- > AOR AR5000 - > Perseus

Running the AOR-AR5000 on a Perseus receiver

With the IF output of the AR5000 you can use all the features of the Perseus for higher frequencies. Just connect the IF out on the rear of the AR5000 via a BNC cable with the Perseus. The first surprise: The spectrum is upside down. So if you tune the AR5000 to 100 MHz and the Perseus to 10.7 MHz, then 101 MHz will apprear at 9.7 MHz

The IF spectrum is broader than one might expect. This is a typical spectrum.
AR5000 IF output spectrum

For exact dB measurements the AGC of the AR5000 should be off. The noise level is between -100 and -110 dB. A 0 dB signal at 10.7 MHz would be...

MHzloss (db)

The AR5000 filters do not affect the IF output.

By the way you can watch mirror frequencies. They move at double speed if you tune the AR5000 and disappear with the attenuator of the AR5000. This is from 93 MHz with ATT 0dB:
intermodulation spectrum

And this is with ATT 10 dB:
no intermodulations with 10dB ATT

Frequency calibration

is tricky because the AR5000 and the Perseus both differ a bit from exact values. The AR5000 frequency seems to be calibrated at 1 MHz. With every MHz the frequency offset rises proportionally by a few Hz (4.2 Hz per MHz with mine). It's easiest to check it with a UHF TV carrier. A 21 MHz SW frequency will do, too. Just put on USB, select Step 0.001 Hz and check a few AM signals.
The Perseus can be calibrated as described in the manual is not stable during the first hour of operation.

Another surprise: USB and LSB shift the frequency spectrum by +/- 1500 Hz.

The Spectrum Laboratory software

An alternative software for the Perseus is the Spectrum Laboratory (SpecLab) (also useful for many other purposes). Speclab with the Perseus is still very "beta".

Follow the installation instructions. You can send the audio to the sound card. I'd recommend selecting:

Quick settings - Image cancelling DC receivers - Perseus 125K

This will start stereo processing. The audible result is on the left ear. With 125K can tune from -62.5 KHz to +62.5 KHz. If you set the VFO frequency to 10.700 MHz and set offset to the AR5000-frequency (for example 48250000 Hz), then the scale will show the real frequency and min/max values can be entered as real.
If you change to mono in I/Q settings and try to let the waterfall continue, frequency bands appeared 100% mirrored around the VFO frequency. Probably a question of detailled settings.

FM demodulation

When I wrote the following lies, neither WRplus software nor the FM+ converter and software were out yet. The first software to demudulate a WFM signal was Spectrum Lab:
Broadcast FM signals (87.5-108 MHz) use 150 to 300 KHz. The result with the original Perseus software limited to 50 KHz is a terribly distorted signal with the Perseus software. With the Spectrum Laboratory Wide-FM demodulation is possible, very experimental though. A fast PC is necessary or you will face many audio drop-outs. With the quick settings, FM demodulation did not work that well. But when I tried to set it all manually, it worked: With my setup the sample rate conversion / resampling was not neccessary. So the Options - audio settings work with any value for sample rate and the "use different sample rate" option can remain unchecked.
Configure settings in Options - audio settings - I/Q input adjustment :

speclab screenshot audio I/O

FFT was bypassed. The "Mono SDR" and "Mono DAC" boxes must be enabled. The bandwidth can be connected to the slider. For 400 KHz sample rate use these settings:
With my PC (dual core / 2GHz) 200 KHz works fine, 400 KHz has regular drop-outs.
With these settings strong mirror frequencies appeared as described above: If you tune to the station's frequency minus KSamples (0,125 or 0,25 MHz) the same signal appears again. But maybe future software versions for the Perseus will implement Wideband-FM decoding, maybe even stereo. Nico is probably working on it. RDS demodulation would be great, but that's only a hope yet.

I wonder if it will ever become possible to decode analogue TV signals. 800 KHz bandwidth could be enough. It doesn't have to real-time. There must be someone found to write the software before the signals disappear.