Cheap stable VFO design

[MrSlack]

So I started with grand plans, to build a nice stable but cheap varactor controlled VFO about a month ago that tuned ~7.000-7.040 MHz. Cheap instantly excludes large air variable capacitors, rare FETs, 10 turn pots and other such things.  This went through several iterations and periods of instability as well as about 4 different varactor types. Turns out varactors are expensive as well.

Basic idea is a simple discrete regulator providing an 8.5v reference voltage for the tuning voltage. This is then used to generate a tuning voltage with coarse/fine controls. The voltage is summed and then there is a simple emitter follower. This provides a voltage to the front end of the oscillator. This is a simple colpitts oscillator which uses a 1n4002's reverse biased capacitance vs voltage curve to generate a voltage controlled capacitance. This is used to trim the oscillator.

Drift is about 20Hz/min and is designed to be relatively stable based on battery voltage. Not wonderful but for the general cheapness, adequate.



Course should read coarse above. It was about 2AM when I scribbled that out...

[nugglix]

Hi!

That really looks simple and straightforward.

Will steal a lot ... 

I just discovered that (LTspice) simulation and reality are not necessarily share the same traits.
To my surprise the used varactor (1SV149) was much better in reality than in the simulation.

My current problem is that the transistor distorts the waveform, so I need a way to limit the gain
when the oscillation has started.
I tried 2 anti-par BAT-85 diodes but that introduced a lot of jitter.
I'll do some more experiments tonight - aka build a new oscillator. 

How does the waveform of your circuit look like?
Iow, do you care about that at all?

Cheers
  Guido

[MrSlack]

Yeah I started this design in LTspice. Reality wasn't anywhere near the simulation output.

1n4002 is surprisingly linear for a 0.8-8.5v tuning range. I tried a few BB series varactors and discarded them in the end.

The output waveform is not very clean as you'd expect.. Peaks look fine but troughs are flat, therefore harmonics. I'm adding a buffer amp to it and a low pass with a 3dB point of about 7.2MHz matched to a 50 ohm output to start with. The biasing of the oscillator transistor could be improved as well to keep it in the linear region. In fact version two of this may have a JFET as they are somewhat less tetchy about temperature as well.

Not a bad idea with the diodes. Might have a play with that.

I'll post a picture of the breadboard when I get a few mins. The very definition of fugly.

[nugglix]

My goal is a LO which allows me to cover AM to 160m ham band.
My first IF is 10.7MHz, so I need an oscillator from 11.2MHz to 14.5MHz.
This turns out to be quite a thing -- at least for me.

I went the JFET route today.
See attached files.

The waveform (fft_osc) is nice, I think.
But the bastard quits oscillation on varactor voltages over 3.6V. 
Something the simulation didn't show.
Will add the buffer tomorrow and see if it helps -- but I doubt it.

The clean spectrum amazed me. I'll try this diode limiter again on my
BJT (BF 959) oscillator. Which has a wide tuning range, but also shows
distortion and amplitude issues.

The 2 blue parts at the lower left are 10µH inductors in parallel.
Ahhh... might be that the Q of the inductors is to low.   
Thanks for helping me already
Will try a T68-2 with 30 windings tomorrow.

[T3sl4co1l]

The diode is shorted by the inductor, and the diode is shorting the inductor during half the cycle...

Tim

[MrSlack]

T3sl4co1l: Not 100% sure about that. The thing is throwing 0.36v p-p which isn't enough to make the diode conduct and perform rectification here. Plus it's reverse biased by 0.8v+. If it was a germanium diode at the end of the tuning range perhaps. I'm sticking a couple back to back to see if that does indeed make a difference however.

nugglix: I can't get something with that wide a tuning to actually stable enough not to screw up the audio after the final IF stage hence why I'm trying for 40m CW only. I did an initial cut that tried terribly to span 20-160m but when I opened the window it blew off enough heat to drift by about 10khz. Perhaps AD9850 can save me in the future there! Yours looks good though - pretty clean.

Just blew up my last JFET so will have to wait until Wednesday for more fun

Fugly (tuning was done on a separate board to start with as it had half a varactor experiment stuck to it):

[nugglix]

OH... 20m to 160m is quite a range.
I wouldn't try that, given that I have problems with a tuning range of 3.3MHz.

Ok, last thing for today: try limiting diodes on BJT oscillator.

Back in a minute...     

[nugglix]

Ha! Success!

Just added the limiting diodes (D3, D4) to the BJT oscillator.
Looks much better (see pics).
I think I'll take this one.

This was a good day at the bench.
Now time to get some sleep.

Cheers!

PS: The amplitude is still enough for a SA612 mixer.

[uncle_bob]

Hi

.... as mentioned above, you have a DC short to ground across your tuning diode. ...

Bob

[edavid]

Haven't you heard about huff & puff stabilizers?

http://www.hanssummers.com/huffpuff.html

[MrSlack]

Ha! Success!

Just added the limiting diodes (D3, D4) to the BJT oscillator.
Looks much better (see pics).
I think I'll take this one.

This was a good day at the bench.
Now time to get some sleep.

Cheers!

PS: The amplitude is still enough for a SA612 mixer.

Looks good. I'm still waiting on my SA612's to arrive. Ordered them a week ago. Considering, as I have a surplus of T37-6's just building a diode mixer and being done with the IC side of things.

Haven't you heard about huff & puff stabilizers?

http://www.hanssummers.com/huffpuff.html

I looked at them myself but have written this off as "too complex". This is designed to be cheap, simple etc. We're getting into DDS/PLL territory there which is a whole different kettle of fish.

Some further thoughts on this:

First idea was to pull the RF energy off the emitter. Unfortunately, not exactly ideal waveform as discussed above (all 100mV/div):



30 mins in a copy of EMRFD and in their progressive "first transmitter" section, they decide to pull it off the base. Quick poke with the scope reveals a cleaner waveform:



So next step is to buffer this, then shove it through a simple 3 pole LC with 3dB at 7.2MHz matched to 50R out. When this is working I'll rebuild it so it's not quite as fugly.

Also borrowing a Tek 494 from a friend on Monday so will have a look at this in the frequency domain too.

[nugglix]

Hi all!

Just a quick follow-up.
Replaced the inductors in the JFET oscillator with a hastily made T68-2 core w/ 25 turns
(~3.5µH). My meter shows a Q of just 30.

Seems to do the job.
Can now tune from ~11.3Mhz to over 16MHz. Which is more than I need.

Will now attach the buffer and do some more measurements.

Link JFET ciruit:
https://www.eevblog.com/forum/rf-microwave/cheap-stable-vfo-design/msg950836/#msg950836

[Radio Tech]

Very nice to see folks building stuff like this. Gives me hope and future plans for myself.

[MrSlack]

Indeed. It's fun doing it this way and that's all that matters

Hi all!

Just a quick follow-up.
Replaced the inductors in the JFET oscillator with a hastily made T68-2 core w/ 25 turns
(~3.5µH). My meter shows a Q of just 30.

Seems to do the job.
Can now tune from ~11.3Mhz to over 16MHz. Which is more than I need.

Will now attach the buffer and do some more measurements.

Link JFET ciruit:
https://www.eevblog.com/forum/rf-microwave/cheap-stable-vfo-design/msg950836/#msg950836

Good stuff. Not used material 2 yet. Got material 6 only.

I just ordered 10 each of T37-6 (final VFO) and FT37-43 (for mixer - need some reactance there!) as well as some more 10k pots and some Altoids tins (this needs to be a pocket rig).

 I increased the tuning range of my VFO by using two 1n4002's. Can now tune the entire 40m band in theory. Buffer circuit time was replaced by garden chores (boo hiss)

[AF6LJ]

Just found this thread this morning a couple of observations.

I would ditch the emitter follower, that is only going to add drift to the circuit.

Your varicap diode shouldn't be a major part of your total C in your tank circuit.
Make it a smaller part of the total C and keep the RF voltage across it to a minimum, if the diode conducts it will dirty up your waveform.

When you have it built and working like you want it to...
Pot it that will make it more stable over temp and less sensitive to mechanical shock.

[MrSlack]

Thanks for the tips - much appreciated.

Definitely going to pot/epoxy the inductor at least on mine - tapping on the table causes drift!

[PA0PBZ]

Definitely going to pot/epoxy the inductor at least on mine - tapping on the table causes drift!

Shout at it and you will have simple FM modulation 

[MrSlack]

That did cross my mind

I tried using candle wax like the old 1970s IF transformer gunk on another experiment a couple of weeks back. Worked as well but buggers up your soldering iron tip if you poke it by accident.

[nugglix]

Hi!

Now everything is in place and the oscillator does it's job nicely.
Wound a new inductor on a T68-2 using 27 windings of #22 (0.63mm) wire.
Meter says Q is 46.

Oscillator goes from 11.2MHz at ~1.5V tuning voltage to 14.5MHz at ~4.5V.
Of course the linearity isn't the best.
So I better add a super-fine-control for sub-millivolt tuning.   

At 11.2MHz the 1st harmonic is at -32dBc, which should be okay.

In my first BJT oscillator I glued the toroid onto the board horizontally.
Not a good idea as Sue said in another thread.
This time I'll try to glue it vertical on some insulating material.
No clue what I can use...

Nice thread, hope it goes on.
Now for something to eat and then some 

Cheers
  Guido

[AF6LJ]

There are things you can do to improve linearity.
Using a small portion of the tuning range is one way, paralleling diodes is another.
When you are below 30mHZ you have a lot working against you, not the least of which is trying to get a few MHZ of tuning range.

Remember also if these are home brew projects you are not constrained to having to work with voltages less than 12V I am more comfortable with supply voltages in the 18-22V range. This was due to my work at Loral corp, I had other constraints current being one of them.

As for potting; everything plays hell with a hot soldering iron when it comes to potting compounds. We always knew how high to set our VCOs before they got potted and it always worked out just fine, but that was high UHF and we used the high priced potting compound. At least the stuff came in a nice shade of Pink.

[nugglix]

Hi!

There are things you can do to improve linearity.
Using a small portion of the tuning range is one way, paralleling diodes is another.
When you are below 30mHZ you have a lot working against you, not the least of which is trying to get a few MHZ of tuning range.

First, let me correct my statement about the tuning range.
The 3.3MHz range is of course from 500kHz to 3.8MHz, which is AM to the end of 80 meters.
For the end of 160m (2.0MHz) I'd need only 1.5MHz tuning range.
So things are getting better already.   

Remember also if these are home brew projects you are not constrained to having to work with voltages less than 12V I am more comfortable with supply voltages in the 18-22V range. This was due to my work at Loral corp, I had other constraints current being one of them.

At the moment I've set the max. voltage for the supply to 12V.
What would a higher supply voltage give me?

The control voltage for the varicaps is between 1V and 5V.
According to the datasheet for my 1SV149 varicaps this is the most linear region.

Looking at the diagrams I have to admit I missed the f/f_max to V_R diagram.
Discovering new things every day...
Will see if I can make use of it. But this is for the next weekend.

... At least the stuff came in a nice shade of Pink.

Manual drift compensation is completely fine...     


Just attached this
   http://www.aliexpress.com/item/0-1MHz-1000MHz-digital-Frequency-counter-meter-tester-Cymometer-8-digits-0-56-LED-display/32410222299.html
to the oscillator. Looks good.
You can set the IF in the counter and then display the RF you're targeting. Nice thing.

Now preparing for a trip to Ikea, need more shelves... 

Thanks for the hints and that you made me look at the datasheet again.



Edit: spelling

[AF6LJ]

Hi!

There are things you can do to improve linearity.
Using a small portion of the tuning range is one way, paralleling diodes is another.
When you are below 30mHZ you have a lot working against you, not the least of which is trying to get a few MHZ of tuning range.

First, let me correct my statement about the tuning range.
The 3.3MHz range is of course from 500kHz to 3.8MHz, which is AM to the end of 80 meters.
For the end of 160m (2.0MHz) I'd need only 1.5MHz tuning range.
So things are getting better already.   

Remember also if these are home brew projects you are not constrained to having to work with voltages less than 12V I am more comfortable with supply voltages in the 18-22V range. This was due to my work at Loral corp, I had other constraints current being one of them.

At the moment I've set the max. voltage for the supply to 12V.
What would a higher supply voltage give me?

The control voltage for the varicaps is between 1V and 5V.
According to the datasheet for my 1SV149 varicaps this is the most linear region.

Looking at the diagrams I have to admit I missed the f/f_max to V_R diagram.
Discovering new things every day...
Will see if I can make use of it. But this is for the next weekend.

... At least the stuff came in a nice shade of Pink.

Manual drift compensation is completely fine...     


Just attached this
   http://www.aliexpress.com/item/0-1MHz-1000MHz-digital-Frequency-counter-meter-tester-Cymometer-8-digits-0-56-LED-display/32410222299.html
to the oscillator. Looks good.
You can set the IF in the counter and then display the RF you're targeting. Nice thing.

Now preparing for a trip to Ikea, need more shelves... 

Thanks for the hints and that you made me look at the datasheet again.



Edit: spelling

Always glad to help out when I can.
I've seen those counter boards around, it is hard to believe they can produce them for that cost.

[MrSlack]

I think that's literally the cost of those things. We have a very abstracted cost over the top of everything whereas in China they have just the material cost. Profit per item isn't as important as profit on a volume of stuff.

My frequency counter, which I will describe in another thread, didn't cost more than that either. Pretty much stolen from EMRFD with some resolution modifications (more digits).

Awaiting postman at the moment who will bring me gifts of toroids ... I hope ... but probably bills.

Just a quick note: with respect to tuning, when my VFO eventually turns into a transceiver, the idea is to skip back to tuning capacitors as I have just obtained a large box of very nice ones cheap from ebay with reduction drives. The transmit and receive frequencies are offset in the mode I'm using so the plan is to have a separate receive and transmit capacitor and switch them in and out on TX/RX. Using some tips on a video from VK3YE, you can zero beat the TX capacitor easily enough with a switch. So tune until you receive a CW signal with the RX capacitor, switch the TX capacitor in, zero beat it and then switch it back. Using a solid state auto TR switch and semi-break in should make it usable after it is set up if it doesn't drift too much.

I'm going very low tech here for ref

[AF6LJ]

Sounds like this is going to work out well.
Keep us posted on your progress.

[MrSlack]

No further progress. T37-6 toroids I ordered on the 23rd of May haven't turned up yet. Some FT37-43's turned up for the mixer so will have a bash at that and recycle the T50-2's for now. FFFfffff ebay and Royal Mail do my head in.

Small diversion as a pile of SMD stuff turned up from CPC. Designed a simple push pull audio amplifier for the receiver. Built it using dead bug through hole. Worked fine first time. Simple gain stage, rubber diode and push pull with some negative feedback. Decided to try and build an SMD version - spent the best part of two hours manually cutting traces on the board with an X-acto and soldering it up. Powered it up with current limiting and nothing happened. So an hour of debugging turned out that some of the MMBT3906's I ordered were actually MMBT3904's. Managed to eventually sort them all out and solder a replacement on.

Small amplifier - after replacement transistor:



Not RF or VFO related but it will be connected after the mixer and AF filter

Edit: just noticed I missed a transistor base solder joint and ALL the tants are soldered on the wrong way round (doh!). Still works though and didn't blow up in my face