VCO circuit design

[Dan Moos]

First, here's the "why" of what I want to do. I repaired my HP 3468a bench meter, but the cal data got lost. I am content if it matches the accuracy of my Siglent 3055, so I plan to use that meter as a transfer standard. Honestly, I don't need a pro level cal.

Ok, with stuff I already have, I can cal almost every function of the meter. The main thing I lack is an AC voltage source in the proper ranges. Since I do some tube stuff, a high voltage adjustable AC supply (say, up to maybe 450 volts or so) could be handy. So I am gonna build one and use it to cal my meter.

My plan is to start with maybe an LM317 based adjustable DC supply, use that to power a VCO, and use a transformer to step up the voltage.

So about the VCO? I know I could do a fixed 60Hz design, but this is for fun too, and I want dumb capabilities I don't need. I'm a hobbyist after all.

So, I'm thinking say, 10s of hertz to maybe a kilohertz. If I ever have to mess with some 400Hz stuff, I'd be covered.  I plan to make the thing mcu controlled too, if that matters.  I probably don't need the purist of sine waves, but in the spirit of excellence, that's a plus.

My gut says Wien bridge of some sort. But I'm open to anything, including a square wave gen that I filter into a sine.  I plan to conjure a simple counter in order to have a readout, and possibly to use as a feedback mechanism to stabilize the frequency.  I am looking for an excuse to play with PLLs too if that is a useful element.

Ideas?

[Kleinstein]

You don't need a VCO, but more like an adjustable sine generator.  A VCO based on the wien bridge or similar state variable oscillator principle was an old style possibilty. An XR2206 based generator is possible too.  A more modern generator option would be a DDS generator as a special chip, µC and software or FPGA plus DAC.  There is also the possibilty to start with a ready made function generator and just add an amplifier for higher voltage / more power.

[Dan Moos]

You don't need a VCO, but more like an adjustable sine generator

Agree.

Of the options you mentioned, I'm leaning Wien bridge. I want to do this fully in discrete hardware (other than ther mcu), and I want stand alone because why not.

[Terry Bites]

there are anlog soultions with triange vco and diode and transistor based "sine shaper" circuits.
I'm an old analog geezer and I fear life's too short for this sort of thing. My trimpots are set for good now.

see ebay "AD9833 Signal Generator" 28-bit precision for $5, what's not to love.
Vomit a few register loads at it, job done. That's the easy bit.
Now about that power output stage......

[Kleinstein]

A wien bridge is relatively easy with a fixed frequency, but difficult to build as a VCO. If at all one could consider the related state variable oscillator with transconductance amplifier (e.g. LM13700). he LM13700 data-sheet includes a plan for a sine wave VCO.

The old style (but now chinese copy) FGEN chip XR2206 is also possible, possibly as a cheap kit from China.

[Dan Moos]

Now about that power output stage......

Not worried about that too much. I'm basically gonna make a low distortion audio amplifier, and feed a transformer to step up the voltage. This, I know how to do...

see ebay "AD9833 Signal Generator" 28-bit precision for $5, what's not to love.

I mean, you're not wrong, but a big part of doing this as a hobby is doing things the hard way 'cause reasons. Yer talking to a guy who was dumb enough to roll VGA video from scratch with 7400 logic.

[Dan Moos]

A wien bridge is relatively easy with a fixed frequency, but difficult to build as a VCO

Well, if I go wien bridge, I won't do VCO. I will probably use some digital pots I have in my stash to adjust the two filter sections. I am hoping that since I only want about 500 Hz of range, I should be able to make that work.

[Kim Christensen]

I know I could do a fixed 60Hz design, but this is for fun too, and I want dumb capabilities I don't need. I'm a hobbyist after all.
So, I'm thinking say, 10s of hertz to maybe a kilohertz. If I ever have to mess with some 400Hz stuff, I'd be covered.  I plan to make the thing mcu controlled too, if that matters.  I probably don't need the purist of sine waves, but in the spirit of excellence, that's a plus.

Since you are doing this for fun: You could use an MCU to generate a low bitrate sinewave, low pass filter it, feed it to an audio amplifier IC, and backfeed that through an audio-output transformer (tube amp type) for the high voltage AC option.

[Kleinstein]

There are plenty of Wien bridge circuit variations around. The main tricky part is the amplitude control part. The old style lamp is not very stable against variations of the bridge gain (e.g. from changing the frequency).  Ideally one could do the amplitude control from behind the amplifier. So one could include the amplifier (and maybe transformer) in the amplitude regulation loop.
The more modern amplitude control uses JFETs or the LM13700.

[Benta]

The Wien bridge oscillators I've seen are "bouncy" when starting up or changing frequency, at least the lamp- or fast NTC stabilized types.
JFET or other stabilizing can work, but tend to increase distortion (linearity and response times are an issue).

With today's really cheap, fast and powerful MCUs, my solution of choice would be generating a PWM output (modulated through a sine-wave lookup table) and simply RC filtering it.
The RC filtering is really easy, as it only needs to take care of the constant PWM carrier frequency, not the resulting sine-wave itself.
Tip: if you possess AoE, the X Chapters, check out:
"4x.25 Ripple Reduction in PWM".

[PCB.Wiz]

So, I'm thinking say, 10s of hertz to maybe a kilohertz. If I ever have to mess with some 400Hz stuff, I'd be covered. 
I plan to make the thing mcu controlled too, if that matters.  I probably don't need the purist of sine waves, but in the spirit of excellence, that's a plus.

My gut says Wien bridge of some sort. But I'm open to anything, including a square wave gen that I filter into a sine.  I plan to conjure a simple counter in order to have a readout, and possibly to use as a feedback mechanism to stabilize the frequency.  I am looking for an excuse to play with PLLs too if that is a useful element.

Filtering a square wave does not sweep well.

The classic Function Generator approach was to generate a triangle wave and then shape/clip that into a sine wave.

If you already plan to make it MCU controlled, the simplest by far, is to get a MCU with a DAC, and generate the sine wave from a table.
You get inherently precise frequency control, and for 10Hz~1kHz interrupt driven DDS is very easy manage and filter. 
Even a PWM DAC is ok at 1kHz max.

Which MCU family will you use ?

[David Hess]

A wien bridge is relatively easy with a fixed frequency, but difficult to build as a VCO

Well, if I go wien bridge, I won't do VCO. I will probably use some digital pots I have in my stash to adjust the two filter sections. I am hoping that since I only want about 500 Hz of range, I should be able to make that work.

An example of how it can be done with a single element for frequency adjustment is shown below.

[dobsonr741]

DDS (AD9833) will give you the freq. variability with the simplest and fastest path to start calibrating.

Building a stable and low distortion Wien without gain control loop is almost impossible. Gain control will give slow settling to the stable state. Wien only makes sense today for very low distortion audio spectrum measurement.

[PCB.Wiz]

First, here's the "why" of what I want to do. I repaired my HP 3468a bench meter, but the cal data got lost. I am content if it matches the accuracy of my Siglent 3055, so I plan to use that meter as a transfer standard. Honestly, I don't need a pro level cal.

Ok, with stuff I already have, I can cal almost every function of the meter. The main thing I lack is an AC voltage source in the proper ranges. Since I do some tube stuff, a high voltage adjustable AC supply (say, up to maybe 450 volts or so) could be handy. So I am gonna build one and use it to cal my meter.

Since this is just for transfer meter calibrate, sub 1kHz, maybe even simpler is to just use a PC sound card.
Plenty of software out there that does sine generate, and this is easily inside that range, then you just need a high voltage amplifier.
https://dinotechno.com/list-best-free-signal-generator-software-for-pc/
 

[Terry Bites]

You'll likely find this thread on Wien design interesting: www.eevblog.com/forum/projects/wien-bridge-project/