[Hackrf-dev] FW: FW: deaf HackRF

Wed Jan 21 15:18:16 EST 2015

Hi Paul,

Thanks for the suggestions.

I have been using a cavity for the NOAA tests & a 'reasonably' band specific
antenna  .

>> I have rerun the test -with HackRF sample rate 2MSPS & 0.25MSPS as you
suggested. 
>I only suggested running the HackRF at 2MSPS and not 0.25MSPS, the 0.25MSPS
was for the >RTL2832,

Sorry - I explained this poorly. The test was run with 2MSPS on HackRF &
0.25 on the RTL.

I have not run the HackRF with GNURadio yet. Can anyone comment on
experience running HackRF in a VM Linux build? I have a Kali image running
using Vmware player but suggestions of alternative distributions would be
welcome.

Regards

Stephen

From: HackRF-dev [mailto:hackrf-dev-bounces at greatscottgadgets.com] On Behalf
Of Paul Connolly
Sent: Thursday, 22 January 2015 12:17 AM
To: hackrf-dev at greatscottgadgets.com
Subject: Re: [Hackrf-dev] FW: FW: deaf HackRF

With Microsoft windows as your OS, you can probably not currently, or
easily, get the best performance out of your HackRF. There are no decimation
options in most(all?) of the windows SDR software, where you trade bandwidth
for extra bits ( Although not exactly the same the idea it is similar enough
to make the concept easier to understand - read:
http://www.atmel.com/images/doc8003.pdf ). Half a bit increase in the
resolution, or lower the in band noise floor by 3dB, every time you half the
bandwidth..  With the right flowgraph in gnuradio-companion, if you correct
for the IQ DC offset and start with 20MSPS, and trade bandwidth for extra
bits,  NOAA reception (12KHz) you could gain 10 extra bits. resolution  Or
60dB, which would make the HackRF slightly better than a RTL2832R for NOAA
reception. But you would need a bandpass filter for in front of the HackRF
for NOAA frequency range to fully realise that 60dB gain.

> I have rerun the test -with HackRF sample rate 2MSPS & 0.25MSPS as you
suggested. 
I only suggested running the HackRF at 2MSPS and not 0.25MSPS, the 0.25MSPS
was for the RTL2832, because you should get slightly better performance with
it as well. 2MSPS is about the lowest samplerate that you should set the
HackRF. As 1.75MHz is the smallest bandpass filter available in the MAX2837
( https://github.com/mossmann/hackrf/wiki/Hardware-Components#block-diagrams
)
$ hackrf_transfer                                       :(
... snip ...
        [-b baseband_filter_bw_hz] # Set baseband filter bandwidth in MHz.
        Possible values: 1.75/2.5/3.5/5/5.5/6/7/8/9/10/12/14/15/20/24/28MHz,
default < sample_rate_hz.

Paul

On 21/01/2015 11:03, Stephen wrote:

Hi Karl & Paul,

Thanks for the comments.

Karl - for both tests the gain details were max gain on both devices. VGA
gain on hackrf to 14db (but adjusting VGA gain makes little improvement to
S/N)

HackRF AMP was OFF for first test.

I don't think a half duplex TX capability is going to significantly affect
receive performance. Conventional 2 way radios use a diode switch to select
/ isolate RX while TX is active & this does not prevent RF performance of
<0.35uv gating sensitivity.

Paul,

I have rerun the test -with HackRF sample rate 2MSPS & 0.25MSPS as you
suggested.

I ran the test in a simpler way, RF level required for carrier to show above
noise floor.

The performance gap closed to 28dB from 32-36db with the amp off & about
16db with the amp on.

Also, to expand on the NOAA RX test I did previously, I tried a variety of
sample rates & amp settings to receive a NOAA satellite - the best being
fractionally above the noise floor. Under same test conditions, the RTL was
able to receive NOAA APT easily.

I think my intent has been misunderstood here - I am interested to establish
if my HackRF is faulty or if this the normal RF sensitivity.

Below is  a previous reference to a deaf HackRF, related to a manufacturing
issue:

http://nine.pairlist.net/pipermail/hackrf-dev/2013-August/000236.html

Is anyone able to provide more details on the fault listed in the link above
(I.E. is it common, measured effect, resolution etc)? 

Thanks

Stephen

From: HackRF-dev [mailto:hackrf-dev-bounces at greatscottgadgets.com] On Behalf
Of Karl Koscher
Sent: Tuesday, 20 January 2015 3:10 PM
To: Paul Connolly
Cc: hackrf-dev at greatscottgadgets.com
Subject: Re: [Hackrf-dev] FW: deaf HackRF

There are also multiple gain settings for the HackRF. I'm not sure what the
gain settings are for an RTL device, but I believe both the rtl and the
tuner have their own gain settings as well.

Keep in mind that there are engineering tradeoffs when designing any device.
The HackRF is designed to be a low-cost TX/RX SDR peripheral with a large
frequency range and high bandwidth. RTL devices are RX-only, have a smaller
frequency range, and significantly less bandwidth.

On Mon, Jan 19, 2015 at 12:51 PM, Paul Connolly  <mailto:eeipcy at gmail.com>
<eeipcy at gmail.com> wrote:

As the sample rate increases, so does the the noise floor. Or with more
bandwidth comes more noise. So giving the numbers without the sample rates
used is not useful. In SDR# for the HackRF you can type in a sample rate in
the sample rate pull down box, try using 2 MSPS. And for the RTL2832 try
dropping the sample rate to 0.25MHz. And repeat your experiment again.

On 19/01/2015 20:25, Stephen wrote:

Hi Tom,

Thanks for the reply.

The RX sensitivity figure I quoted is dBm - 'Decibels above a MilliWatt'.
-95dBm is 95 db below a MilliWatt, wheras -127dBm is 127 db below a
MilliWatt.

This test was conducted with an RF signal generator.

Regarding the test, to register minimum scale on the SDR# spectrum, the
HackRF required -95dBm , the RTL2832 required only -127dBm.

Expressing this as microvolts instead of dBm, the equivalent level in
microvolts (to register the same scale on SDR#) are:

0.1uV  for the RTL2832 

3.5uV for the hack RF

Accordingly, The HackRF requires considerably more signal to register the
same scale on SDR#.

I discovered this performance limitation when  I attempted to try to receive
NOAA weather satellites, I found HackRF barely registered, wheras the
RTL2832 registered a strong signal for the same satellite pass  on the same
antenna.

I appreciate you are trying to help & I don't mean to be rude by
contradicting you,  but the HackRF I have is definitely 'deaf'.

Regards,

Stephen

From: Tom Buelens [mailto:tom.buelens at gmail.com] 
Sent: Monday, 19 January 2015 8:23 PM
To: Stephen
Subject: Re: [Hackrf-dev] deaf HackRF

Hi Stephen,

I might be mistaken but I actually think the numbers you mention show that
the HackRF is better at receiving the signal.

You see, an attenuation of -105 dBm is resulting in a smaller signal then
-69dBm. Please also see here:

http://en.wikipedia.org/wiki/DBm

Cheers,

Tom

On Mon, Jan 19, 2015 at 2:36 AM, Stephen   <mailto:refsmmat at gmail.com>
<mailto:refsmmat at gmail.com>  <mailto:refsmmat at gmail.com>
<refsmmat at gmail.com> wrote:

Hi,

Ive just bought a HackRF, but I have found  RF performance is poor. Compared
to a $12 USB RTL2832 SDR, it is deaf by about 32-36dB.

The tests were done with an HP 8922 test set & an unmodulated carrier at
137.5MHz. The test was the generated RF signal required for both devices to
achieve the same scale level in SDR#.

RTL SDR                  HackRF     Difference

Scale1   -127dBm              -95dBm               32dB

Scale2   -115dBm              -79dBm               36dB

Scale3  -105dBm               -69dBm                36dB

The tests were done with max gain on both devices, but with the HackRF AMP
off.

I have seen from posts that others have had this problem - related soldering
of RF switches in the manufacturing process.

Can anyone provide further details on the fix  or suggest a resolution
please?

Thanks,

Stephen

_______________________________________________
HackRF-dev mailing list
HackRF-dev at greatscottgadgets.com
https://pairlist9.pair.net/mailman/listinfo/hackrf-dev

_______________________________________________
HackRF-dev mailing list
HackRF-dev at greatscottgadgets.com
https://pairlist9.pair.net/mailman/listinfo/hackrf-dev

_______________________________________________
HackRF-dev mailing list
HackRF-dev at greatscottgadgets.com
https://pairlist9.pair.net/mailman/listinfo/hackrf-dev

_______________________________________________
HackRF-dev mailing list
HackRF-dev at greatscottgadgets.com
https://pairlist9.pair.net/mailman/listinfo/hackrf-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://pairlist9.pair.net/pipermail/hackrf-dev/attachments/20150122/29981b05/attachment.html>