DIAL SCALE LINEARITY -- Spreading out the Frequencies for the 15-10 BITX Rig

Click on the image for a better view

The 15-10 Rig has been performing very well, pulling in a lot of DX contacts on both bands.  But there is one thing that has bothered me:  The way the transceiver tunes.  It can be a bit difficult getting an SSB station tuned in properly.  At first I thought this was caused by a lack of lubrication on the variable cap that I've been using (out of an old QF-1), but it turned out that this was not the cause.  The problem is something that Pete Juliano has lamented several times:  LC style analog VFOs have a tendency to have the frequencies "bunched up" at one end of the tuning range.  In other words, the tuning range is far from linear.  I was having trouble tuning stations on on the portion of the band where the frequencies were bunched up.  I did some quick measurement and found that on this side of the capacitor's tuning range, one turn of the dial would move the frequency about 100 kHz -- that is far too much.  On the other end of the capacitor moved  only 22 kHz with one turn of the dial (as I recall this is close to the recommended 20 kHz per dial rotation).  Clearly I had a lot of the dreaded bunching up.  This was what was making tuning difficult. 

I had built a pretty standard Colpitts FET VFO.  I had a 6.6uH coil, and a 9-135 pF variable cap in series with a 68 pF fixed cap.  I was pleased that the VFO worked, and I put it in the circuit.  Only later did the bunching up shortcomings become apparent. 

I decided to build another VFO, this time paying attention to DIAL SCALE LINEARITY. 

I turned to the excellent Bandspread Calculator on Bob Weaver's Electronic Bunker web site:  http://electronbunker.ca/eb/BandspreadCalc.html

I plugged in the frequency range that I needed and the values for my variable capacitor.  I calculated Cs which was the combined capacitance of the feedback and coupling capacitors.  Finally, I had to make a decision about the nature of my variable cap:  was it a Midline-Center Cap or was it a Straightline Capacitance cap.  I consulted with Bob and he suggested that it might be somewhere between the two.  I got out some graph paper and measured it -- it looks to me like a Straight Line Capacitance cap, with the capacitance varying linearly with changes in in the rotation of the shaft. 

It looked fairly linear, so I selected "Straightline Capacitance."  Bob's calculator predicted a much better dial scale linearity (see the picture at the top of this blog post). 

I then built the oscillator stage in LTSpice using the values called for by Bob's calculator: 

It worked well in LTSpice: 

So I built it in the real world.  I didn't have the exact values for the padder and trimmer caps, so I use values that were close. 

Using the frequency counter in my Rigol 'scope, I again measured the frequency change for each movement of the shaft. 

Here are the results: You can see that the bunching up has been largely eliminated.  Frequency change for a 20 degree (not %) movement at one end of the capacitor's range is essentially the same as it is on the other end of the range. 

I will continue to play around with the padder and trimmer cap values to get this VFO where I want it.  I may also have to opt for less frequency range in order to get closer to the desired 20 kHz per dial turn value.  I will also have to play around with the additional capacitance that will be switched in to move the VFO down a bit to the range needed for the 10 meter band. 

The bottom line here is that Bob's bandspread calculator is very useful in figuring out how best to avoid the dreaded bunching up of frequencies that can -- sometimes -- come with the use of analog LC VFOs.  The display of Dial Scale Linearity that appears at the end of each calculation is really brilliant, and allow for an instantaneous look at how changes in the various parameters will affect the linearity of tuning.  This is a really wonderful tool for the homebrewer. 

Thanks Bob Weaver!