EMRFD Message Archive 6218
Message Date From Subject 6218 2011-04-27 08:24:38 Tim K7HFD oscillator simulations and real-life tank measurements Spent some time playing with the K7HFD oscillator on the bench over the weekend.
Also stuck it into LTSPICE to see what it thought.
The energy levels in the tank in the simulation are extremely impressive. 1 Amp peak current, 70 V peak voltage, and that's with an inductor Q of 200 (inductor ESR of 0.25 ohms).
But simulation aside, how do I measure tank voltage and currents in real life with say a scope? I can make a 2 MHz J310 Hartley on the bench and hook my 10:1 scope probe up and see 30V or 60V (depending on feedback tap) pk-pk on the tank but I suspect the voltage was larger before I loaded it down with the scope probe.
These voltage levels are high enough that I wonder if I'm bending the rules by using 50V ceramic NP0's.
I'm not intentionally spoiling inductor Q or anything but at the same time I'm not taking exceptional steps (like really thick wire or super oversized toroids or anything) to keep the inductor Q high. Just, say, a T68-6 with 22 gauge wire.
With a 100 MHz scope and a 10:1 scope probe, how much do I load the tank down if I sample voltage across the entire tank? 50 ohm tank impedance seems typical with K7HFD and maybe 2x to 4x higher in J310 Hartley design. Am I loading it down any less if I measure at a tap at 1/10 the turns or 1/4 the turns?
It seems like a current probe would be nice too - especially one that won't spoil circuit Q - but I don't have the bucks. I am surprised at oscillator circulating currents modeled to be in the amp range and to be honest I don't actually believe it :-). Is there a cheap substitute for RF I can make with ferrite toroids?
Tim N3QE6219 2011-04-27 08:36:22 Jim Miller Re: K7HFD oscillator simulations and real-life tank measurements for the current probe how about a single wire through a toroid for the
primary then a multiple turn secondary loaded by 50ohms. that should scale
down by the square of the turns ratio to minimize the loading on the
primary.
do i have that right?
jim ab3cv
[Non-text portions of this message have been removed]6220 2011-04-27 08:36:34 chuck adams Re: K7HFD oscillator simulations and real-life tank measurements On Wed, 2011-04-27 at 13:48 +0000, Tim wrote:
>... snip snip ...
> Spent some time playing with the K7HFD oscillator on the bench over
> the weekend.
>
> Also stuck it into LTSPICE to see what it thought.
>
> The energy levels in the tank in the simulation are extremely
> impressive. 1 Amp peak current, 70 V peak voltage, and that's with an
> inductor Q of 200 (inductor ESR of 0.25 ohms).
>
> Tim N3QEI had thought about this before, but since you have a working circuit,
>
>
may I suggest the following?
Find an IR thermal measurement instrument. I have a cheap one that
I bought a harbor freight some time back just for the purpose of
measuring case temperatures of transistors before and during operation.
Was/am going to do several such experiments to see if I can find
something useful.
The IR temperature probe won't load the electronics.
So how about measuring the coil temperature over a period of time to
see just how much it heats up. Maybe even try several wire diameters
to see just how much the coil heats up.
I hope I'm not out on a limb here in suggesting this experiment.
FYI
chuck, k7qo6221 2011-04-27 08:51:36 kb1gmx Re: K7HFD oscillator simulations and real-life tank measurements 6223 2011-04-27 11:44:28 Wes Re: K7HFD oscillator simulations and real-life tank measurements Hi Tim, and group,
One thing you can do is to start with a good capacitor of small value, but of high Q. For example, a 1 pF NP0 ceramic with a suitably high breakdown voltage would be good. If you run through the numbers, you will find that this cap in series with 50 Ohms looks like a pretty high resistance when the combination is hung across a tank. For example, this 1 pF 50 Ohm combo looks like 5 Meg in parallel with 1 pF at 10 MHz. The 1 pF can be attached to a coax connector on the oscillator breadboard and then attached to a 50 Ohm instrument with cable. The instrument can be an analyzer, power meter, or 50 Ohm terminated scope.
You might not know the exact value of the "1 pF" capacitor. Not a problem. Use the test set up described on one channel of a 50 Ohm terminated scope. Then put a 10X probe on the other channel. The comparison will give you a calibration for the 1 pF path. You can also note the impact of the 10X probe as it is connected and disconnected at the tank.
If the 1 pF is still too much loading, some shunt C across the 50 Ohm end will further decrease the loading. Calibrati6225 2011-04-28 10:40:05 Glen Re: K7HFD oscillator simulations and real-life tank measurements Tim,
Have been fooling with oscillators using ceramic resonators recently.
Some have pretty high unloaded Q - in the 800 ballpark. Was trying to
see what "tank" voltages could be measured with a 'scope, and
suspected, like you, that the 10X probe was degrading RF voltage and
giving a lower reading. So tried it this way instead...
Set the 'scope trigger for "single-sweep", with trigger level close
to the peak tank voltage. Then stab the probe onto the tank. The
envelope of the RF wave exponentially falls to the new level, loaded
by the probe losses. But at the moment of stabbing, you get a
somewhat better picture of unloaded peak tank voltage. You STILL get
an attenuated picture, especially when 10X probe capacitance is an
appreciable fraction of tank capacitance.
I like Wes' suggestion of a really small probe capacitor to a 50 ohm
resistor. This would be good with a 50-ohm terminated scope, rather
than a 1X or 10X probe. The tiny series cap is better than all but
the really pricey probes. However, calibration is a bit of a pain.
You can still use the single-sweep method, if you suspect loading.
A half-ohm current probe can easily be made with a ferrite bead,
wound with 10 turns of small wire. A 50 ohm resistor shunts the
ten-turn winding, and goes to the 'scope. Thread the wire you wish
to probe through the bead (this is a one-turn primary winding).
A tenth-ohm current probe would require a 10-ohm loading resistor.
That K7HFD oscillator can give some unstable results, not because of
tank heating, but because of VHF ringing caused by the transistor
turning on and off so hard and quick - especially with tapped coils
and the small # of collector link turns. Perhaps its one reas