SWR meter readings after a mistuned amplifier. Looong.

[The DB]

I ran an experiment for a topic on another forum. I am posting what I did and the results here for comments.

Namely:
Did I miss anything?
Is there something I can do to further test and refine the results and conclusions made?
Thoughts in general?

---

Setup:
An SWR meter, later referred to as the "test meter".
The MFJ-949E, more commonly known as the "MFJ Delux Versa Tuner II".
Varying test (or "dummy") loads to simulate different tunes on antennas.
A few short pices of RG-8X coax used to connect everything together. None of these coax lengths exceeds three feet.
Also of note is the SWR meter built into the tuner mentioned above. While not directly part of the test as it is before the variable tuning of the tuner, I did notice a correlation between this SWR meter and the actual test SWR meter so I am including it here as well. I am referring to this as the "secondary meter".

These are connected as such. The RF generator, which is an old RCI-2950 which dates from before the DX models began. This is set to deliver a 10 watt deadkey. The mic gain on the radio and the Astatic Night Eagle "lollipop"microphone have been turned all of the way down to minimize to eliminate the effects of modulation as much as possible.

After that there is a coax run using the above listed coax from the radio to the input of the MFJ-949E antenna tuner. In this tuner, before the tuning circuit, is the secondary meter mentioned above. After the secondary meter is the tuner itself. After the tuner is another piece of coax which goes to to test meter. On the output of the test meter I am plugging in the test "dummy" loads directly.

Test 1: 50 ohm test "dummy" load.
The test meter between the tuner and the 50 ohm test load showed a perfect match no matter what I did.
There were no changes in the SWR on the line after the tuner no matter how the tuner was tuned.
NOTE: This is the only test I ran before. At the time I only had a 50 ohm test "dummy" load available to me. Things have changed since then, now I have more test loads to play with.

Secondary: As I have secondary notes below, I came back to add this secondary note. The readings on the secondary meter ranged from a perfect match to off the charts. There is no correlation between the two meter readings here as the test meter always shows a perfect match.

Test 2: 75 ohm test load, produces a 1.5:1 SWR between the feedline and the load.
Running through all 12 tap points of the inductor, and on each setting adjusting both variable capacitors from their minimums to maximums and about everything in between in relation to each other for each and every tap point I do indeed have a changing SWR. However, the SWR is never higher than 1.5:1 SWR on the test meter, but I was able to get it down to a match on the test meter.

Secondary: While not directly related to the experiment at hand, on this test I also noticed a correlation between readings on the test meter and the secondary meter. When the readings on the test meter are at their lowest the readings the readings on the secondary meter are at their highest. This is also true in reverse, as one meter shows better readings the other shows worse readings. As one goes up the other goes down.

Test 3: 100 ohm test load, produces a 2:1 SWR between the feedline and the load.
Running through the settings of the tuner as I did above the same thing happened with one exception. The maximum SWR present on the test meter was never above 2:1.

Secondary: The same secondary effect happened here.

Test 3: 150 ohm test load, produces 3:1 SWR between the feedline and the load.
Running through the settings of the tuner as I did above the same thing again happened with one exception. The maximum SWR present on the test meter was never above 3:1.

Test 4: 250 ohm test load, produces a 5:1 SWR between the feedline and the load.
There is a caveat with this test. The test meter only reads to 3:1 SWR. It does have a red scale measurement area beyond that 3:1 SWR that I used for this test. I made my own mark for reference.
While I don't have the scale available to me there is still a definite high reading that is never exceeded on the test meter. This reading is clearly above the 3:1 marked reading but is not full scale.

Secondary: No change in the correlation between the test and secondary meters. Everything is as it happened above.

Test 5: 475 ohm test load, produces a 9.5:1 SWR between the feedline and the load.
This test has the same caveat as test 4. I have also made a mark for this test load for reference purposes. That point, which is further up the scale than the mark in test 4, is still not full scale deflection.
The marked high point below is still never exceeded.

Secondary: No change in the correlation between the test and secondary meters. Everything is as it happened above.

Conclusion:
An antenna tuner (or the output tuning circuit in an amplifier) can in fact change the SWR as viewed from an SWR meter after it on the feedline. However, the effect it has on that SWR is limited to the mismatch that exists between the feedline and the load (antenna). Never once did the SWR on the test meter exceed the SWR of the mismatch between the load and the antenna. As the test meter maximum readings are limited to the readings from the mismatch between the feedline and the load (antenna), the tuning circuit can only make the measured readings on the test SWR meter lower, therefore a mismatch in the tuning of an amplifier output circuit can only make an SWR reading on a meter after it appear better than it actually is.

Further, while tuning an antenna tuner, or for that matter an amplifier output tuning circuit, to the lowest possible reading on an SWR meter that is in line after it only increases the SWR that exists before the tuning circuit. Any SWR that exists before the tuning circuit in an amplifier only has one place to go, and that is back into the amplifier itself. I'll let you draw a conclusion on what happens from there.

One further note. If you happen to have a perfect match between the feedline and the antenna, the test SWR meter will ALWAYS show a match, no matter how your amplifier load control or antenna tuner is set.


The DB

 

[Shockwave]

I could be wrong but it sounds like you may not be recalibrating the forward scale on the SWR meter as you make big changes in the forward power by misadjusting the match. This is the only way I can explain the results you are seeing. Forward and reflected power will drop as a result of the mismatch between the transmitter and load but the standing wave ratio should remain stable baring any other variables.

 

[The DB]

I could be wrong but it sounds like you may not be recalibrating the forward scale on the SWR meter as you make big changes in the forward power by misadjusting the match. This is the only way I can explain the results you are seeing. Forward and reflected power will drop as a result of the mismatch between the transmitter and load but the standing wave ratio should remain stable baring any other variables.

I did, in fact, recalibrate the forward scale every time I made a change to the tuner and took a measurement using the meter marked as "test meter". I had to, every change, no matter how large or small made a difference that required an adjustment to the forward needle to get an accurate reading.

Running a quick test with the 100 ohm test load, it is possible, if you don't adjust the forward scale on the test SWR meter every time, to get the tested SWR to go up higher than any reading I took while using said test load in the test before. I have seen readings, which should be no higher then 2:1 show up at over 3:1. However, once you adjust the forward scale and recheck the reading drops back down to a 2:1 or less.

On the meter marked as "secondary meter" I did not have to do that as it is of the duel needle variety as built into the MFJ Versa II tuner. I wish I had a stand alone meter such as that for the meter after the tuner, would have made the process much quicker overall...


The DB

 

[The DB]

Before I ran the test I thought as you did Shockwave.

Here is my current line of thought on what is happening. The feedline is not matched to the generator (in this case the tuner), but being a feedline instead of a raw load (such as an antenna or a dummy load) it instead transforms the complex impedance that is presented by the tuner as the signal travels down the feedline. This happens whenever the complex impedance presented at the output of the tuner does not perfectly match the input impedance of the feedline it is connected to.

Also remember, an antenna tuner tunes not just the antenna but the entire circuit that exists beyond it, including the feedline.

As the signal travels over the feedline, its properties (effectively R and X) are transformed. It is this transformed complex impedance (which changes as you adjust the tuner) that is matched (or mismatched) to the load (antenna). As the complex impedance presented to the load varies (it isn't a constant 50 ohm zero reactance impedance), the resulting reflections generated also vary.

Something to remember, presenting a mismatched source into a feedline is not the same as presenting a mismatched source into a load. A feedline instead acts as an impedance transformer. We already know this transformation happens, look at the example of a quarter wavelength transformer which works on the exact same principle, as does the 12'th wavelength transformer.


The DB

 

[vkrules]

Thank you for conducting this test and sharing your finding even though it contradicted your original train of thought. Well done

 

[Shockwave]

Before I ran the test I thought as you did Shockwave.

Here is my current line of thought on what is happening. The feedline is not matched to the generator (in this case the tuner), but being a feedline instead of a raw load (such as an antenna or a dummy load) it instead transforms the complex impedance that is presented by the tuner as the signal travels down the feedline. This happens whenever the complex impedance presented at the output of the tuner does not perfectly match the input impedance of the feedline it is connected to.

Also remember, an antenna tuner tunes not just the antenna but the entire circuit that exists beyond it, including the feedline.

As the signal travels over the feedline, its properties (effectively R and X) are transformed. It is this transformed complex impedance (which changes as you adjust the tuner) that is matched (or mismatched) to the load (antenna). As the complex impedance presented to the load varies (it isn't a constant 50 ohm zero reactance impedance), the resulting reflections generated also vary.

Something to remember, presenting a mismatched source into a feedline is not the same as presenting a mismatched source into a load. A feedline instead acts as an impedance transformer. We already know this transformation happens, look at the example of a quarter wavelength transformer which works on the exact same principle, as does the 12'th wavelength transformer.


The DB

Something else is going on here that is related to the specific amount of power being applied to the mismatched load that is tricking the 50 ohm meter. I ran a very similar test to what you just described using the tank output of a FL-2100 amp as the means to shift the source impedance. As you described, when the load impedance is other than 50 ohms, changes in the source impedance appear to show a drop in VSWR.

What I found is that you cannot directly connect the change in source impedance to the drop in VSWR when the load in not 50 ohms. You can however connect the amount of power being delivered into the mismatched load to these changes in VSWR. When I shifted the load control away from where it was tuned into 50 ohms, the 1.5 VSWR I started with dropped. However, when the drive on the exciter was increased to compensate for the loss in power through the mismatch, the VSWR returned right back to 1.5:1 even though the tank (source) was still tuned away from 50 ohms.

 

[The DB]

Something else is going on here that is related to the specific amount of power being applied to the mismatched load that is tricking the 50 ohm meter. I ran a very similar test to what you just described using the tank output of a FL-2100 amp as the means to shift the source impedance. As you described, when the load impedance is other than 50 ohms, changes in the source impedance appear to show a drop in VSWR.

What I found is that you cannot directly connect the change in source impedance to the drop in VSWR when the load in not 50 ohms. You can however connect the amount of power being delivered into the mismatched load to these changes in VSWR. When I shifted the load control away from where it was tuned into 50 ohms, the 1.5 VSWR I started with dropped. However, when the drive on the exciter was increased to compensate for the loss in power through the mismatch, the VSWR returned right back to 1.5:1 even though the tank (source) was still tuned away from 50 ohms.

This appears at least at first glance to match up with what I have found, with the exception that I did not have the ability at the time or thought of increasing the power (dead key wise) to make up for that loss in power. I was running the radio with the analog power dial turned all of the way up. I can turn the power down and when there is a drop in SWR on the test meter I can raise the power level in an attempt to compensate for said loss looking for the effect you described.

I do find the suggestion that increasing the power is all that is necessary to bring the SWR presented back into line as odd. From what I understand on the subject, changing the power input should not change SWR at all, the ratio between VMAX to VMIN that make up the SWR reading should still remain consistent. However, on further thought, to my knowledge this is assuming there is a 50 ohm source impedance transferring power into a 50 ohm feedline impedance with an SWR meter designed to work at said 50 ohm impedance as well. I'm not yet sure how changing the source impedance to be potentially something other than the expected 50 ohms will affect this result.

When I get a chance I'll hook everything up again and begin at a lower power to see if I can confirm the results that you are describing.


The DB

 

[The DB]

Running through the test again I am getting completely different results from when I did it this morning. Nothing changed. The same meters, the same coax, I didn't even unplug anything short of the test loads to plug in a new one before this round of testing. I did put a test antenna as someone (on another forum) asked what would be the effect if reactance were added, and that had a similar result to this morning (it was also done in the morning).

My problem is the completely different (new) results are perfectly in line with what I expected in the first place. As much as I hate to say it, either I missed something the first time around, or this time I did something different without realizing it. Another thing I hate to say is the comment on not re-adjusting the forward control on the test SWR meter made by Shockwave is as close as I can come up with to being a likely cause of the differing results.

Two completely different test results from the same test, they can't both be right, right?

Now I am frustrated with this, so now I am going to disconnect everything and put it away, and in a few days if not a week or two, when I am completely awake and alert, hook everything back up and run the test again from scratch, being extra cautious and methodical every step of the way.


The DB

 

[Shockwave]

Don't get discouraged DB because I've been able to confirm everything you said so far. I just did not believe this was a result of any change in source impedance because that does not line up with RF theory regarding source impedance not effecting VSWR.

My tests did not use a resistive load. I mistuned my antenna tuner to create the 1.5:1 VSWR with reactance. I performed specific tests to isolate the differences between changes in source impedance and changes in power delivered to the mismatched load.

By changing the load setting on the amp, this altered its output impedance and caused the mismatched 1.5:1 VSWR to go down. I had the drive set so the amp was only producing about 100 watts carrier as I varied the output tank impedance. Whenever the drive was increased to compensate for the loss in power delivered to the load through the mismatch, the VSWR returned to the 1.5:1.

To reconfirm source impedance was not the cause, I turned off the mismatched power amp producing 100 watts into the 1.5:1 VSWR and set the 50 ohm exciter to 100 watts. It too produced the same 1.5:1 VSWR.

The fact creating a mismatch at the source can only lower the VSWR reading combined with this VSWR always returning to the same reading when the forward power is adjusted to compensate for loss in the mismatch shows the effect is not related to source impedance but more connected to the power level being applied to the mismatched load.

We are told that VSWR is a ratio between forward and reflected power so that once the meter is recalibrated for a different power setting the VSWR should remain the same into the same load. Part of the problem may be the 50 ohm meter is losing accuracy as the load impedance deviates away from 50 ohms.

 

[The DB]

Don't get discouraged DB because I've been able to confirm everything you said so far. I just did not believe this was a result of any change in source impedance because that does not line up with RF theory regarding source impedance not effecting VSWR.

My tests did not use a resistive load. I mistuned my antenna tuner to create the 1.5:1 VSWR with reactance. I performed specific tests to isolate the differences between changes in source impedance and changes in power delivered to the mismatched load.

By changing the load setting on the amp, this altered its output impedance and caused the mismatched 1.5:1 VSWR to go down. I had the drive set so the amp was only producing about 100 watts carrier as I varied the output tank impedance. Whenever the drive was increased to compensate for the loss in power delivered to the load through the mismatch, the VSWR returned to the 1.5:1.

To reconfirm source impedance was not the cause, I turned off the mismatched power amp producing 100 watts into the 1.5:1 VSWR and set the 50 ohm exciter to 100 watts. It too produced the same 1.5:1 VSWR.

The fact creating a mismatch at the source can only lower the VSWR reading combined with this VSWR always returning to the same reading when the forward power is adjusted to compensate for loss in the mismatch shows the effect is not related to source impedance but more connected to the power level being applied to the mismatched load.

We are told that VSWR is a ratio between forward and reflected power so that once the meter is recalibrated for a different power setting the VSWR should remain the same into the same load. Part of the problem may be the 50 ohm meter is losing accuracy as the load impedance deviates away from 50 ohms.

Whenever something like this happens I always take time to think about what happened, why I think it happened, and what I can do to make the test better/more reliable when I redo it. I never jump in to redoing such a test that I did twice with completely different results quickly. I have an error somewhere, I realize it, and my goal at this point is to remove that from occurring again. I called it frustrated earlier, annoyed is a more accurate description.

I do appreciate your effort at confirming/denying my results. An experiment is only worthwhile if the results are reproducible.

I do have a question.

I mentioned this earlier as well, but I'm finding it odd that the power level is changing the SWR on the test meter. From my knowledge of transmission line theory, the ratios between forward and reflected power remain the same no matter how much power is in the system. Therefore, as I understand it, raising the power should not change the readings on the SWR meter. What do you think the cause of the change in power changing your readings is? Could it have something to do with using an SWR meter with a mismatched (as far as the meter is concerned) forward signal?


The DB

 

[Road Squawker]

...
Two completely different test results from the same test, they can't both be right, right?


sounds like your external meter may be affected by stray RF @ the higher power levels

 

[The DB]

sounds like your external meter may be affected by stray RF @ the higher power levels

Actually, those tests were at the same power level. I was looking to find an SWR dip I found before as that is where the initial experimenting on the second round would take place and couldn't find one no matter how I set the antenna tuner. Whenever I would reset the forward control the SWR would come back into line.


The DB

 

[Shockwave]

I think it has more to do with using a 50 ohm line section in a system we know is not close to 50 ohms due to the intentional mismatch. I agree the ratio should remain the same. Perhaps Bob may have more insight but I'm convinced it is not related to source impedance. The fact I can reproduce the same results you saw the first time suggests your second test may have been flawed.

 

[bob85]

i have not tested the idea with 75ohm line, with 50ohm line my lp-100 shows no change in vswr when i mistune an amp or tube pa hf set,
my first thought was DB could be using a long piece of coax with loss thats putting his vswr meter detector diodes in the none linear part of their conduction curve.

 

[The DB]

i have not tested the idea with 75ohm line, with 50ohm line my lp-100 shows no change in vswr when i mistune an amp or tube pa hf set,
my first thought was DB could be using a long piece of coax with loss thats putting his vswr meter detector diodes in the none linear part of their conduction curve.

I used two pieces of coax in the setup, both were about three feet of RG-8X. When I redo it I will add a third piece between the test meter and the load, this is to satisfy the person on another forum who sparked the test to begin with. I don't see it as making any difference.


The DB