6M: Question on matching low impedance 20 ohms j12 to 50 ohm coax

[K1JJ]

Hola,

I wonder if someone can suggest a way to match an unbalanced  20 ohm j12 reactive load to an unbalanced 50 ohm coax using an inductor and capacitor or some other LC combination?

I am trying to make a 1500w VARIABLE matching network for 6M that sits out on the tower in a box - to go from the harnessed eight phased yagis to the one 50 ohm hardline.

Would a coil tapped at the bottom for 20 ohms and tapped at the top for 50 ohms with a variable cap across it be the right approach or is there some other way?

I do not want to use coax transformers or toroidal baluns because they are not able to finely tune out the reactance as easily as an LC circuit. I expect the array's match to drift over time because of weather, age, etc, and it will need re-tweaking from time to time.

I was thinking of a simple copper coil and a stanadrd air variable with close plate spacing due to the low voltage/ low impedance application.  I need to build six of these for various arrays, so want to keep it cheap and simple.

Ideas?
T

[WA1GFZ]

Tom,
Check out Rick Measures site and look at his tuner. A couple pages in gives the formula for each kind of network. This way you can match the network you need to the parts you have. The simple one would be a L network with the cap on the 50 ohm side. A Pi would be more parts but better harmonic rejection. A tapped coil would also work but I would think the L value could quite low. I would avoid bread slicers unless they connect stator to stator eliminating current on the spring connection. I.E. a 2 section cap so both sections end up in series with the rotor as a common putting them in series.

[Tom WA3KLR]

Tom,

“(+)j12”; you are stating positive by convention, which is inductive.  I know that the MFJ-259B analyzers, if that’s what you used, display the reactance but not the sign:

I came up with 76 pf. for the capacitor value in parallel with the 50 feed point to ground, then for the series inductance to your 20 Ohm reactive load:

39 nH, if your j12 is inductive (+j12).
114 nH, if your j12 is capacitive (-j12).

50 Ohms - - >         - - - - [ 39 or 114 nH] ---- > 20 Ohms +/- j12 Ohms
                      76p                                     
                     
                      GND                                                  GND

[K1JJ]

Tom,

“(+)j12”; you are stating positive by convention, which is inductive.  I know that the MFJ-259B analyzers, if that’s what you used, display the reactance but not the sign:

I came up with 76 pf. for the capacitor value in parallel with the 50 feed point to ground, then for the series inductance to your 20 Ohm reactive load:

39 nH, if your j12 is inductive (+j12).
114 nH, if your j12 is capacitive (-j12).

50 Ohms - - >         - - - - [ 39 or 114 nH] ---- > 20 Ohms +/- j12 Ohms
                      76p                                     
                     
                      GND                                                  GND

Hi Tom,

You are one of the few guys I know who is either intuitive or carefully listens and thinks out a solution like it was your own problem. Yes, the MFJ doesn't tell  + or -  j and your double suggestion of possible values is much appreciated.

So looks like I can tune the right capacitor value and have to experiment with the right coil to find the + or - j . The coil is gonna be very small, which is FB. This will be sporting up on the tower...

BTW, normally I would have needed about 25 connectors to do just one eight Yagi array. I will eventually have six arrays, each fixed covering 60 degrees of six directions,  so that's 150 connectors needed! (These would be N type or PL-259, etc)  So, instead of spending the rest of my life servicing them as the weather caused problems, I instead carefully simulated a 50 ohm connection and soldered all the connections without connectors on this first array - experimental.

If the connections exhibited no impedance bumps, I should have seen the final two 50 ohm cables coming together to produce 25 ohms j0. But I ended up with 20 ohms j13. That's not too bad I figure and proves the connections at 50 mhz are acceptable. Considering what I saved in connectors, time, future maintainence and reliability, I'm happy.

Thus, the need for the match tuning. Since the array will now feed into a matched hardline all the way to the shack, (using your matching system) it may prove to be a good system overall loss-wise.

Tnx again, OM and I will let ya know the results. And TNX Frank for your suggestion of using an L network. Yes, that is the best choice. I use one to go from the 50 ohm 6M driver into the GG 8877 linear  40 ohm?j?  cathode with 1:1 swr results.

For a 6M 1500W signal, what precautions wud you take for the cap plate spacing, construction and the wire diameter for the connections and coil?  It would be nice to use vac variables, but I'd need six for the whole project, so will stick wid bread-slicers. OK Frank on the rotor connection precautions. I will be careful there as usual.

73,
T

[WA1GFZ]

Tom why not add a quarter wave of RG11 to transform 50 ohms to 100 ohms then tie them in parallel.

.3 UH inductors in the L network sounds like you need to find some silver plated stuff next week...

[K1JJ]

Tom why not add a quarter wave of RG11 to transform 50 ohms to 100 ohms then tie them in parallel.

.3 UH inductors in the L network sounds like you need to find some silver plated stuff next week...

Yes, I thought of that Frank and also considered a 1/12 wave matching section using a pair of 50 ohms in parr and a 50 ohm. That wud do a 2:1 also.

But there is some reactance in there to tune out and I don't know if it is L or C at this point.

I'm  considering trying Tom/KLR's values and using a 100pf quality doorknob to ground and then play with a small L to fine tune it. This wud eliminate the variable breadslicer. Remember that I need to build six units, so want it simple, but variable to some degree.

I think I need to climb up with a "kit" of parts and JS around to find the best combo using the MFJ-259B.

Yesterday I climbed up and down the tower and wired up the harness for the second array facing NE for Maine/Europe. So, it appears I'm definately gonna go thru wid this big project using eight fixed Yagi arrays in six different directions. (48 Yagis - sigh)

The WX looks good for Hosstraders - Gary, Steve/Huz and John JN will probably be here on Thurs. You might stop by in the late afternoon and check it out. We plan to have your favorite, shushi for dinner - join us!

T

[Tom WA3KLR]

Tom,

I did a simulation of the L-C matching network with 1500 Watts:

The voltage across the 76 pf. shunt input capacitor is 387 V. peak.  An air spacing of 0.015” would be adequate for this.  The current through the capacitor is 6.6 A. rms.

The current through the series inductor is 8.7 A. rms.

Although the following wouldn’t be easy to adjust, it would be reliable once done.  I was picturing a mini-box 4 x 5 x 3” deep or the next larger size with 2 SO-239/N connectors on the top about 4” apart.  A piece of copper sheet cut to about 3 x 4” and connected to the input connector center pin.  Space the sheet about 0.1” from the side wall of the box with 2 insulated spacers.  Tweak the spacing by hand.

It’s going to be tough to have something low loss enough for the current at 51 MHz. and easy to bend too.  Cut 2 copper straps about ½” wide and 4” long. Connect to the 2 SO-239/N center pins.  Overlap the 2 straps and solder together arched as necessary.  Can bend them some also.  You can try different widths, or heavy wire, or tubing if necessary to home in on the inductance.

It would be best to come up with a connector plug with some ¼ watt resistors with lead length adjusted to duplicate your measured feedpoint impedance and then tweak the box in close to 1.0:1 SWR on the work bench first.

[WA1GFZ]

I wouldn't use door knobs Tom,

[K1JJ]

OK on the homebrew cap, Tom. That is a possible solution once I give it some more thought about construction.

I'm wondering about Frank's idea using a pair of 1/4 wave 75 ohms coax cables in parallel to go from 25 to 50 ohms.

Now, what if I lengthened or shorted this 1/4 length somewhat?  I really need to go from 20 j12 (+ - j?)   to 50 ohms. I wonder if I could pull the resistance up while pulling the reactance down. It would be by luck not knowing if the 12 ohms is Xc or Xl at this point, I think 

Any estimates of success to achieve 50 j0 if the reactance was (+ or -)  and the 75 ohm 1/4 wave matching section is lengthened or shortened?

hmmm... Maybe if I climb up there with a two foot longer than 1/4 wave matching section and trim it to see the trend...

T

[WA1GFZ]

Tom,
I was able to pull my 75 meter array with a series of jumpers in the shack I had a number of patch cables that I selected to dial the SWR in. I would think your long feed line to the shack could be tuned at the rig to get a good match.

[K1JJ]

Tom,
I was able to pull my 75 meter array with a series of jumpers in the shack I had a number of patch cables that I selected to dial the SWR in. I would think your long feed line to the shack could be tuned at the rig to get a good match.

The 2.5 to 1 match needs to be done at the output of the phasing harness going into the hardline, on the tower. This way there is a 1:1 swr for the entire 400' hardline run to the shack. Matching it in the shack buys a matched rig, but results in a 2.5:1 swr loss in the feedline. That adds up on 6M.   

The good thing about the harness is the power splits evenly to the eight Yagis, so the coax harness loss is only the equivalent to about 80' of RG-213....which is the path to a single Yagi. So, if the 400' hardline can be perfectly matched to this harness, it's not a bad system loss-wise overall.

T

[K1JJ]

Tom,
It’s going to be tough to have something low loss enough for the current at 51 MHz. and easy to bend too.  Cut 2 copper straps about ½” wide and 4” long. Connect to the 2 SO-239/N center pins.  Overlap the 2 straps and solder together arched as necessary.  Can bend them some also.  You can try different widths, or heavy wire, or tubing if necessary to home in on the inductance.

Tom, could you explain this a little more?

Are the two SO-239's for the output to the two 50 ohm cables of the harness?  And a third SO-239 is for the hardline input, right.

So why are you using TWO coils... one for each feedline? Why not tie the two SO-239's together and use just one coil from the cap to this junction?

Or maybe you meant a double thickness single coil and connected as I described above.
I just want to be sure in case you have a better idea I'm missing.

This plan sounds like a reasonable way to make up six matching boxes and keep things strapping. Actually I plan to use one hardline to feed two back to back arrays that are individually switched on with vacuum relays.  So there will be two matchers and two relays in each box.

T

[Tom WA3KLR]

Tom,

Two connectors only.  Two pieces of overlapped strap to form only one arched, adjustable inductor.