Discussion:
[Hackrf-dev] Sensitivity
McDonald, J Douglas
2015-01-20 14:28:02 UTC
Permalink
I have not measured numerically the sensitivity of these tyoe devices, but
have comapred them. In actual use the seinsitivity of the HackRF and
the RTL dongles are similar, using both SDR$ and Gnuradio.

Both devices are limited by a too-low number of digitized bits.
It is extremely important that the gains be set correctly. You have to
have the HackRF's multiple gains split correctly or it can limit even in an
early stage, resulting in all sorts of image problems, etc.

Try experimenting with widely varying signal levels and gain settings.

I find that for signals well prefiltered performance is extremely similar,
but if not, the HackRF is vastly (no exaggeration) inferior. Note that
"well" implies really great rejection of unwanted signals from 0.3
to 10,000 Mhz. I have no idea what is happening above the
max frequency spec of either, but for HackRF I suspect the
worst in terms of intermodulation problems. Computers with the current rage RF networking
running can seriously overload these gizmos. So use prefilters and
be SURE that they are rejecting all the microwave stuff you are
actually using. Test them. If you are working below a GHz, your filter
may not be killing enough up in the upper reaches.




Doug McDonald
Iluta V
2015-01-20 21:17:15 UTC
Permalink
Dear McDonald,

Thank you for sharing this, and could you suggest more about filters:

Which filters are best suited for RX in sub 1 Ghz area, since on Gnu Radio
3.7.7 version there are 14 different filters.

I made a simple grc file consisting of Source - Throttle - Sinks, and how
the best both types of Filters would fit there?

Since signal quality in outdoors is by far better than indoors with all
other interference around, just for me it means sitting in a car in minus
degrees outside to catch quality signals, until laptop batteries die.

Which would be prefilters you mentioned, and maybe you could have a sample
for prefilter usage?

Does it make sense to add filters for RTL as well, to get even better
signal quality, or your suggestion to use prefilters and filters for HackRF
only?

(sorry, just I'm not a specialist in IT or SDRs at all).

Best regards,

Iluta
Post by McDonald, J Douglas
I have not measured numerically the sensitivity of these tyoe devices, but
have comapred them. In actual use the seinsitivity of the HackRF and
the RTL dongles are similar, using both SDR$ and Gnuradio.
Both devices are limited by a too-low number of digitized bits.
It is extremely important that the gains be set correctly. You have to
have the HackRF's multiple gains split correctly or it can limit even in an
early stage, resulting in all sorts of image problems, etc.
Try experimenting with widely varying signal levels and gain settings.
I find that for signals well prefiltered performance is extremely similar,
but if not, the HackRF is vastly (no exaggeration) inferior. Note that
"well" implies really great rejection of unwanted signals from 0.3
to 10,000 Mhz. I have no idea what is happening above the
max frequency spec of either, but for HackRF I suspect the
worst in terms of intermodulation problems. Computers with the current rage RF networking
running can seriously overload these gizmos. So use prefilters and
be SURE that they are rejecting all the microwave stuff you are
actually using. Test them. If you are working below a GHz, your filter
may not be killing enough up in the upper reaches.
Doug McDonald
_______________________________________________
HackRF-dev mailing list
https://pairlist9.pair.net/mailman/listinfo/hackrf-dev
JMG
2015-01-22 08:37:10 UTC
Permalink
Sensitivity measurements

Test results for the noise figure NF of the HackRF, obtained at a
frequency of 435MHz:

NF = 22dB LNA=OFF
NF = 10.8dB LNA=ON
NF = 1.1dB LNA=ON, additional external LNA "LNA4ALL"

RTL-SDR, tuner 820T or 820T2
NF = 3.7dB

The HackRF really is extremely "deaf" when the internal preamplifier is
turned off. Turning it on and installing
a good external amplifier brings NF down to a very good value.
Applying so much gain at the preamplifiers in front of the mixer will
degrade the intermodulation properties.
Adding an external LNA will block the transmitter path to the antenna.

The noise figure is the fundamental parameter to specify the sensitivity
of a receiver. All the other, more common,
sensitivity specifications can be derived from it. For example NF=3dB
will give an input sensitivity of
S=-174dBm+NF+48dB=-123dBm for typical NBFM signal. The factor "48dB"
somewhat depends on the test parameters.
Michael Ossmann
2015-01-23 21:24:34 UTC
Permalink
Could you please share the details of your noise figure test method?

Thanks,

Mike
Post by JMG
Sensitivity measurements
Test results for the noise figure NF of the HackRF, obtained at a
NF = 22dB LNA=OFF
NF = 10.8dB LNA=ON
NF = 1.1dB LNA=ON, additional external LNA "LNA4ALL"
RTL-SDR, tuner 820T or 820T2
NF = 3.7dB
The HackRF really is extremely "deaf" when the internal preamplifier is
turned off. Turning it on and installing
a good external amplifier brings NF down to a very good value.
Applying so much gain at the preamplifiers in front of the mixer will
degrade the intermodulation properties.
Adding an external LNA will block the transmitter path to the antenna.
The noise figure is the fundamental parameter to specify the sensitivity
of a receiver. All the other, more common,
sensitivity specifications can be derived from it. For example NF=3dB
will give an input sensitivity of
S=-174dBm+NF+48dB=-123dBm for typical NBFM signal. The factor "48dB"
somewhat depends on the test parameters.
_______________________________________________
HackRF-dev mailing list
https://pairlist9.pair.net/mailman/listinfo/hackrf-dev
i***@gmail.com
2015-01-23 21:29:05 UTC
Permalink
Mike,

When are you going to respond to those of us who received defective
units? I have sent you a number of emails, and posts here as well, and
you have yet to respond. What's the deal? Please respond with RMA
instructions.
Post by Michael Ossmann
Could you please share the details of your noise figure test method?
Thanks,
Mike
Post by JMG
Sensitivity measurements
Test results for the noise figure NF of the HackRF, obtained at a
NF = 22dB LNA=OFF
NF = 10.8dB LNA=ON
NF = 1.1dB LNA=ON, additional external LNA "LNA4ALL"
RTL-SDR, tuner 820T or 820T2
NF = 3.7dB
The HackRF really is extremely "deaf" when the internal preamplifier is
turned off. Turning it on and installing
a good external amplifier brings NF down to a very good value.
Applying so much gain at the preamplifiers in front of the mixer will
degrade the intermodulation properties.
Adding an external LNA will block the transmitter path to the antenna.
The noise figure is the fundamental parameter to specify the sensitivity
of a receiver. All the other, more common,
sensitivity specifications can be derived from it. For example NF=3dB
will give an input sensitivity of
S=-174dBm+NF+48dB=-123dBm for typical NBFM signal. The factor "48dB"
somewhat depends on the test parameters.
_______________________________________________
HackRF-dev mailing list
https://pairlist9.pair.net/mailman/listinfo/hackrf-dev
_______________________________________________
HackRF-dev mailing list
https://pairlist9.pair.net/mailman/listinfo/hackrf-dev
Michael Ossmann
2015-01-23 22:24:27 UTC
Permalink
I apologize for the lack of response. I will get a replacement to you
right away. Please reply off-list to confirm your shipping address.

Mike
Post by i***@gmail.com
Mike,
When are you going to respond to those of us who received defective
units? I have sent you a number of emails, and posts here as well, and
you have yet to respond. What's the deal? Please respond with RMA
instructions.
Post by Michael Ossmann
Could you please share the details of your noise figure test method?
Thanks,
Mike
Post by JMG
Sensitivity measurements
Test results for the noise figure NF of the HackRF, obtained at a
NF = 22dB LNA=OFF
NF = 10.8dB LNA=ON
NF = 1.1dB LNA=ON, additional external LNA "LNA4ALL"
RTL-SDR, tuner 820T or 820T2
NF = 3.7dB
The HackRF really is extremely "deaf" when the internal preamplifier is
turned off. Turning it on and installing
a good external amplifier brings NF down to a very good value.
Applying so much gain at the preamplifiers in front of the mixer will
degrade the intermodulation properties.
Adding an external LNA will block the transmitter path to the antenna.
The noise figure is the fundamental parameter to specify the sensitivity
of a receiver. All the other, more common,
sensitivity specifications can be derived from it. For example NF=3dB
will give an input sensitivity of
S=-174dBm+NF+48dB=-123dBm for typical NBFM signal. The factor "48dB"
somewhat depends on the test parameters.
_______________________________________________
HackRF-dev mailing list
https://pairlist9.pair.net/mailman/listinfo/hackrf-dev
_______________________________________________
HackRF-dev mailing list
https://pairlist9.pair.net/mailman/listinfo/hackrf-dev
Stephen
2015-01-24 02:10:09 UTC
Permalink
Hi Mike,
I have also contacted you regarding the performance of my HackRF.
My former work (some time ago) was test/repair of radio equipment & I still
have access to RF test equipment.
I have performed RX sensitivity tests on the HackRF extensively on its own,
as well as doing performance comparisons with an RTL-SDR 820T and attempts
to receive NOAA passes.

The HackRF's RF performance has consistently been inferior to the RTL-SDR
820T - irrespective of configuration / software used.

Can I send my HackRF to you for your assessment?
Regards,
Stephen





-----Original Message-----
From: HackRF-dev [mailto:hackrf-dev-***@greatscottgadgets.com] On Behalf
Of Michael Ossmann
Sent: Saturday, 24 January 2015 9:24 AM
To: ***@gmail.com
Cc: hackrf-***@greatscottgadgets.com
Subject: Re: [Hackrf-dev] Sensitivity

I apologize for the lack of response. I will get a replacement to you right
away. Please reply off-list to confirm your shipping address.

Mike
Post by i***@gmail.com
Mike,
When are you going to respond to those of us who received defective
units? I have sent you a number of emails, and posts here as well,
and you have yet to respond. What's the deal? Please respond with
RMA instructions.
Post by Michael Ossmann
Could you please share the details of your noise figure test method?
Thanks,
Mike
Post by JMG
Sensitivity measurements
Test results for the noise figure NF of the HackRF, obtained at a
NF = 22dB LNA=OFF
NF = 10.8dB LNA=ON
NF = 1.1dB LNA=ON, additional external LNA "LNA4ALL"
RTL-SDR, tuner 820T or 820T2
NF = 3.7dB
The HackRF really is extremely "deaf" when the internal
preamplifier is turned off. Turning it on and installing a good
external amplifier brings NF down to a very good value.
Applying so much gain at the preamplifiers in front of the mixer
will degrade the intermodulation properties.
Adding an external LNA will block the transmitter path to the antenna.
The noise figure is the fundamental parameter to specify the
sensitivity of a receiver. All the other, more common, sensitivity
specifications can be derived from it. For example NF=3dB will give
an input sensitivity of S=-174dBm+NF+48dB=-123dBm for typical NBFM
signal. The factor "48dB"
somewhat depends on the test parameters.
_______________________________________________
HackRF-dev mailing list
https://pairlist9.pair.net/mailman/listinfo/hackrf-dev
_______________________________________________
HackRF-dev mailing list
https://pairlist9.pair.net/mailman/listinfo/hackrf-dev
Paul Jones
2015-01-24 07:03:07 UTC
Permalink
Is there any easy way to check for defective units? I bought a pair of units from the kickstarter which I suspect might have the same problem. Using every and any antenna I've got I can barely pick up fm radio. I live in the suburbs of Sydney so there are plenty of stations around, but the only one I get any sort of reception on is 96.1, which is the strongest. (I'll take a screenshot when I get my laptop, just to check the gain settings are right). I've also tried scanning GSM frequencies but can't see anything unless I make a call on my phone. Wifi I can just make out above the noise.
Is there anything to look for on the PCB that could be reflowed? I've got all the stuff to do everything except BGA.

Paul.
Post by Stephen
-----Original Message-----
Behalf Of Michael Ossmann
Sent: Saturday, 24 January 2015 9:24 AM
Subject: Re: [Hackrf-dev] Sensitivity
I apologize for the lack of response. I will get a replacement to you right
away. Please reply off-list to confirm your shipping address.
Mike
Post by i***@gmail.com
Mike,
When are you going to respond to those of us who received defective
units? I have sent you a number of emails, and posts here as well,
and you have yet to respond. What's the deal? Please respond with
RMA instructions.
Post by Michael Ossmann
Could you please share the details of your noise figure test method?
Thanks,
Mike
Post by JMG
Sensitivity measurements
Test results for the noise figure NF of the HackRF, obtained at a
NF = 22dB LNA=OFF
NF = 10.8dB LNA=ON
NF = 1.1dB LNA=ON, additional external LNA "LNA4ALL"
RTL-SDR, tuner 820T or 820T2
NF = 3.7dB
The HackRF really is extremely "deaf" when the internal
preamplifier is turned off. Turning it on and installing a good
external amplifier brings NF down to a very good value.
Applying so much gain at the preamplifiers in front of the mixer
will degrade the intermodulation properties.
Adding an external LNA will block the transmitter path to the antenna.
The noise figure is the fundamental parameter to specify the
sensitivity of a receiver. All the other, more common, sensitivity
specifications can be derived from it. For example NF=3dB will give
an input sensitivity of S=-174dBm+NF+48dB=-123dBm for typical NBFM
signal. The factor "48dB"
somewhat depends on the test parameters.
_______________________________________________
HackRF-dev mailing list
https://pairlist9.pair.net/mailman/listinfo/hackrf-dev
_______________________________________________
HackRF-dev mailing list
https://pairlist9.pair.net/mailman/listinfo/hackrf-dev
_______________________________________________
HackRF-dev mailing list
https://pairlist9.pair.net/mailman/listinfo/hackrf-dev
Mike Carden
2015-01-24 09:16:19 UTC
Permalink
G'day Paul.

Can you share with us the OS and SDR software that you are using with your
hardware?

Ta,
MC
Paul Jones
2015-01-29 08:34:01 UTC
Permalink
I originally noticed the problem when going through the SDR lessons (Great job there BTW !)
I’ve since upgraded the firmware to the latest and installed gnuradio using pybombs.
I’ve captured two screenshots from SDR# and Gqrx. I’ve adjusted the gains to get the best signal-noise ratio that I can.
I’ve moved house in the last few weeks to a much more “quiet” area with better signal strength so now I can at least tune other stations besides 96.1 and get something more than static. I think one of my previous neighbours had a switching power supply that was defective, and flooding the spectrum with harmonics.

I’m not sure if attatchments are allowed here, so here are some screenshots on flickr
https://www.flickr.com/photos/***@N00/15773951373/in/photostream/
https://www.flickr.com/photos/***@N00/16367983856/in/photostream/

Thanks,
Paul.


From: Mike Carden [mailto:***@gmail.com]
Sent: Saturday, 24 January 2015 8:16 PM
To: Paul Jones
Cc: Michael Ossmann; hackrf-***@greatscottgadgets.com
Subject: Re: [Hackrf-dev] Sensitivity

G'day Paul.

Can you share with us the OS and SDR software that you are using with your hardware?

Ta,
MC

Loading...