Thanks again to all. I'm going to take charge of this problem first
of all by rebuilding the project, using direct wiring, using the
layout recommended by QST, June 1967. Using this method, I can fit
everything into the just the lid section of my existing box and I
won't have to remake the tube socket holes in a new chassis. Based
on what everyone has said, I ought to get rid of all that loose wiring
anyway, prior to making any tests with a scope, which I don't yet
have. I want the thing to be well-constructed, to eliminate all that
unpredictability we've been talking about. I will also use only my 50
ohm dummy load for testing until everything's as good as I can make
it.

On 8 Oct 2003 05:40:24 -0700, (Ashhar Farhan)
wrote:

(John Sandin) wrote in message ...
Regarding the 3-watt transmitter I'm building (and having trouble
with), here's an update.
...

john,
a receiver will probably report all the harmonics of the crystal. so
that is not really a big deal. almost all the harmonics will also
sound pretty powerful and equal to each other in the HF receiver
because the recevier's AGC will boost the lowering outputs to a higher
level.
the output at 3.5mhz can be easily explained by a receiver shielding
deficiency. what i would really suggest to you is to TAKE CHARGE. i
mean, figure out exactly what is going and how.
do you have access to an oscilloscope? a single oscilloscope trace is
really worth a thousand blind trails. failing which,
i suggest that you should first make yourself some test instruments. i
would highly recommend a wavemeter. it requires about 5 components (a
couple more if you want to build a GDO-cum-wavemeter). A single coil
should cover most of the HF band of interest to you. then ...
rebuild the transmitter from scratch, keep everything at 50 ohms
impedance. refer to EMRFD for guidlines. then ...
don't push too much gain in each stage. It is actually easier to debug
and build a transmitter with more stages having lower gain than trying
to push maximum out of fewer components. finally ...
shoot for about 7 watts of output. An 80 cent IRF510 should easily
give you that much. use the wavemeter and the a power meter to measure
the harmonic outputs. But frankly even if you are pushing out 50% of
the energy in fundamentals, then political correctness aside, you
should be able to make contacts. QRP of single watters are best left
to old war horse like W7EL. it is tough for a novice to get through on
a single watt being rock-bound.
i would also suggest that you hook up with a ham down the block to be
able to monitor your signal. then another across the city ... then
another across the state...
- farhan


-John Sandin KC0QWE

Remove the "T" to respond by e-mail

John Sandin wrote:

Regarding the 3-watt transmitter I'm building (and having trouble
with), here's an update.

I removed the lowpass filter and loosened the wiring, and for a while
it seemed to help. The transistor didn't get as hot. But the output
stayed the same (less than 1 watt). But then I noticed, as I've
noticed before, that the voltage measured at the dummy load doesn't
always drop when the key is released. And the transistor stays hot
even after the key activity has stopped.

I did as one poster suggested. I removed the antenna from my
receiver, tuned around while transmitting, and discovered the
following:
(Crystal is 7110 kHz)
at 3554 kHz, signal is S8
at 7110 kHz, signal is S6
at 14218 kHz, signal is S6
at 21326 kHz, signal is S7
at 28433 kHz, signal is S5

Even I can see that this isn't good.

This IS good - it is what you should expect with the LP filter
removed. (In fact, you may very well "see" these frequencies
even with the LP filter installed.) You're "seeing" the harmonics
with the receiver. It proves the thing is producing oscillation
in the Q1 stage, which is a desired effect. But it doesn't tell
you anything about self oscillation or the "health" of the Q2
circuit. If you disconnect the base of Q2 and the receiver doesn't
hear those frequencies any more, or hears them at a lower level,
that proves that the Q2 stage was amplifying.

What is of interest is the fact that you saw voltage across the
dummy load, after the key was released. That is bad. Q2 should
be cut off by the 33 and 56 ohm resistors with no signal on L2,
and even if it was shorted, there should be no voltage across
the dummy load. That suggests that Q1 must still be producing
a signal, or self oscillation in the Q2 circuit. Check the
capacitor values in the Q1 circuit - make sure it is a 10 uf
electrolytic, and the 220 is 220 pF. The resistance to ground
from the base of Q2 should be about 90 ohms. The components
connected to the base of Q2 should have short leads, and you
want to keep wires and components connecting to the Q2 collector
away from the Q2 base and all of the components to the left
(on the schematic) of Q2.

And the emitter of Q2 connects
to the same trace as everything else that's grounded in the circuit,
and it's not a very big trace.

I'm probably going to rebuild this rig, without the circuit board.
Many thanks to all who gave me advice on this. Any further comments
are also much appreciated.

I put the schematic on my website. The design calls for 28 VDC for a
5-watt output, but I am using 12, which is also ok according to the
article.

http://www.kcnet.com/~oyster/transmi...ansmitter.html

-John Sandin KC0QWE

Remove the "T" to respond by e-mail

John Sandin wrote:

Regarding the 3-watt transmitter I'm building (and having trouble
with), here's an update.

I removed the lowpass filter and loosened the wiring, and for a while
it seemed to help. The transistor didn't get as hot. But the output
stayed the same (less than 1 watt). But then I noticed, as I've
noticed before, that the voltage measured at the dummy load doesn't
always drop when the key is released. And the transistor stays hot
even after the key activity has stopped.

I did as one poster suggested. I removed the antenna from my
receiver, tuned around while transmitting, and discovered the
following:
(Crystal is 7110 kHz)
at 3554 kHz, signal is S8
at 7110 kHz, signal is S6
at 14218 kHz, signal is S6
at 21326 kHz, signal is S7
at 28433 kHz, signal is S5

Even I can see that this isn't good.

This IS good - it is what you should expect with the LP filter
removed. (In fact, you may very well "see" these frequencies
even with the LP filter installed.) You're "seeing" the harmonics
with the receiver. It proves the thing is producing oscillation
in the Q1 stage, which is a desired effect. But it doesn't tell
you anything about self oscillation or the "health" of the Q2
circuit. If you disconnect the base of Q2 and the receiver doesn't
hear those frequencies any more, or hears them at a lower level,
that proves that the Q2 stage was amplifying.

What is of interest is the fact that you saw voltage across the
dummy load, after the key was released. That is bad. Q2 should
be cut off by the 33 and 56 ohm resistors with no signal on L2,
and even if it was shorted, there should be no voltage across
the dummy load. That suggests that Q1 must still be producing
a signal, or self oscillation in the Q2 circuit. Check the
capacitor values in the Q1 circuit - make sure it is a 10 uf
electrolytic, and the 220 is 220 pF. The resistance to ground
from the base of Q2 should be about 90 ohms. The components
connected to the base of Q2 should have short leads, and you
want to keep wires and components connecting to the Q2 collector
away from the Q2 base and all of the components to the left
(on the schematic) of Q2.

And the emitter of Q2 connects
to the same trace as everything else that's grounded in the circuit,
and it's not a very big trace.

I'm probably going to rebuild this rig, without the circuit board.
Many thanks to all who gave me advice on this. Any further comments
are also much appreciated.

I put the schematic on my website. The design calls for 28 VDC for a
5-watt output, but I am using 12, which is also ok according to the
article.

http://www.kcnet.com/~oyster/transmi...ansmitter.html

-John Sandin KC0QWE

Remove the "T" to respond by e-mail