Bird Wattmeter - What's Inside??

[WBear2GCR]

Curious.

Think I knew that answer at one time... what exactly is inside a Bird Wattmeter, besides the meter itself? The slug, wazzinat? What makes the slug directional?
What makes it frequency sensitive?

Why/how is it different than a "calibrated" SWR indicator?

Anyone got a general purpose schematic of the thing?

How does it differ from a "thermocouple RF ammeter"?
Would a "thermocouple RF ammeter" actually be more or equally accurate in terms of determining real world power (especially into non-50 ohm loads)??

Anything else to know?

Thinking in general about "power meters"...

                _-_-bear

[WA1GFZ]

you can download it off bama

[flintstone mop]

The Bird IS the WORD!!!!
It's just a rugged commercial grade wattmeter that can get beat around, and give an accurate reading into 50 ohms. The two-way service shops use those for the final word on the health of a mobile two-way radio or trunking base station. In my rich days I purchased a Bird and the peak reading mod and that is my wattmeter standard for my station. I have several elements for HF up to 2500watts and a couple for VHF and UHF Ham bands. I think the plug-in elements contain a diode. Not sure what makes the magic happen in a Bird. I have seen them in many B'cast stations to constantly monitor the RF output of the station and to monitor any SWR. When your running 50KW of AM B'cast or espcially KW's of FM or even TV, higher SWR can be disasterous to the feedline or the transmitter. FIREWORKS!!!!!!!!!
Fred

[WBear2GCR]

Tanks...

I downloaded the patent, and read two "inside the slug" websites.

Any comments on RF Ammeters out there?


              _-_-bear

[Jim, W5JO]

RF ammeters require or have a thermocouple inside that heats causing a small current fed to the meter.  Accurate about mid to high scale usually.  They are sort of like an expanded scale voltmeter in that they are inaccurate at low scale readings.  For instance don't use a 15 amp meter to read the current on a 100 or 500 watt transmitter. 

[AB2EZ]

Bear

Interesting questions... that I had wondered about myself.

The patent and the also the following Web page (with lot's of good pictures and the schematic which is attached below) are excellent references:  http://www.repeater-builder.com/projects/bird-element-tour/bird-element-tour.html

The slug is a combination of a physically small directional coupler and a diode peak detector. The output of the slug is the forward or reverse envelope of the (modulated or constant) rf signal.

I looked at the output of a Bird slug with a scope one day, and concluded the the time constant of the peak detector is short enough to follow the modulation on my AM output signal. I believe (but did not confirm with measurements) that you could use any type of peak detector in combination with the Bird meter, and get a reasonably accurate measurement of power throughout the range of the meter... provided you run the output of the peak detector through a circuit (an operational amplifier or a passive circuit) that yields an output impedance that is significantly higher than 1500 ohms, and which produces 30 uA of current at full scale. The folks who produce peak-detection converters for Bird watt meters design their peak detectors to have a 68000 ohm output impedance, a 1500 ohm input impedance (simulating the Bird meter) and unity current gain.

What makes the Bird slug directional? Well, looking at the photographs and the description in the patent specification, the slug is just a physically small version of a standard directional coupler. This is particularly the case for HF and VHF slugs... where the wavelength is long enough to mask any subtleties in the physical design of the slug. The voltage across the diode peak detector is the sum of the outputs of two sensors: a small wire loop (which responds to the rf current passing through the line unit that the slug is plugged into) and a capacitive voltage divider (which responds to the rf voltage between the center conductor and ground of the line unit that the slug is plugged into)*. As in any directional coupler of this type, the output of the rf current sensor (the small wire loop) changes polarity for a wave traveling left-to-right vs. a wave traveling right-to-left. The output of the capacitive voltage divider does not change polarity for a wave traveling right-to-left vs. a wave traveling left-to-right. If you add the output of the current sensor (the small wire loop) to the output of the voltage divider... and assuming the levels of these two are made to be equal (by adjusting the length and the position of the loop within the slug, which affects both the strength of the current sensor and the ratio of the capacitive voltage divider) you get zero response to a wave traveling left-to-right (i.e., the two sensor outputs cancel). This is why the slug behaves as a directional coupler.

*The small wire loop serves not only as an rf current pickup, but also (in conjunction with the center conductor of the line unit) as part of one of the capacitors of the capacitive voltage divider.
Best regards
Stu

[WA1GFZ]

Stu,
I had a rig problem and didn't get a chance to answer your RF current transformer question. I run the whole coax through the core but only terminate the shield at one end to produce a faraday shield in the transformer. Since one end of the shield floats the center conductor still couples to the core. My coupler is in a mini box with connectors a couple inches apart. I have a hunk of teflon coax between the connectors with the shield only terminated to one connector. It is cut about 1/4 inch short on the other end.  The T68 core sits on the outer jacket. I have #26 wire secondary. I  got the idea from an old ARRL SSB handbook showing the Collins Watt meter schematic. I scaled up the power and use two meters.
minor rig problem Clamp tube decided to shut me down. I think the levelpot is open. Just yanked the tube till I get around to fixing it.  gfz

[WBear2GCR]

Nice description Stu!
I did not pick up at all on the capacitive voltage divider when reading the patent.
Sheds a whole new light on the design.

GFZ, which handbook year?
Sounds interesting.

               _-_-bear

[AB2EZ]

Bear: Thanks for the kind words!

Frank: Thanks for following up on providing the info on your "pickup" (hopefully your transmitter problem is easy to fix). The design is very interesting (to me) and very relevant to this discussion of the Bird slug design.

A couple of observations on the pickup design you are using:

1. The design takes advantage of a problem (i.e., turns what is normally a problem into an opportunity) we all encounter from time to time in our shacks. The open shield creates a ground loop that bypasses the core (i.e., the shield current goes around the core through the metal box). What would normally be a source of "rf in the shack" (if it were not confined to the small metal box) becomes the useful signal that the pickup responds to, and which is proportional to the current flowing through the center conductor.

2. Separately, as you mentioned, the short length of braid (connected only at one end of the box) serves as a Faraday shield... which will significantly reduce the the capacitive coupling between the center conductor and the secondary winding of the transformer. This prevents the sensor from acting as a combination of a current sensor and a voltage sensor (by essentially eliminating the undesired capacitive voltage divider). Current flows through the braid to charge of up the capacitor formed by the center conductor and the braid... but much of this current does not flow through the core because the piece of braid is long compared to the length of the core. Thus the current flowing on the piece of braid does not cancel the current flowing through the center conductor (from the perspective of magnetic coupling to the secondary winding of the transformer).

3. In the Bird slug, the capacitive voltage divider is used to advantage, whereas in this sensor, it is essentially eliminated by the Faraday shield.

Best regards
Stu

[WA1GFZ]

Stu, my clamper tube pulled the final screen very low. The tube is bad or the pot that sets bias. The pot is very hard to turn so may be bad. I just pulled the tube to get back on.

The Schematic of the Collins watt meter was in the old SSB handbook maybe from around 1970. I use 2 meters rather than a switch. Also set it for 2 KW full scale. It tracks my Bird 43 quite well 160 to 10 meters. The current loop is through the case of the pickup so the faraday shield does not cause SWR problems.

[WBear2GCR]

Frank, any chance you could scan it and post it or just email me a copy?

I'm interested in trying that design out.

Wondering how they generate a log output?
I presume one needs a log output to drive a linear reading meter? 
Hmmm... same question on the Bird!

                _-_-bear

[WA1GFZ]

Man, where is that book. I have not looked at it in years. Like most other watt meters the meter is calibrated log to handle the diode output in a voltage or current sensor. Search on RF watt meters might be easier. Versions of this circuit have been for years. I'm sure some in handbooks.
You will need a log to linear converter if you want a linear meter display. QEX did a project on that a year or so ago with a digital display.
I'll keep my eyes open and get back to you if I find a good article.

[AB2EZ]

Bear
Frank

You can find a diagram/schematic of the Collins directional coupler here (on the Collins Collectors Association web site): http://www.collinsradio.org/archives/manuals/312B-4-5_3rd-ed-09-62_.pdf

See page 10 of 20 of the .pdf of the manual for the Collins 312-B4.

Note that the Collins directional coupler consists of a toroidal transformer (T1) serving the role of a current sensor, and capacitors C1/C3 and C2/C4 serving in the roles of capacitive voltage dividers (voltage sensors). Also shown is the Faraday shield that protects the secondary of T1 from capacitively coupling to the primary of T1. Collins used two (2) voltage sensors, and two diodes, for symmetry.

Note that the diode peak detectors in the Collins coupler (and the diode peak detector in the Bird slug as well*) are essentially linear envelope detectors. I.e., the current or voltage output of the peak detector is (approximately) proportional to the envelope of the monitored rf signal. The non-linear characteristic (voltage drop ~ log of the current) of the diode has a small effect, because the voltage drop across the diode is generally a small fraction of the voltage across the capacitor of the peak detector... except when the monitored rf signal is at the very low end of the observable range.

*I connected the output of a Bird line unit with a 1kW Bird slug to my scope... terminating it in a 1500 ohm resistor to simulate the load of the Bird meter. I applied the output of my KW-1 to the line unit... modulating 100% with a 1kHz sine wave. The output of the Bird sensor (as observed on the scope) was a nice sine wave, which tracked the envelope of the modulated rf signal very closely. I observed both the output of the Bird line unit, and the rf signal (using a sniffer) on my dual-trace scope... and the output of the Bird superimposed nicely on the top half of the modulated rf envelope.

If you want to connect the sensor to a meter that reads the voltage (or current), then you have to square the reading to determine the rf power being observed. The Bird meter face is marked to display the power directly, and the marked scale also takes into account the small deviation of the output of the peak detector in the Bird slug from being linearly proportional to the envelope of the rf signal being monitored. It also takes into account the behavior of the meter movement (rotation v. applied current).

Best regards
Stu

[WBear2GCR]

For those following along in the Libreto, patent 4493112 ought to be the Breuner patent on the Collins unit.

And N8LP Larry Phipps has a download .pdf describing his coupler, the text also includes a discussion of some of the issues with similar designs.

It is seeming to me that within reason either the Bird type coupler or one of the several toroidal couplers are sufficient.

There seem to be some yin/yang variations depending on accuracy, frequency extremes, and power handling.

Not sure in my mind if any of these things matter unless one is particularly interested in all of the above, or any one of the above issues. For me, it seems that a reasonably accurate coupler that will cover the 160 - 10 mtr bands with good relative accuracy is sufficient.

I guess the other question that remains unanswered relates to the AD application circuit using a special precision log amp (iirc), but the point is that it uses a direct voltage divider for the sensing. So, I'm wondering if it is sufficient to merely detect the voltage to determine the power - or are they assuming the current through the resistors detects the current?

If so, is the AD app idea "good enough"?

            _-_-bear

[WA1GFZ]

I think the transformer had 32 or 36 turns CT of wire on one layer. I used a T68 red iron core. Don't use 1n82s I used 1N5711. Disconnect it from the antenna while not in use. lightning will pop the reflected power diode.

[WD8BIL]

Here's a simple circuit to use that helps with peak measurements in a wattmeter.
Since the meter itself limits the ability to follow the peaks, a peak detector with a meter driver is necessary.

I've used this for more than 20 years. A single quad op-amp is all ya need.

[WBear2GCR]

I was referring to this:

http://amfone.net/Amforum/index.php?topic=11845.0

the AD8307 chip...

I was wondering about the way it senses the RF wrt to the factors we've just discussed, that being current & voltage sensing in the coupler.

Put it another way, using this circuit as the conversion method, would it be best to use a "Collins" type or Bird type coupler with it, or would one get satisfactory results with the simple method shown?? And why?

Hmmm... just speculating now, but does this simple pick-up method show allpower - forward and reflected combined??

Showing ignorance... trying to make heads and tails of it.

            _-_-bear

[AB2EZ]

Bear

Think of it this way:

The coaxial cable has two waves traveling through it. One wave is traveling in the forward direction, and one wave is traveling in the reverse direction. At any point along the coax, if you measure the voltage between the center conductor and the shield of the coax, it will be of the form

voltage = v(1) sin [2pi x f x t] + v(2) sin [(2pi x f x t) + theta]; where

v(1) is the amplitude of the forward-traveling wave
v(2) is the amplitude of the reverse traveling ("reflected") wave
f is the frequency
t is time

and (very important): theta is the phase difference, at that point along the coax, between the forward wave and the reverse wave.

I.e., the voltage you measure between the center conductor and the shield, at any given point along the cable, is simply the sum of the voltage produced by the forward wave (at that point along the cable) and the voltage produced by the reverse ("reflected") wave (at that point along the cable).

If the SWR is unity, then v(2) is zero (i.e. no reverse traveling wave)... and the voltage measured at any point along the coax (using a peak detector) will be just v(1).

However, if the SWR is not unity, then the voltage you measure at any point along the coax will be larger or smaller than v(1)... depending upon the value of theta (the phase difference between the forward and the reverse wave) at that point.

The largest voltage you will measure (at some point along the coax... if it is long enough to find such a point) is v(1) + v(2).  The smallest voltage you will measure (at some point along the coax... if it is long enough to find such a point) is v(1)-v(2). The ratio: [v(1)+(v(2)] / [v(1) -v(2)] is called the standing wave ratio (SWR). Again, when there is no reverse wave, i.e., when v(2)=0, then the SWR=1. [v(1) + 0] / [v(1) -0] = 1, for any value of v(1).

The box (including the IC inside), shown in the description that you referred to, senses only the voltage between the center conductor and the shield of the coaxial cable... at the point along the cable where it is attached. Thus, it does not produce a measurement of either the forward power or the reflected power... but only the "power" that corresponds to the voltage it senses at that point along the cable: "power"= [voltage x voltage]/50 ohms. 

As you know, the IC produces an output that is proportional to the log of the measured voltage... and thus produces an indication of the "power" directly in dB. [Note: log [voltage x voltage/50] = 2 log [voltage] - log[50]; so when I say that the IC produces an output that is proportional to the log of the voltage, I am also saying that it produces an output that is proportional to the log of the "power"]

Note that this "power" reading does NOT correspond to the sum of the power in the forward wave and the reflected wave. Depending on where along the cable you place the box, it could correspond to  [(v(1) + v(2)] x [(v(1) + v(2)] / 50; or it could correspond to [(v(1) - v(2)] x [(v(1) - v(2)] / 50; or it could correspond to anything in between these two values.

If you want to know the power in the forward wave, you need a directional coupler which is set to sense only v(1). If you want to know the power in the reverse wave, then you need a directional coupler that is set to sense only v(2). The Collins coupler (which has a port that senses only v(1) and a port that senses only v(2)) or any directional coupler (but not including any internal peak detector) can be used with the IC you referred to... once you delete the peak detectors, and once you calibrate it to take into account the coupling ratio of the directional coupler.


Stu

[WD8BIL]

Bear; that is not a directional power coupler. That circuit will "measure" total power with no regard to direction. The result will be some result of power vector addition/cancellation/phase shift/swr influences.

Bill: What's inside a Bird ?

1- 30ua meter
1- 7/8" dia. 50 ohm line section
1- cable to connect the two.

That's it..... unless you got the peak reader then they add the peak detector/meter driver, a switch and batteries.

Bird Electronics was established in 1942. Work on their patented ThruLine design began in the late 40's culminating in the introduction of the now famous Model 43 wattmeter in 1952.

A short history from their website

[W2JBL]

     the other day at work i dropped a Bird 43 on the frozen/packed snow at a transmitter site and watched as it skidded 50 yards downhill before smacking the fence at Mach 1.5... still works fine.