400KHz to 33MHz Wideband VFO

John_1981 wrote:Your 2 transistor common base oscillator is very interesting, I’m going to have to give it a go. I expect the with the additional gain very small values of feedback capacitors will work. At first I thought it was a cascode configuration but I see both transistors are in common base.

Yes. Add more 2N3904s here and there to add more and more reverse isolation. No RF transistors needed but you would have to use more 2N3904s as their FT is only 350MHz but their < 4pF base impedance really shines at VHF. 

It should be mega stable for those looking for HF signal generators or even LO for their mixer stage. Please add your own buffer to isolate the oscillator from antenna or output stage but keep all transistors 2N3904. They are cheap and easily available everywhere.

Your 2 transistor common base oscillator is very interesting, I’m going to have to give it a go. I expect the with the additional gain very small values of feedback capacitors will work. At first I thought it was a cascode configuration but I see both transistors are in common base.

Wow! Very nice.

Great! Build it on a single sided copper clad board! 

Okay this two transistor circuit would make a very stable FM transmitter as the tank circuit is isolated from antenna and feedback capacitor is only 1pF which presents a reactance of 1K on the FM band. It then feeds a common base isolation amplifier to provide the oscillator transistor with the necessary feedback. The LC tank is again isolated quite a bit.

A third buffer before the antenna would be good. The audio is applied directly to the base of the oscillator with some pre-emphasis.

A LtSpice simulation of this TX would be nice. 

A practical use I could think of is 1 to 100 MHz signal generator.

More transistors could be added to isolate the tank circuit even further from the active devices. Removing DC through the inductor should increase oscillator stability even further.

No, the only thing i'm interested in now is the tuning range and a flat output across the tuning range. As its just a lash up on a breadboard i don't expect stability to be fantastic. 

With the varicaps removed from circuit only the 10s of Hz digit flickers on my frequency counter at around 5.145MHz. With 1000 pF across one half of the coil winding I get 1.4 MHz so the range is good.

Excellent. Thank you! 

Did you measure the stability?

John_1981 wrote:Another promising oscillator circuit with wide bandwidth and flat output level across the range. This time its a BJT Hartley oscillator. Here it goes:

PNP transistor. Emitter via 470R resistor to +1.6 V (LED shunt regulator again). Base to one end of a center tapped coil (1:1) via a 33K resistor and 10pF capacitor in parallel for DC bias and feedback. Centre tap of coil to ground. Collector to other end of the winding. Pair of back to back varicap diodes coupled via a 1nF capacitor to the collector terminal and 22pF capacitor coupling from collector to a FET source follower with 1k source resistor. 

I'm getting 2.47-4.62MHz across the FET source resistor. Output is flat across the range at about 1.5V P-P and looks to be a reasonably clean sine wave. 

The base bias resistor can be varied between 6.8k-470k with little effect on the output waveform. Higher values (470k) give more output but with some distortion at the HF end of the range.  Lower values give lower output level but with less distortion. 

Coil is about 40T bifiliar on a T50-7 toroid.

Could you draw it and click a picture with your cell phone with flash on?

Another promising oscillator circuit with wide bandwidth and flat output level across the range. This time its a BJT Hartley oscillator. Here it goes:

PNP transistor. Emitter via 470R resistor to +1.6 V (LED shunt regulator again). Base to one end of a center tapped coil (1:1) via a 33K resistor and 10pF capacitor in parallel for DC bias and feedback. Centre tap of coil to ground. Collector to other end of the winding. Pair of back to back varicap diodes coupled via a 1nF capacitor to the collector terminal and 22pF capacitor coupling from collector to a FET source follower with 1k source resistor. 

I'm getting 2.47-4.62MHz across the FET source resistor. Output is flat across the range at about 1.5V P-P and looks to be a reasonably clean sine wave. 

The base bias resistor can be varied between 6.8k-470k with little effect on the output waveform. Higher values (470k) give more output but with some distortion at the HF end of the range.  Lower values give lower output level but with less distortion. 

Coil is about 40T bifiliar on a T50-7 toroid.

Thank you. I'll try Franklin next. 

This was created on LtSpice in 2020.

If Franklin is stable < 10MHz  then 30MHz * 3 = 90MHz Xtal source mixed with 5.5MHz FM modulated Franklin should be a stable FM TX with max + - 3KHz drift. Use fixed np0 in series with a trimmer for 95.5MHz fine tuning and modulate it with two IN4007 in parallel for + - 75KHz deviation fed by LM386 and an opamp for active preemphasis. 

I have tried the Vackar and it did look promising. Its a shame the LC circuit is not ground referenced. That said, I have spent the last couple of months experimenting with lots of different oscillator configurations using both FETs and BJTs and always end up coming back to the BJT Franklin. The thing I like about it is the wide tuning range, very low current consumption and clean output waveform. It is easy to get going without any of the component values being critical. As I am going to be using it with a PLL drift doesn't matter much for the time being. AGC can be applied by varying the voltage that the base bias supply is derived from.

Without drawing the circuit, I use 2 PNP Si transistors. Both bases to -ve via a 470K resistor. Each base to opposite collector via a 10pF capacitor. One collector straight to -ve, other collector via the LC circuit to -ve. Both emitters tied together then via a 6.8K resistor to +ve. Power supply is about 2.5V using a green LED and 1K resistor as a shunt regulator. Output buffered by a FET source follower. (J112 with 680R source resistor). At the moment it covers 2.5-4.4MHz. 

2x 1SV149 varicap diodes back to back and connected directly across the tank circuit. No distortion provided a low supply voltage is used for the oscillator. 

I am now working on the PLL side of things using a MC145157 serial load device. Because the N and R values can be set to anywhere between 3 and 16,383 with this device the output frequency can be made very close to 1kHz steps whilst still having a reference frequency around 10kHz so loop filtering is made easier. 

With a 12MHz crystal,  for an output of:
3500kHz R=1008, N=294 (ref=11.9kHz) with zero error
3501kHz R=1491, N=435 (ref=8.048kHz) with an error of 6Hz. 

Thanks again to Ivan for the PLL divider program!

Oh. Have you ever tried Vackar VFO? 
Is it more stable than a Franklin?

Similar circuit to what I have been using. When the transistors are connected with the base DC coupled to the collector I found that i can only get a clean sine wave output with very low DC supplies, around 0.7-0.9V but perhaps your AGC circuit takes care of this. 
I used PNP transistors (BC558) so that the coil is ground referenced to make it easier to connect to varicap diodes. I tried an AGC circuit but abandoned it as the variation in output level was low and will not matter over the narrow range I require