6T9 Parasitic Continues

[Carl WA1KPD]

Attached is a better scan of the schematic I followed with the changes. Components in red were added to address parasitic. I have also attached a pic of the actual wiring.

Since my first post I have
Bypassed the fil leads _ (they were twisted).
Added a separate .01 uF capacitor to ground at the cathode of the crystal oscillator.
Wound a new parasitic choke
Grounded pin 6- Internal shield.

Several other observations.

The peak is in one spot on the plate cap, it is not squirrely. Acts just like a resonant action- just in reverse.
The note sounds real nice in a rcvr. It is on one spot on the rcvr not all sorts of odd oscs in the rcvr.

I checked the rig with a freq counter reads dead on the crystal freq on peak.

If I take the tank coil and ground one end I get a nice DIP. When I go back to PI net, feeding the dummy load directly (This avoids running wire back under the chassis and theoretically removes stray coupling from that source from the cap output I get the peaking again. So this is somehow related to the PI net

Mike I did not measure the grid current but the peak is so large that I think there is enough grid current. I will measure it

WD8BIL I will "create" a centertap with a pair of resistors. 220 Ohms +/- from each filament leg to ground will do it. 6.3v/200= .032 amps     .032 x 6.3 = @.2 watts


Any other ideas?

Thanks and 73 fer now
Carl /KPD

[AB2EZ]

Carl

Yes... "the difference between theory and practice is much larger in practice than it is in theory..."

A few more things to check:

1. Does the dc voltage on the grid of the tetrode section change (i.e., does it become more negative) when you resonate the output pi network? This would indicate that additional r.f. from the output is getting back into the input (or that the output of the oscillator section is changing).

2. Does the screen voltage on the tetrode section of the tube go up when you resonate the pi-network?

3. Is the loading introduced by the pi-network sufficient? For a tetrode, the plate current is not very sensitive to the plate voltage, unless the plate voltage drops below the screen voltage. If the loading is insufficient, then the r.f. voltage that develops across the pi network may not be sufficient to cause a dip in the plate current at resonance. The loading capacitor (1000 pf max) seems adequate for 80 meters and 40 meters if the pi network inductor has the appropriate value. However, grounding the output side of the inductor would significantly increase the loading.

The RF load impedance seen by the tetrode, at resonance, is proportional to L x L / R*; where L is the inductance of the pi network inductor, and R* is the effective output load resistance, taking into account the effect of the loading capacitor. I.e., increasing the loading capacitance reduces R*... which, in turn, increases L x L / R*... which, in turn, increases the loading on the tetrode.

R* = R / [1 + (2pi x f x C x R)**2]; where f is the r.f. frequency, C is the value of the loading capacitor, R is the resistance of the physical load placed at the output (50 ohms in the case of a 50 ohm dummy load), and " (**2) " means square the value inside the parenthesis.


Stu

[AB2EZ]

Carl

An additional thought:

Your plate current meter is measuring (as most do) the sum of the plate current and the screen current. If the plate current dip is not very pronounced (due to insufficient loading from the pi-network), then it is possible that the screen current is increasing enough, when you resonate the tank circuit, to make the total current peak (rather than dip) at resonance.

Try moving the 15k screen resistor to the other side of the meter.

Stu

P.S. If you are going to plate modulate this rig, as indicated, the bypass capacitor from screen to ground will have to be reduced in value... but that is not important at this time. The screen bypass capacitor should be either .001 uF or .002 uF, depending upon how high you want the 3dB modulation roll off  frequency to be. A .01 uF screen bypass resistor, used in conjunction with a 15k screen resistor, would cause the 3dB modulation roll off frequency to be about 1 kHz.

[KL7OF]

The meter is hooked up backward

[The Slab Bacon]

Carl,
        After taking a close look at the skiz, here is my $.02 worth. I see that there is NO tuned circuit between the crapstal oscillator and the final amplifier. This could be causing some wierd type of global feedback wrapping around BOTH sections of the tube. If this is happening you may be getting an increase in grid drive to the final as you swing it through resonance, thus causing the plate current to rize as you tune it.

You have got to put a scope on the outpoot and take a good look at what is coming out of it. see if you are getting 2 signals at once. Also JS a plate current meter in the oscilators plate circuit and see what it is doing while the pa plate current is rizing at resonance. That could be telling you that you need a little more isolation between the oscilator and the PA.

Also try tuning the final to the second harmonic of the crapstal. (use an 80m crapstal and tune the pa to its second harmonic on 40m) See what this looks like on the scope. If you see that the harmonic and the fundamental are close to the same amplitude, You may need to rethink your design out a little bit.

                                                                                The Slab Bacon

[Carl WA1KPD]

OK Stu, Slab Mike, Johnny et al

1. Does the dc voltage on the grid of the tetrode section change (i.e., does it become more negative) when you resonate the output pi network?

It becomes slightly more positive. it goes from -2.7 to -1.7 in resonance

2. Does the screen voltage on the tetrode section of the tube go up when you resonate the pi-network?
It drops from 232V to 209V

it is possible that the screen current is increasing enough, when you resonate the tank circuit, to make the total current peak (rather than dip) at resonance.
Try moving the 15k screen resistor to the other side of the meter.

Same peak, just lower current

Also tried changing the RFC in the osc to a higher value- no luck

Slab- Interesting comment You have got to put a scope on the outpoot and take a good look at what is coming out of it. see if you are getting 2 signals at once. Also JS a plate current meter in the oscilators plate circuit and see what it is doing while the pa plate current is rizing at resonance. That could be telling you that you need a little more isolation between the oscillator and the PA.

Slab-The osc plate current rised from about 3.1 to 4 ma as the plate peaks. I presume your theory is that I should not see any change in osc plate current if it is sufficiently isolated.  One version of this rig in Handbook did have a tuned circuit. I have attached it as it also shows the modulator I plan to use.

I'm going to try changing the location of several components to isolate the two circuits a bit more physically

Thanks- I'll keep you posted.

73 KPD

[AB2EZ]

Carl

Going back to the comments by Mike (WU2D)...

The (negative) voltage on the grid appears to be very low (not enough grid drive). If you don't have enough drive on the grid of the tetrode, then it won't be operating in Class C.

If the tetrode is operating in class C, its plate current will be a sequence of narrow pulses... whose repetition frequency is the fundamental frequency. Therefore the plate current normally will contain a lot of harmonic content. The tuned pi network presents a high impedance (the desired load impedance) to the fundamental frequency component of the plate current... but not the harmonics. Therefore, it is not obvious that resonating the tank circuit will cause a dip in the total average current. However, if the loading is high enough, somehow (this is a very non-linear circuit... and therefore difficult to analyze without a computer simulation) the total average plate current usually dips when the pi-network is resonated.

If the grid drive to the tetrode is too low to start with, the plate current will contain a different mix of harmonics (the current is never cut off, and the plate current is also richer in lower order harmonics)... and this may impact on the "usual" behavior... i.e., we may not see the expected dip at resonance...
 
According to the specifications on the NU7P web site, the tetrode portion of a 6T9 should have -8 volts DC bias on the grid.

Somehow, the triode oscillator is oscillating at only a small fraction of its applied plate voltage... which suggests that there is something wrong with the operation of the oscillator. Either that, or the coupling capacitor between the plate of the oscillator and the grid of the tetrode is open. If it is working properly, the voltage swing on the plate of the triode should be a significant fraction of the plate voltage.

Stu

 

[AB2EZ]

Turning more attention to the oscillator:

The load on the oscillator is the parallel combination of several things... including the RFC.

Looking again at the data sheet on NJ7P's web site... the triode's plate current is specified at only 1.5 mA... and the output resistance of the triode is 45k ohms. If the total of all the parallel loads on the triode is (for example) 10k ohms, then one would expect a 15 volt p-p swing if the p-p current is 1.5 ma.

Now, looking at the schematic... one thing I would suspect as presenting too low an impedance is the RFC. Theoretically, 1 mH => a reasonably high impedance (>24 k ohms at 3.885 MHz)...  but I suspect that the r.f. choke is behaving like a capacitor (well past the self-resonant frequency) with a much lower impedance. If its capacitance is 10 pf (implying a self-resonant frequency of 1.6 MHz)... then its impedance at 3.885 MHz is only ~4000 ohms. Thus, the load on the triode oscillator plate would be less than 4000 ohms. 4000 ohms x 1.5 ma of p-p plate current = 6 volts of p-p plate swing => ~ 3 volts peak plate swing around its average value.

You may need to try substituting a tuned circuit for the r.f. choke to increase the load on the oscillator...

Stu

[w3jn]

Carl, before you do ANYTHING get out your scope and take a look at what's going on.  Troubleshooting without a scope is like Ray Charles driving in Manhattan at rush hour    You're blind and you're gonna bounce off a lot of stuff un-necesarily.

One comment - the plate lead to your tank (the red wire I presume) is pretty close to the crystal socket.

[AB2EZ]

Even simpler:

Substitute a 20k ohm, 5 watt resistor for the r.f. choke. The average plate current of the triode, flowing through this resistor will lower the average plate voltage on the triode by only 1.5 ma x 20,000 ohms ~ 30 volts. At the same time, it will provide the load you need for the oscillation to build up to a sufficient value.

Bypass the lower end of this resistor with the existing .01 uF capacitor to ground (probably not needed... but just to be sure)

Stu

[The Slab Bacon]

FWIW, you might also have a neutralization issue. when an rf amp is properly neutralized the plate current should dip at max outpoot and the grid current should show little change as the plate is swung through resonance. The skiz shows no neutralization whatsoever.
this could also be causing some of your proclem as the final may well be trying to oscilate. Excessively low grid drive to the final could also cause similar problems. with both the osc and pa in the same bottle Isolation could be a real issue at radio frequencies and neutralization could become much more critical. also like Stu said you may not have enough grid drive to drive the tetrode section into class-C operation. Dont forget that with grid leak biass only on the pa, you must drive it into class C operation. But like John said and I said before, get yer scope out and see what's happening.

                                                                               the slab bacon