/topic/ultra-cheap-sdr/RH4neus2dX0/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to ultra-cheap-s...@googlegroups.com.

[jdow]

What do performance monitors say about CPU usage with 1, 2, and 3 WebSDR
instances? You may be oversubscribed at the CPU cycles bank.

{^_^}

[Glenn Pearson]

I haven't checked but it is unlikely. It's 8 cores @ 4Ghz each.
I had wondered if it was a serial number issue. I had one dongle for a few months before ordering three more directly from China.
I thought maybe the issue want that I was running three but two from the same batch. Alas the issue was exactly the same when using all dongles from the same batch. ie. two work but three doesn't.
The message from rtl_tcp is 'worker cond timeout' then signal caught etc. I checked the source code and the message only appears once, in tcp routine. I've not had chance to explore it properly due to family commitments. I'll get it working eventually though 😉

Glenn
M0PEA

Cheshire Computer Repairs Ltd.
Hoole
Chester
CH2 3EW

Tel: 01244 458779
Tel: 01606 339308
Tel: 01928 798161
Tel: 0151 353 9982
Mob: 078 456 45870

www.cheshirecomputerrepairs.co.uk

[Glenn Pearson]

I've just checked and with two dongles all cores are idling @ 25% or less.
I noticed that I'm using libusbx and not libusb. It's standard for Ubuntu/Debian these days.

I'll have another look at the source code when I can.

Thanks for your input.

Glenn

Cheshire Computer Repairs Ltd.
Hoole
Chester
CH2 3EW

Tel: 01244 458779
Tel: 01606 339308
Tel: 01928 798161
Tel: 0151 353 9982
Mob: 078 456 45870

www.cheshirecomputerrepairs.co.uk

> On 30 Aug 2014, at 18:13, jdow <jd...@earthlink.net> wrote:
>

[jdow]

This is sort of obvious but.... Do you have all the rtl_tcp instances on
different port numbers? (Two on the same IP seems impossible. But....)

{^_^}

[Glenn Pearson]

They're all using the local host 127.0.0.1 and sequential port numbers 1234,1235,1236 etc.

Cheshire Computer Repairs Ltd.
Hoole
Chester
CH2 3EW

Tel: 01244 458779
Tel: 01606 339308
Tel: 01928 798161
Tel: 0151 353 9982
Mob: 078 456 45870

www.cheshirecomputerrepairs.co.uk

[jdow]

In theory that should work unless localhost is somehow a bottleneck. Hm, try
reducing the buffer size. And if your Ethernet configuration allows it setup
jumbo packets. (I'm grasping at straws here presuming you've done most of the
usual troubleshooting like making sure each one works alone.)

Can you run three dongles at once without using the TCP connections?

Oh - are you on a gigabit net or a 100 megabit net? 2.4 Msps means 2*2.4*8
megabits per second or 38.4k bytes per second. Three times that is well over 100
mbps.

{^_^}

[Glenn Pearson]

Hi,
I just checked and I can run rtl_fm on a third dongle while the other
two work happily on rtl_tcp...
My LAN is Gigabit. I also checked and the loopback or localhost has no
errors. I'm not so sure I can set jumbo packets for loopback but will check.
I've had look at the rtl_tcp source but my C isn't great these days. I
can't see what's causing the timeout.

I'll have another go later this week.

Glenn

On 30/08/14 20:24, jdow wrote:
> In theory that should work unless localhost is somehow a bottleneck.
> Hm, try reducing the buffer size. And if your Ethernet configuration
> allows it setup jumbo packets. (I'm grasping at straws here presuming
> you've done most of the usual troubleshooting like making sure each
> one works alone.)
>
> Can you run three dongles at once without using the TCP connections?
>
> Oh - are you on a gigabit net or a 100 megabit net? 2.4 Msps means
> 2*2.4*8 megabits per second or 38.4k bytes per second. Three times
> that is well over 100 mbps.
>
> {^_^}
>
> On 2014-08-30 12:16, Glenn Pearson wrote:
>> They're all using the local host 127.0.0.1 and sequential port
>> numbers 1234,1235,1236 etc.
>>
>>
>>
>>

>>> On 30 Aug 2014, at 20:03, jdow <jd...@earthlink.net> wrote:
>>>
>>> This is sort of obvious but.... Do you have all the rtl_tcp
>>> instances on different port numbers? (Two on the same IP seems
>>> impossible. But....)
>>>
>>> {^_^}
>>>
>>>> On 2014-08-30 11:07, Glenn Pearson wrote:
>>>> I haven't checked but it is unlikely. It's 8 cores @ 4Ghz each.
>>>> I had wondered if it was a serial number issue. I had one dongle
>>>> for a few months before ordering three more directly from China.
>>>> I thought maybe the issue want that I was running three but two
>>>> from the same batch. Alas the issue was exactly the same when using
>>>> all dongles from the same batch. ie. two work but three doesn't.
>>>> The message from rtl_tcp is 'worker cond timeout' then signal
>>>> caught etc. I checked the source code and the message only appears
>>>> once, in tcp routine. I've not had chance to explore it properly
>>>> due to family commitments. I'll get it working eventually though 😉
>>>>
>>>> Glenn
>>>> M0PEA
>>>>
>>>>
>>>>
>>>>

>>>>> On 30 Aug 2014, at 18:13, jdow <jd...@earthlink.net> wrote:
>>>>>
>>>>> What do performance monitors say about CPU usage with 1, 2, and 3
>>>>> WebSDR instances? You may be oversubscribed at the CPU cycles bank.
>>>>>
>>>>> {^_^}
>>>>>
>>>>>> On 2014-08-30 01:36, Glenn Pearson wrote:
>>>>>> Interesting video, thanks.
>>>>>> Ii think you may be right about rtl_tcp. I run three or more
>>>>>> instances and they all start fine, waiting for a connection. It's
>>>>>> only when the websdr server attempts to connect to the third
>>>>>> dongle that the signal caught message appears on one of the
>>>>>> already connected rtl_tcp processes.
>>>>>> I have made sure they're on different USB controllers but the
>>>>>> problem persists.
>>>>>>
>>>>>> I will investigate further.
>>>>>>
>>>>>> Thanks,
>>>>>> Glenn
>>>>>>
>>>>>>
>>>>>>
>>>>>>

[Adam Nielsen]

> I just checked and I can run rtl_fm on a third dongle while the other
> two work happily on rtl_tcp...

Sounds like USB isn't the problem then, although rtl_fm might be using
less bandwidth. Can you try rtl_tcp with reduced bandwidth and see
whether that lets you get three dongles running?

> My LAN is Gigabit. I also checked and the loopback or localhost has
> no errors. I'm not so sure I can set jumbo packets for loopback but
> will check.

I don't think the loopback interface is going to be the problem as it
only uses CPU and not any NICs. As long as your CPU usage isn't high,
enabling jumbo packets won't help, and at any rate on my Linux machine
the localhost interface is set to accept packets up to 64k which is
much larger than a 9k jumbo packet.

Perhaps you could confirm that your network really is capable of
gigabit speeds, with a utility like ttcp or similar? Just in case
something is unexpectedly reducing it to 100Mbps or less.

Sorry perhaps you already covered this, but did you say you only have
the problem accessing the dongles remotely? Or do you start seeing
problems when you are accessing the rtl_tcp instances from the local
machine too? Because if it's a problem when using localhost then of
course it's not going to be a network bandwidth issue.

Cheers,
Adam.

[Glenn Pearson]

Hi Adam,
The loopback interface is indeed 64kb packets. The websdr server also runs on the same machine. So it is unlikely to be a network issue.
I've tried reducing the sample rate to 256kHz for all three dongles. It makes no difference.
I have read somewhere that requesting data from rtl_tcp immediately after setting the frequency or sample rate can cause this issue.
I'll try to set the freq etc at the command line before starting the server. Although the server will do it again when it's started.
Will report back.

Thanks,

Glenn

[Glenn Pearson]

Hi All,
I installed the same OS and all the updates onto a spare PC this week. It's a pretty crappy Pentium 4 HT but it runs ok.
I plugged in three RTL dongles and started up three instances of RTL_TCP and then tried the websdr64 server. Same result as my main server. One or more instances of RTL_TCP exited.
But when I reduced the sample rate and tried again it worked. Eureka!
It's still not sorted on my main server but at least I know it's not some glitch in RTL_TCP that's causing the issue. I'll add some more USB ports, via PCI cards, to the server and try again.

Thanks all,
Glenn

[Adam Nielsen]

> But when I reduced the sample rate and
> tried again it worked. Eureka! It's still not sorted on my main
> server but at least I know it's not some glitch in RTL_TCP that's
> causing the issue. I'll add some more USB ports, via PCI cards, to
> the server and try again.

Very interesting! Thanks for getting back to us with an update. Let
us know if the extra USB ports via PCI work (and whether they are the
same or a different chipset to your onboard ones.)

Cheers,
Adam.

[Glenn Pearson]

Hi Adam,
It didn't work 😞
The PCI card showed up, in lsusb -t, as a separate 480Mbps controller. I made sure the first two dongles were on different internal controllers and they worked as before.
I then attached the third dongle to the new controller, also 480Mbps.
As soon as I fired up the websdr the RTL_TCP process communicating with the dongle sampling at 2.8Msps fell over.
Maybe I should reduce the sample rate... It's only so wide because it covers 10 & 11m bands.
In fact I could actually run it as two separate dongles and two different bands (assuming I can get a third working). It may improve reception too. I feel the SSB sounds distorted and had a few comments about the CW sounding the same on this band. It may well be because the dongle is running at it's absolute limit.

Rant over,
Glenn

[Siegfried Jackstien]

The "limit" is 2.4 mps

Thats is what I could find out here on 3 different pcs

If you wanna use many dongles then its best to change the dongles serial
numbers ...

Dg9bfc

Sigi


> -----Ursprüngliche Nachricht-----
> Von: ultra-c...@googlegroups.com [mailto:ultra-cheap-

> s...@googlegroups.com] Im Auftrag von Glenn Pearson
> Gesendet: Sonntag, 14. September 2014 16:29
> An: ultra-c...@googlegroups.com
> Betreff: Re: [ultra-cheap-sdr] Re: Unable to run more than two RTL-SDR

> dongles at the same time.
>

> Hi Adam,
> It didn't work ??

> --
> You received this message because you are subscribed to the Google Groups
> "Ultra Cheap SDR" group.
> To unsubscribe from this group and stop receiving emails from it, send an

[Glenn Pearson]

Hi sigi
I have 2.8Msps working here http://ccr.linkpc.net:8901
But I think it's too high. If I get chance tomorrow I plan on trying 3 again but having all of them at 2Msps.
If it fails again ill try changing the serials.
Will let you know.

Glenn

[Adam Nielsen]

Hi Sigi,

> If you wanna use many dongles then its best to change the dongles
> serial numbers ...

I'm curious - why does changing the serial numbers make a difference?
I'm not sure Linux makes use of the serial numbers, unless you have
custom udev rules.

Cheers,
Adam.

[Glenn Pearson]

I finally have it working with 3 dongles. See ccr.linkpc.net:8901
I altered the 10m band from 2.8Msps to 2Msps and added a third band at 1Msps.
They're all on different USB buses running at 480Mbit.
I plan on adding a fourth this weekend.... If it works I'll be amazed.
They all have the same serial number BTW. So that's not an issue.

Will report back with more info after more experimentation.

Glenn

[Glenn Pearson]

Hi All,

I think I have it solved. 

Two days ago I managed to entirely screw up the webSDR.  After having three dongles running for the past few days I thought I'd adjust the frequencies and make a couple of slight changes. Big mistake. I couldn't get three running again and have up in disgust. 

However, keen to see it running again, I downloaded the latest rtl tools library from osmocom and compiled it. Amazingly the RTL_TCP now seems rock solid. 

I currently have four dongles running with no stability issues at all. It even stops and restarts the RTL_TCP thread automatically if I restart the webSDR server!

Thank you osmocom and all on here that offered advice. 

Regards,

Glenn

--
You received this message because you are subscribed to a topic in the Google Groups "Ultra Cheap SDR" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ultra-cheap-sdr/RH4neus2dX0/unsubscribe.
To unsubscribe from this group and all its topics, send an email to ultra-cheap-s...@googlegroups.com.

[Adam Nielsen]

> I downloaded the latest rtl tools
> library from osmocom and compiled it. Amazingly the RTL_TCP now seems
> rock solid. I currently have four dongles running with no stability
> issues at all. It even stops and restarts the RTL_TCP thread
> automatically if I restart the webSDR server! Thank you osmocom and
> all on here that offered advice.

Glad to hear you finally got it all working! Thanks for letting us
know your solution.

Cheers,
Adam.

[Glenn Pearson]

Hi All,
After seemingly resolving my problem I now have another.
If I attempt to run more than 4 dongles I now have the following issue;
Using device 3: Generic RTL2832U OEM
Found Rafael Micro R820T tuner
Tuned to 100000000 Hz.
listening...

Use the device argument 'rtl_tcp=127.0.0.1:1238' in OsmoSDR (gr-osmosdr) source
to receive samples in GRC and control rtl_tcp parameters (frequency, gain, ...).
client accepted!
set freq 51000000
Failed to submit transfer 3!
set sample rate 2048000
set freq correction 0
worker cond timeout
Signal caught, exiting!
comm recv bye
Signal caught, exiting!
all threads dead..
listening...

Anyone know what this means?

Thanks in advance,
Glenn

[jdow]

That happens in the code that attempts to start the reading process. In
specific it happens after the program THINKS that _rtlsdr_alloc_async_buffers
has successfully completed. (Guys, when you have something returning an
error code it might be useful to test it.) The next step is to fill out
some of the overhead for each of the buffers created and initiate the
transfer in libusb. Specifically it completed three submissions and the
fourth (index 3 starting with 0) cannot start. Since there is no print
out of the return value from libusb_submit_transfer we don't have any
idea why it would have failed.

Are you sure data source machine has enough memory to run all the rtl_tcp's
that you're asking for? Is there a memory quota on your account? The only
other thing that makes sense at the moment is you're asking for more CPU
cycles than it's ready to give.

You could go into librtlsdr.c and add an error check on the output value
for _rtlsdr_alloc_async_buffers().
Change:
_rtlsdr_alloc_async_buffers();
to:
r = _rtlsdr_alloc_async_buffers();
if ( r < 0 )
{
fprintf(stderr, "Failed to create buffers!\n");
async_status = RTLSDR_CANCELING;
break;
}
end:

A few lines further down change:
fprintf(stderr, "Failed to submit transfer %i!\n", i);
to:
fprintf(stderr, "Failed to submit transfer %i %d!\n", i, r );
end:

That will give you more data on what failed.

{^_^} Joanne

> --
> You received this message because you are subscribed to the Google Groups "Ultra
> Cheap SDR" group.

> To unsubscribe from this group and stop receiving emails from it, send an email
> to ultra-cheap-s...@googlegroups.com

> <mailto:ultra-cheap-s...@googlegroups.com>.

> To post to this group, send email to ultra-c...@googlegroups.com

> <mailto:ultra-c...@googlegroups.com>.

[jdow]

My error below. Change the "break;" to "return r".

{^_^}

[Glenn Pearson]

Hi Joanne,
I've not tried altering the source yet.
The server has 16GB RAM and no account limitation. The processor is pretty much idling too.
I did come across someone with a similar problem and it was due to libusb. The developers commented that it would be fixed in the next release. I have the current approved version but there is a later version. I may attempt installing it this week.
Forgive my ignorance but will I need to install libusb-dev and then recompile the rtlsdr tools?

Thanks in advance,
Glenn

[jdow]

On 2014-09-30 04:16, Glenn Pearson wrote:
> Hi Joanne,

> Forgive my ignorance but will I need to install libusb-dev and then recompile the rtlsdr tools?

That would give you more information about the failure. But if somebody who
works with the source has duplicated the problem you can probably sit back and
relax. (I can't help - I'm doing my SDR work on Windows. I'm agnostic that way.)

{^_^} Joanne

[Glenn Pearson]

Hello again,
I finally have 6, yes SIX, dongles running on the one machine.
I had to use 'rtl_tcp -d 4 -p 1238 -b 3 -n 1000' to fire up rtl_tcp.
It seems that reducing the buffers to 3 worked but the waterfall display on the webSDR seemed slow.
Doubling the number of max number of linked list buffers from 500, default, to 1000 seems to make up for reducing the buffers.
I was listening to an SSB contest on 4m recently and it worked very well.
In fact, they are all running from a single USB 2.0 port on the same powered hub.
I'll keep on adding bands until I run out of dongles... Let's see how far I can push this

Thanks everyone for your help,
Glenn - M0PEA
www.cheshiresdr.co.uk

[Siegfried Jackstien]

Use different ports on your machine and get more power out of the setup

Dg9bfc

Sigi



> -----Ursprüngliche Nachricht-----
> Von: ultra-c...@googlegroups.com [mailto:ultra-cheap-
> s...@googlegroups.com] Im Auftrag von Glenn Pearson

> Gesendet: Mittwoch, 1. April 2015 18:43

> An: ultra-c...@googlegroups.com
> Betreff: Re: [ultra-cheap-sdr] Re: Unable to run more than two RTL-SDR
> dongles at the same time.
>

> Hello again,
> I finally have 6, yes SIX, dongles running on the one machine.
> I had to use 'rtl_tcp -d 4 -p 1238 -b 3 -n 1000' to fire up rtl_tcp.
> It seems that reducing the buffers to 3 worked but the waterfall display
> on the webSDR seemed slow.
> Doubling the number of max number of linked list buffers from 500,
> default, to 1000 seems to make up for reducing the buffers.
> I was listening to an SSB contest on 4m recently and it worked very well.
> In fact, they are all running from a single USB 2.0 port on the same
> powered hub.
> I'll keep on adding bands until I run out of dongles... Let's see how far

> I can push this :o<http://www.rtl-
> sdr.com/forum/images/smilies/icon_e_surprised.gif>

>
> Thanks everyone for your help,
> Glenn - M0PEA
> www.cheshiresdr.co.uk
>
>
> On Tuesday, September 30, 2014 at 6:16:40 PM UTC+1, jdow wrote:
>
> On 2014-09-30 04:16, Glenn Pearson wrote:
> > Hi Joanne,
>
> > Forgive my ignorance but will I need to install libusb-dev and
> then recompile the rtlsdr tools?
>
> That would give you more information about the failure. But if
> somebody who
> works with the source has duplicated the problem you can probably
> sit back and
> relax. (I can't help - I'm doing my SDR work on Windows. I'm
> agnostic that way.)
>
> {^_^} Joanne
>
>

> --
> You received this message because you are subscribed to the Google Groups
> "Ultra Cheap SDR" group.
> To unsubscribe from this group and stop receiving emails from it, send an

> email to ultra-cheap-s...@googlegroups.com.
> To post to this group, send email to ultra-c...@googlegroups.com.

[Glenn Pearson]

Hi Sigi,
I realise I can / should use more than one port.
I calculate I'm currently using about 200mb/s at the moment.
USB 2 can handle 480mb/s so I should be able to have 12 dongles on one band..... although I doubt it would work I may give it a go.
That said I only have 2 spare dongles now. I better order a few more.

73's,
Glenn

[Siegfried Jackstien]

See to get some with the r 8020t2 tuner in it ... they really work a tiny
bit better ... I had one loaned from a friend here to make some a/b tests
... and yes the t2 tuner is a tiny bit more sensitive (I dunno why ... just
can say my results)

With original drivers a higher signal from dab and dvbt stations (presto pvr
used) ... with fewer dropouts on low signals and some more stations found
(on lower signal end and not worth receiving in dvbt or dab) ... but for dx
fm hunting exactly that low signal stations are that what you are hunting
for...

So ... if you see only a few bucks difference ... then buy the ts tuner
dongles

With the newest drives (where you can select different gain settings for
lowest imd or highest sensitivity) you can squeeze the last tiny signals out
of the noise ... sadly the sampling chip can not go to higher bit rate :-(

....

I am thinking about buying an sdrplay ... uses the same tuner chip as the
new fcdpp from howard (mirics tuner) ... but uses also mirics backend and
gives 8 megs of bw ... shortwave included without upconverter etc ...

Yes I can buy a sag full of rtl dongles for the sdr plays price .. but I
wanna HAVE ONE ... (I will maybe sell the fcdpp then ... who knows!?!)

Greetz

Sigi

Dg9bfc

> -----Ursprüngliche Nachricht-----
> Von: ultra-c...@googlegroups.com [mailto:ultra-cheap-
> s...@googlegroups.com] Im Auftrag von Glenn Pearson

> Gesendet: Sonntag, 5. April 2015 19:50

> An: ultra-c...@googlegroups.com
> Betreff: Re: [ultra-cheap-sdr] Re: Unable to run more than two RTL-SDR
> dongles at the same time.
>

> Hi Sigi,
> I realise I can / should use more than one port.
> I calculate I'm currently using about 200mb/s at the moment.
> USB 2 can handle 480mb/s so I should be able to have 12 dongles on one
> band..... although I doubt it would work I may give it a go.
> That said I only have 2 spare dongles now. I better order a few more.
>
> 73's,
> Glenn
>
> On Thursday, April 2, 2015 at 11:22:15 PM UTC+1, DG9BFC Sigi wrote:
>
> Use different ports on your machine and get more power out of the
> setup
>
> Dg9bfc
>
> Sigi
>
>
>
> > -----Ursprüngliche Nachricht-----

> > Von: ultra-c...@googlegroups.com <javascript:> [mailto:ultra-
> cheap- <javascript:>
> > s...@googlegroups.com <javascript:> ] Im Auftrag von Glenn Pearson

> > Gesendet: Mittwoch, 1. April 2015 18:43

> > An: ultra-c...@googlegroups.com <javascript:>

> > Betreff: Re: [ultra-cheap-sdr] Re: Unable to run more than two
> RTL-SDR
> > dongles at the same time.
> >
> > Hello again,
> > I finally have 6, yes SIX, dongles running on the one machine.
> > I had to use 'rtl_tcp -d 4 -p 1238 -b 3 -n 1000' to fire up
> rtl_tcp.
> > It seems that reducing the buffers to 3 worked but the waterfall
> display
> > on the webSDR seemed slow.
> > Doubling the number of max number of linked list buffers from 500,
> > default, to 1000 seems to make up for reducing the buffers.
> > I was listening to an SSB contest on 4m recently and it worked
> very well.
> > In fact, they are all running from a single USB 2.0 port on the
> same
> > powered hub.
> > I'll keep on adding bands until I run out of dongles... Let's see
> how far
> > I can push this :o<http://www.rtl-
> > sdr.com/forum/images/smilies/icon_e_surprised.gif

> <http://sdr.com/forum/images/smilies/icon_e_surprised.gif> >

> > email to ultra-cheap-s...@googlegroups.com <javascript:> .

> > To post to this group, send email to ultra-c...@googlegroups.com

> <javascript:> .

> > Visit this group at http://groups.google.com/group/ultra-cheap-sdr

> <http://groups.google.com/group/ultra-cheap-sdr> .

> > For more options, visit https://groups.google.com/d/optout

> <https://groups.google.com/d/optout> .