LINRAD VIDEOS

LINRAD VIDEOS
(June 11 2024)

This is my video channel on YouTube The videos you can find there are listed according to cathegories below. The most recent videos may be unlisted.

Installation on different platforms.
Setup and calibration.
Setup and tests at large bandwidths.
Getting the best CPU performance from your system.
Interference fighting.
Tests, comparisons and modifications of different SDR hardware.
The transmit side of Linrad.
Use Linrad as a test instrument.
How to use Linrad processing functions.
Artifacts in Linrad.
Other videos (non-Linrad.
Sideband noise for nerds only (non-Linrad.

Installation on different platforms.

Use the linrad installers. |||| [linrad-installers-windows7] For linrad-04.00 and later.
Windows XP. |||| [linrad-install-windows-xp] How to install the Linrad executable.
Linux X11. |||| [linrad-install-linux-x11] How to install Linrad from source using the latest (unstable?) version from the Linrad repository.
Linux in terminal mode with svgalib. |||| [linrad-install-svgalib] How to install Linrad from source using a released version.
Linux in terminal mode with fbdev. |||| [flinrad] How to install Linrad from source using the latest (unstable?) version from the Linrad repository.
Windows XP |||| [linrad-install-win-repo] How to install Linrad from source using the latest (unstable?) version from the Linrad repository.
Linux Fedora 20 |||| [linrad-libs-fedora20] How to install libraries for different hardware.
Linux Ubuntu 10.04 |||| [linrad-install-x11oss] How to install 4Front OSS. An alternative sound system.
Linux Suse 13.1 |||| [linrad-install-suse] How to install Linrad.
The Afedri USB interface |||| [afedri-usb-interface] How to use the USB interfac of Afedri SDR.
Zadig [xp64libusb] About using Zadig to install USB drivers under Windows.
Using the Airspy. |||| [airspyUSB] About the Airspy. 10 MHz is very fast on USB 2...
Gain modes |||| [GainModes] Gain modes in Linrad explained.
Libraries under Linux. |||| [install32on64] This video shows how to copy 32 bit Linux libraries to 64 bit Linux installations.
Linrad on a Raspberry Pi. |||| [RaspberryPi] Running an rtlsdr at a bandwidth of 2.2 MHz. It matters which screen is selected. Plain X11 is slow, MIT-SHM is better and FBDEV, the front buffer device is fast.
Raspberry Pi2 with Perseus. [perseus-raspi] Install from scratch. First the operating system, then Linrad with the Perseus. En excellent system for battery powered operation.
Install Linrad04-13 on Fedora 24. [install04 13] Install from the repo.

Setup and calibration.

Softrock with a function generator. |||| [softrock-setup-and-cal] Installation and setup in Linux.
Softrock with a proper pulse generator. |||| [softrock-cal2] Description of a proper pulse generator. Use it to evaluate performance of the calibration using the function generator. Calibrate with the proper pulse generator.
IQ+ with a proper pulse generator. |||| [linrad-linux-install-cal-iqplus] Installation, setup and calibration under Linux.
IQ+ with a multi-pulse generator. |||| [linrad-windows-install-cal-iqplus] Installation, setup and calibration under Windows XP.
The WSE converters. |||| [win calibrate wse] Installation, setup and calibration under Windows XP.
The rtlsdr USB dongle. |||| [win-rtl-install-cal] Installation, setup and calibration under Windows XP.
SDR-IP. |||| [win-sdrip-setup-cal] Installation, setup and calibration under Windows 7.
Maya 44. |||| [fourchan01cal] Here the problem with the Maya 44 soundcard is demonstrated. Linrad versions up to 04-07 can not be calibrated.
Maya 44 is here compared to a Delta 44. |||| [Delta44vsMaya44] An IQ+ is used to feed both the Delta44 and the Maya 44 so they can be compared directly.
Peculiarities with the Maya44. |||| [Delta44vsMaya44blanker] With linrad-04.08 and later the Maya 44 can be properly calibrated, but there is some ambiguity in the system. Sometimes the hardware starts with a different time delay between channels.
IQ+ vs WSE |||| [iqpluswse] Here an IQ+ is compared to a wse converter chain and calibration is demonstrated for both.
Calibrating the rtlsdr. |||| [rtluwv] Here an rtlsdr dongle is calibrated when run at 2 MHz on a fairly slow computer, an Intel Core 2.
Calibrating time delay between channels. |||| [phasecal] Here Linrad is calibrated to absorb the phase angle variation with frequency that is caused by a time delay between channels.

Setup and tests at large bandwidths. (Relative to the CPU used.)

bladeRF setup. |||| [linux bladerf install] Installation from the repo under Debian wheezy. In February 2014 10 MHz bandwidth was possible which is demonstrated by the reception of cellular phones on 950 MHz.
bladeRF setup wide. |||| [bladerf fm] Installation from the repository and setup under Linux, Fedora 17. This video shows the reception of the entire FM band 88-108 MHz and recording on the hard disk at a bandwidth of 25 MHz.
bladeRF 460-465 MHz. |||| [bladerf-460to465mhz] Setup under Windows 7 to receive 460 to 465 MHz with a 5 MHz filter and 8 MHz sampling rate. There are image spurs and overtone spurs, but with appropriate oversampling there would be no alias spurs. Performance is limited by reciprocal mixing, the images can be removed by calibrating Linrad.
rtlsdr wideband FM on Pentium4. |||| [rtlsdr-pentium4] The WFM mode in Linrad is not optimized for CPU efficiency, but nevertheless it is good enough to be useful on a Pentium 4 computer.
Using rtl-sdr on a slow computer with flinrad. |||| [pentium3 flinrad] This video shows performance on a 650 MHz Pentium 3. Today that is an obsolete computer, but there are low end low cost ARM computers with similar or better performance.
Many USB dongles. |||| [multi-rtl] This video shows a Compaq 6510b running with 4 rtlsdr USB dongles at 2 MHz. It also shows a desktop computer running 8 rtlsdr dongles at 2.4 MHz plus a PCIe-9842 at 200 MHz.
Frequency calibration for recorded files. |||| [recording-cal] This video shows how to set parameters to correct frequency errors in recorded files.

Getting the best CPU performance from your system.

Effects of setting CPU affinity. |||| [cpu-affinity] This video demonstrates that the improvement is small and explains why the routine for setting CPU affinities for heavily loaded threads is disabled in Linrad.
100 MHz bandwidth on different operating systems. |||| [pcie9842-200mhz] Linrad is run with a PCIe9842 under several operating systems. It is quite clear that performance differs significantly between the operating systems.
Hyperthreading. |||| [hyperthreading] It is not always a good idea to use hyperthreading to get twice as many CPU cores that sometimes might have 50% of normal CPU power.
OS differences. |||| [os-differences] Operating systems differ under high CPU load.
Win 7 peculiarities. |||| [Win7peculiarity] Reducing CPU load may sometimes cause problems because the operating system decides it is OK to reduce the clock frequency for some CPU core(s) while that is not acceptable. (Or maybe there is some other problem.)
High speed recording and playback. |||| [SDR record and playback at 25 and 40 MHz bandwidth] A BladeRF with a B200 transverter is used to record and play back the entire FM band.
Pentium IV with 2.5 MHz bandwidth. |||| [Airspy-One-on-Pentium-IV] Here SDRsharp and Linrad are used with an Airspy One. The Pentium IV is a very slow computer for processing a bandwidth of 2.5 MHz.
Linrad on Windows 10. |||| [win10] Windows 10 will never forget any mistake. (as far as I know.)
Airspy timing. |||| [airspy-timing] This video shows the time delay from antenna to loudspeaker.
Using a GPU fast FFT 1. |||| [gputest] Testing the graphics adapter in Linrad using OpenCL and clFFT. Performance is far below expectations and it seems like clFFT is using the motherboard CPU(s) rather than the NVIDIA card(s)
Using a GPU for fast FFT 2. |||| [clfft] In this video the performance of a slow and a fast NVIDIA graphics adapter is compared.
Using a graphics adapter for fast FFT 3. |||| [gt610] The GEFORCE GT 610 is tested for speed of Fourier transforms on Intel systems with Linrad using clFFT and OpenCL. Results are discouraging.

Interference fighting.

Sensors Part 1. |||| [sensors-part1] The first video in a series of four. It is about elimination of conducted interference on HF.
Sensors Part 2. |||| [sensors-part2] The second video in a series of four. It is about elimination of conducted interference on HF.
Sensors Part 3. |||| [sensors-part3] The third video in a series of four. It is about elimination of conducted interference on HF.
Sensors Part 4. |||| [sensors-part4] The last video in a series of four. It is about elimination of conducted interference on HF.
PLT interference??? |||| [plt1] This video shows impulse noise removal in Linrad. It is a challenge to DX listeners and radio amateurs to provide wideband recordings with worse problems that could be used to make a better video than this one.
Noise blanking on MW. Part 1. |||| [MWpulses] Linrad and Perseus.
Noise blanking on MW. Part 2. |||| [perseus night] perseus.exe v5 at night.
Noise blanking on MW. Part 3. |||| [hdsdr night] HDSDR at night.
Noise blanking on MW. Part 4. |||| [linrad night] linrad at night.
Noise blanking on MW. Part 5. |||| [perseus day] perseus v5 in daytime.
Noise blanking on MW. Part 6. |||| [hdsdr day] HDSDR in daytime.
Noise blanking on MW. Part 7. |||| [linrad day] linrad in daytime.
Noise blanking on MW. Part 8. |||| [linrad day net] linrad in daytime. Experiments with a narrowband blanker in a second instance over the network.
Interference fighting on 144 MHz. |||| [qrm144] Noise blanker and spur removal.
Lawnmover robot. |||| [lawnmower] A strong interference source. Can be removed both in the time domain with blankers and in the frequency domain with notches.
saq |||| [saq] Pulse interference that can be removed both both in the time domain with blankers and in the frequency domain with notches.
Adaptive phasing. |||| adapt.mp4"> [adapt] The automatic phasing in linrad maximises the signal within the selected passband for a two channel receiver. In cases when the dominating signal is an interference the orthogonal channel would be interference-free. A second instance of Linrad can pick it up.
Strong powerline noise. |||| [blanker] The Linrad noise blanker fightng strong powerline noise on 144 MHz.
FM splatter killing 1. |||| [fmsplatter] Here the processing in Linrad to reduce/remove FM splatter is demonstrated.
FM splatter killing 2. |||| [fmdx] Installation and setup of Linrad under MS windows for splatter reduction of wideband FM signals on 88-108 MHz.
FM capture rate. |||| [fmcapture] When two FM stations are present on the same frequency, only the stronger one can be heard and only if the signals are very near in signal level one can hear interference. The effect depends strongly on the bandwidth.
Notch filters in Linrad |||| [pst] People interested in NDBs (Non Directional Beacons) often have problems with carriers close to the signal of interest. Linrad iffers several ways ti mitigate problems.

Tests, comparisons and modifications of different SDR hardware.

Several SDRs compared on 88-108 MHz. |||| [many-fm] ELAD FDM-S1, Mirics MSi3101 (MSi001+MSi2500), rtlsdr with R820T as well as with E4000 in original Osmocom mode as well as in Linrad sensitivity mode. FunCube Pro plus and PCIe-9842 with a converter. The general conclusion is that the cheap dongles are reasonably good and that it will be a good idea to have some different types. In situations where one dongle fails another might be OK. None of them is perfect...
PCIe-9842. |||| [pcie9842] Here the high dynamic range of a PCIe-9842 is demonstrated. The card samples 14 bit at 200 MHz.
BladeRF direct A/D input on. |||| [bladerf j61] A bladeRF can sample two channels with 12 bits at 40 MHz. This video compares it to a PCIe-9842 that samples 14 bits at 200 MHz corresponding to about 15.2 bits at 40 MHz. The bladeRF is used on the J61 connector directly into the A/D converters. A simple check of saturation vs noise floor shows that the bladeRF is very good, but real life tests show something else. There is a huge problem with digital feedback.
rtlsdr dongles compared with SDR#. |||| [donglecmp] USE HD QUALITY (720p) and "Full screen." Linrad is pixel oriented!
rtlsdr with SDR# and Linrad. |||| [donglecmp2] USE HD QUALITY (720p) and "Full screen." Linrad is pixel oriented! The tuner chips E4000, R820T and FC0013 differ a little in sensitivity and significantly in spurious responses. The linearity mode in Linrad shows some advantage over the original Osmocom mode due to a different gain distribution.
Work bench for comparisons of USB dongles |||| [MVI0017] Setup for the blocking and IM3 comparison below.
Compare dongles |||| [donglecmp3] Dynamic range of rtlsdr dongles compared with Linrad. The above setup was used. USE HD QUALITY (720p) and "Full screen." Linrad is pixel oriented!
Sensitivity mode |||| [e4000 sensitivity mode] The sensitivity mode for the E4000 tuner. USE HD QUALITY (720p) and "Full screen." Linrad is pixel oriented! This mode requires an external filter, but it provides better performance in the close range by having a lower noise floor.
Linearity of AD converters. ||||^ [ADCintermod] Here two units that sample at a high speed are compared. The Perseus SDR and the Adlink PCIe-9842 digitizer. Performance differs significantly!!
ANAN-100D |||| [anan100d-dac] Sideband noise measured directly on the DAC when transmitting with the ANAN 100 D.
Compare tuners |||| [SDRcompareTunersWide] SDRs that use tuners, rtlsdr, BladeRF, Airspy, SDRplay and others are not designed for a very large dynamic range within the range of their IF filters. As a consequence these SDRs have a mediocre performance as wideband receivers where we have both strong and weak signals within the passband.
Compare tuners. |||| [SDRcompareTunersNarrow] Here tuners are compared for usage when the interfering signal is within the visible passband. For example 144 MHz with a suitable filter between the antenna and the tuner.
Compare tuners. |||| [SDR on 88-108 FM part1] Compare tuners for usage on the FM band. Part 1, blocking and reciprocal mixing.
Compare tuners. |||| [SDR on 88-108 FM part2] Compare tuners for usage on the FM band. Part 2, intermodulation at low signal levels.
Compare tuners. |||| [SDR on 88-108 FM part3] Compare tuners for usage on the FM band. Part 3, intermodulation at normal signal levels.
The IF filter in R820T. |||| [R820T with IF filtering] Here it is demonstrated that setting a narrow IF filter can improve the performance of rtlsdr dongles as FM receivers by 30 dB and more.
Softrock Ensemble RXTX. |||| [softrock] Here it is demonstrated that the input transformer does not isolate for RF voltages on the screen of the antenna cable. A capacitor will fix the problem.
The Softrock Ensemble RX. |||| [softrockRX] Optimizing the receiver hardware and eliminating ground loops.
Tests on the SunSDR2. |||| [sunsdr2] Receiver dynamic range and transmitter sideband noise.
Tests on the SDRplay. |||| [sdrplay] The unit suffers from USB noise.
A modification on the SDRplay. |||| [sdrplay2] One way to remove the USB noise.
A modification on the Airspy mini. |||| [airspymini] One way to remove the USB noise.
The two-tone test. |||| [im3dr] Why the two-tone test is not the appropriate figure of merit when ranking receiver after dynamic range performance. Part 1.
The two-tone test. |||| [antenna-im3dr] Why the two-tone test is not the appropriate figure of merit when ranking receiver after dynamic range performance. Part 2.
Compare Airspy/Spyverter and SdrPlay. |||| [airspy-sdrplay]
rtlsdr-com vs Airspy/Spyverter |||| [rtlsdr-com-airspy] The 8-bit rtlsdr.com compared with the 12 bit Airspy/Spyverter on HF.
A FT1000 is compared with a Perseus. |||| [ft1000perseus] Testing on a real antenna on 14 MHz.
DX PATROL: [DXpatrol] The unit tested suffers badly from USB noise.
Compare HF receivers part 1. |||| [rx1compare] Perseus, Airspy+Spyverter, BladeRF+B200, BladeRF direct ADC input.
Compare HF receivers part 2. |||| [rx2compare] Perseus, SDR-IP, Excalibur, Colibri and SDR-14.
Compare HF receivers part 3. |||| [airspy-sdrplay2] The Airspy plus Spyverter combination compared to SDRPlay.
Compare HF receivers part 4. |||| [rx3compare] RTL-SDR.com, Funcube Pro plus, SDRplay, Airspy+Spyverter, Afedri NET and FDM-S1.
Modify the Softrock Ensemble. |||| [rxdynrange_softrock] Better dynamic range with a better soundcard requires modifications.
Testing soundcards on Softrock. |||| [softrock_soundcards1] M-Audio:Delta 44, IDT(Sigmatel):D5400XS motherboard,SIIG:CE-SA0011 and StarTech:Virtual 7.1 USB.
Testing soundcards on Softrock. |||| [softrock_soundcards2] Xonar:Essence STX and ESI:Maya 44 Xte.
Testing soundcards on Softrock. |||| [AudigySE] Creative: Audigy SE.
Testing soundcards on Softrock. |||| [softrock_soundcards3] Lynx:Lynx Two(mod), Terratec: DMX 6fire PCI and M-Audio:Revolution 5.1.
Testing soundcards on Softrock. |||| [softrock_soundcards4] Terratec:DMX 6fire USB, AKAI:EIE pro, Creative:SB Live!, Creative:SB Live! 5.1 Digital and Komplete: Audio 6.
Testing soundcards on Softrock. |||| [softrock_soundcards5] M-Audio:M-Track Quad, Steinberg:UR12 and Creative:SB Live! External USB.
Testing soundcards on Softrock. |||| [softrock_soundcards6] Terratec:DMX 6fire USB, Lynx:Lynx Two (mod), M-Audio:M-Track Quad and Komplete:Audio 6.
Testing soundcards on Softrock. |||| [softrock_soundcards7] Terratec:DMX 6fire USB, Lynx:Lynx Two (mod), M-Audio:M-Track Quad and Komplete:Audio 6.
Testing soundcards on Softrock. |||| [softrock_soundcards8] M-Audio:Delta 44 (mod), Creative: Audigy SE, Realtek: X9DAI motherboard, IDT (Sigmatel):D5400XS motherboard and SIIG:CE-SA0011.
Testing soundcards on Softrock. |||| [softrock_soundcards9] Xonar:Essence STX, Creative:SB Live!, Creative:SB Live! 5.1 Digital and Creative:SB Live! External USB.
Problems with Maya 44. |||| [Maya44] Maya 44 has a problem in that there is a time shift between the two stereo channels. This causes a problem when this soundcard is used with an IQ+ direct conversion SDR.
The Anglian 3 transverter 1. |||| [anglian_rx_reciprocal] Reciprocal mixing on the receive side.
The Anglian 3 transverter 2. |||| [anglian tx] The transmit side. AM sideband noise noise is detected.
The Anglian 3 transverter 3. |||| [anglian-rx-mod] Here it becomes clear that the SPF5043 is a big problem.
The Anglian 3 transverter 4. |||| [PSA4vsSPF] Here the PSA4-5043 is compared to the SPF-5043 for AM sideband noise.
The Anglian 3 transverter 5. |||| [AM_sideband_noise] The AM sideband noise of many 144 MHz amplifiers is evaluated.
The Anglian 3 transverter 6. . [reduceAMnoise] Input decoupling, or better AF feedback reduces the AM noise modulation from a SPF-5043.
The Anglian 3 transverter 7. |||| [anglian3_ip3] Evaluation of third order intermodulation on the receive side.
The Anglian 3 transverter 8. |||| [testing-MMIC] Testing different MMICs for AM noise with different AF decoupling on the input pin.
The Anglian 3 transverter 9. |||| [adl5535] Testing the adl5535 for AM noise modulation.
The Anglian 3 transverter 10. |||| [anglian LO] Some experiments on the LO of the Anglian 3 transverter.
The Anglian 3 transverter 11. |||| [anglian116mhz] Tests on the 116 MHz LO in the Anglian 3 transverter.
Dynamic range on 7 MHz. 1. |||| [compare7mhz] These receivers: FT1000, IC706MKIIG, Colibri, Afedri NET and Softrock Ensemble (modified) are compared.
Dynamic range on 7 MHz. 2. |||| [rx4compare] The tested receivers are SDRplay, FUNcube Pro +, Airspy with Spyverter, Airspy HF+, Afedri SDR-Net, ELAD FDM-S1, ICOM IC-706MKIIG and Microtelecom Perseus.
Dynamic range on 7 MHz. 3. |||| [rx5compare] More comparisons between receivers on 7 MHz. SDRplay RSP1, FUNcube Pro+, Airspy with Spyverter, Airspy HF+, Afedri SDR-Net, ELAD FDM-S1, ICOM IC-706MKIIG and Microtelecom Perseus.
Dynamic range on 7 MHz. 4. |||| [rx7compare-part1] More comparisons between receivers on 7 MHz. Perseus, Afedri NET, SDRPlay RSP1, Airspy+Spyverter, Airspy HF+ Dual and Airspy HF+ Discovery.
Dynamic range on 7 MHz. 5. |||| [rx7compare-part2] More comparisons between receivers on 7 MHz. Perseus, Afedri NET, SDRPlay RSP1, Airspy+Spyverter, Airspy HF+ Dual and Airspy HF+ Discovery.
Dynamic range on 7 MHz. 6. |||| [rx7compare-part3] More comparisons between receivers on 7 MHz. Perseus, Afedri NET, SDRPlay RSP1, Airspy+Spyverter, Airspy HF+ Dual and Airspy HF+ Discovery.
Dynamic range on 14 MHz. |||| [rx7compare-part4] Comparisons between receivers on 14 MHz. Perseus, Afedri NET, SDRPlay RSP1, Airspy+Spyverter, Airspy HF+ Dual and Airspy HF+ Discovery.
Second order intermodulation. |||| [rx7compare-part5] Comparisons between receivers: Perseus, Afedri NET, SDRPlay RSP1, Airspy+Spyverter, Airspy HF+ Dual and Airspy HF+ Discovery.
Performance on 7 MHz |||| [rx6compare] Several SDRs are compared with the same signals on the antenna connector: Perseus, Afedri NET, SDRPlay RSP1, Airspy+Spyverter, Airspy HF+ Dual and Airspy HF+ Discovery.
Colibri from EXPERT Electronics. |||| [colibri_noise] How to cure a noise problem.
Colibri and Airspy compared on 144 MHz. |||| [colibri144airspy] A ColibriDDC and an Airspy are compared on 144 MHz.
Softrock Ensemble. The LO-3*fmod spur 1. |||| [softrock_spur] Showing the mechanism behind the spur.
Softrock Ensemble. The LO-3*fmod spur 2. |||| [softrock_mixer] About the LO - 3*Fmod spur in the Softrock Ensemble.
Sound Blaster X-Fi Surround 5.1 Pro |||| [xfi51pro] Testing with a Softrock Ensemble.
Airspy HF+ vs Perseus. |||| [hfplus vs perseus] The Airspy HF+ is compared with a Perseus SDR on 14 MHz.
Airspy on 144 MHz. |||| [environment] Out-of-band signals may degrade perfpormance. Here the 88 -108 MHz band in my rural location is demonstrated.
Airspy on 144 MHz. |||| [airspies] An Airspy One is compared with an Airspy HF+ on 144 MHz.
Filter on 7 MHz. |||| [hfplusfilter] An Airspy HF+ is compared to an Airspy One with a Spyverter 7 MHz with real antenna signals. Inserting a simple band pass filter for 7 MHz makes a big improvement for the HF+.
UDAC4 tested with WSE. |||| [udac4wse] A prototype of the UDAC4 is tested with the WSE converters. The UDAC4 is designed for use with direct conversion software defined radios.
UDAC4 tested with Softrock. |||| [udac4wse] A prototype of the UDAC4 is tested with Softrock Ensemble.
Ground loops in direct conversion SDR. |||| [groundloop] A prototype of the UDAC4 is compared to a DMX6Fire USB for sensitivity to ground loops and magnetic fields.
HF+ unboxed. |||| [hfplusunboxed] The Airspy HF+ board is tested for common mode interference outside its box at 100 kHz.
HF+ preselector. |||| [hfpluspreselector] A pre-production preselector for the Airspy HF+ is tested. It has not been put in production because of inadequate dynamic range in the switches that are used to select which filter to use.
The S-meter in SDR Console. |||| [consolemeter] Old analog receivers as well as most SDR software provide an S-meter that is essentially a power meret. When the bandwqidth is changed the reading of the noise floor changes. The SDR Radio Console is different.
The HF+ preselector. Part 1. |||| [hfpluspresel1] The Airspy HF+ is still under development. This video compares old and new versions of the HF+. More to follow.
The HF+ preselector. Part 2. |||| [hfpluspresel2] The Airspy HF+ is still under development. This video compares old and new versions of the HF+. More to follow.
The HF+ preselector. Part 3. |||| [hfpluspresel3] The Airspy HF+ is still under development. This video compares old and new versions of the HF+. More to follow.
RSP1 software. |||| [rsp1] Sdrplay RSP1 with different software: SDRuno and Linrad.
Airspy HF+: Dual, Mini and Discovery. |||| [airspy-hw] Airspy Mini, Airspy HF+ Dual Port and Airspy HF+ Discovery are compared on the FM band.

The transmit side of Linrad.

Tx setup. |||| linradtx1.mp4"> [linradtx1] The principles of the speech processor.
Speech processing. |||| SpeechProcessor.mp4"> [SpeechProcessor] Setup in SSB mode: Speech processing on voice signals.
Speech processing. |||| SpeechProcessorIC706.mp4"> [SpeechProcessorIC706] Here it is demonstrated how adding a speech processor in front of a conventional SSB transmitter can reduce splatter and increase the average transmitted power for a better readability at low signal levels.
Soundcard timing. |||| [4Front OSS vs ALSA] Linrad allows both OSS and ALSA. Here Ubuntu 15.04 is used to compare the sound systems. Provided that Pulseaudio is disabled ALSA and 4Front OSS are both good.
Softrock transceiver. |||| [Ensemble01] First test of Softrock Ensemble RXTX as a transceiver in QSK.
Softrock Tx. |||| [Ensemble02] The Softrock Tx in "medium power mode."
Softrock Tx. |||| [Ensemble03] The Softrock Tx in "QRP mode."
Softrock Rx. |||| [Ensemble04] Screening a Softrock against electric fields.
Softrock Tx. |||| [EnsembleTxSoundcards] The Softrock in Tx mode with different soundcards.
Softrock Tx. |||| [Ensemble05] The Softrock Tx in high power mode.
Softrock transceiver. |||| [FirstSoftrockQSO] My first QSO with Softrock and Linrad. Here the Softrock Ensemble RXTX is used without any modification. Keying on the paddle input and TR switch via USB.

Use Linrad as a test instrument.

S-meter calibration. |||| [amplitude-calibration] Here the S-meter is calibrated with a rtlsdr dongle using the information that the typical noise figure for a dongle with the E4000 tuner is 8 dB.
Blocking dynamic range (BDR). |||| [pcie9842-dynrange]"> A single very strong signal will degrade performance due to an increased noise floor on other frequencies. This video show how to measure BDR for a Linrad system.
How to measure signal to noise ratio. |||| [measure-signal-to-noise] This video shows how to measure (S+N)/N on a weak signal by use of Linrad.
Reciprocal mixing. |||| [perseus jitter] Here Linrad is used to measure reciprocal mixing in a Perseus receiver.
IM2. |||| [second_order_im] How to measure second order intermodulation. PCIe-9842, Perseus, SDR-IP, Excalibur, WSE and FDM-S1 are compared.
Correlation spectra. |||| [corrspectra] Two SDR-IP locked to the same 10 MHz reference allow measurement of very low sideband noise levels without the need of any notch filters. Part 1.
Correlation spectra. |||| [corrspectra2] Two SDR-IP locked to the same 10 MHz reference allow measurement of very low sideband noise levels without the need of any notch filters. Part 2.
Correlation spectra. |||| [AfedriCorrelation] The two channel Net Afedri can be used to produce correlation spectra. Such spectra show the combined noise sidebands of a test signal and the sampling cloc. A big capacitor can improve the sampling clock. Part 1.
Correlation spectra. |||| [netafedrinoise] The two channel Net Afedri can be used to produce correlation spectra. Measure sideband noise down to -150dBc/Hz. Part 2.
Precision noise measurements |||| [precnoise] An rtlsdr dongle can be used to measure noise floor levels fairly rapidly because of the large bandwidth.
Checking im3 in hybrids and other things. |||| [hybrids] When we measure distorsion it is important to know that the test system is significantly better than the test object. This video shows how to verify. Hybrids, attenuators, resistors, capacitors, connectors - almost anything can cause problems.
Sideband noise in oscillators |||| [sidebandnoise] Phase noise sidebands and amplitude noise sidebands can be the limiting factors for system performance. When created by the same proces, i.e. flicker noise current in a transistor amplitude and phase noise are likely correlated. Some common design errors in oscillators are demonstrated.
Sideband noise in oscillators. Part 2. |||| [sidebandnoise2] The performance of the Linrad phase noise analyzer is compared to measurements with a Rohde & Schwarz FSUP signal analyzer. Several oscillators are measured and some interesting observations are found.
Allan deviation. |||| [allan] Evaluation of Allan deviation is added in Linrad. This video shows an attempt to verfy the correctness of the Linrad code by comparisons with results from a HP5370A with TimeLab.
Allan deviation. |||| [allan2] Various measurements of Allan deviation and comparisons of results between TinyPFA with TimeLab, Linrad and Linrad with TimeLab. An error by a factor 1.4, the square root of two is found an corrected in Linrad.

How to use Linrad processing functions.

Recorded files under Linux X11 |||| [linrad-wav-x11] Download files from the internet and play in Linrad. This example shows a little about medium wave AM and how to use notches and synchronous AM detection.
Receiving spectral broadened CW |||| [eme10ghz] This video is about receiving Morse coded signals on 10 GHz when doppler spread is significant.
Use of recordings. |||| [long mw recording] An overnight recording of the entire medium wave spectrum is made here. It is demonstrated how Linrad can draw a waterfall from the entire recording in a fairly short time and how to select interesting time intervals to repeat over and over again when looking for specific things.
MAP65 with rtlsdr. |||| [map65 with linrad and rtlsdr] Here it is demonstrated how Linrad can use an rtlsdr dongle to sample at 1536 kHz, downsample to 96 kHz and send the signal for decoding in MAP65.
Linrad over the Internet. |||| [utwente] Here Linrad is used together with VAC to process signals received over the Internet.
Panadapter. |||| [ic706panadapter] Here an rtlsdr dongle is used as a panadapter on an IC706MKIIG.
Sensitive waterfall. |||| [senswaterf] Here Linrad is set up to produce waterfall graphs with high sensitivity for extremely weak signals.
Sensitive waterfall. |||| [p4senswaterf] Here it is demonstrated on a Pentium IV how to set parameters that allow extreme sensitivity with a low CPU load.
Coherent processing on non-directional beacons. |||| [ndb-1min] Dig 6 dB deeper into the noise. In a first stage coherent processing is used to improve S/N by 3 dB by using both sidebands of an A2 modulated NDB. In a second stage another 3 dB is gained by coherent processing of the 400 Hz modulation tone.
Coherent processing on non-directional beacons. |||| [ndb-vj] Dig 20 dB deeper into the noise! By use of everything we know about NDBs we can gain about 20 dB compared to listening on one sideband with an optimum CW filter. This video explains how to do it - but someone would have to write some relatively trivial software to use this in full.
Netafedri with two coherent RF channels. |||| [netafedri]
rtlsdr and the Linrad noise blanker. |||| [rtluwv]

Artifacts in Linrad.

IIR filter for output. |||| [linrad_iir_filter] There is an IIR filter to remove aliasing spurs after resampling the output. This video shows the effect of changing from 32 bit float to 64 bit float for the IIR filter.

Videos about sideband noise for nerds only. (non-Linrad).

High power notch filter. |||| [nerds1] Problems with high power notch filters.
High power notch filter. |||| [nerds3] Sideband noise at high power levels. Up to 0.5W.

Noise in quartz crystals. |||| [nerds2] At high power levels crystals create sideband noise.
Noise in quartz crystals 1. |||| [nerds36] The sideband noise in a crystal filter is found to be of the order f-1.6, proportional to the frequency offset to power -1.6 by interferometry.
Noise in quartz crystals 2. |||| [nerds46] The sideband noise in a crystal filter is measured by how much it degrades the performance of a professional ULNO. The result is the same that was found by interferometry in nerds36.
Noise in quartz crystals 3. |||| [nerds47] The sideband in individual crystals are measured by interferometry. Crystals are found to be VERY different. Some are good and some are very bad with respect to sideband noise.
Noise in quartz crystals 4. |||| [nerds48] Evaluating crystals for use in oscillators. Part 1.
Noise in quartz crystals 5. |||| [nerds49] Evaluating crystals for use in oscillators. Part 2.
Noise in quartz crystals 6. |||| [nerds60] Flicker noise in a crystal filter.

Noise in attenuators. |||| [nerds4] Sideband noise in attenuators and in a MOS-FET amplifier.
Noise in attenuators. |||| [nerds28] Sideband noise in an attenuator is shown to be AM noise.
Noise in attenuators. |||| [nerds29] The AM sideband noise of an attenuator is measured from 1 Hz to 50 kHz. Slope is 1/f.
Noise in attenuators. |||| [nerds34] The sideband noise of several attenuators is measured by interferometry.

Noise in amplifiers: AMC-147. |||| [nerds30] The noise sidebands of AMC-147 are measured by interferometry.
Noise in amplifiers: SPF-5043. |||| [nerds31] The noise sidebands of SPF-5043 are measured by interferometry.
Noise in amplifiers: BUK75150. |||| [nerds32] The noise sidebands of BUK75150 are measured by interferometry.
Noise in amplifiers: 4xJ310. |||| [nerds33] The noise sidebands of 4xJ310 are measured by interferometry.
Noise in amplifiers: UTC-565-1. |||| [nerds37] The noise sidebands of UTC-565-1 are measured by interferometry.
Noise in amplifiers: QBH-1297. |||| [nerds38] The noise sidebands of QBH-1297 are measured by interferometry.
Noise in amplifiers: 5xJ310. |||| [nerds39] The noise sidebands of 5xJ310 are measured by interferometry.

Noise in capacitors 1. |||| [nerds5all] Sideband noise in 10 pF capacitors.
Noise in capacitors 2. |||| [nerds6] Sideband noise in 47 pF capacitors.
Noise in capacitors 3. |||| [nerds7] Sideband noise in 1 nF capacitors.

Noise in resistors 1. |||| [nerds8] Sideband noise in 22 ohm resistors.
Noise in resistors 2. |||| [nerds9] Sideband noise in 50 ohm resistors.

Selective amplifiers 1. |||| [nerds10] Series resonance crystal filters with grounded base bipolar transistors.
Selective amplifiers 2. |||| [nerds11] Parallel resonance crystal filters with JFET source followers.
Selective amplifiers 3. (First design) |||| [nerds12] A high power selective amplifier with source and emitter followers.
Selective amplifiers 4. (First design) |||| [nerds13] Balancing the gate to source capacitance improves the noise performance of the input source follower.
Selective amplifiers 5. (First design) |||| [nerds14] The effects of a "bootstrap" capacitor on the source follower.
Selective amplifiers 6. (First design) |||| [nerds15] Sideband noise at 10 kHz offset.
Selective amplifiers 7. (Second design) |||| [nerds16] Here the design of the crystal filter is described.
Selective amplifiers 7. (Second design) |||| [nerds17] Here the design of the low noise input stage is described.
Selective amplifiers 8. (Second design) |||| [nerds18] Here the design of the high power output stage is described.
Selective amplifiers 9. (Second design) |||| [nerds19] This is about thermal noise in the output stage.
Selective amplifiers 10. (Both designs) |||| [nerds20] The oscillators "first design" and "second design" are mixed with schottky diodes in different ways and the resulting low frequency signal is analyzed for noise.
Selective amplifiers 11. (Both designs) |||| [nerds21] Level 23 Schottky diode mixers are used with triplexers to see the combined noise of the oscillators "first design" and "second design".
Selective amplifiers 12. (Both designs) |||| [nerds22] Level 23 Schottky diode mixers are used with triplexers to see the combined noise of the oscillators "first design" and "second design". Here various complications are demonstrated.
Selective amplifiers 13. (Both designs) |||| [nerds23] Here level 29 Schottky mixers are used to see the combined noise of the oscillators "first design" and "second design".
Selective amplifiers 14. (Both designs) |||| [nerds25] Oscillators are phase locked and noise is evaluated by interferometry.
Selective amplifiers 15. (Both designs) |||| [nerds26] Oscillators are phase locked and noise is evaluated by an improved interferometric setup.
Selective amplifiers 16. (Second design) |||| [nerds27] Two oscillators of the second design are phase locked and noise is evaluated by interferometry.
Selective amplifiers 17. |||| [nerds35] The sideband noise in my selective amplifiers is found to be of the order f-3, proportional to the frequency offset to power -3 by interferometry.
Selective amplifiers 18. (Third design) |||| [nerds50] Part 1. Uses selected crystals.
Selective amplifiers 19. (Third design) |||| [nerds52] Part 2. Evaluate and reduce flicker noise and flat noise floor.
Selective amplifiers 20. (Third design) |||| [nerds53] Part 3. More about the sideband noise.
Selective amplifiers 21. (Third design) |||| [nerds53] Part 4. More about the sideband noise.
Selective amplifiers 22. (Third design) |||| [nerds55] Part 5. The origin of the noise??
Selective amplifiers 23. (Third design) |||| [nerds56] Part 6. An event in crystal ageing.
Selective amplifiers 24. (Third design) |||| [nerds58] Finding shortcomings in the design.
Selective amplifiers 25. (Third design) |||| [nerds61] Close range AM nose suppressed.
Selective amplifiers 26. (Third design) |||| [nerds66] Noise at large separations evaluated with crystal notch filters.
Selective amplifiers 27. (Fourth design) |||| [nerds67] The properties of a 3rd overtone crystal for 10 MHz.
Selective amplifiers 28. (Fourth design) |||| [nerds68] A low noise amplifier for a 3rd overtone crystal at 10 MHz with J310 in a common source configuration.
Selective amplifiers 29. (Fourth design) |||| [nerds69] Investigations on the low noise amplifier with 3 parallel J310 in a common souirce configuration
Selective amplifiers 30. (Fourth design) |||| [nerds70] Investigations on the low noise amplifier with 3 parallel J310 in a common gate configuration and some tests on ferrite cores.
Selective amplifiers 31. (Fourth design) |||| [nerds71] Investigations on the low noise amplifier with 21 parallel J310 in a common gate configuration.
Selective amplifiers 32. (Fourth design) |||| [nerds72] Adding a second amplifier stage.
Selective amplifiers 33. (Fourth design) |||| [nerds73] The first amplifier stage is optimized for a 50 ohm load.
Selective amplifiers 34. (Fourth design) |||| [nerds74] The second amplifier stage is optimized to fit the first stage, a third stage is added and the unit is run as an oscillator.
Selective amplifiers 35. (Fifth design) |||| [nerds75] Experiments with bipolar transistors to use for the first amplifier stage.
Selective amplifiers 36. (Fifth design) |||| [nerds76] A 3rd overtone crystal with a single BF240 in common base configuration as the first amplifier stage.
Selective amplifiers 37. (Fifth design) |||| [nerds77] The second amplifier with 6 parallel BF240. Crystal noise seems to dominate at close range.
Selective amplifiers 38. (Fifth design) |||| [nerds78] The fifth design running as an oscillator.
Selective amplifiers 39. (Fifth design) |||| [nerds79] Close range phase noise seems to be caused by the crystal only and to increase with temperature.
Selective amplifiers 40. (Sixth design) |||| [nerds80] Two series connected crystal filters to make the effective Q of the amplifier higher.
Selective amplifiers 41. (Sixth design) |||| [nerds85] The sixth design. More investigations. Ferrite cores 4C65 add noise while iron powder Amidon material 6 does not - or at least significantly less.
Selective amplifiers 42. (Sixth design) |||| [nerds86] More investigations. The load impedance on the grounded gate FETs has to be kept fairly low.
Selective amplifiers 43. (Sixth design) |||| [nerds87] More investigations. Could not restore the initial good performance but found a sensitivity to infrasound.
Selective amplifiers 44. (Sixth design) |||| [nerds88] More investigations. Trying to eliminate sensitivity to infrasound but failed. The box must be kept closed to keep air pressure variations outside.
Selective amplifiers 45. (Sixth design) |||| [nerds89] More investigations. The advantage of using two crystals is small. 3 dB only below 2 Hz offset and no advantage above 5 Hz offset.
Selective amplifiers 46. (Seventh design) |||| [nerds91] A complete rebuild of the fourth design to improve the close range phase noise. FETs are replaced by bipolar transistors, but the real problem turned out to be the stiffness of the walls of the box made from PCB laminate.
Selective amplifiers 47. (Sixth design) |||| [nerds92] Trying to optimize, various problems investigated. In the end the worst problem was contact resistance noise in a trimmer capacitor.
Selective amplifiers 48. (Sixth design) |||| [nerds93] More problems investigated. Found a problem with the interferometer, a mistake as you can see in nerds94.
Selective amplifiers 49. (Sixth design) |||| [nerds95] More experiments with the oscillator but paused after I looked at the noise from a rubidium source.
Selective amplifiers 50. (Sixth design)|||| [nerds98] Teflon dielectric trimmer capacitors cause noise bursts.
Selective amplifiers 51. (Sixth design)|||| [nerds99] More power in the channel 2 reference oscillator has improved phase noise.
Selective amplifiers 52. (Sixth design).|||| [nerds100] Various problems. Performance good for a short time, then noise below 0.15 Hz offset increased again. Reference oscillators now locked to a LPRO-101 rubidium stabilised oscillator to make studies of stability in time scales above 5 seconds easier. PLL not yet optimised.

Allan deviation and phase noise. |||| [nerds96] Various problems detected and corrected. Cross talk between oscillators in a correlation setup, stability of PLL loops, agreement or not between Linrad and measurements with TimeLab and a HP5370A. Stability of GPS diciplined oscillators vs rubidium stabilized.
Allan deviation. |||| [nerds97] The TinyPFA is much better than the HP5370A for use with TimeLab for evaluation of Allan deviation. Here various experiments with Allan deviation are demonstrated.

Noise in oscillators 1. |||| [nerds45] The sideband noise of a Wenzel 501-04538 ULNO is measured in my setup. Agreement with earlier measurements with a Rohde & Schwarz FSUP is fairly good.
Noise in oscillators 2. |||| [nerds57] Phase noise and amplitude noise may differ significantly.
Noise in oscillators 3. |||| [nerds59] The Wenzel 501-25420 phase noise. At close range my setup has problems that are shown here.
Noise in hybrids. |||| [nerds65] Interferometric study of sideband noise in hybrids.

Schottky diode mixers. |||| [nerds24] Here I try to improve Schottky diode mixers - but fail.
Schottky diode frequency doublers. |||| [nerds40] Interferometric study of sideband noise in Schottky diode frequency doublers.
Schottky diode frequency doublers. |||| [nerds41] The origin of sideband noise in Schottky diode frequency doublers.

Active frequency doublers. |||| [nerds42] Sideband noise in active frequency doublers studied with notch filters.
Active frequency doublers. |||| [nerds43] The 20 MHz notch filter used in nerds42 is improved and the J310 doubler is re-measured.
Active frequency doublers. |||| [nerds44] The sideband noise of two active frequency doublers is measured in an interferometer.

Correlation spectra 1. |||| [nerds51] About the noise floor in correlation spectra.
Correlation spectra 2. |||| [nerds62] Correcting a 2 dB error present in the Linrad signal analyzer phase/amplitude sideband noise up until September 19 2020.
Correlation spectra 3. |||| [nerds63] A couple of problems and how to manage them.
Correlation spectra 4. |||| [nerds64] More problems and how to manage them.
Correlation spectra 5. |||| [nerds81] The 4-channel soundcard fails. Mostly about soundcard problems.
Correlation spectra 6. |||| [nerds82] Modifications for the SDR-IP to use two of them for sideband noise measurements.
Correlation spectra 7. |||| [nerds83] Measurements using M-Track Quad and SDR-IP Now that my UADC4 has failed.
Correlation spectra 8. |||| [nerds84] In sideband noise, phase noise and amplitude noise may be correlated. An amplitude limiter can reduce phase noise significantly in such cases.
Correlation and flicker noise. |||| [nerds94] The limits of sideband noise measurements. Improving the isolation between the 10 MHz oscillators does not help. Attenuators in the interferometer are checked for flicker noise.

Power supply noise |||| [nerds90] Noise on the power supply affects the phase noise of my oscillators. While investigating this I found why the signal path from the mixers to the soundcard has to be DC-coupled (or use very big capacitors.)

To SM 5 BSZ Main Page