Hi Frank,

I am sending your message through the group a 2

^{nd}time, as I saw no comments after the first run… it’s getting to be a busy time for folks, I’d guess…Maybe we can catch a reply this time…

I have purchased a new SG-239 tuner within the last week. After

hooking it up to an antenna and to a test setup I have come up with a

couple of questions regarding its behavior.

Test setup: I have an IC718 transceiver which I hooked up, through a

MFJ941E operated in the bypass mode therefore utilizing the VSWR and

forward power meter feature, to the SG-239. The coupling cable

between The IC718 and MFJ VSWR meter is one foot of RG58. The output

of the VSWR meter is hooked up to the SG239 utilizing a one foot long

length of RG58 with 7 ferrite beads slipped over the outside of the

cable. The antenna connections on the SG-239 is connected with "pig

tail" clip leads on one end of a two foot long RG58 cable to a 100

watt dummy load.

Test results: The SG-239 matched to the dummy load with typical VSWR

readings of < 1.5:1 as measured on both the internal IC-718 SWR meter

and the MFJ VSWR meter except for a couple of bands. Initially a

match on the 20 meter band was attempted using the IC-718 in the AM

mode with 40 watt carrier. The SG-239 "clicked" a few times and

refused to match and VSWR was > 5:1. A second attempt was made on 40

meters and the SG-239 relays began "clicking" and a 1.1:1 match was

quickly achieved. I returned to 20 meters and tried again and this

time the SG-239 found a match that was < 1.5:1. I tried the other

bands 80 through 10 meters and achieved a good match on all bands

except 15 meters. The best match available on 15 meters was 3:1, this

was in spite of the observation of the VSWR meter during the tuning

process that several very good VSWR readings were found but were

skipped over in favor of a 3:1 match.

Antenna setup:

(Message over 64 KB, truncated)- Hi, Frank.

The SG-239 (and all the other SGC tuners) works just fine feeding a

doublet fed with ladder line. As a matter of fact, I have one in

service now. I too have seen my 239 act "nervous" at times for no

obvious reason, but it always seems to work after a couple of tries

at matching.

> Some antenna impedances presented to the SG-239 that I'm trying to

match:

>

These are some extremely high impedances, and would be difficult to

> 7.2 MHz: 93 -j 1616

> 14.2 MHz: 88 -j 147

> 18.1 MHz 279 -j 646

> 21.4 MHz 78 +j 252

> 28.4 MHz 73 +j 353

match with most any kind of tuner. The limited choice of circuit

constants available on the SG-239, especially the load side

capacitance, makes it unlikely you will be able to match these

loads. Looking at the schematic of the SG-239, we can see the range

of adjustments of the transmitter side capacitance is binary

combinations of 51pF, 100pF, 200pF, 400pF, 800pF, and 1620pF (51pF

to 3171pF in increments of 51pF). The series L range of adjustments

is binary combinations of 0.125uH, 0.25uH, 0.5uH, 1uH, 2uH, 4uH, and

8uH (0.125uH to 15.875uH in steps of 0.125uH). The load side range

of adjustments is binary combinations of 50pF, 100pF, 200pF, and

390pF (50pF to 740pF in 50pF steps).

I ran your stated impedances through the ZIZL spreadsheet calculator

and then looked at the match with the nearest approximation the SG-

239 could produce, with the following results:

40 meters -- No match possible. The extremely high capacitive

reactance value means you would need essentially a series inductor

to match the load. The calculated Pi network from ZIZL's "T-Pi-L"

calculator function shows you would need a 5pF output capacitor on

the load side to match as a L network. The minimum available step

for the SG-239 is 50pF. With this value selected, it's impossible to

get a match within a 2:1 VSWR. Choosing a Pi network gives

essentially the same results.

20 meters -- This would probably work OK. I empirically determined

the best match using available values would occur with a Pi network

configuration having an input side capacitance of 151pF, a series L

of 1.25uH, and a load side capacitance of 50pF. This would give a

transmitter port load impedance of 38.7 -j0.7 with a VSWR of 1.3.

Not coincidentally, a 59 foot piece of feedline is reasonably close

to being a full wavelength on 20 meters, and would come close to

repeat the actual antenna feedpoint impedance (presumably 73 +j10 or

so in free space) at the tuner end.

17 meters -- No match. Again, the limited choice of load side

capacitance in the 239 has you foiled. A Pi net with a load side

capacitance of 18pF is required.

15 meters -- This **might** work. An L net with a load side

capacitance of 61pF is required. This load can matched using the

50pF output cap step with a series L of 2.125uH. This results in a

transmitter port impedance of 105 +j1.4 with a VSWR of 2.1.

Depending on the actual physical reality, you may get lucky here.

10 meters -- No good again. We need a load side capacitance of 34pF;

the minimum 50pF is way too high to work.

It appears you have selected an unfortunate length of ladder line

for your antenna. If possible, I would recommend changing the line

length in small increments to see if you can get a more tractable

set of terminal impedances. At the higher frequencies, it won't take

much change to make a dramatic difference in the impedance presented

to the tuner at the end of the feedline.

The impedance values you quote look like they were derived from some

sort of calculated model. Actual measurement of such high impedances

in the field is essentially impossible. If this is the case, I would

advise playing around with the modeled feedline length to see if you

can find a compromise length that will work physically and present a

less bizarre load to the tuner, at least on some bands. Also, as

your experience shows, models seldom correspond to the real world.

The fact that your tuner found matches on all bands except 15 and 12

meters bears this out. Try pruning the feedline a little and see

what happens. It's going to be tedious, but I think you can

eventually find a compromise that will work better than your out-of-

the-box installation.

73...

Randy, W8FN - Test results: The SG-239 matched to the dummy load with typical VSWR

readings of < 1.5:1 as measured on both the internal IC-718 SWR meter

and the MFJ VSWR meter except for a couple of bands. Initially a

match on the 20 meter band was attempted using the IC-718 in the AM

mode with 40 watt carrier. The SG-239 "clicked" a few times and

refused to match and VSWR was > 5:1.I believe I would just concentrate on getting the tuner to tune my antennas. A 50 ohm dummy load obviously doesn't need tuner because it already has a near 1:1 match and this sometimes can confuse an autotuner. So concentrate on what the tuner will do with an actual antenna connected before you start to worry about it tuning something that doesn't need tuned. Also, if the radio has a built in swr bridge, use it and skip the external one to help eliminate some additional unwanted losses.

72,

Rick McKee, KC8AON, Ohio - www.angelfire.com/electronic2/qrp

Monthly QRP Field Day Group - http://groups.yahoo.com/group/MQFD/

With God all things are possible <>< - Thanks to all who answered.

W8FN, I really liked your description as "nervous" as to how the

SG-239 acts sometimes. It seems that it finds a good match, however

it seems to forget where it was sometimes and tries again with poorer

results. I guess that is why they have the "lock" feature which I

have not been able to implement yet.

You are correct in assuming my impedances are from computer programs.

I used EZNEC and the ARRL program "TLW" to determine the approximate

impedances. I was somewhat surprised to see in the December issue of

QST in the "The Doctor is IN" column that the same two programs were

recommended for deriving antenna impedances.

The TLW program also includes a "tuner" program to figure the

efficiency of different impedances which can be matched with certain

matching networks. The Pi network used by SGC is included and the

limitations of a 50 pF capacitor on the output end of the matching

network can be entered into the calculation. After looking at the

"perfect matches" such as 63' for 20 meters I choose a compromise

match for all desired bands of operation of 59'.

The surprising thing is matches at 80 and 40 meters, which are

indicated to have high losses, are the most stable with the SGC-239

with a 20 meter dipole antenna.

Thanks again,

Frank

--- In Smartuners@yahoogroups.com, "w8fn1" <w8fn@...> wrote:

>

> Hi, Frank.

>

> The SG-239 (and all the other SGC tuners) works just fine feeding a

> doublet fed with ladder line. As a matter of fact, I have one in

> service now. I too have seen my 239 act "nervous" at times for no

> obvious reason, but it always seems to work after a couple of tries

> at matching.

>

> > Some antenna impedances presented to the SG-239 that I'm trying to

> match:

> >

> > 7.2 MHz: 93 -j 1616

> > 14.2 MHz: 88 -j 147

> > 18.1 MHz 279 -j 646

> > 21.4 MHz 78 +j 252

> > 28.4 MHz 73 +j 353

>

> These are some extremely high impedances, and would be difficult to

> match with most any kind of tuner. The limited choice of circuit

> constants available on the SG-239, especially the load side

> capacitance, makes it unlikely you will be able to match these

> loads. Looking at the schematic of the SG-239, we can see the range

> of adjustments of the transmitter side capacitance is binary

> combinations of 51pF, 100pF, 200pF, 400pF, 800pF, and 1620pF (51pF

> to 3171pF in increments of 51pF). The series L range of adjustments

> is binary combinations of 0.125uH, 0.25uH, 0.5uH, 1uH, 2uH, 4uH, and

> 8uH (0.125uH to 15.875uH in steps of 0.125uH). The load side range

> of adjustments is binary combinations of 50pF, 100pF, 200pF, and

> 390pF (50pF to 740pF in 50pF steps).

>

> I ran your stated impedances through the ZIZL spreadsheet calculator

> and then looked at the match with the nearest approximation the SG-

> 239 could produce, with the following results:

>

> 40 meters -- No match possible. The extremely high capacitive

> reactance value means you would need essentially a series inductor

> to match the load. The calculated Pi network from ZIZL's "T-Pi-L"

> calculator function shows you would need a 5pF output capacitor on

> the load side to match as a L network. The minimum available step

> for the SG-239 is 50pF. With this value selected, it's impossible to

> get a match within a 2:1 VSWR. Choosing a Pi network gives

> essentially the same results.

>

> 20 meters -- This would probably work OK. I empirically determined

> the best match using available values would occur with a Pi network

> configuration having an input side capacitance of 151pF, a series L

> of 1.25uH, and a load side capacitance of 50pF. This would give a

> transmitter port load impedance of 38.7 -j0.7 with a VSWR of 1.3.

> Not coincidentally, a 59 foot piece of feedline is reasonably close

> to being a full wavelength on 20 meters, and would come close to

> repeat the actual antenna feedpoint impedance (presumably 73 +j10 or

> so in free space) at the tuner end.

>

> 17 meters -- No match. Again, the limited choice of load side

> capacitance in the 239 has you foiled. A Pi net with a load side

> capacitance of 18pF is required.

>

> 15 meters -- This **might** work. An L net with a load side

> capacitance of 61pF is required. This load can matched using the

> 50pF output cap step with a series L of 2.125uH. This results in a

> transmitter port impedance of 105 +j1.4 with a VSWR of 2.1.

> Depending on the actual physical reality, you may get lucky here.

>

> 10 meters -- No good again. We need a load side capacitance of 34pF;

> the minimum 50pF is way too high to work.

>

> It appears you have selected an unfortunate length of ladder line

> for your antenna. If possible, I would recommend changing the line

> length in small increments to see if you can get a more tractable

> set of terminal impedances. At the higher frequencies, it won't take

> much change to make a dramatic difference in the impedance presented

> to the tuner at the end of the feedline.

>

> The impedance values you quote look like they were derived from some

> sort of calculated model. Actual measurement of such high impedances

> in the field is essentially impossible. If this is the case, I would

> advise playing around with the modeled feedline length to see if you

> can find a compromise length that will work physically and present a

> less bizarre load to the tuner, at least on some bands. Also, as

> your experience shows, models seldom correspond to the real world.

> The fact that your tuner found matches on all bands except 15 and 12

> meters bears this out. Try pruning the feedline a little and see

> what happens. It's going to be tedious, but I think you can

> eventually find a compromise that will work better than your out-of-

> the-box installation.

>

> 73...

> Randy, W8FN

>