[Elecraft] Measuring SWR (Long)

Tue Apr 9 16:39:36 EDT 2019

My friend Al has described some interesting things about SWR measurement here. But not all power/swr meters behave this way. In fact the Elecraft W2 meter is in this class. The designer added an ingenious little circuit to the diodes to bias them into their linear region. What the smeans is that even at low power the W2 is giving accurate measurements not just of reflected power, but for forward power as well. this is important since the effect that Al describes is valid not just for reflected power, but also forward power at low levels.

There is one other things to watch for that actually negates Al’s suggestion to always tune for lowest reflected power. As an ATU tunes, it provides different loads to the driver. This will greatly affect not just the reflected power, but forward power measurements as well, to the point that we may measure very low reflected power, but the forward power is very low as well. Thus when we tune we really do look for both maximum forward power in conjunction with minimum reflected power.

One last point. The directional coupler measurements in the KPA500, KAT500 and KPA1500 are designed to handle large power levels since the devices are meant for high power. We don’t really try to be accurate at low levels. Indeed the KPA500 won’t even indicate power below 25 watts and the KAT500 needs something like 10 or 20 watts to work properly. This allows us to be accurate at high power levels and also avoid having to be concerned with the diodes workin in their non-linear mode. In other words, on these products you can pretty much believe the power levels the device is showing when it is showing them.

Al, thanks for a nice write-up.

73!
Jack, W6FB

> On Apr 9, 2019, at 12:42 PM, Al Lorona <alorona at sbcglobal.net> wrote:
> 
> There's always a lot of discussion about measuring SWR, low SWR values, SWR lights not lighting up, etc. Here's something that might give you a better 'feel' for SWR.
> 
> Imagine that you measure your forward power at 100 W and your reflected power at only 1 W. You'd probably be very happy about this. Congratulations, your SWR would be 1.22 to 1.
> 
> Pause for a moment and let it soak in that an SWR of 1.22 is fabulously good. Once you get to this point, below which we're dealing with reflected power that's less than 1% of your power, or four-hundredths of a dB, it's not worth doing any more to your antenna system to improve it. No trimming, cutting, raising, lowering, hanging stuff from it, adding remote tuned things, nothing. You're done for the day and can now get on the air!
> 
> If the reflected power were 4 W, which would still look pretty small on the meter, the SWR would be 1.5 to 1.
> 
> A lot of people might be bothered by an SWR of 1.5, but really this is still very good and it's probably not worth going back up to the antenna to mess with it any more. Most importantly, the station on the other end cannot possibly hear the difference between your 1.2 and 1.5. 
> 
> At this point you may be saying, "Yeah, but my rig/amplifier/other thing isn't happy unless the SWR is below 1.5 to 1, so I would still have to do more work."
> 
> Okay, that's fine, but my point is that the absolute number 1.5 nonetheless represents a system that is working quite well.
> 
> Let's allow the reflected power to increase all the way to <gasp> 10 watts! At that point, your SWR would be about 2 to 1. Sounds pretty bad, but surely if you were stranded in the desert and had to use your KX3 and a wire to get help, you'd be happy to have a match this good.
> 
> Now, let's look at a certain type of SWR meter found in virtually every HF rig out there that uses a directional bridge, a diode as a half-wave rectifier, a filter capacitor, a resistor, and a readout of some kind - analog meter, digital display, or other.
> 
> These types of SWR meters are very common. There's one in the K2 (it's actually in the KPA100), one in the K3, and probably every other rig out there. And those Birds, MFJs and Daiwas have them, too.
> 
> This type of circuit turns RF voltage into a DC current, usually using either a germanium or Schottky diode, because these diodes have the best sensitivity due to their low forward voltage drops: 0.3 V for the germanium, and 0.4 V for the Schottky.
> 
> Let's say that our directional bridge, when 100 W is in the forward direction, delivers 10 V to the forward Schottky diode. In our example above of 1 watt reflected, the reflected diode would see 1.0 volts from the reflected port of the bridge. Since this is above the diode threshold voltage, the diode would be happy and operating in its linear region where we'd get about 1 volt out for 1 volt in. The meter would correctly read 1.22 to 1.
> 
> Now say you have your Elecraft rig set for TUN PWR = 10 W, so when you tune up, forward power is 10 watts, and if the SWR is the aforementioned 1.2 to 1, the forward power would produce a little more than 3 volts at the diode and the 0.1 watts reflected would produce about 0.3 volts. Here is where we run into trouble.
> 
> Since a Schottky diode doesn't fully conduct until we exceed its 0.4 volt threshold, expecting to get 0.3 V out for 0.3 V in isn't quite realistic. The diode's output voltage won't be zero, but it'll be smaller than expected, according to the nonlinear region around its "knee". I measured a Schottky diode and got about 0.09 V at that level. The meter would indicate an SWR of... 1.06 instead of the actual 1.22. This reading is false.
> 
> There are some things that can be done to the circuit to compensate for this error but I'm not sure how many rigs go through the trouble. Suffice it to say that measurement of very low SWR has a large amount of uncertainty because the detectors get more and more nonlinear as the thing they're trying to measure (reflected power) gets smaller and smaller. Due to this effect (and also due to directivity) every meter of this kind has a point below which SWR measurements are probably wrong.
> 
> When we see posts here by folks puzzled as to why low SWR measurements from two different instruments don't agree, this undoubtedly is one of the reasons.
> 
> I chose these scenarios deviously, to illustrate a kind of corner case, but I think you get the idea that errors always exist in any power measurement like SWR. You might say, "But my SuperDuper wattmeter is calibrated by aliens in a metrology lab deep inside a secret mountain in Nevada and I would bet my life on it." Great, cool, congratulations. But I'm afraid every last power meter has errors stemming from mismatch, directivity, nonlinearity and drift.
> 
> If you simply tune for minimum reflected power you should be good to go, and don't get too hung up on if the meter says "1.2" or "1.5" or "1.0". Remember, once you get down into this zone it's very likely good enough.
> 
> Al  W6LX
> ______________________________________________________________
> Elecraft mailing list
> Home: http://mailman.qth.net/mailman/listinfo/elecraft
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:Elecraft at mailman.qth.net
> 
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html