SSB Carrier Supression

[Nigel Goodwin]

At this time I don't have any copper clad board. I'm waiting for 6 experimenter boards to arrive. I also don't have any coil wire.
I can get these locally I guess so next week I'll go buy them.

You can get wire out of any old transformer, TV scan coils are an easy choice as well.

There are plenty of ways of building circuits without PCB's or copper board, such things are relatively modern innovations

For now, sim is fun because osc. are new for me to "design".
The colpitts will be easier to simulate.

It's also easier to build as well.

 

[kchriste]

At this time I don't have any copper clad board.

A few pages back in this thread is SV's example of using tuna cans to make a "tunable" oscillator.
I think the sardine cans, being sort of squarish, look better myself.

 

[Mikebits]

A few pages back in this thread is SV's example of using tuna cans to make a "tunable" oscillator.
I think the sardine cans, being sort of squarish, look better myself.

Might smell worse though...

 

[Space Varmint]

I'm sorry flat5. I did not get notified be Email. You can Email me direct if you want. mrgone2lunch@Gmail.com

Anyway, if you are using air coil it must be big. I think I used like no. 22 gauge wire and wound about 20 turns around a 3 quarter inch diameter pill bottle. It is a Hartley oscillator and the advantage over Nigel's suggested Colpitts is the Colpitts is split stater capacitors which means they will appear in series and therefore will not have the same capacitance range as a single tuning cap. So less frequency range will be the result. The caps C4 & C5 are for fine tuning. If C5 is very small like 15pf, then C4 won't be felt by the oscillator circuit much, and makes a very minute fine tune adjustment.

Just replace R2 & R3 with one 100 ohm resistor. That will be fine and it will run down to less than 3 volts supply voltage.

So what you want to do is to scrape the insulation off the coil on one side so that you can move the tap up and down the coil. Then you can see it oscillate and can choose a good solder point for the broadest range using your variable cap. C3 is the main tuning cap. I think the best position for power on start-up is 1/3rd from cold side of coil. But best range should be around center. The trick to making them work and broadest range is of course your capacitive reactance ratio to coil inductive reactance. So try some different coils.

Nigel is right. It is just one coil with a tap. When I use ferrite cores I will use something like no 36 gauge wire and fold the wire over about 1/3rd and twist the wire down a couple of inches, then use some sand paper to remove much of the enamel coating from the wire and then tin it with solder. That will be the tap.

 

[JimB]

It is a Hartley oscillator and the advantage over Nigel's suggested Colpitts is the Colpitts is split stater capacitors which means they will appear in parallel and therefore will not have the same capacitance range as a single tuning cap.

WHAT??

Where did split stator capacitor come from?
You dont need a split stator capacitor in a colpits oscillator.

Have a look here: colpitts oscillators

JimB

 

[duffy]

Dump the simulator, get the soldering iron and some real components, connect them up on a breadboard* and try it for real.

I agree completely. The simulator's waveforms rarely match up with the actual circuit when there's more than a few (very few) components, and especially so when you are working with RF.

 

[Space Varmint]

WHAT??

Where did split stator capacitor come from?
You dont need a split stator capacitor in a colpits oscillator.

Have a look here: colpitts oscillators

JimB

OK man. You see the two capacitors C1 & C2? Well, if you are interesting in making the Colpitts into a VFO, then they will be variable and if you make only one variable it will not change the frequency nearly as much as making them both variable. Hence...split stator or ganged capacitors.

Anyway, flat5. Look back a couple of posts to see some details on the Hartley.

And....Happy Easter gents....

 

[Space Varmint]

There are several reasons I like a Hartley. One is they are high Z. Therefore they do not require much power and this makes them run very cool, temperature wise. Heating causes drift. They actually are spectrally pure if done right.

They really are quite forgiving on the placement of the tap and they start up reliably as long as you are in the ball park with your ratios, capacitive to inductive reactance.

One thing you almost always have to do with a Hartley though, is because of it's very low power consumption, you must buffer it. Some sort of emitter follower arrangement is fine. I will not even discuss the Darlington other than to say, they have always worked well for me.

I will say, I have made every type of oscillator many many times and I am so set on the superior perfomance of the Hartley that I went out of my way to come up with a VXO design based on the Hartley configuration rather than using the more common Pierce oscillator. In a PLL we know that we are not so concerned with drift though due to digital control of the center frequency. If an oscillator is prone to drift in a PLL it will create high phase noise and phase sideband levels. But we also want some power to come from the VCO section. I thought an Armstrong would be ideal for this, but in practical application it was absolutely the worst! Again the Hartley reigned superior even though I think most use a Colpitts probably due to it's simplicity of design. They claim it has a broad bandwidth, yet I have found that the error dc voltage would cause more of a spread in frequency using the Hartley, ratio wise as in voltage level from phase detector & loop filter to frequency deviation of the VCO. This enables higher levels of filtering out the phase sidebands in the loop filter.

Also it is much cheaper to build a coil with a tap, than it is to buy a split stator capacitor.

 

[Nigel Goodwin]

OK man. You see the two capacitors C1 & C2? Well, if you are interesting in making the Colpitts into a VFO, then they will be variable and if you make only one variable it will not change the frequency nearly as much as making them both variable. Hence...split stator or ganged capacitors.

C1 and C2 aren't the tuning capacitors, they are the feedback capacitors, that rather useless partial diagram doesn't show the tuning capacitor which is directly across the coil and is only a single gang one.

 

[Nigel Goodwin]

Also it is much cheaper to build a coil with a tap, than it is to buy a split stator capacitor.

And a little easier to build a coil without a tap, and no more expensive at all for exactly the same single gang tuning capacitor

 

[Space Varmint]

C1 and C2 aren't the tuning capacitors, they are the feedback capacitors, that rather useless partial diagram doesn't show the tuning capacitor which is directly across the coil and is only a single gang one.

Without referencing anything, I am sure I have seen C1 & C2 as ganged variable caps. These would be used for your main tuning or broad tuning range. There most certainly are other ways to tune the Colpitts, but they will not be as broad.

As far as a tapped coil as opposed to an inefficient single variable cap in a Colpitts. I can make a tapped coil for less than a penny. That includes the solder....chop chop!

 

[Nigel Goodwin]

Without referencing anything, I am sure I have seen C1 & C2 as ganged variable caps. These would be used for your main tuning or broad tuning range. There most certainly are other ways to tune the Colpitts, but they will not be as broad.

Any differences would be slight, and certainly using ganged feedback capacitors to tune a colpitts would be extremely unusual (I don't think I've ever seen it, or even seen it suggested before?) - how broard a tuning range are you looking for? - aren't we talking amateur bands here?.

As far as a tapped coil as opposed to an inefficient single variable cap in a Colpitts. I can make a tapped coil for less than a penny. That includes the solder....chop chop!

But, as everyone has said, single variable tuning on a colpitts isn't inefficient, it's perfectly efficient and works exceedingly well. An added advantage of the colpitts is that band switching is easier, only one pole to switch rather than two.

Nothing wrong with a Hartley, but you appear to be dismissing Colpitts for only imaginary reasons.

 

[Space Varmint]

But, as everyone has said, single variable tuning on a colpitts isn't inefficient, it's perfectly efficient and works exceedingly well. An added advantage of the colpitts is that band switching is easier, only one pole to switch rather than two.

Nothing wrong with a Hartley, but you appear to be dismissing Colpitts for only imaginary reasons.

I am not dismissing Colpitts. As a matter of fact that is usually what I see in PLL. I think most engineers have dismissed Hartley in PLL when my own personal experience has shown a Hartley to be better in many respects but not to any over whelming degree.

I don't see your point on the band switching but what I am saying on the Colpitts single variable cap is that the two capacitors are seen as series capacitors. Remember that the true method of telling what kind of oscillator is being used it to follow the AC feedback path. So two capacitors in series is equal to less than one of the values of the smallest capacitor. The capacitor value of both capacitors must be equal for the broadest tuning range because the range of capacitance will be largest if the capacitors are matched.

With the Hartley, the range of capacitance will be equal to the variable capacitor in the tank circuit. This is a major reason I like them in PLL because you can reduce the loop gain thereby keeping the phase sidebands low. But of course with the speed of processors today, I think DDS is becoming vastly more popular unless there are significant reduced power consumption demands or cost is a major concern or you are using straight VFO.

 

[Space Varmint]

Here's a good website basically describing what I'm talking about with the ganged series capacitors:
**broken link removed**

 

[flat5]

The ratio of the two caps is not critical. You want the stage gain to be a little over one to sustain oscillation only. In practical circuits one variable cap will be enough, not counting a trimmer.
Then there is the series tuned colpitts. In this circuit the feedback caps are only used for feedback.
Is this not correct?

and thank you SV for the additional info on your circuit.
I would like to know a ballpark figure for the total reactance of the inductor. It makes choosing one easier for any frequency desired.

 

[Space Varmint]

The ratio of the two caps is not critical. You want the stage gain to be a little over one to sustain oscillation only. In practical circuits one variable cap will be enough, not counting a trimmer.
Then there is the series tuned colpitts. In this circuit the feedback caps are only used for feedback.
Is this not correct?

and thank you SV for the additional info on your circuit.
I would like to know a ballpark figure for the total reactance of the inductor. It makes choosing one easier for any frequency desired.

Well, what you do if you want to calculate it, is to choose a center frequency, then choose a variable capacitor and calculate the Xc at that frequency by the formula XC=1 / 2pi X fC. Then you will match the inductive reactance XL by XL =2pi X fL.

Really, unless I am using a ferrite core which uses different formulas depending on the AL value of the core, then I will just play with it. The physical construction example I gave you should be a good starting point and you must consider the stray capacitance's and so forth depending and the type of circuit board you are using and so forth.

The best method is to play with it...lol. I usually calculate up everything when I am done. Then if I am using a printed board from a board house I can be sure that circuit replicas will closely within tolerance, match.

 

[Nigel Goodwin]

Here's a good website basically describing what I'm talking about with the ganged series capacitors:
**broken link removed**

Don't know about it been a 'good' website, but that circuit is the wierdest looking colpitts I've ever seen - googling round it looks to be based on the original valve colpitts design.

Colpitts oscillators are usually built using a common collector circuit, or for VHF/UHF using common base (like all the many FM bug circuits).

 

[Space Varmint]

Don't know about it been a 'good' website, but that circuit is the wierdest looking colpitts I've ever seen - googling round it looks to be based on the original valve colpitts design.

Colpitts oscillators are usually built using a common collector circuit, or for VHF/UHF using common base (like all the many FM bug circuits).

It's the AC feedback path that determines the type of oscillator. Doesn't have anything to do with transistor or tube configuration.

 

[Nigel Goodwin]

It's the AC feedback path that determines the type of oscillator. Doesn't have anything to do with transistor or tube configuration.

No, but certain configurations are used for certain reasons - and the diagram you posted the link to was a particularly strange one, that makes life difficult

 

[Space Varmint]

No, but certain configurations are used for certain reasons - and the diagram you posted the link to was a particularly strange one, that makes life difficult

It looked fine to me. You Brits do everything differently anyway.