Jerry, This is very helpful, especially your explanation of the right side of T3. ?? I think I'll play with your idea of bypassing T3 and see what it does to the stability of the power out.???
And thanks for the hint about the Private button, never noticed it before.
/John

Getting to your primary puzzle, I doubt the right side is trying to be resonant.
The datasheet says the right side only has 2 turns, whereas the left side has 14 turns,
So the inductance is far less, your inductance meter speaks the truth.
I'm guessing the cap at C44 would work about as well if it were 0.1uF instead of 470pf.
Would be an interesting experiment to tack a cap in parallel, see if performance changes.
It does sort of look like a series resonant circuit, so we might think it would not conduct
unless at the resonant frequency.? However, the output of that transformer is a voltage source
with our 3.5mhz sine wave, the 470pf cap at C44 does not see an inductance to interact with.

John,

I prefer to have these conversations within the forum
Often there's somebody else with the extra info needed to resolve an issue.
As Ryan says, people come to this forum because they want to see a technical discussion.??
I spent a career in digital design, much of this RF stuff I am only now figuring out
by hanging out in these forums.? It's how we learn, both by hearing what others have
to say, and by trying to put into coherent words that what we think we know.

Private email can be sent by going to a forum post on the website such as
? ??/g/qrptech/message/397
Then click "Reply", fill in your message, then click "Private" on the bottom right, and "Send".
Me, I may not respond, I prefer to have these discussions within the forum.

Anyways, here's a datasheet plus a brief description of the transformer used at T3 and T5
? ??
? ??
The left side has 14 turns, nominally 4.5uH, Diz says it's adjustable between 3.2 and 5.7 uH
Internally, there's a 47pF cap.
With just the 47pF cap (plus a few pf of stray capacitance due to board traces and such)
in parallel with the mid-range setting of 4.5uH on the coil, it resonates at 10.7mhz.

As you surmise, with a 470pf cap at C43 in parallel with the internal 47pf cap,
it can be parallel resonant at our target frequency of around 3.5mhz.
A parallel resonant circuit is a short circuit to all but the resonant frequency,
so only our 3.5mhz energy goes on through the transformer to show up at the secondary.

Getting to your primary puzzle, I doubt the right side is trying to be resonant.
The datasheet says the right side only has 2 turns, whereas the left side has 14 turns,
So the inductance is far less, your inductance meter speaks the truth.
I'm guessing the cap at C44 would work about as well if it were 0.1uF instead of 470pf.
Would be an interesting experiment to tack a cap in parallel, see if performance changes.
It does sort of look like a series resonant circuit, so we might think it would not conduct
unless at the resonant frequency.? However, the output of that transformer is a voltage source
with our 3.5mhz sine wave, the 470pf cap at C44 does not see an inductance to interact with.
If you didn't want this filter in place, could remove C43, T3, and C44, drive the base
of Q6 through a series DC blocking cap from a clean source of 3.5mhz (sine or square wave).

At T5, we have our roughly 4.5uH inductor in parallel with the internal 47pf plus the
external 470pf of C64.? As with T3, the parallel resonance does not allow our desired
3.5mhz energy to pass through, so the 3.5mhz must find its way out toward the base of Q7 instead.
What T5 does allow through is DC into Q6 from the power supply.
To make T5 broadband it could be replaced with an RF choke just like L4, be sure to remove C64

Anyways, if I am correct, you pretty much had it figured out on your own.


Yup, the ride from Lewiston to Enterprise gives you a couple hundred sharp curves
on the Rattlesnake and Buford Grades where you cross the Grande Ronde.
Good road to practice your cornering on a bike, assuming you don't find a log truck.
We're off grid in a straw bale house about 3 miles west of the top of that Buford Grade,
but would take us a 30 minute drive to get there.? It's rough country.

Jerry, KE7ER


On Thu, Dec 12, 2019 at 05:05 PM, John Roland wrote:

Jerry,
Thanks for all your response and interesting ideas.?? My frustration beyond the fluctuation in the output over about 100 kcs, is that I don't really understand how the filter works.?? I think see a parallel resonant circuit comprised of the left side of T3 and C43 plus the internal capacitor which I assume shunts out of band frequencies to ground but passes desired frequencies (3.5 mHz).?? I'm having trouble with the right side of T3.?? My inexpensive Chinese inductance meter gives me a very low value of inductance between the 2 pins on the right side of T3 (pins 4 and 6 from the datasheet)? (.16 uH)? which would resonate with C44 at around 18 mHz if I'm using my online calculator correctly.?? At which point I decide that must not be how it works..... ?
Sorry to drag the group through this, but I couldn't find your email address.? I'm in QRZ and visible in the directory for this group.
And yes you're just over the hill from me.? I grew up in Lewiston and Enterprise was a frequent day trip on the motorcycle 50 years ago.??
/John

Jerry,
Thanks for all your response and interesting ideas.?? My frustration beyond the fluctuation in the output over about 100 kcs, is that I don't really understand how the filter works.?? I think see a parallel resonant circuit comprised of the left side of T3 and C43 plus the internal capacitor which I assume shunts out of band frequencies to ground but passes desired frequencies (3.5 mHz).?? I'm having trouble with the right side of T3.?? My inexpensive Chinese inductance meter gives me a very low value of inductance between the 2 pins on the right side of T3 (pins 4 and 6 from the datasheet)? (.16 uH)? which would resonate with C44 at around 18 mHz if I'm using my online calculator correctly.?? At which point I decide that must not be how it works..... ?
Sorry to drag the group through this, but I couldn't find your email address.? I'm in QRZ and visible in the directory for this group.
And yes you're just over the hill from me.? I grew up in Lewiston and Enterprise was a frequent day trip on the motorcycle 50 years ago.??
/John

The 5'th harmonic of that VFO at 5*4.4=22mhz will mix with the third harmonic
of the 8mhz local oscillator at 3*8=24mhz to give a spur at 24-22=3mhz.?

Most CW transmitters are broadband all the way out through the final.
Being class C, the final creates plenty of harmonics, but the output lowpass filter (L5,L6 in this case)
should be sufficient to clean that up.

To make a long story short, I am have a problem with the output power varying a couple watts over the tuning range depending on where I peak the filter (T3 T5).? I built the rig using ugly construction and have tried to improve the layout of the transmitter once but no joy.?? Now that the crummy weather has returned to the PNW, its down off the shelf again and I'm looking at the filter as the culprit.?? I'm thinking I have probably made an error in the assembly of that section but can't see my problem.? If you have any experience with this circuit and could elmer me through this, I'd be very grateful.??
Here's a link to the original article.??
If up to the challenge, please contact me off list.
/John? W7SWB