I traced the low sensitivity problem I was having with the
SW40+ to Q1. Good audio on the source and low audio
on the drain. Had some MPF102's on hand and replaced
same. Problem solved! Now, no matter where I put the
probe around the bench, I have to turn my signal generator
off to not hear it. The keying thump also went away. Guess
that was the clue all along.

I tacked an 8.2pf NPO (had one) to the board, across C1 and
can now tune T1 for a peak.( K8IQY changed C1 to 68pf to get
a peak and Glen Leinweber mentioned that he gets T1 to
peak with core turned near the end.)

I am building a SW-40+ (with the QRP-L Elmer 101 project). I got the
transmitter working with your help. Thanks! It puts out a nice, clean
2+ watts. About 30 V peak-to-peak. Very nice signal.

Then I couldn't wait any longer, and went ahead and finished the kit!
Therefore, I can't really post this to QRP-L yet. However, I have a
problem. Now when I transmit, I get very little power out of the
transmitter. (I haven't even tried to see if the receiver works yet.)
I only get about 150mv peak-to-peak out when keying the transmitter.
I am trying to understand the TR switch circuitry, and the 4 diodes
(D7 - D10). It almost looks like the diodes will limit the voltage.
Then I calculated the impedance of C40 to be 500 ohms at 7MHz, so I guess
if I can get enough current flow through RFC3, the combined IR drop
across C40 and the voltage drops from the two diodes should still be OK.
It isn't. Do you have any ideas what could be wrong? I did install the
potentiometer at J1 and peaking T1 does affect the transmitter output
very slightly. RFC3 is not "open". I have not installed the tuning
potentiometer at J2 yet. but still have the jumper from pin 2 to 3.

Any help would be much appreciated! I will wait for the group if you
would rather.

- Craig, AA ZZ


Craig, (6/22/98)
C40 and RFC3 form a series-resonant tuned circuit, which resonates
around 7MHz. Its low-Q, so doesn't need to be trimmed. So even though you
calculate 500ohms for C40, the actual impedance between the antenna PI
filter and the RX input (T1) is very low.
With a properly working transmitter, D7-D10 will conduct on the
transmitted peaks, limiting the >30v to about 2.6v p-p. When the diodes
conduct (while transmitting), their impedance is VERY low. This destroys
the Q of C40/RFC3, and you can't think of these two components as a series-
resonant circuit anymore. Now RFC3 acts as a choke of 440 ohms impedance
going into the receiver.
What does the transmitter see? It mostly sees a 47pf capacitor going to
ground - the diodes provide a ground-path for a large portion of the
7MHz cycle. You could ground the diode-end of C40 and the transmitter
notice (but then the receiver wouldn't get any signal).

In troubleshooting your transmitter, I'd look elsewhere for the
trouble, not at this part of the TR circuit. I'd be looking around Q6...
at D12, D6.
I'd want to know if Q6 is getting a good base-drive signal. That'd
tell me if the trouble is before the final. If there IS a good 7MHz signal
at the base, I'd be taking a close look around Q6, D12, L2, and the PI-
network to the antenna. Hope you've got a dummy load on, while testing.

- Glen

Glen, (6/22/98)
Thanks for the help! Here is the latest in the saga.
First of all, yes, I do have a dummy load connected.
I checked the voltage at the base of Q6 and get a nice, solid 2 v p-p at
The voltage at the collector of Q6 remains a meager 150 mv p-p. This pointed
me to suspect D12. I lifted one end of this diode so I could check it with
my Ohm meter. I get about ~74 ohms(??) in one direction and "infinite" in
the other direction. I think this indicates that it is OK, but it may still
be the culprit. I don't have an easy way to test the reverse "zenner"
action that this diode is supposed to provide. I don't have another 33v
zenner handy, but will get one if you think that is still a prime suspect.
I haven't tried transmitting with D12 lifted. I don't think it would hurt
anything, and didn't dare!

I tried grounding the diode end of c40, and the transmitter output still
remains the same.
I have checked all the DC voltages that are indicated on Dave's schematic,
and all are good.
L2 appears to be good; the 150mv p-p is found on both sides of L2.

The thing that is so confusing to me is that the transmitter WAS WORKING
FINE until I added the receiver and TR circuitry. This makes me suspect a
component that was ADDED. Not necessarily, I know.

Is it correct that I should see about 30v p-p from the collector of Q6 to
ground? (Same as the output of the PI network.)

Thanks again, Glen.

- Craig, AA ZZ

Craig, (6/23/98)
Since you've got D12 lifted, try the transmitter. Just be sure
your dummy load is connected: the purpose of D12 is to protect the final
amp from rouge loads.
74 ohms seems a little low for measuring the
impedance of the zener. But some ohmmeters display strange stuff when
measuring diodes, depending on their current source, and voltage scaling.
Since it DOES measure infinite resistance the other way, I'm with you - I
think it's probably OK.

Looks like you've got good base drive on the final. Its looking
more and more to me like the final amp (Q6) is cooked somehow. Before
pulling it out, do a really careful visual inspection on top of the
board, and on the bottom of the board around Q6, L2, D12, C36, C37, C40
and L3. Look for open circuits and shorts. I'd do this check under a
really bright light and a magnifying glass. I've found more than a few
faults this way: things like hair-line cracks in traces, or lifted pads
or filamentary shorts between traces.
Measure the DC voltage on the heat-sink tab of Q6: should be +12V DC.
Measure the base drive right at the base lead of Q6: you may have
measured it elsewhere, and somehow its not getting into Q6.
Also, make sure the emitter lead is well-grounded too - its not a check
that Dave has asked for, but the final amp sure won't work if Q6's emitter
isn't at ground.
Yes, the peak-to-peak voltage at Q6's collector should be in the
range of 20-30 volts, unless you've got the drive pot turned way down.

The next step after doing all these checks is to pull Q6, and test it
with your ohmmeter. De-soldering a three-legged transistor like this
is tricky: the printed circuit pads take a beating. Its an area where
currents are high, so you don't want to beat-up these pads too much.
Sometimes its better to cut Q6 out rather than de-soldering
it. Then you can clear the three holes one-by-one.
Good luck, and keep in touch

Glen, (6/24/98)
YES! It works! The bad guy was Q6. I changed it, and now the transmitter
is putting out a nice, clean 2+ watts again. Getting almost 30v p-p on
output! Now on to testing the receiver.

Glen, thanks so much for your time and your patience.

In retrospect, debugging a problem like this has helped me understand the
circuit more, and that is why I am building this. I really was confused
though, and I really appreciate your depth of knowledge and willingness to

>My sw-40+ tunes from 7.0010 to about 7.036. When I tune in the low
part(7.0010-7.0017), I hear a high frequency whine. It is now very loud.
a bit over the background noise. What is causing it?
Where does it comes from?<<

If the whine is tunable, and is at its lowest pitch at the low end of the
band. It's a 'birdy', or spurious mixing product. The product detector is
sensitive not only to its 4 Mhz LO, but to oddharmonics of that LO signal as
well. As you tune to 7.000, the 4th harmonic of the (3.000 mhz) LO beats with
that spurious 12 Mhz product detector response to produce an audio output.

If the pitch of the whine isn't tunable, it may be an intermod (IMD) product
and can be eliminated by backing down the 'gain' control a bit. If that
doesn't take care of it, it's an IF response, and can be minimized by 1)
grounding the crystal cases ans 2) buttoning up the board in an enclosure.
The diagnostic for this involves touching the crystal cases with a fingertip
to see if this makes the whine louder.

>>I will reverse the tuning pot wires... Right now, it is set in a reverse
>>way (increasing freq. by going counterclockwise).
>I will probably do the same for the gain. Louder counterclockwise...

>>Did Dave reverse things on his drawing coming with his enclosure? Or did I
simply reversed the pots? (solder posts up instead of down...)
>>Reply to the list... others are interested...

My original hookup illustration was incorrect- it's been fixed. (That presumes
that 'clockwise' should correspond to'increasing frequency', of course.)

First, I had to play around with L1 to get the desired range. Any suggestions
on how to "cast it in concrete" to make sure it doesn't drift due to shock and
vibration? What material do you people use, and how do you go about
pouring/installing it without getting goop all over the rest of the

Second, on receive I've got considerable hum in the background even with the
volume control all the way down. Any way to get rid of it?

Third, when I turn off the power, the rig gives a loud squeal through the
headphones -- loud enough to be quite uncomfortable if you're wearing the
phones. Again, any way to get rid of this electronic flatulence?

The answers.......

Paul et.al.,

1. Use clear finger nail polish on the toroid(s). Cheap and it works

2. On the hum. You did not say what you were using for P/S. Battery?
Otherwise check the DC output from other P/S's. P/S = Power Supply.

3. The "I'm turning off patented NN1G feature" of the SWL-XX and SW-XX+
series is due to the audio section becoming an oscillator as the
voltages and currents wind down. See Elmer200 series later for
some reasons why this happens. :-)

D13 smokes on power up after completing section 13


Thanks for your replies earlier today. I started by removing L2 to
separate the caps from the other stuff. It was not the caps. I removed
D13, but had no resistance between either pad and ground. Must be a
short someplace. I removed Q6 completely (the pain!) and checked both
D13 pads to ground with the same result. Ooh, that was scary. The only
connection I saw was from the collector of Q6 to one side of C36.
Looked again, and noticed that the same pad (of C36) is connected to the
ground plane. This certainly doesn't look right. No, it can't be
right. It looks like it's time to break out the Xacto knife to make the
little penninsula into an island... then start putting things back
together. You were both right about C113 not being blown up ;-)

Thanks again. I'll probably melt solder early tomorrow morning. Ah,
what a way to start the day!

>In Mike's final Elmer101 installment, he warns against keying the rig with a
>poor SWR at the antenna connector, indicating it could wipe out the final
>transistor. I've seen this same warning many times before.
>Could somebody please explain to me the mechanics of what's going on? Why
>should the transistor care what kind of SWR is out there? What happens in
>terms of emitter/collector voltages and currents that destroys the device?

Going to take a crack at answering. Haven't smoked a rig
this way myself, so am going to apply some theory....

When you've got a 50ohm resistive load at the antenna jack, then Q6 sees
close to 50 ohm NON-REACTIVE as well. This is because the PI-filter
consisting of C37, L3, C38, L4, C39 transforms the 50-ohm antenna.....
into a 50-ohm impedance at its other end (at the transistor). This is
true only at one frequency...7MHz.
Everything runs great this way.

Now suppose you pull the antenna off....
The PI network was designed for 50 ohm load. Now the load is infinite.
Q6 sees a whack of reactance now, when it used to see about 50 ohms.
Or suppose you short the antenna out....
You've STILL got a whack of reactance at Q6's collector, although
it's value will be different than the case where load is infinite.

I modeled Q6 as a SPST switch (to ground) that opens and closes at
a 7 MHz. rate. What's the switch see on its collector? A whack of
reactance...perhaps inductive reactance from L2. This would look
like a flyback switch (like a car's spark-plug coil and distributor):

During the time that Q6 is ON, current climbs (linearly) in L2, as
current passes thru L2, thru Q6 to ground. Then Q6 turns OFF. L2's
current goes to zero, and the voltage across L2 swings w-a-y above
the supply voltage. Bad news for Q6, which can't take a big over-
voltage on its collector. D12 saves the day by clipping voltage
excursions on Q6's collector to +33 volts.

I once saw +160 volt spikes on the collector of a transistor
operating in a circuit like this (with a 12v DC supply). The circuit
became unstable because of the high (and reactive) load impedance,
and turned into a low-frequency oscillator with these huge,
sharp spikes. Luckily, the device survived.

When you've got a proper antenna load, the load resistance
"damps" the wild voltage swing, so that it doesn't even get high
enough to cause D12 to conduct. In this case SWR is low, because
you've got a load close to 50 ohms.