link coupling

[wa3qns]

Hi Everyone,
I have a little problem and maybe someone can help me out.
Something tells me from what I have read this won't be an
easy fix.  With only 7 or 8 watts you just can't waste any.
I recently bought that one tube transmitter kit from
Pastime Projects for use on 40 meters.  The ONLY antenna I
am able to put up is a 50' random wire. The antenna output from
my Kenwood is connected by a 3' jumper of 50ohm coax to a
simple L network just inside the window, the output of the L
network goes thru a 4" strip of wood which the window is closes
on.  The wire then goes from the window down to a small tree 50'
away.  A counterpoise runs down the wall.  The shack is on the
second floor. The output of this transmitter is "LINK" coupled
and I have not been able to find any simple way to couple my
50' wire to this transmitter.  Any ideas appreciated.
Thanks
Joe
wa3qns

[N2DTS]

I would think you can just ground (counterpoise) one side of the link, the other goes to the antenna.
Or, you can put a tuning cap in series with the link,, or hook the lick up to the tuner.
Tube rigs with link coupling do not care about swr if the feedline is good, or does not exist...

The antenna tuner will likely just waste power...
 

[kb3ouk]

On the side of the coil that is grounded you might want to have a loading cap between the coil and ground, something like the tuning cap out of an AM broadcast receiver should work, you just need some way to adjust the loading to make sure you are feeding all of the power to the antenna, since the antenna coil on that transmitter is fixed and there's no way to adjust the antenna loading. On some transmitters that have a swinging link, it's smple to adjust the loading, just swing the link in to or out from the plate coil, but since that uses a fixed coil, you can use a cap to do the same thing (I've saw some designs that used both, a swinging link and a loading cap).

[W3GMS]

Your wire is of such length that the wire looks voltage fed when your on 40M.  The link is likely a 50 ohm link.  To transfer power effectively I would use your simple L matching network to transform the low impedance output of the transmitter to a higher impedance that your antenna presents.   

If you would use the wire on 80M it would look current fed and the link would work much better directly into the antenna, but not the case when your on 40.  Voltage fed is good since your losses will be lower since a current fed end fed requires a good ground system to work against otherwise the radiation resistance goes up considerable. 

A simple L matching network especially at low power pretty much allows you to be not concerned with RF losses in that network. 

Joe, GMS       

[wa3qns]

Thanks for all your replies, sounds like I was making this a bigger problem than it is however
I have read link coupling is not the way to go if you have other options.  All I have ever
dealt with from the 60's has been pi networks and 50ohm coax.
Glenn, NN8G the fellow offering these kits high lighted the following in the original
QST article after he heard my comments on links.
"If room for only a short length of wire is available for the antenna, say 40 or 50 feet
it is best to connect its end to one antenna post and a good ground to the other as
shown in Fig. 6-41C.  Here again some experimentation will be necessary to determine
the optimum size of L3."  So I hope to have some success with this one way or the other.
Next time it will be a pi output.
If all goes well I'll post an update on all the DX I've worked:o)))
Thanks & 73
Joe

Joe- GMS
Long time no hear. Tnx for your input as an RF engineer I am not.
I can follow schematics and built almost anything, I design little
to nothing.

[VE3LYX]

 I recently played with link vrs pi on a hb project. The switch to a pi brought instant success and much better adjustability. It took only a few minutes and a second variable C to wire it in and the problem I had been stubbornly fighting vanished . It bothered me to capitulate but I can see now why the Pi is so popular.
BTW My homebrew antenna tuner is a pi as well. It could not be more versitle I think . Allows me to run dipole or long wire single feed with ease. 
My 2 Cs worth
Don

[W3GMS]

Thanks for all your replies, sounds like I was making this a bigger problem than it is however
I have read link coupling is not the way to go if you have other options.  All I have ever
dealt with from the 60's has been pi networks and 50ohm coax.
Glenn, NN8G the fellow offering these kits high lighted the following in the original
QST article after he heard my comments on links.
"If room for only a short length of wire is available for the antenna, say 40 or 50 feet
it is best to connect its end to one antenna post and a good ground to the other as
shown in Fig. 6-41C.  Here again some experimentation will be necessary to determine
the optimum size of L3."  So I hope to have some success with this one way or the other.
Next time it will be a pi output.
If all goes well I'll post an update on all the DX I've worked:o)))
Thanks & 73
Joe

Joe- GMS
Long time no hear. Tnx for your input as an RF engineer I am not.
I can follow schematics and built almost anything, I design little
to nothing.

Hi Joe,
Yes its been a very long time and hope you have been well.  That sounds like a fun project your working on and I am sure you will work lots of DX with it.  I have been hanging out on 40CW quite a bit and maybe we will work on good old CW!

73,
Joe, GMS

[w7fox]

Joe,

Your L network will work fine hooked to the link.  What you need is an output indicator.  A sensitive meter with a diode across it an a couple wires for an antenna are all you need, just tune for max.  If you peg the meter, use shorter wires for the meter antenna.  Or, you could use the same thing I used with my arc5 as a novice, a flashlight bulb with a couple of wires soldered on. with alligator clips.  You clip across a few feet of your random antenna, and move the wires closer together if it gets too bright.  Have fun.

Chris

[W3RSW]

If you plan to stay on 40 , I think I'd shorten the active antenna wire to 33 ft. Or so, amd as mentioned place a 365 pf variable cap in series with the link output and operate as a quarter wave.  The cap allow some play simulating changing the number Of turns on your output link coil.

This should work well if your counter poise is also 32 or 33 ft. 

Hey I like your fly trap alarm.  Great use for a limit switch.  Just don't be doing anything extracurricular when pesky squirrels or a heavy breeze comes jumping around. 

[wa3qns]

Rick,
That little sonalert went off twice to let me know my big dx antenna was down.  Because people walk
their dogs back there I can't afford someone tripping on my downed antenna.  I also have a web cam
set up to detect motion and over the last 6 months I have some interesting pics.  Tnx for the antenna
suggestion, when we first moved here 15 yrs ago I did have it tied off to my neighbors tree 35 feet away.
Tree died (maybe from RF overdose!) had to plant my very own tree, then it died, planted tree #2, it
looks sick! Can you see a pattern here?  Anyway that 50' hunk of wire got rave reviews all over Europe
when I worked PSK so I don't want to change much.
Chris,
Tnx for your handy tips, I am thinking this dog might hunt after all, what with all this positive feedback
I've gotten here.
73 Guys
Joe

[w7fox]

Joe,

Answering your post brought back memories.  I certainly didn't have a sensitive meter when I was 12, but I knew where I could find a flashlight bulb.  That explains why none of the flashlights around the house worked when I was a kid.

Chris

[wa3qns]

I have a flashlight fetish, gotta stop and look at them. Have a couple dozen of those little LCD jobs
that Harbor Freight gives away.  When most other flashlights are useless those are still giving
useable light.
I bet Don would like to know that I'm a Francophile, fell in love with the language through music
abt 20 years ago still can't hold a conversation.  Went to NYC to see Patricia Kaas in 2000.  Zaz is
my current fav.
Viva La Francais...
Joe

[wa3qns]

My first tube kit not going well.  I have no idea if these old FT243 40 mtr
xtals are oscillating or not, but I'm really hung up on the power supply ground
to the xmitter.  The transformer is a Thordarson T13RO8  Secondaries: 
350-0-350 90mA; 5V 3A; 6.3V 3.3A; 2.5V 6A. I think I'm confused because
with wood there isn't any chassis ground.  Their schematic shows one side
of the filament grounded where I'm taking the 6 volt heater voltage straight
to the tube, no ground. Filament lights up nice!  I know now I shouldn't
have put 400 volts on that 6V6 but considering I had zero output is
probably OK.  I guess a big resistor will solve the over voltage problem.
The HV center tap (ground) is connected to nothing on the xmitter, that
can't be right, can it?
Tnx for and all replys..........
Joe

[K1JJ]

Joe,

That's right, it won't work until you connect all the grounds together.

If you have no chassis to connect all the "grounds" to then simply run a heavy wire to ALL points that show the ground symbol.   A metal  chassis is just a convenient and low impedance path for ground connections.  A heavy wire will do about the same thing in your case. Sometimes a heavy strip of copper called a "ground buss" or "busbar," is used.

For best electrical stability, make the RF grounds on the oscillator especially short in length within the RF circuit itself - and connected to the same point if possible. The DC power supply ground leads can be longer to satisfy parts placement.

Tom, K1JJ

[wa3qns]

Success at last! (well sort of) Have a solid clean 7 watts into 50 ohm dummy load.  Thanks
one and all, funny how everything has to be grounded to work isn't it;o) One problem remains.
I used the Kenwood to load and tune my wire thru the L-network then attached xmitter, touched
the key heard one little chirp on the Kenwood then nothing. Plugged in another xtal, no problem,
meaning one 40 mtr FT243 xtal is now history! So, unfortunately I'm still faced with this link
coupling nightmare.  I do not intend to blow anymore rocks.  I really don't care to mess with that
output coil, the turns were all hot glued so I'm thinking about dumping the link for a pi network. 
Thanks again.
Joe
wa3qns

[WU2D]

Joe,

For loading adjustment with a link, it is handy to take the coupling a bit over  optimum by adding an xtra turn or two on the link and inserting a variable between the ground side of the link and ground. Now the tuning and loading can be optimized much like a Pi network. For 40M, an ordinary 365 pF variable will work fine. 

You might get a few more Watts out.

Mike WU2D

[wa3qns]

Tnx for the suggestion Mike.  My second try was successful, no cracked xtals and a solid
7 watts output "BUT" am changing link to pi output tonight, will post results here later.
I wish I knew how I screwed up the first time I put that key down but I did!
In market now for a receiver as I will not use my Kenwood.
ciao
Joe

[wa3qns]

These are the changes I made in the Pastime Kit. The pi output
works but I believe the original circuit has a bit higher output.
Most circuits use some value screen resistor but it's use doesn't
seem to be of any value because the orig. circuit works very
well without it. Something I changed has lowered the output
power by a watt or two.  Could it be the screen resistor? I would
appreciate it if someone would explain this resistors function.
Tnx much,
Joe

[WU2D]

Well for one, the 6V6 will not be happy with 400V on the screen since its maximum rated screen voltage is 285 V. It may damage the screen or shorten the tubes life.

The maximum plate voltage is 315V by the way. Running the plate at 400V is typical ICAS treatment and as long as it does not arc over, the plate dissipation in CW is so low that you can get away with it.

More importantly, the screen is very happy with 200-250V on it and the crystal current will be lower and it likely will key cleaner!

[Al AE6CM]

touched the key heard one little chirp on the Kenwood then nothing. Plugged in another xtal, no problem,
meaning one 40 mtr FT243 xtal is now history!

The oscillator circuit you have is called a "Tri-tet" and it is historically (and notoriously) known as a "crystal cracker".  It was invented in 1933 at a time when crystals were relatively large slabs of very expensive quartz and with maximum current ratings of over 100mA, and when hard-to-drive triodes were the tube of choice for a ham's final amplifer.  So the high-harmonic-output Tri-tet circuit was hailed as one of the best things to come along since "sliced quartz".  Now a ham could get by with just one expensive 80 meter crystal that would also work on the 40 and 20 meter bands.  A 40 meter rock could be used up to 10 meters with enough drive for a triode final even without a buffer stage.  Ham bands at the time were all harmonically related.  Then, as crystals became more available at reasonable prices and high-power-sensitivity tetrode and pentode RF power tubes became the norm for finals the Tri-tet's popularity waned.
 
To keep a tri-tet from cracking crystals, the two important things to do are 1) keep the screen voltage under 150 volts (use large value screen resistor) and 2) make the resonant frequency of the cathode L-C tank no lower than 1.5 times the crystal frequency.  The L-C resonant  frequency has not much to do with output power but everything to do with crystal current.

All that said and despite its historical acuracy, it is a poor circuit choice if it is only going to be used at the crystal's fundamental frequency.

GL, Al

BTW, the #49 lamp should be connected directly in series with the crystal to indicate crystal current, but be aware the the bulb will not show any glow until the current reaches about 30mA.  That may be to late for the new "micro-sized" crystals.

[WU2D]

The 6V6 as a Colpitts oscillator was the transmitter in a famous circuit, the Paraset. My 6V6 does about 6 Watts out. Realistically, your 8 Watts is about the max. I plugged a 6W6 into the socket and got 9 Watts out and cracked my first crystal.  One of the reasons that the 6V6 was abandoned as an oscillator, was its tendency to damage crystals. The 6AG7 is a far better choice in this regard.

[W3RSW]

]Here are the links again for why the 6AG7 or more modern minature version released by GE, the 6CL6 works so well as a single tube oscillator/ amplifier or oscillator as a driver for a larger final output tube, say a 807 or 6146.

http://faculty.frostburg.edu/phys/latta/ee/6cl6xmtr/6cl6why6cl6.html

Also explains why an electron coupled oscillator using the 6AG7 puts very little strain on the crystal and also amplifies the oscillation. Two stages in one tube, so to speak.

These links have been on am.fone several times and ought to be memorized.