USRP B200 host BW with USB 3.0/2.0

GNU Radio
#1

I understand the given values of host bandwidth for each protocol:

USB 2.0 8 (MS/s @ 16-bit I/Q)

USB 3.0 61.44 (MS/s @ 16-bit I/Q)

However I would like to process 40MHz of analogue BW in GNURadio over
USB
2.0

I fully understand a continuous reception is not possible to manage
since
it would require 40 IQ MS/s and I am limited to 8MS/s.

However, is it possible to take snapshots of 40MHz over the time so that
I
am able to receive bits of 40MHz with USB 2.0 which are not continuous
in
time?

Are there buffers in the FPGA that manage this kind of operation?

The idea is to receive the high speed 40MHz in the FPGA and then send it
over the USB 2.0. When it is accomplished we would be ready to receive
another 40MHz sample block.

I am afraid the only alternative is to receive 5 times 8MHz (each at
8Ms/s)
so that I get 40MHz of analogue BW in 5 different
readings/reconfigurations
of my GNURadio flowgraph. However I ask just in case…

Many thanks in advance,

Jorge.

#2

On 01/29/2015 01:48 PM, Jorge G. wrote:

I understand the given values of host bandwidth for each protocol:

USB 2.0 8 (MS/s @ 16-bit I/Q)

If you go down to 8-bit I&Q, you will get twice that amount, if that’s
any help.

However, is it possible to take snapshots of 40MHz over the time so that
I am able to receive bits of 40MHz with USB 2.0 which are not continuous
in time?

Are there buffers in the FPGA that manage this kind of operation?

None big enough for anything useful.

It seems like attaching a USB3.0 connector would be simpler than running
5 USRPs? Not that I would complain about the sales :slight_smile:

M

#3

I understand his question that he just wants to let the whole thing pump
out
samples, and he takes only a portion of it. I have no idea if this works
in
a controlled manner, but at least for spectrum display this works just
fine,
the lost samples are not a big issue.

Ralph.

#4

Martin,

The alternative of a USB 3.0 is not having 5 USRPs. It is just reading 5
times at different centrer frequencies in order to get 40MHz information
although that information will not be gathered at the same time.

For my application I just need to “see” the sprectrum. I do not need to
demodulate data so that it would be fine to proceed that way. However i
would rather get 40MHz in a row.

Many thanks,
Jorge

#5

How is that possible since only 8MHz can be placed over USB 2.0 ?

On one hand I have a B200 which is supposed to deliver 56 MHz.

On the other hand, for my final application, I am thinking of a low-cost
host (such as BeagleBone Black) which I am afraid only has USB 2.0

My question was about getting 40MHz BW over USB 2.0 without tuning the
central frequency of the USRP several times.

Regards,

Jorge

On 29 January 2015 at 16:35, Ralph A. Schmid, dk5ras [email protected]

#6

I made tests with the bladeRF at 30 MHz bandwidth, and I was able to see
the spectrum just fine, not matter if I used a USB2 or a USB3 cable. No
hacks, no switching chunks of the spectrum, just using a standard
receiver program. The lost samples are not a problem as longs as you do
not want to listen in.

Ralph.

From: [email protected]lid
[mailto:[email protected]lid] On Behalf Of
Jorge G.
Sent: Thursday, January 29, 2015 4:07 PM
To: Martin B.
Cc: GNURadio D.ion List
Subject: Re: [Discuss-gnuradio] USRP B200 host BW with USB 3.0/2.0

Martin,

The alternative of a USB 3.0 is not having 5 USRPs. It is just reading 5
times at different centrer frequencies in order to get 40MHz information
although that information will not be gathered at the same time.

For my application I just need to “see” the sprectrum. I do not need to
demodulate data so that it would be fine to proceed that way. However i
would rather get 40MHz in a row.

Many thanks,

Jorge

On 29 January 2015 at 14:07, Martin B. <[email protected]
mailto:[email protected] > wrote:

On 01/29/2015 01:48 PM, Jorge G. wrote:

I understand the given values of host bandwidth for each protocol:

USB 2.0 8 (MS/s @ 16-bit I/Q)

If you go down to 8-bit I&Q, you will get twice that amount, if that’s
any help.

However, is it possible to take snapshots of 40MHz over the time so that
I am able to receive bits of 40MHz with USB 2.0 which are not continuous
in time?

Are there buffers in the FPGA that manage this kind of operation?

None big enough for anything useful.

It seems like attaching a USB3.0 connector would be simpler than running
5 USRPs? Not that I would complain about the sales :slight_smile:

M

#7

To be honest, I have no idea :slight_smile: However I did not yet test this with the
B210, as now USB3 works flawless. In the beginning the BladeRF was not
compatible with my USB3 setup for whatever reason, so I had to stay with
USB2 for a while until those issues were sorted out both in BladeRF
software and VMware.

Let’s see…FPGA loading really takes a while through a USB2 cable…HDSDR
is starting…master_clock_rate=56e6…

Tadaaa – here we go: http://dk5ras.dyndns.org/screenshots/B210_USB2.png
with an USB 2 cable on an USB2 port.

And in comparison http://dk5ras.dyndns.org/screenshots/B210_USB3.png
with an USB3 cable on an USB3 port of the same PC.

Same portion of the spectrum, same receiver location (just with a short
wideband Delock 88451 antenna, sitting on my table), similar picture…

Ralph.

From: Jorge G. [mailto:[email protected]]
Sent: Thursday, January 29, 2015 4:44 PM
To: Ralph A. Schmid, dk5ras
Cc: Martin B.; GNURadio D.ion List
Subject: Re: [Discuss-gnuradio] USRP B200 host BW with USB 3.0/2.0

How is that possible since only 8MHz can be placed over USB 2.0 ?

On one hand I have a B200 which is supposed to deliver 56 MHz.

On the other hand, for my final application, I am thinking of a low-cost
host (such as BeagleBone Black) which I am afraid only has USB 2.0

My question was about getting 40MHz BW over USB 2.0 without tuning the
central frequency of the USRP several times.

Regards,

Jorge

On 29 January 2015 at 16:35, Ralph A. Schmid, dk5ras <[email protected]
mailto:[email protected] > wrote:

I made tests with the bladeRF at 30 MHz bandwidth, and I was able to see
the spectrum just fine, not matter if I used a USB2 or a USB3 cable. No
hacks, no switching chunks of the spectrum, just using a standard
receiver program. The lost samples are not a problem as longs as you do
not want to listen in.

Ralph.

From: [email protected]lid
mailto:[email protected] [mailto:discuss-gnuradio-bounces+ralph
mailto:discuss-gnuradio-bounces%2Bralph [email protected]
mailto:[email protected] ] On Behalf Of Jorge G.
Sent: Thursday, January 29, 2015 4:07 PM
To: Martin B.
Cc: GNURadio D.ion List
Subject: Re: [Discuss-gnuradio] USRP B200 host BW with USB 3.0/2.0

Martin,

The alternative of a USB 3.0 is not having 5 USRPs. It is just reading 5
times at different centrer frequencies in order to get 40MHz information
although that information will not be gathered at the same time.

For my application I just need to “see” the sprectrum. I do not need to
demodulate data so that it would be fine to proceed that way. However i
would rather get 40MHz in a row.

Many thanks,

Jorge

On 29 January 2015 at 14:07, Martin B. <[email protected]
mailto:[email protected] > wrote:

On 01/29/2015 01:48 PM, Jorge G. wrote:

I understand the given values of host bandwidth for each protocol:

USB 2.0 8 (MS/s @ 16-bit I/Q)

If you go down to 8-bit I&Q, you will get twice that amount, if that’s
any help.

However, is it possible to take snapshots of 40MHz over the time so that
I am able to receive bits of 40MHz with USB 2.0 which are not continuous
in time?

Are there buffers in the FPGA that manage this kind of operation?

None big enough for anything useful.

It seems like attaching a USB3.0 connector would be simpler than running
5 USRPs? Not that I would complain about the sales :slight_smile:

M