[Radio1950]

Thanks again Jeremy for your analysis.

The Issues
I have been concentrating on three issues, the VCO spurious components around the centre frequency as seen in the demodulated 1090 Hz, the mains hum on the AF out, and the AF harmonic distortion.

By the way, the 1090 Hz is chosen to avoid possible aliasing in some Test Equipment, and to allow visual distinction on spectrum plots with any 1000Hz Ref signal from the Phase Detector and PPL/VCO. I should have used the standard 820Hz, but I am an old radar man, and old habits ...

VCO Sidebands
Without a Spectrum Analyser, nor an AF Wave Analyser, (not even an N&D Set!), I have been trying out new methods of visualising the VCO problem components, so that I can carry out live experiments and tweaks.

Plotting a new audio spectrum with Audacity is rather tiresome for the testing phase, as it is not live, and does not have sufficient frequency resolution, anyway, but is a good final guide.

I was using my Siemens D2055 Selective (ie Tuneable) Level Meter monitoring the RX IF Out at 455 KHz, and which has a 20 Hz filter. I was tuning off the centre frequency, and watching the sidebands, 15 dB down from centre, on the analogue meter.

I have also been observing the IF Out on a CRO, looking particularly at the spurious signals as showing as 1000 Hz AM.

I used IF Out, because it is before any final demodulation to AF, and should only contain VCO and mixer products, and perhaps a very little AM from the DC supplies.

The RX IF Out is useful.
I have measured its output at -3dBm75 ie in 75 ohms (abt 1.5 V pp) and measured its Zs at about 17 ohms. Deo gratia; JRC used an emitter follower here as driver.

To get an indicative value for the overall AM on the IF signal, and expressed in pseudo percent AM, I used my other 2022 sig gen at 455 KHz, and observed both sig gen and IF Out on a CRO.

Varying the 2022 AM mod level to get the same visual mod as the IF, I observe that there is about 2% AM, and, as it is a slightly distorted sine, there are other products.
There are measurement uncertainties here, especially at the bottom of the mod capabilities of the 2022, but it does provide a guide.
Reversing the 2018 and 2022 sig gen roles gave the same result.
Photo 1 shows this for RX1.

Interestingly, RX2 gives the same observed AM mod level, and with a similar mod wave shape on the IF.

No matter what I do, I cannot reduce the VCO spurious further, without more extensive mods or circuit change.
I again experimented with earth loops, and shielding, and seeing if any inductors etc were susceptible to mains effects.

Adding the extra 10M resistor across C316 in the loop filter area modifies the results, but is academic.
Removing that resistor, and adding another 4M7 capacitor, is little better.
The loop, as designed, is stable enough with no problems all through the VCO range.
I have restored my RX to standard.

All components here are original, except C83 0.47Mfd is now ceramic.
I think it is likely that circuit degradation which provokes VCO spurious will be limited to certain capacitors, and not semis, inductors nor the metal resistors. That is my experience.
I have replaced all relevant capacitors with new types.
RX2 has noticeable FM on its demod signals, as seen on a CRO.

Audio Distortion
I spent some time on the audio stages, to determine if the distortion was able to be improved, and to see if the distortion was from the VCO, the DC supplies, or from degraded components.

The Speaker Output has a very little more AM distortion (ie Harmonic Distortion) than the Line Out.
The AF Line Out looks very clean.

Unfortunately JRC has used a non-standard method to drive this output.
It is just fed from the wiper of a 5K trimpot off the AF driver output.

Its Zs is therefore variable. I found that the trimpot at 80% travel gave a useful compromise of AF level and Zs with approx -2dBm600 (approx 0.7V RMS, or 2.0 V pp) with a Zs approaching 600 ohms.

I cannot discern where the various components of the AF distortion are coming from, as the levels are too low.

I “see” on the CRO, and by experience, that the RX AF distortion on SSB is less than 4%.

I calculate the THD at 1.5% from the harmonic levels on the spectrum.

Improvement is not worthwhile. In professional coms specs (not HiFi) we normally used a 5% overall limit on RX distortion, and JRC specify less than 10%, so it is fine.
If my theory is correct, the contribution of VCO “noise” is small compared to audio non-linearities in the RX “back end” after the IF stages.

The FM component on the AF output, both Line Out and Speaker Output is now so low that only spectrum analysis reveals it.
It is not discernible on any other test equipment that I have, and certainly cannot be heard.

RX Setup for tests
10.505 MHz except as indicated, USB, BW 6KHz, Slow AGC, ANL OFF, RF Gain max, ATTN OFF, , PBT centre, Delta F OFF, VFO Int.
Signal Input level -53 dBm50, (500uV P.D.) mod off.


Recording For Analysis
I have upgraded my method of recording to WAV files.
I now use RX Line Out feeding a 600 ohm 2:1 shielded transformer (actually an old AF hybrid) and have an 8 dB resistive attenuator on the output, then feeding the MIC input of an AF to USB converter dongle into a (ASUS UX330) laptop using Audacity, and running on battery. Hopefully all this minimises unwanted signals.

Wrangler
The transformer has no bushes on the assembly screws.
Its copper strap is over the bobbin, but does not cut the flux, ie, it does not go through the core, but is wrapped over the top and bottom of the laminations.

I cannot see if it is bonded end to end, but I would think it is likely.

Also, hard to see in the photo, there is a ferrous metal strap going all around the laminations, at 90 deg to the windings, and bonded at the strap ends, on the bottom.
If I use a wire test loop fed to a CRO, there is just a little external field, and less than I would have expected.
There is no variation in the observed field as I approach the screws, but the methodology is not exact.

Jeremy,
The VCO drive swing is 5.50 V for freq xx.000KHz, 7.35 V for xx.500KHz, 11.9 V for xx.999KHz.
It swings 10 MHz from 24 to 34 MHz, then is divided by 10, as you are aware.

Some more WAV files as PDF for you, if you still have time, with details in file names. Next post due forum limitations.

Please put that Agilent analyser in your will for me please.

And some spectrum plots for everyone’s interest.

Temporary Break
I am approaching the end of this stage of my investigation.
We are going away on hols for a few weeks, away from the workshop at least.
I may have an update later, and with some maintenance tips specific to this RX, picked up along the way.


Many thanks to all contributors, especially Jeremy.