G1HBE

Better than Nothing.

More Huffing and Puffing!

 

May 2012.  Another pointless project to keep me out of the pub. This one is a tunable VHF oscillator which combines a genuine VFO 'feel' and the stability of a crystal. It is based on the 12 MHz Huff & Puff VFO which is mentioned briefly on H&P page 1. To get it up to VHF, three methods were considered:

1. Multiplication.  A bit of a non-starter, as it multiplies the tuning rate, the step size and any phase jitter present on the signal. Extensive filtering is required to make sure the final output is free of unwanted multiplication products.

2. Direct mix-up. By mixing the output of the 12MHz VFO with a cystal multiplier and filtering the output, the oscillator signal can be placed anywhere as long as you don't mind plenty of filtering. As the mixing is done in the actual signal path, you have to watch out for unwanted products and filter them out.

3. Phase locking.  This is almost the reverse of (2) above. A voltage-controlled VHF oscillator generates the wanted frequency directly, and its frequency is controlled by the 12 Mhz VFO via a phase detector. To compare the two frequencies, the VHF signal has to be mixed down to 12 MHz to match the VFO.  The mixing takes place out of the signal path so the risk of sproggies is greatly reduced. You can see the idea in the block diagram down the page. Shades of the old VHF Minisynth!


I started with option 2 and just grabbed a nice 'round figure' xtal out of the box (50MHz) and multiplied by 3 to produce 150Mhz.  This would make the 'mix up' appear on 162-ish MHz.
Some time(!) was spent cleaning up the multiplier. My normal cleanliness standard is that no sprogs should be worse than -60dBc, but I hit a brick wall at -45. I suspect that leakage around the final 150Mhz filter was to blame, but I wasn't prepared to do a re-build so I ploughed on regardless. Bad move, as the mix with the VFO produced a small forest of unwanted signals!
Extra filtering of the VFO to remove its harmonics helped, but the whole thing was beginning to look distinctly half-baked so I started again, this time opting for number 3 on the list.

The VCO was my old friend the common-base oscillator. These are usually well-behaved and can give a nice wide tuning range without too much faffing around. It was built in one of Alan Melia's small tin boxes and I managed to fit in a MAR1 buffer as well. Tuning it with a pot showed it to be fairly stable.
Now I needed a mixer-downer to get the frequency down to 12MHz for the phase detector. A slightly bigger box from Alan was used to contain these two devices. The mixer was the same NE612 chip from the mixer-upper in the previous attempt and the PD was a 74HC4046.
To cut a long story short, it locked up first time! Well, it did after I'd realised I hadn't taken pin 8 of the PD to deck...
The lock was solid and the VCO followed the VFO nicely, but there was quite a lot of noise around the base of the peak on my analyser and the ultimate noise level never got below -50dB. Attention to the phase detector time constant and decoupling put this right.  Unwanted outputs were few and far between, but there were some fast-tuning carriers which were caused by harmonic radiation from the VFO mixing with stray signals from the 150MHz multiplier. The 20MHz clock oscillator in the Huff&Puff box was also getting in on the act, but again better screening and filtering fixed it. Overall noise level was now just 2dB worse than my Marconi 2024 sig gen and no sprog was worse than -75dBc. Most of them had disappeared down to noise level.

So there it is. I can't claim any originality except that my VFO is stabilised using Huff & Puff, providing xtal-like stability, free knob tuning and click-less 30Hz steps, whereas all other designs I've seen leave the VFO free running and accept a small amount of drift.

 

 

 


Block diagram. The 160 MHz VCO mixes with the 150 MHz from the xtal osc to produce an IF around 10 to 12 MHz. This goes into one half of the phase detector and the signal from the manually-tuned VFO goes into the other. Any difference between the two is corrected by an error signal (orange track).

 

 

 

 

 

 

 

 

 

 

 

If I keep on buying Alan's tin boxes at this rate, I'll be invited to his staff dances! Here's the PLL huffalator on test. The VFO is at the front and the boxes contain the xtal osc/multiplier, the mixer/phase detector and the VHF VCO.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Output spectrum spanning 50 KHz each side. The yellow trace is the Huffalizer and the blue one is my Marconi 2024 sig gen for comparison. Not much in it!